Chapter  79:  Diffusion and Dissemination of Evidence-based Cancer Control Interventions

Effectiveness of Cancer Control Interventions

Smoking cessation interventions found to be effective in this review include: brief advice by a health care professional, office prompts (reminder systems, telephone counseling either as a single intervention or in combination with other interventions, and individual smoking cessation counseling), and media campaigns. Effective multi-component interventions include office reminders combined with physician training, with or without patient education.

There is some evidence that physician education in dietary counseling is an effective dietary intervention. However, there is no consistent evidence of effectiveness of other health care provider-directed interventions. Interventions directed at individuals that were shown to have some effect in producing dietary change include: tailored interventions; multiple interventions; and provision of multiple contacts and environmental interventions. Media campaigns may result in increased knowledge and awareness of behaviors to reduce risks.

Interventions that have been consistently shown to be effective for increasing mammography are: invitations or mailed reminders, office system interventions (i.e., prompts), and financial barriers interventions—especially when a multi-component strategy is undertaken that combines behavioral and cognitive interventions.

Effective interventions to promote uptake of cervical cancer screening include office systems (computer or manual chart reminders), and invitations and reminders to individuals. There is limited evidence of effectiveness for educational materials, telephone counseling, removal of financial barriers, media campaigns, and advice from health care providers.

There is inadequate data regarding effective interventions for the control of cancer pain. Promising interventions include the transmission of patients' self-reported pain scales to oncologists, pain education for nursing staff, and the use of daily pain diaries.

Strategies to disseminate cancer control interventions

Fifteen primary studies were identified in the systematic review of dissemination strategies for smoking cessation activities. The majority of these used nonrandomised designs to evaluate the dissemination strategy. There was no strong evidence of effective dissemination strategies for smoking cessation interventions.

Train-the-trainer approaches improve knowledge and awareness of the smoking cessation issues among health care providers, but there is no evidence they impact on advice to quit smoking or on smoking cessation rates. Several studies evaluated the use of educational facilitators to disseminate smoking specific information, or information about multiple preventive behaviors. Overall, preventive services appear to be increased but the specific impact on smoking cessation activities is more uncertain. Several studies examined the importance of different media sources for recruitment of patients or their families to use the Cancer Information Service (CIS). Media awareness campaigns, in particular those using television, are important strategies to disseminate information about CIS help lines.

There are few studies evaluating dissemination strategies to promote the uptake of a healthy diet. Seven studies were included in the review of strategies to promote an adult healthy diet. Many of the studies identified are primarily descriptive rather than evaluative. The use of educational facilitators to promote the use of office systems by health care providers shows some promise at improving the provision of preventive services in community practices. One additional study, using peer educators in the worksite, did demonstrate some short-term increase in dietary fruit and vegetable intake.

Only six studies were identified that examined dissemination strategies for mammography. These were predominantly targeting health care providers. Several studies examined the use of educational facilitators to promote the use of office systems. They provide mixed results for disseminating office system interventions. Overall, there is insufficient evidence to conclude that any dissemination strategy is effective at increasing mammography.

There were four studies identified in the review of dissemination of cervical cancer screening interventions. These studies examined the use of educational facilitators to disseminate office systems (for health care providers), or media awareness campaigns to disseminate information about the CIS. These studies all examined cervical cancer screening as one of a number of topic areas. Educational facilitators appeared to increase overall indicators of preventive care. However, there was no statistically significant increase in cervical cancer screening rates. There is no evidence for effective strategies to disseminate interventions to promote cervical cancer screening.

Three studies were included in a review of strategies to disseminate interventions for control of cancer pain. There is a lack of research examining dissemination of interventions to promote effective pain control. Dissemination of a treatment algorithm for pain management produced only short-term change in provider adherence. Cancer pain role-modeling programs were shown to improve process measures such as knowledge and education of other health care providers. However, no information is given on integrating pain assessment into clinical practice.

Reviews of smoking cessation interventions as part of preventive healthcare services in family practice

Two reviews examined adult smoking cessation interventions as part of a larger review of the delivery of preventive healthcare services in family practice.^{58, 65} Ashenden et al. undertook a systematic review with meta-analysis of randomized controlled trials (RCTs) evaluating the effect of lifestyle advice provided by family practitioners, in changing patient behavior.⁵⁸ Lifestyle advice was defined as advice provided in the general practice setting to quit smoking, make dietary changes, reduce alcohol consumption, and exercise more. It was mainly verbal advice but could include written material. More intensive advice included additional followup by appointment, telephone, or letter. Data were analyzed separately for each topic.

The review by Mullen et al. included RCTs and non-RCTs evaluating patient education and counseling.⁶⁵ Education and counseling in this setting were advice on preventive health behavior to apparently healthy individuals seen in a clinical setting in a developed country. The analysis was broken down into three groups: smoking/alcohol, nutrition/weight, and other behaviors.

Reviews of multiple adult smoking cessation interventions

Four reviews of multiple adult smoking cessation interventions were identified.^{60, 61, 71, 72} Two comprehensive US public health reports on smoking cessation were identified.^{60, 61} These included systematic reviews of both single and multiple interventions. The Public Health Service (PHS) Clinical Practice Guideline is a comprehensive guideline that includes meta-analysis, addressing the broad issue of tobacco dependence treatments and practices.⁶⁰ These include a spectrum of cancer control interventions, along with pharmacologic interventions to assist patients attempting smoking cessation. A second report by Hopkins et al., presented a series of systematic reviews contained in the Task Force on Community Preventive Services (TFCPS) report on evidence regarding interventions to reduce tobacco use and exposure to environmental tobacco smoke.⁶¹

A third review examined research related to family physician-assisted smoking cessation interventions.⁷¹ This review analyzed biochemical measures of tobacco abstinence at six and 12 months. Pharmacologic interventions such as NRT were evaluated, as well as interventions to promote the uptake of smoking cessation. This makes comparison with the findings of other reviews more difficult. The fourth of these reviews examined the literature concerning smoking cessation interventions among African Americans.⁷² Methodologically weak, the authors of this review specifically state they did not attempt to include all studies, but have included major studies to provide some indication of the patterns of results and comparisons of African American and Caucasian smokers and quitters. It has been discussed separately in a section on target populations.

Reviews of adult smoking cessation single interventions

Nine reviews evaluated adult smoking cessation single interventions.^{59, 62–64 66–70, 73} Bains et al. undertook a review of community- or population-based interventions that involved the use of financial incentives or competitions (e.g., vacation to Disney World) to promote participation in community-based smoking cessation programs.⁵⁹ The studies included in this review utilized pre—post-test comparisons, post-test measurements only, or quasi-experimental designs with comparison with state-wide or other population controls. Worksite incentive programs were specifically excluded from this review. However, the use of financial incentives and competitions to increase participation in worksite smoking cessation programs were the focus of another review by Matson et al.⁷⁰

One review by the Cochrane Tobacco Control Group focused primarily on the effectiveness of training healthcare professionals to provide smoking cessation interventions.⁶⁴ This review updated a prior systematic review by Silagy et al.⁷³ Training was generally provided in a group setting (e.g., workshop or tutorial), including lectures, videos, role plays, and discussion. Minimal contact strategies were emphasized. Meta-analysis was not utilized.

The remaining six systematic reviews in this category examined interventions directed at individuals. Two reviews examined healthcare provider advice to stop smoking: one directed to physicians⁶⁷ and one directed to nurses.⁶⁶ Advice was defined as verbal instructions from the healthcare professional with a “stop smoking” message. The review directed to nurses included brief interventions as well as higher-intensity interventions.⁶⁶ A second publication of this review was also identified.⁷⁴

Several reviews addressed some aspect of patient counseling.^{62, 68, 69} Two reviews examined the effectiveness of telephone counseling for smoking cessation.^{68, 69} Stead and Lancaster⁶⁸ conducted a review with meta-analysis of 23 RCTs and quasi-randomized trials in which the primary question evaluated reactive or proactive telephone counseling. Proactive counseling includes studies in which a counselor initiates one or more calls to provide support in making an attempt to quit, whereas reactive counseling is provided by helplines that may offer information, recorded messages, personal counseling, or a mixture of components. Lichtenstein et al.,⁶⁹ undertook a meta-analysis of 13 studies evaluating proactive telephone counseling. A third systematic review with meta-analysis examined RCTs and quasi-randomized trials of more intensive individual behavioral counseling by a smoking cessation counselor.⁶² Counseling was defined broadly, based upon more than 10 minutes of face-to-face contact, rather than as the use of any specific behavioral approach.

The final paper in this category was a systematic review with meta-analysis examining studies in which the primary question addressed self-help interventions for smoking cessation.⁶³ Self-help interventions were any manual or program to be used by individuals to assist a quit attempt not aided by healthcare providers, counselors, or group support. Forty-five studies examined self-help interventions alone or in combination with NRT, additional educational materials or a video, provision of a helpline or telephone contact, or individualized materials.

The types of studies included in these 15 systematic reviews varied considerably. Some focused only on RCTs, whereas others included a variety of other study designs. Not all the systematic reviews used meta-analysis to combine the data. There was also variation in the outcomes assessed. Some reviews focused on process measures such as the proportion of patients in whom smoking cessation is discussed, whereas others looked at outcomes such as smoking cessation rates. Reviews reported point prevalence (percentage of smokers not smoking at time of assessment) and others reported on the proportion who had remained continuously abstinent. The time at which assessments were undertaken also varied (six months vs. 12 months). Some reviews relied on self-report of smoking cessation, while others utilized biochemical testing to confirm this. For these reasons, it was not possible to quantitatively combine the results of the systematic reviews. It also created difficulties in standardizing the reporting of findings of the systematic reviews in this report.

Healthcare provider-directed interventions

1. Physician training

   Three systematic reviews addressed the issue of training physicians to provide smoking cessation interventions.^{61, 64, 71} Lancaster et al.⁶⁴ reported that training health professionals to provide smoking cessation interventions had a measurable effect on professional performance but there was no strong evidence that it changed smoking behavior. They reported that trained health professionals were 1.5 to 2.5 times more likely to counsel patients about smoking than controls. However, six of eight studies found no significant effect of practitioner training on smoking cessation rates. One other study reported sustained abstinence rates at 12 months of 8.8 percent for the trained group vs. 6.1 percent and 4.4 percent for control groups.

   The TFCPS Community Guide⁶¹ evaluated 16 studies of provider education interventions. Determination of smoking status was increased by median of 8 percent (range 0.1 to 35 percent). Ten studies reported on provider delivery of advice to quit smoking. The median increase in provider advice to quit was 2.2 percent (range -5 to 73 percent) in comparison with control practitioners. Only two studies reported on smoking cessation rates. These studies reported a 1.7 percent and 5.3 percent increase in smoking cessation in comparison with controls. This review concluded there is insufficient evidence of the effectiveness of provider education alone as too few studies evaluate the effect on tobacco use cessation. Ritvo et al.⁷¹ concluded that physician training has a modest effect on smoking cessation, but this effect is not quantified.

   PHS clinical practice guideline recommended that all clinicians and clinicians in training be trained in effective tobacco use treatments.⁶⁰ They stated that this recommendation was based on a review of the literature. However, data to support this recommendation were not included.

2. Office system prompts

   Two reviews addressed the issue of office prompts and reminders.^{60, 61} These reviews demonstrate that office reminder systems produce significant increases in tobacco use documentation and physician advice to quit. The PHS guideline⁶⁰ reported that tobacco use identification systems led to a significant increase in delivery of smoking cessation interventions by physicians. Among nine studies included, office systems led to an increase in the use of smoking cessation interventions (odds ratio [OR] 3.1, 95 percent CI 2.2 to 4.2). Smoking cessation rate increased (OR 2.0, 95 percent CI 0.8 to 4.8). However, this difference was not statistically significant.

   The TFCPS Community Guide⁶¹ evaluated seven studies assessing provider reminder systems and concluded that provider reminder systems were effective at increasing delivery of advice to quit. These included efforts to identify patients using tobacco products and efforts to prompt physicians to discuss or advise patients about cessation. Provider reminder systems resulted in improved process measures. These included a median absolute increase in documentation of smoking status of 32.5 percent (range 26 to 57.6 percent) and a median absolute increase of 13 percent (range 7 to 31 percent) in delivery of advice to quit smoking were reported. Only one study evaluated abstinence (4 percent absolute increase).

3. Audit and feedback

   Hopkins et al.⁶¹ reviewed three studies evaluating the effectiveness of provider assessment and feedback with tobacco-using patients. These studies utilized retrospective assessment of provider performance in the identification of tobacco use status, delivery of advice to quit, or a combination of both as an intervention to motivate providers. There was a median absolute increase in provider recognition of patient tobacco use of 21 percent (range 13 to 39 percent). No study evaluated provider advice to quit, or cessation rates. They concluded there was insufficient evidence to recommend provider feedback as an effective intervention to increase uptake of smoking cessation.

Individual-directed interventions

1. Telephone counseling

   Three systematic reviews were identified addressing some aspect of telephone counseling.^{60, 68, 69} They concluded that proactive counseling was an effective intervention for smoking cessation. The PHS clinical practice guideline on smoking cessation examined the effectiveness of proactive telephone counseling as part of a systematic review of psychosocial treatments for tobacco dependence.⁶⁰ A meta-analysis of 26 studies found that proactive counseling significantly increased the smoking cessation rates (OR 1.2, 95 percent CI 1.1 to 1.4) in comparison with no intervention. Estimated abstinence rates were increased from 10.8 to 13.1 percent.

   Lichtenstein et al.⁶⁹ reported the results of a meta-analysis of 13 studies of proactive telephone counseling as part of a systematic review of telephone counseling services. Outcome data was reported as both short (three to eight months) and long-term (12 to 18 months) self-reported abstinence rates. Two studies were eliminated from the short-term meta-analysis because of significant heterogeneity. No sensitivity analysis was performed. There was a significant increase in short-term abstinence rates (OR 1.34, 95 percent CI 1.19 to 1.51). Two additional studies were removed from the analysis of long-term followup because of heterogeneity. The effect size appeared to decrease with longer-term followup (OR 1.20, 95 percent CI 1.06 to 1.37). No estimate of the absolute change in smoking rates was provided.

   Stead and Lancaster⁶⁸ undertook a systematic review with meta-analysis of RCTs and quasi-randomized trials of proactive and reactive telephone counseling. The authors concluded that proactive telephone counseling can be effective but the effect size is uncertain. There was significant heterogeneity among 10 trials of proactive counseling versus minimal intervention so the data were not pooled. In two trials the quit rate was lower in the intervention group than the control. Three trials observed significantly increased quit rates in comparison with controls (absolute increased quit rates 2 percent, 3.4 percent, 8 percent). Three trials observed increased quit rates between 1 to 2 percent, but these were not statistically significant, and two additional trials observed nearly identical quit rates. Four trials evaluated the addition of telephone counseling to a face-to-face intervention. There was no evidence that this increased quit rates (OR 1.08, 95 percent CI 0.87 to 1.34). Similarly, there was no evidence that the addition of telephone counseling to NRT improved quit rates (OR 1.08, 95 percent CI 0.82 to 1.43). Three trials evaluated the provision of a helpline to self-help materials. One trial observed a significant increase in quit rates from 4 to 6.6 percent. Two other trials showed no benefit. The review concluded there was uncertainty about the incremental benefit of telephone counseling in combination with a face-to-face intervention.

2. Healthcare provider advice to individuals to quit smoking

   Five reviews evaluated the impact of healthcare provider advice to quit smoking.^{58, 60, 66, 67, 71} These reviews provide convincing evidence of the importance of health professional advice to stop smoking. At a minimum, health professionals should aim to advise all patients who smoke to stop. The PHS clinical practice guideline on tobacco use and dependence identified seven trials evaluating brief advice (modal length < 3 minutes) by a physician to quit smoking compared with controls.⁶⁰ Brief advice was associated with a significant increase in abstinence rates (OR 1.3, 95 percent CI 1.1 to1.6). This equates to an estimated increase in smoking abstinence from 7.9 to 10.2 percent. Separate analyses examined intensity of contact (43 studies), total contact time (35 studies), and number of contact treatment sessions (45 studies) strengthening the dose-response relationship. Abstinence rates increased significantly with intensity of contact: (OR 1.3, 95 percent CI 1.01 to1.6) for minimal counseling of less than three minutes, (OR 1.6, 95 percent CI 1.2 to 2.0) for low-intensity counseling of three to10 minutes, (OR 2.3, 95 percent CI 2.0 to 2.7) for high-intensity counseling of more than 10 minutes. Abstinence rates also increased significantly with total amount of contact time: (OR 1.4, 95 percent CI 1.1 to1.8) 1 to 3 minutes, (OR 1.9, 95 percent CI 1.5 to 2.3) 4 to 30 minutes, (OR 3.0, 95 percent CI 2.3 to 3.8) 31 to 90 minutes, (OR 3.2, 95 percent CI 2.3 to 4.6), and (OR 2.8, 95 percent CI 2.0 to 3.9) >300 minutes, and with increasing number of treatment sessions: (OR 1.4, 95 percent CI 1.1 to 1.7) 2 to 3 sessions, (OR 1.9, 95 percent CI 1.6 to 2.2) 4 to 8 sessions, and (OR 2.3, 95 percent CI 2.1 to 3.0) more than 8 sessions.

   Two reviews were undertaken by the Cochrane Tobacco Control Group.^{66, 67} Both reviews reported pooled ORs, but not abstinence rates. Silagy and Stead⁶⁷ undertook a systematic review with meta-analysis of RCTs evaluating physician advice to stop smoking. Sixteen studies evaluated brief advice to stop smoking (single consultation < 20 minutes duration). Brief advice was associated with a significant increase in smoking cessation compared with no advice (OR 1.69, 95 percent CI 1.45 to 1.98). They estimated that this equated to a 2.5 percent increase in absolute smoking cessation rates. Trials using direct comparisons of more intensive versus brief advice showed a benefit from more intensive advice (OR 1.44, 95 percent CI 1.23 to 1.68), although there was evidence of heterogeneity among trials and the results were not robust to sensitivity analysis. The effect size was greater for patients at high risk of smoking-related diseases and if followup visits were scheduled. Using indirect comparisons, there was insufficient evidence to show that more intensive interventions were significantly more effective than brief interventions.

   Rice et al.⁶⁶ examined the effectiveness of RCTs evaluating nursing interventions for smoking cessation. Interventions were grouped into low- (advice to stop smoking) and high-intensity (initial contact more than 10 minutes). Sixteen trials compared nurse interventions with usual care. Smokers offered advice by a nursing professional were significantly more likely to quit smoking than those who received usual care (OR 1.50, 95 percent CI 1.29 to 1.73). High- intensity interventions appeared no more effective than low-intensity interventions.

   Two additional systematic reviews examine the topic of healthcare provider advice to stop smoking.^{58, 71} Ashenden et al.⁵⁸ identified 23 studies of physician advice to quit smoking compared with no advice. Two studies were excluded because of heterogeneity. Brief advice was associated with an increase in abstinence rates compared with a no-advice control group (OR 1.32, 95 percent CI 1.18 to 1.48). The odds of quitting were greater with more intensive interventions than with brief interventions (OR 1.46 vs. 1.27). However, no statistical comparisons were made. Ashenden et al.⁵⁸ concluded it was necessary to provide advice to 35 smokers to produce one quitter. Meta-analysis was not performed in the systematic review by Ritvo et al.⁷¹ They reached a similar conclusion that physician advice is an effective intervention to promote smoking cessation.

3. Self-help and educational materials

   Two systematic reviews examined patient self-help materials.^{60, 63} Lancaster and Stead⁶³ included forty-five trials in their review of self-help interventions. Pooled ORs were reported in the absence of absolute smoking cessation rates. In comparison with no intervention, self-help produced a small but significant increase in abstinence rates (OR 1.23, 95 percent CI 1.02 to 1.49). However, when self-help materials were added on to other interventions such as brief contact (self-help materials given directly to individuals but not in context of formal advice to stop smoking), healthcare provider advice to quit, or nicotine replacement therapy, there was no evidence of increased quit rates. Enhancements to self-help materials, such as additional written materials or a video, did not significantly increase quit rates. The individual tailoring of self-help materials on the basis of responses to baseline questionnaires or the addition of proactive telephone counseling to self-help were associated with increased quit rates. Similarly, Fiore et al.⁶⁰ reported a small increase in abstinence rates from self-help strategies compared to a control group (OR 1.2, 95 percent CI 1.02 to 1.3).

   Mullen et al.⁶⁵ conducted a systematic review of patient education and counseling for three groups of preventive behaviors. The analysis grouped studies concerning smoking and alcohol together. The overall weighted effect size for smoking and alcohol was 0.61 (95 percent CI 0.45 to 0.77) standard deviation units. It is not clear from the review what outcome measure this refers to, or the time point at which it was assessed.

4. Social network

   The PHS clinical practice guideline for tobacco use and dependence examined social support as part of a review of counseling and behavioral therapies.⁶⁰ They reported three main aspects of support: the training of patients in support solicitation skills; the prompting of support seeking; and clinician-arranged outside support. In comparison with no counseling, studies evaluating social supports were associated with modest increases in abstinence rates (OR 1.5, 95 percent CI 1.1 to 2.1).

5. Financial incentives

   Three systematic reviews evaluated the effectiveness of financial incentives or competitions on quit rates.^{59, 61, 70} Bains et al.⁵⁹ undertook a systematic review of the impact of financial incentives in population-based smoking cessation programs. Seventeen studies of various methodological designs were included. There is no convincing evidence that incentive programs influenced participation or quit rates.

   Matson et al.,⁷⁰ evaluated the impact of financial incentives and competitions on participation and quit rates in worksite smoking cessation programs. Fifteen studies were identified. Three of eight studies with appropriate controls demonstrated that incentives and competitions increased participation in the worksite program. Five studies showed increased quit rates, but the magnitude of this effect was not quantified. One study found that competitions increased quit rates in addition to incentives.

   The TFCPS community guide⁶¹ included only one study on smoking cessation contests. This showed a 13 percent participation in the contest and a 3.3 percent cessation rate at six months compared with a group of smokers given general health education materials. The community guide concluded that there was insufficient evidence to assess the effectiveness of cessation contests given that there was only one study.

6. Other types of interventions

   Several reviews have examined the effect of patient counseling to stop smoking.^{60, 62, 71} There is evidence to recommend the use of trained smoking cessation counselors and individual behavioral therapies as effective smoking cessation interventions. Lancaster and Stead⁶² undertook a systematic review of smoking cessation counseling from a trained smoking cessation counselor not involved in routine clinical care. Individual counseling significantly increased the odds of quitting (OR 1.55, 95 percent CI 1.27 to1.90). This result was robust to a sensitivity analysis. There was no additional benefit observed from more intensive counseling in comparison with brief counseling.

   The PHS clinical practice guideline⁶⁰ examined counseling and behavioral therapies. They found that abstinence rates were significantly improved by five strategies in comparison with abstinence rates in a no-counseling group: (OR 1.5, 95 percent CI 1.3 to 1.8) general problem solving skills; (OR 1.3, 95 percent CI 1.1 to 1.6) intra-treatment support (providing support during a smokers direct contact with a clinician); (OR 1.5, 95 percent CI 1.1 to 2.1) extra-treatment support (intervening to increase social support); (OR 2.0, 95 percent CI 1.1 to 3.5) rapid smoking; and (OR 1.7, 95 percent CI 1.04 to 2.8) other aversive smoking procedures. The PHS guideline did not directly address the issue of brief versus more intensive behavioral counseling.⁶⁰ Ritvo et al. also concluded that cognitive-behavioral counseling was one of three key strategies in physician- assisted smoking cessation strategies, although there was no measure of effect size.⁷¹

Access-enhancing interventions

1. Financial barriers

   The TFCPS community guide identified five studies evaluating the effectiveness of reducing out-of-pocket expenses for tobacco cessation therapies.⁶¹ All five studies reduced or eliminated patient costs for NRT. All studies observed an increase in the use of cessation therapies (range 6.5 to 28 percent). Four of the studies observed increased smoking cessation (2 to 11 percent). Therefore, there is evidence to recommend reducing out-of-pocket expenses.

2. Media education campaigns

   The TFCPS community guide examined the evidence of effectiveness of media campaigns.⁶¹ Fifteen studies were included in the systematic review. All of the studies evaluated a media campaign co-coordinated or concurrent with other interventions. There is strong evidence that media campaigns associated with other interventions are effective in reducing tobacco consumption. The median increase in cessation rate was 2.2 percent (range -2 to 35 percent). Three studies evaluated statewide tobacco consumption in response to media campaigns. Per capita consumption declined between 9 to 20.4 packets per capita per year (relative decrease 9.8 to 17.5 percent).

3. Policy level interventions

   One review on a policy level intervention to reduce tobacco consumption was identified.⁶¹ Seventeen studies were included examining the impact of increasing the unit price of tobacco products. There was strong evidence that increasing price through taxation would reduce tobacco consumption. The median estimate from these studies was that a 10 percent increase in price would result in a 4.1 percent decrease in population consumption (range 2.7 to 7.6 percent).

4. Multicomponent interventions

   The TFCPS community guide⁶¹ reported a review of multi-component interventions involving provider reminders and provider education with or without patient education. Twenty studies evaluated a provider reminder system and provider education. There was a median 20 percent (range 5.2 to 60 percent) increase in provider advice to quit and 4.7 percent (range -1 to 25.9 percent) increase in abstinence rates with followup between five weeks to 12 months (median 10 months). A subset analysis was performed of thirteen studies evaluating combinations of provider reminders and education, plus patient education. The median provider advice to quit was 22 percent, with a median cessation rate of 5.7 percent. The report recommended the use of these multicomponent interventions.

   The TFCPS Community Guide⁶¹ also conducted a review of multicomponent interventions that included patient telephone support. Thirty-two studies were included. In all studies, telephone counseling was combined with additional interventions such as patient education, provider delivered counseling, NRT, or a smoking cessation clinic. Cessation rates from -3.4 to 23 percent (median 2.6 percent) are reported. The report concluded that there was strong evidence that telephone counseling was effective when implemented with other interventions. These conclusions are somewhat discordant with the meta-analysis of telephone counseling from Stead et al.⁶⁸ who found no increase in the odds of quitting in studies of telephone counseling in addition to face-to-face interventions or NRT. These differences may be explained in part by the fact that more studies were included in the TFCPS review and the data were not combined quantitatively. Therefore, there is uncertainty about the incremental benefit of telephone counseling in combination with a face-to-face intervention.

5. Target populations

   One review specifically addressed smoking cessation among African Americans.⁷² Multiple literature sources were searched. However, only major studies were included in the review “to provide some indication of the patterns of results and comparisons with white smokers and quitters”. A wide range of health-related interventions were evaluated in African American church communities. These included less intensive interventions such as smoking cessation counseling and self-help materials, as well as more intensive interventions including sermons on smoking cessation, smoking cessation counseling, and spiritual stop-smoking tapes and guides. There was a trend to greater effectiveness from the more intensive interventions, but this did not reach statistical significance. Quit rates among church attendees appeared to be greater than among the non-attendees (10.6 percent vs. 5.9 percent). In community programs, Pederson et al.⁷² concluded that there was no difference in quit rates between African and Caucasian Americans.

   The PHS clinical practice guideline also examined the evidence for smoking cessation interventions in ethnic minorities.⁶⁰ There was no consistent evidence that specific targeted cessation programs resulted in higher quit rates in these groups than did generic interventions of comparable intensity. A range of interventions were shown to be effective, including NRT, clinician advice, counseling, tailored self-help materials, and telephone counseling. There were differences in smoking prevalence between white and racial/ethnic minorities. In addition, some racial/ethnic groups had inadequate access to primary care. These factors may be more of a barrier to effective smoking cessation interventions.

Reviews of healthy diet interventions as part of preventive healthcare services in family practice

Ashenden et al.⁵⁸ undertook a systematic review of RCTs evaluating the effect of lifestyle advice provided by family practitioners, in changing patient behavior. Lifestyle advice was defined as advice provided in the general practice setting to quit smoking, make dietary changes, reduce alcohol consumption and exercise more. Advice was mainly verbal but could include written material. More intensive advice included additional followup by appointment, telephone, or letter. Data were analyzed separately for each topic; meta-analysis was done only for smoking.

The review by Mullen⁶⁵ included RCTs and uncontrolled studies evaluating patient education and counseling. Education and counseling in this setting were advice on preventive health behavior to apparently healthy individuals seen in a clinical setting in a developed country. The analysis was broken down into three groups: smoking/alcohol, nutrition/weight, and other behaviors.

Glanz⁸⁹ reviewed twenty-five studies regarding medical school curricula and physician's knowledge, attitudes, and practices related to nutritional care, with a focus on prevention of coronary heart disease through cholesterol control.

Reviews of multiple healthy diet interventions

Ammerman et al.⁸⁵ conducted a comprehensive review for the US AHRQ on the efficacy and effectiveness of behavioral interventions in promoting dietary change. In particular, they attempted to find (1) evidence for one intervention, alone or in combination, that is more effective than another in modifying diet to increase fruits and vegetables and reduce fat; (2) evidence for the efficacy of dietary interventions by population subgroup (ethnicity and gender); and (3) conclusions about cost-effectiveness of these types of interventions. A total of 92 studies were included that were RCTs or non-randomized trials conducted in any setting, with any age group except infants, healthy or high-risk populations (non-institutionalized), and any intervention (education, counseling, support groups, classes, etc.), with consumption of fruit, vegetables, or dietary fat as outcomes. Outcomes were reported in meta-analysis, a “difference-in-deltas” approach, and “summary of significant findings” approach.

The purpose of the review by Brunner et al.⁸⁶ was to assess the effectiveness of dietary advice designed for primary prevention of chronic disease. They identified 17 suitable RCTs to include in a meta-analysis. The populations were often within primary care, but were also in specialty clinics or worksites. The dietary interventions included individual advice, phone and mail support, small group meetings, classes and demonstrations, and educational materials. Analysis was reported by dietary fat as a percentage of food energy, serum cholesterol, urinary sodium, and systolic blood pressure.

A review of the effectiveness of community-based interventions to increase fruit and vegetable intake in people four years of age and older found a total of 15 studies.⁸⁷ Only five studies were relevant to adults and these designs were RCTs, cohort analytic studies, and interrupted time series studies. No meta-analysis was done. The interventions consisted of worksite education, formation of community coalitions, and tailored, individual education.

In a series of reviews within one publication, Contento et al.⁸⁸ aimed to discover which elements across effective interventions (if any) were successful, and to make subsequent implications for nutrition education programming, research, and policy. Two-hundred and seventeen articles were included; 117 were on adult healthy diet, but the methods or quality were not reported. Interventions included individual education or counseling, in-service of healthcare professionals and food-industry workers, mass media campaigns, worksite programs, point-of-purchase interventions, and intensive nutrition education programs for low-income families. No meta-analyses were performed.

The final review in this section was specifically concerned with dietary interventions delivered within healthcare settings to women.⁹¹ Forty-five studies were identified, of which 19 were on physical activity alone, 14 were dietary alone (10 RCTs), and 12 combined diet and physical activity (eight RCTs). A mean effect size was calculated for outcomes. This review attempted to look at differences by ethnicity.

Healthcare provider-directed interventions

The provision of interventions to change healthcare provider behavior has, as a final goal, improvement in patient behaviors. Education of providers and system supports, such as computerized reminder systems, or automatic recall of patients, are some of the interventions directed to healthcare providers. Three reviews included at least some primary studies of healthcare provider interventions;^{88, 89, 91} however, none of these indicated the quality of the included primary studies within the report. In the Wilcox et al.⁹¹ review, five of the 14 studies of dietary interventions involved resident-physician or physician education in addition to their usual training; improvements in patient knowledge were found, but without subsequent significant impact on patient weight reduction. The Wilcox et al.⁹¹ review did not analyze comparisons of different types of interventions (e.g., provider- vs. patient-directed); only comparisons of single vs. multiple interventions were reported: diet-only and combined interventions were equally effective in reducing dietary fat (mean effect size 0.13, 95 percent CI 0.08 to 0.18, vs. 0.11, 95 percent CI 0.05 to 0.17, using the Pearson Product Moment Correlation r). Three primary studies that specifically included other ethnic groups were included; the review indicated that interventions that were targeted and tailored could produce significant effects, but that more research was needed.

Glanz and Gilboy⁸⁹ found that attention to nutrition in medical school curricula and to continuing education for practicing physicians had increased modestly, with some improvement in attitudes about dietary interventions. The review raised important awareness of other determinants of physician behavior beyond education: that of reduction of barriers, such as lack of time, payment issues, and co-ordination of care.

The Contento et al.⁸⁸ review considered education of paraprofessionals and professionals. They found that additional educational input of paraprofessionals resulted in increased knowledge in before-after studies. Expectations of program supervisors, positive attitudes towards work, and knowledge of teaching-learning strategies were identified as characteristics of paraprofessionals most closely associated with positive program outcomes. Continuing education of physicians was found to be effective in changing patient dietary behaviors, but only when physician perceptions were altered regarding patients desire for and ability to follow dietary advice, and when reimbursement issues were addressed.

Individual-directed interventions

All nine included reviews presented information about individual-, or consumer-directed interventions such as education, counseling, and healthcare provider advice to individuals or community groups. Ammerman et al.⁸⁵ found 92 primary studies focused on fruits, vegetables, and dietary fat intake in a wide variety of settings, age groups, and populations. Interventions included workplace, community and patient education, mass media campaigns, mailed interventions, peer education, and combined strategies. They concluded that dietary interventions were consistently associated with a decrease in dietary fat and saturated fat (estimated as 7.3 percent reduction in percentage of calories from fat in the intervention groups vs. control), and an increase in fruit and vegetable consumption (0.6 servings per day).

Similarly Brunner et al.⁸⁶ reviewed 17 trials of dietary interventions of at least three months duration in diverse participants (e.g., volunteers, employees, breast cancer patients). Those studies achieved a reduction of 2.5 percent in percentage of calories from fat (95 percent CI -3.9 to1.1).

Mullen et al.'s⁶⁵ review considered any patient education or counseling intervention of preventive behavior with healthy individuals in clinical settings. The review included alcohol, nutrition/weight, and other behaviors. The interventions more often targeted patients with elevated risk; included education, support, some behavioral approaches, and often included more than one contact; and they were less likely to involve self-monitoring, media channels or to include followup longer than 30 days. The overall weighted effect size for nutrition and weight studies in Mullen et al.'s review of 17 included studies was 0.51 standard deviation (SD) units (95 percent CI 0.20 to 0.82); 65 percent of the variance was accounted for in a regression model; followup of at least 30 days, higher score on behavior change support factor, and use of self-monitoring were all significantly associated with higher effects.⁶⁵ The combination of results regarding nutrition and weight probably underestimated the effect size, as weight is so difficult to change and maintaining the loss is even more difficult.

Four reviews reported on worksites as a channel for messages regarding individual behavior change.^{85, 87, 90, 92} Multiple interventions, compared with no intervention or usual employee programs, were common in these primary studies and included peer educators, printed educational materials, cafeteria posters, and food labeling. These interventions were associated with at least short-term increases in fruit and vegetable intake (servings per day) and reduction of dietary fat.

Lifestyle education (smoking, diet, alcohol and exercise) was the target of the review by Ashenden et al.⁵⁸ Ten RCTs were included about healthy diet and they tried to alter fat, fiber, lipid levels, blood pressure, and/or weight reduction. The meta-analysis and the results focused primarily on the smoking results. However, they concluded that brief or intensive advice, often with written materials and support, provided by general practitioners, had a modest and variable effect on lifestyle improvements.⁵⁸

The Wilcox et al. review⁹¹ of all types of interventions in healthcare settings supports Ashenden et al.'s conclusion. In addition, that review was able to do further comparisons and concluded that intervention effects on dietary behaviors were generally larger for samples with a mean age over 50 years and for studies with less than six months followup.

Use of behavioral interventions,^{65, 82, 88, 90} tailored interventions,^{82, 88} multiple interventions,^{65, 82, 87, 88, 90} provision of multiple contacts,^{82, 87, 88} and environmental interventions^{82, 87, 88} were more likely to be effective in producing dietary change. Interactions with food (community kitchens, community gardens, taste testing) were promising interventions to increase fruit and vegetable intake and reduce fat intake.⁸⁵

Media education campaigns

Mass media campaigns were the subject of parts of two of the reviews.^{87, 88} The primary studies included great variation in channels, media and intensity of the campaigns, and compared the outcomes with before-after samples or with outcomes in communities with no intervention. They resulted in increased knowledge and awareness of behaviors to reduce risks, particularly when the campaign was based on the audience planning of the campaign (such as social marketing). Behavior change has been associated with a highly targeted and focused message⁹¹ and was often part of multiple intervention strategies.⁸⁷

Reviews of screening mammography as part of preventive health care services in family practice

Two reviews were found that addressed preventive health care services in family practice.^{94, 102} Balas et al. undertook a systematic review with the use of meta- analysis of RCTs evaluating the effect of prompting physicians to change clinical practice.⁹⁴ Prompting interventions were clinician prompt, alert, or reminder in the study group and no prompt in the control group and included such preventive services as Papanicolaou (Pap) testing, mammography, influenza vaccination, pneumococcal vaccination, tetanus vaccination, and fecal occult blood testing. The analysis was broken down by preventive care topic.

Shea et al. conducted a systematic review of RCTs using meta-analysis to assess the overall effectiveness of computer-based and manual reminder systems in ambulatory settings directed at preventive care.¹⁰² The analysis was broken down into three categories: preventive practice generically and by prevention topic specifically.

Reviews of multiple screening mammography interventions

Nine reviews of multiple mammography screening interventions were identified.^{96–98, 100, 102, 104, 105, 107, 108} Two of the nine systematic reviews used-meta analysis to combine data that addressed the use of multiple interventions (i.e., combined behavioral and cognitive interventions) to provide pooled estimates of effectiveness.^{98, 100}

Bonfill et al. conducted a systematic review of interventions for increasing the participation of women in community breast-screening programs.⁹⁵ The review assessed the use of single interventions (invitations and reminders, education, and home visits) to recruit women as well as a combination of interventions to enhance recruitment (one or more of the above mentioned interventions). The main outcome measure for this overview was attendance for a mammogram among women in the intervention groups.

Yabroff et al.,¹⁰⁷ Mandelblatt et al.,¹⁰⁰ Yabroff et al.,¹⁰⁸ and Sin et al.¹⁰⁴, all conducted systematic reviews addressing the use of multiple interventions to promote the uptake of screening mammography. The results, however, are not presented by intervention type (i.e., mailed reminders, physician prompts, or audit and feedback); rather they are grouped into specific classification categories for the purpose of analysis. Both Yabroff et al.¹⁰⁷ and Mandelblatt et al.¹⁰⁰ used the following intervention classification scheme (1) behavioral interventions, (2) sociological interventions, and (3) cognitive interventions. Sin and St. Leger,¹⁰⁴ however, classified interventions into either (1) person-directed, (2) social-network-directed, or (3) multi-strategy for the purpose of analysis. All reviews included RCTs,^{100, 104, 107} and one included descriptive studies.¹⁰⁴ Yabroff et al.¹⁰⁸ compared broad categories of inreach and outreach strategies and presented results under further classification within these categories as behavioral, sociological, and cognitive, making data extraction by specific intervention difficult.

A sixth review⁹⁸ addressed the effectiveness of combined intervention effects on women with historically lower mammography screening rates. These groups consisted of women who were disproportionately older, poorer, and of racial-ethnic minorities; had lower levels of formal education; and lived in rural areas. Combinations of access-enhancing and system-directed interventions were examined.

The Jepson report⁹⁶ included systematic reviews of multiple different interventions. No meta-analytic techniques were utilized and results are narratively presented by primary study. Results were synthesized broadly by intervention topic for a wide range of preventive services.

Reviews of single screening mammography interventions

Nine reviews evaluated single screening mammography interventions.^{94–96 99, 101–103, 106, 108} Bonfill et al. conducted a systematic review of interventions for increasing the participation of women in community breast screening programs.⁹⁵ Interventions such as letters of invitation, mailed educational materials, and phone calls were all addressed. RCTs were included for this review.

Two reviews addressed interventions to promote screening mammography as part of a larger review of preventive services.^{94, 102} Balas et al. conducted a systematic review of prompting physicians to improve preventive care.⁹⁴ RCTs (n=33) that compared physician prompting in the study group with a control group with no intervention were included. The effects of prompting on selected procedures (i.e., fecal occult blood testing, mammography, Pap testing) were presented separately. Shea et al.¹⁰² assessed the overall effectiveness of using computer- or manual- reminder systems in ambulatory settings directed at preventive care. RCTs or concurrent led control trials in which the control group received no intervention were included for analysis. Sixteen studies were included in the review.

Mandelblatt et al.⁹⁹ also addressed the use of interventions to enhance physician breast cancer screening delivery. In this review physician reminder, other office systems, audit with feedback, and physician education were all addressed.

One additional review addressed the use of mailed patient reminders on mammography screening.¹⁰⁶ Sixteen published studies were included for meta-analysis. Patient reminders included generic vs. tailored and letters with set appointments vs. no appointment.

The review by Yabroff et al.¹⁰⁸ classified interventions broadly under the following headings: behavioral, sociological, and cognitive interventions. Results are presented by these specific category types and not by single intervention (i.e., mailed reminders, physician advice) making it difficult to assess the effect size of a specific intervention.

The Jepson et al. report⁹⁶ and Ratner et al. review¹⁰¹ included systematic reviews of several single interventions. No meta-analytic techniques were utilized for the Jepson et al. report, and results were narratively described by primary study. Results were synthesized broadly by intervention topic for a wide range of preventive services (i.e., mailed reminders seem to be effective overall). The Ratner et al. review incorporated meta-analysis to identify factors that influenced the effectiveness of interventions in increasing women's use of screening mammography programs. Included in this review were interventions such as mailed materials, physician reminders, telephone counseling, and patient education.

Finally, the review by Shekelle et al.¹⁰³ was a systematic review to determine the best strategies for early detection and prevention currently covered by Medicare and to asses interventions designed to improved screening in several prevention topics (mammography, Pap testing, colon cancer screening, immunization etc). Included in this review were interventions such as financial incentives, patient and provider reminders, organization changes, patient and provider education, and feedback.

Healthcare provider-directed interventions

1. Physician Training

   No systematic review exclusively addressing physician training on promoting uptake of screening mammography was identified. However, three systematic reviews, which were all rated highly methodologically, addressed physician training by including primary study data.^{96, 99, 100, 103} The Jepson et al. report (2001)⁹⁶ included four such studies (three RCTs and one controlled trial) that evaluated the impact of educational sessions, printed materials, and educational outreach visits targeted towards health care providers. The review suggests a small increase in the uptake of screening tests in the intervention group when compared with the control group. The report states that relative risks (RRs) were not calculated due to the lack of data. However, one RCT evaluating a day-long education session for eight screening procedures (Pap, mammography, breast self exam, cholesterol screening, etc.) reported that physician education intervention ultimately increased the proportion of women having a mammogram (p<0.01).⁹⁶

   Mandelblatt and Kanetsky concluded that there was a paucity of controlled trials addressing medical education strategies.⁹⁹ Within this review, six controlled trials included educational strategies, two used education as the major intervention, and the remainder included educational components with other concurrent interventions. One community-based study, which the review identified as having sufficient data to calculate confidence intervals, found a significant increase in mammography rates. The difficulty with much of the literature in this area is that reports focusing on educational strategies are limited by the inability to separate the impact of the education component exclusively from other strategies.

   In 1999, Mandelblatt and Yabroff reviewed provider-targeted interventions to increase screening mammography.¹⁰⁰ In this review, the author classified educational sessions as a cognitive intervention and did not report interventions separately. The review concluded that cognitive intervention strategies improved mammography rates by 18.6 percent (95 percent CI, 12.8 to 24.4). All types of interventions targeted at providers were effective in increasing mammography rates (behavioral, sociological, and cognitive).

   The review by Shekelle et al. reported effectiveness of provider education designed to improve the use of mammography screening (OR 2.26, 95 percent CI 1.81 to 2.82). The report found that the intervention with the greatest number of studies was patient reminders, followed by patient education.

2. Office System Prompts

   Ten reviews addressed the issue of office system prompts and reminders to promote uptake of screening mammography.^{94, 96, 97, 99–105} Four reviews combined all office-system prompts (chart reminder, and computer and manual reminders).^{96, 100, 104, 105} Three reviews addressed specific office-system prompts and included computer-generated, non-computerized, front-of-chart reminders, and those that had an alternative delivery method.^{94, 97, 99} Only one review specifically evaluated office-system prompts involving computer-based clinical reminder systems.¹⁰² All of the reviews were consistent in their findings and suggested a positive effect on screening utilization.

   Snell and Beck¹⁰⁵ state generally that office-system interventions increased compliance with cancer screening (d [the average amount of change in standard deviation units achieved by individuals in a treated group vs. the change achieved by members of a control/comparison group for a particular study] +0.1705, 95 percent CI +0.16 to +0.18). The results in this review were not presented separately by specific office-system intervention and also included other screening practices such as Pap smears, fecal occult blood tests, and rectal exams.

   Shea et al conducted a systematic review of RCTs to evaluate computer-based clinical reminder systems in preventive care.¹⁰² The review found that computer-based reminder systems improved breast cancer screening (OR 1.88, 95 percent CI 1.44 to 2.45).

   The overview by Balas et al.⁹⁴ included 33 studies that looked at improving preventive care by prompting physicians. Fourteen studies pertained specifically to mammography and evaluated all types of office-system prompts combined. The results demonstrated the effect of prompting (presented as a rate difference) to be 11.5 percent, (95 percent CI 7.1 to 16.0).

   Kupets and Covens conducted a systematic review of interventions for improving both cervical and breast cancer screening and presented results separately by topic.⁹⁷ The review examined physician-based, patient-based, and the combination of the two strategies. The report concluded that the six studies that included computer-generated reminder systems suggest this to be an effective intervention. The delivery of mammography to patients improved by an absolute rate of 6 to 30 percent. The author noted that no benefit was seen from the use of an information sheet or reminder placed on the front of chart.

   Mandelblatt et al.^{99, 100} evaluated the use of interventions designed to increase physician screening for breast cancer. Provider-targeted interventions included reminder or office system prompts^{99, 100} or other physician-reminder systems.⁹⁹ Results for both reviews were similar. The review concluded that physician-based interventions can be effective in increasing screening use. In particular, behavioral interventions (reminders, office system prompts) increased screening by 13.2 percent (95 percent CI 7.8 to 18.4).¹⁰⁰

   As part of a larger review, Jepson et al.⁹⁶ identified five RCTs that evaluated the effectiveness of physician reminders in increasing uptake. All three reported an effect of the intervention, but one was only a small cluster RCT. Jepson et al. reports that one good-quality RCT reported the mean mammography completion rate was 47.9 percent vs. the control 34.6 percent, which was statistically significant (p value was not reported).

   The review by Sin and St Leger reported findings under the broad categories of person directed, social network directed, and multi-strategy.¹⁰⁴ There was mention of office system prompts; however, the results were reported narratively and did not provide enough data to make definitive conclusions regarding this intervention.

   The review by Shekelle et al.¹⁰³ addressed office-system prompts as part of a larger review of screening services. Office-system prompts specific to mammography were not presented separately; however, the authors reported that office-system prompts, in the form of physician reminders, were shown to be effective and calculated the OR to be 1.59 (95 percent CI 1.36 to 1.86).

3. Audit and Feedback

   No systematic review was identified which exclusively focused on audit and feedback as an intervention to promote the uptake of screening mammography. However, four reviews were identified that addressed the effectiveness of audit and feedback on mammographic screening practices^{96, 97, 99, 103, 105} as part of a larger review.

   Snell and Buck conducted an overview, which combined cervical and breast cancer screening, that showed effect sizes for physician-directed intervention studies (n=4) for audit with feedback alone to be (d +0.2826, 95 percent CI +0.2155 to +0.3498)¹⁰⁵. The results for audit and feedback combined with other physician-directed interventions are reported in the multi-strategy interventions section.

   Mandelblatt and Kanetsky⁹⁹ identified three trials, which demonstrated a 15 to 24 percent increase in rates of mammography. It was reported that long-term effects of audit and feedback were not evaluated in any of the research. Similarly, Kupets and Covens⁹⁷ reported that two studies identified by their search showed an improvement in the delivery of mammograms, with an absolute increase in uptake of 14 to 30 percent.

   The review by Shekelle et al.¹⁰³ addressed feedback as part of a larger review of screening services. The authors reported that feedback was shown to be effective and calculated the OR to be 1.49 (95 percent CI 1.24 to 1.80).

4. Opinion Leaders

   No studies utilizing opinion leaders exclusively to promote the uptake of mammographic screening interventions were identified.

Individual-directed interventions

1. Invitations or reminders

   The most widely investigated interventions to promote the uptake of mammographic screening are patient-directed interventions (mailed invitations, mailed reminders). Twelve reviews addressed invitations and/or reminders to promote the use of mammographic screening.^{95–99, 101–104 106–108} Of the 12 systematic reviews, four^{98, 106–108} combined data to provide results from meta-analysis.

   Only one review identified by our search conducted a meta-analysis that looked specifically at mailed patient reminders for mammographic screening.¹⁰⁶ The review by Wagner evaluated the effectiveness of mailed reminders and included 16 published articles that were pooled for meta-analysis. The review reports that among studies in which controls did not receive any type of reminder, women who received reminders were approximately 50 percent more likely to attend a mammography (OR 1.48, p <0.001). In addition, the author stated that tailored letters were found to be more effective than generic reminders (OR 1.87, p <0.05). In a cost-effectiveness analysis that accompanied the main study, it was found that the cost per women screened ranged from $0.96 to $5.88.

   Sin and St Leger undertook a systematic review of studies that evaluated invitations or reminders under the broad categorization of “person-directed interventions” and used mammographic screening rates as the outcome measure.¹⁰⁴ Of the 20 studies focusing on the person-directed interventions that were included in the review, 14 were randomized trials. Appointments on the invitation letter increased uptake of screening mammography (86 percent) compared with open-ended invitations (76 percent). No additional benefit was observed from the addition of a letter from a general practitioner.

   Yabroff et al. conducted a systematic overview of patient-directed interventions to promote mammography use.¹⁰⁷ They found that 13.2 percent more women who received a behavioral intervention (i.e., telephone or mailed reminder) attended mammography compared with usual care.

   Bonfill et al. reported that a letter of invitation inviting women into a community breast-screening program was one of the most effective interventions (OR 1.66, 95 percent CI 1.43 to 1.92) in order of effect, of the interventions they examined⁹⁵.

   Kupets and Covens⁹⁷ found conflicting results for the four studies identified in their review. The overview did not exclusively focus on mailed invitations or reminders; rather it included four studies that addressed this intervention. Three of the four studies identified in the overview did not show a statistically significant improvement for breast cancer screening. In fact, one study produced a negative effect of the intervention on screening in the study arm of 10 percent, although not statistically significant. One study showed a 10 percent increase in breast screening, indicating a number needed to intervene (NNI) of 10 patients, meaning one out of every 10 patients will respond to a screening reminder.

   Jepson et al.⁹⁶ included 57 studies in a qualitative systematic review that examined invitations or reminders for several screening topics. The results are presented across screening topics; however, specifically for mammography, some evidence was reported for the effectiveness of reminders for mammograms. Studies included in this report showed conflicting evidence. The Jepson et al. report also provided information narratively for telephone reminders vs. mailed reminders or physician advice.

   Like the Jepson et al. review, the Shekelle et al. review included patient reminders as an intervention as part of a larger review of screening topics. The authors reported an OR of 2.57 (95 percent CI 2.22 to 2.98).

2. Telephone counseling

   No systematic reviews were identified that addressed the use of telephone counseling exclusively for the uptake of mammographic screening. However, several reviews addressed this intervention as a component of a larger review or as part of a multicomponent intervention, and data was not extractable for this review.

3. Health care provider advice

   One review addressed the use of health care provider advice⁹⁶ and uptake of mammographic screening; however, the results were not provided separately from other interventions to promote uptake and were not extractable for this review.

4. Educational materials

   Seven studies addressed the use of educational materials as an intervention to promote the uptake of mammographic screening.^{95, 96, 99, 101, 103, 104, 108} Three of the six studies identified by our search classified educational materials into a broader category of cognitive interventions (i.e., individually-tailored education as a component of a letter),¹⁰⁸ or patient-directed interventions,¹⁰⁴ or they combined the results with reminders,⁹⁹ making it difficult to summarize results for the effectiveness of educational materials alone as an intervention to promote uptake.

   The Health Technology Assessment Report by Jepson et al. in 2000⁹⁶ did not focus exclusively on the use of educational materials to promote uptake. Within this report nine studies were identified. The report stated that there was no effect of printed materials versus an active control group in two RCTs in which RRs could be calculated. Jepson et al. also stated that four of five RCTs showed no increase in mammography uptake from educational telephone calls when compared with a control group.

   Bonfill et al.'s review⁹⁵ identified only one study that compared the effects of sending educational materials with no intervention as part of a much larger review. It was reported that there was evidence that mailed educational materials was an effective strategy (OR 2.81, 95 percent CI 1.96 to 4.02) for recruiting women into a community breast-screening program. The results from the review by Ratner et al.¹⁰¹ were not presented individually by intervention topic. Rather, the individual studies were listed narratively in a table format and did not provide enough data to make conclusions regarding the effectiveness of mailed patient education as an intervention to promote the uptake of screening mammography. Within this review, the PRECEDE model was used as a framework to make distinctions among various interventions. These models revealed that more recent studies (those conducted from 1990 to 1996) were associated with higher screening rates, and those designed to target older women (minimum age 50 to 65 years) and set in clinics exhibited smaller screening rates.

   The Shekelle et al. review¹⁰³ found that patient education interventions were effective and reported an OR of 1.31 (95 percent CI 1.31 to1.52). The descriptions of studies are reported narratively and overall conclusions are done across screening topics.

5. Social network

   Of the 15 reviews included in the mammography chapter, only one review addressed and reported social network findings separately with some analysis.¹⁰⁷

   Yabroff and Mandelblatt identified nine studies that pertain to social-network interventions, however, they classified several types of social-network interventions (i.e., community peers, friends, lay health advisors or media representations) loosely under the heading “sociologic interventions” to increase mammographic screening. These patient-directed sociological interventions were grouped for analysis and showed an improved utilization of mammographic screening by 12.6 percent (95 percent CI 7.4 to 17.9).

   Within the Sin and St Leger review, social-network-directed interventions are addressed, but findings are presented in a narrative overview of the included studies and provides no extractable data.¹⁰⁴

6. Other types of inventories

   The Shekelle et al. review addressed organizational change as an intervention to promote screening mammography.¹⁰³ They reported that organizational change was consistently one of the most (or the most effective) interventions at increasing the use of preventive services. The OR for improving mammography was 2.26 (95 percent CI 1.81 to 2.83). Organizational change was the most heterogeneous intervention and was often combined with reminder letters.

   The overview conducted by Bonfill et al.⁹⁵ identified two studies evaluating the use of a home visits to promote the uptake of mammographic screening. No statistically significant differences between the use of home visit and control (no intervention) were found.

   The use of a patient-initiated touch-sensitive computer system was reviewed by Kupets and Covens⁹⁷ 2001. Only one study was included in the report. A computer was placed in the waiting area of the physician's office that was accessible to all patients. The review stated that this intervention was effective for improving mammography with an absolute increase of 9 percent, with an NNI of 11 patients.

Access-enhancing interventions

1. Financial barriers

   Four reviews were identified that addressed some aspect of removal of financial barriers to promote uptake of breast cancer screening,^{96, 98, 103, 107} Legler et al. undertook a systematic review to determine which types of mammography-enhancing interventions were most effective for diverse populations.⁹⁸ The results were presented under a broad topic of access-enhancing strategies, which included removal of financial barriers, mobile vans, vouchers, etc. Results from this meta-analysis showed the strongest categories of mammography-enhancing interventions were access-enhancing interventions (18.9 percent increase in mammography use). The author states that the most impressive effects were not for single-category approaches but for combinations of interventions with an effect size of 26.9 percent (95 percent CI 9.9 to 43.9; 9 studies) for the combination of access-enhancing and individual-directed interventions and 19.6 percent (95 percent CI 8.2 to 30.6; 5 studies) for the combination of access-enhancing and system-directed interventions⁹⁸.

   Yabroff and Mandelblatt found two patient-targeted interventions that used financial incentives to try to increase the uptake of mammographic screening. Both studies showed an increase in mammography use, but a meta-analysis was not conducted due to the small number of studies available.

   In the overview conducted by Jepson et al., they identified three studies that focused on the removal of financial barriers and reported the results narratively by study (free or reduced-cost tests, transportation, or postage). Of the three studies, two were RCTs and one was a quasi-RCT. The author stated that one well-designed RCT was identified that examined the use of free screening by voucher, and showed it was very effective in increasing mammographic screening (RR 4.28, 95 percent CI 1.91 to 9.60). The other two studies also showed a significant effect of the intervention and were targeted at minority women; no other information was provided.

   The overview by Shekelle et al. includes meta-regression analysis for financial incentives as part of a much larger review of preventive services. The authors reported that this intervention was effective in promoting screening mammography with an OR of 3.57 (95 percent CI 2.36 to 5.40).

2. Access barriers

   The findings for access barriers are summarized under the removal of financial barriers heading. However, one review¹⁰⁴ highlighted one small before-and-after study that suggested that bus transport for Asian women (unpublished data) from the health center to the screening center could increase uptake (46 percent before intervention compared with 73 percent after intervention).

3. Media campaigns

   Three systematic reviews that addressed the use of media campaigns to promote the uptake of breast cancer screening were identified.^{96, 98, 107} Two of the three reviews provided results separately by intervention and estimated the effect size of the intervention.^{96, 107} The systematic review by Legler et al included six studies that were specific media campaigns used to promote breast cancer screening.⁹⁸ They reported the pooled incremental benefit for media campaigns was 1.3 (95 percent CI 1.0 to 1.8).

   The overview conducted by Jepson et al. identified one RCT that assessed the effectiveness of mass media interventions compared with community interventions for increasing uptake of mammography and found that there was a significantly lower uptake of mammography by women in the media-promotion towns when compared with the community intervention towns (p<0.001). No details of baseline estimates were provided. They reported a controlled trial that compared a mass media campaign vs. a control to increase breast and cervical screening practices. The intervention had no effect on patients being up-to-date on any of the tests, but did increase knowledge of screening tests and the intention to have a mammographic screening test.

4. Policy level interventions

   No studies that addressed the use of policy level interventions to promote the uptake of mammographic screening exclusively were identified.

5. Multi-strategy interventions

   Eight overviews that addressed the use of multi-strategy interventions to promote the uptake of mammographic screening were identified.^{96–98, 100, 102–105} Of the eight reviews, three included meta-analysis.^{98, 100, 103} Combined interventions appeared to have an increased effect on increasing uptake of mammographic screening.

   Shea et al.¹⁰² addressed the use of combining both manual and computer- generated reminders to improve physician delivery of breast cancer screening. The review included sixteen RCTs of which eleven were exclusively focused on mammography. When comparing the use of manual reminders to the combination of computer-generated and manual reminders, the adjusted OR was 1.42 (95 percent CI 1.02 to 1.97, p=0.04) when all preventive categories were combined for analysis. The review states that these findings suggest an additional benefit from combining the two interventions compared with manual reminders alone. The results for mammography were not presented separately.

   Mandelblatt and Yabroff¹⁰⁰ found that in interventions that combined cognitive (i.e., audit and feedback and educational sessions and materials) and behavioral strategies (i.e., reminder or office-system prompt) to reach providers, the combined effect was a 21.0 percent increase in mammography rates (95 percent CI 8.8 to 33.6) in contrast to usual care. When cognitive and behavioral strategies were combined and targeted to both patients and providers within communities, the analysis showed no improvement (1.1 percent increase, 95 percent CI -6.8 to 9.0).

   Legler et al.⁹⁸ also addressed the combination of interventions in the same fashion as Mandelblatt and Yabroff. The interventions were categorized according to Rimer's intervention typology (See Table 1, Chapter 1). The review stated that the strongest combination of approaches used were access-enhancing and individual-directed strategies and resulted in an estimated 27 percent increase in absolute mammography use (95 percent CI 9.9 to 43.9) and found that the combination of access-enhancing and system-directed combination showed a 20 percent increase (95 percent CI 8.2 to 30.6). Similarly, Snell and Buck¹⁰⁵ reported that greater success was found for interventions targeting the physician both during and outside the patient visit (d +0.1222 during visit, d +0.1849, both d +0.3375). Screening behavior improved when the physicians were the target of more than one, but not more than three interventions (d +0.1360, d +0.2495, d +0.6829, d -0.0058).

   Sin and St Leger¹⁰⁴ conducted a systematic review of interventions to increase breast screening uptake. Studies were included if uptake was the outcome measure of the intervention and if relevant to the United Kingdom (UK) screening program. Interventions were broadly categorized as “social-network-directed”, “person-directed”, or “multi-strategy”. Only one multi-strategy intervention was identified by their search. The use of clerical help to check addresses of non-attendees and a reminder letter in 93 inner city practices was evaluated. Results showed a subsequent uptake of 58.5 percent compared with 53.8 percent before the study. The author stated that the findings were circumstantial, and any real increase probably had a limited role in inner-city practices.¹⁰⁴

   The review by Kupets and Covens⁹⁷ addressed the use of multi-strategy interventions for mammography and cervical cancer screening, specifically audit and feedback combined with computer-generated reminders and physician computer-generated reminders combined with patient-reminder cards. One study for the use of audit and feedback combined with computer-generated reminders was identified by the review. The report stated that there was no difference in the delivery of breast screening between the arm of a physician-reminder letter alone and the combination of audit and feedback and reminder. Four studies looking at the combined effects of computer reminders for physicians and patient-reminder cards showed mixed results within the review. Two studies indicated an increased effect of the combined interventions for screening mammography (15 to 20 percent), while the other two studies showed no additive effect with the use of multiple interventions on screening practices.

   Jepson et al.⁹⁶ summarized combined interventions aimed at physicians and/or patients for several preventive services. For mammography, three studies were identified that evaluated the effect of physician reminders combined with individual letters, and all of the studies reported a statistically significant effect of the intervention when compared with control groups. Two controlled trials compared audit and feedback aimed at the physician with education and a request form for mammography. No statistically significant difference was seen in uptake rates between the intervention groups as compared with controls (45.9 to 49 percent vs. 47 to 56 percent).

   Finally, the Shekelle et al. review¹⁰³ presented narratively some of the studies that looked at multi-strategy interventions and based conclusions on multi-strategy interventions across several screening topics. For this reason, we could not extract enough data from the review to present results for mammography specifically.

Reviews of interventions to promote uptake of cervical cancer screening

Several systematic reviews have been undertaken to determine which interventions are effective to promote the uptake and/or delivery of cervical cancer screening (Summary Table 7). A comprehensive systematic review by Jepson et al.⁹⁶ (2001) of the determinants of screening uptake and interventions for increasing uptake was identified. The review included RCTs, controlled trials, cohort studies, or case-control studies where there was a prospective time barrier between the measurement of determinants and the uptake of screening. They identified 190 studies that met inclusion criteria of which 130 were RCTs. For the purposes of this report, we consider only those studies included in the analysis of cervical cancer screening (n=12, eight of which were RCTs). Two systematic reviews addressed the interventions under the general headings of patient- vs. physician- directed interventions and present the results by grouping rather than by each specific intervention (i.e., reminder letter or telephone call). The remainder of the reviews addressed specific intervention topics. Three looked specifically at the effectiveness of office-system prompts (general practitioner [GP] reminders) on cervical cancer screening.^{102, 113, 118} Within this subset, one addressed computer-generated reminders exclusively¹⁰². Two other overviews addressd the effectiveness of reminders aimed at both GPs and patients ¹¹⁴ or mailed reminders to patients.¹¹⁵ Finally, the review by Shekelle et al.¹⁰³ is a systematic review to determine the best strategies for early detection and prevention currently covered by Medicare and to assess interventions designed to improved screening in several prevention topics (mammography, Pap testing, colon cancer screening, immunization, etc.). This review included interventions such as financial incentives, patient and provider reminders, organization changes, patient and provider education, and feedback.

Reviews of single interventions to promote the uptake of cervical cancer screening

The majority of evidence for interventions aimed at increasing cervical cancer screening uptake was in the intervention category of office-system prompts. The Austin,¹¹³ Shea,⁹⁴ and Balas¹⁰² systematic reviews addressed the effect of physician prompts for increasing cervical cancer screening as part of a larger systematic review addressing preventive healthcare services. Two of the reviews used meta-analytic techniques^{102, 113} and all of them included only RCTs.

Tseng et al.¹¹⁵ undertook a systematic review, with meta-analysis, using RCTs that involved the use of patient letter reminders to promote cervical cancer screening.¹¹⁵ A total of 10 randomized trials were included for meta-analysis. A second review also addressed the effectiveness of patient reminders¹¹⁴ and included physician reminders as well. This review also included a total of 10 studies in the meta-analysis, with two studies on reminders for GPs alone, four studies on patient reminders, and four dealing with a combination of both.

Reviews of multiple interventions to promote uptake of cervical cancer screening

A comprehensive report by Jepson et al.⁹⁶ on the determinants of screening uptake and interventions for increasing uptake was identified. This report included a broad spectrum of interventions, including healthcare provider-directed interventions (office-system prompts such as a chart reminder and physician training), individual-directed (mailed reminders or invitations to patients for a mammogram, educational materials teaching patients about the benefits of mammography, and healthcare provider advice) and multi-strategy interventions (combination of two or more interventions). The report included 190 articles of which 130 were RCTs. Twelve studies related specifically to cervical cancer screening, and eight were RCTs. No meta-analysis was conducted.

Healthcare provider-directed interventions

1. Physician training

   Three systematic reviews address the use of physician training to promote physician uptake of cervical cancer screening.^{96, 103, 105} The Jepson et al. review⁹⁶ identified four studies that addressed educational interventions to promote the uptake of screening tests. The review reported that two studies addressed Pap testing specifically. One evaluated a one-day seminar, four followup bulletins during the following year, and notes on Pap smear techniques. Only 43 percent of physicians randomized to the intervention attended the seminar, and the average number of Pap smears performed per practice was 40.5 percent in the intervention group and 46.1 percent in the control group. The review states that one other controlled trial reported that an educational outreach visit and educational session by a medical doctor resulted in an increase in uptake of cervical cancer screening (7 percent compared with 2.9 percent in the control group).

   Within the Snell and Buck¹⁰⁵ review, two controlled studies were found that a community intervention (which involved a mass media component) in combination with a GP workshop or educational session was effective in increasing uptake of Pap smears. Although this review addressed physician training, results were presented as a multi-component strategy. The results are outlined in the multi-strategy section of this chapter.

   Shekelle et al.¹⁰³ reviewed intervention strategies for early detection and prevention currently covered by Medicare to assess interventions and found that personalized reminders (which was the intervention with the greatest number of studies) are more effective than generic ones. The effectiveness of interventions to improve the use of clinical preventive and cervical cancer screening were: patient financial incentives OR 3.12 (95% CI 2.62–3.72); patient reminder OR 1.84(95% CI 1.67–2.02); organizational change OR 2.65 (95% CI 2.26–3.12); provider education OR 1.59 (95%CI 1.29–1.97); provider reminder OR 1.40 (95%CI 1.27–1.54); feedback OR 1.12 (95% CI 0.97–1.30), and patient education OR 1.53 (95% CI 1.30–1.82). This review also addressed interventions designed to approve influenza and pneumoicoccal immunization rates, mammography rates, cervical smear cytology (Pap Test), and colon cancer screening.

2. Office-system prompts

   The most abundant information regarding the effectiveness of interventions to promote uptake of cervical cancer screening is available in the area of office systems (i.e., prompts in the form of chart reminders directed at healthcare professionals). In total, eight systematic reviews addressed the use of office systems to promote uptake of cervical cancer screening.^{94, 96, 97, 102, 103, 105, 113, 114} Of the eight reviews, four used meta-analysis.^{94, 103, 113, 114}

   In the reviews by Kupets and Covens⁹⁷ and Shea et al.,¹⁰² computer- generated reminders (either a computer generated reminder placed on the front of the chart compared with control arm of no intervention or a manual reminder placed on the front of the chart) were shown to be effective. The Kupets and Covens⁹⁷ review identified several studies addressing office-system prompts (i.e., computer-generated reminders, audit and feedback, manual reminder placed on chart) and determined that computer-generated reminders to physicians were proven effective for improving delivery of preventive healthcare to patients. The review stated that of the six studies included in the review, three studies showed significant improvements in cervical cancer screening (9 to 30 percent) with an NNI of 3 to 10 physicians. Conversely, Shea et al.¹⁰² found that computer-generated reminders were not effective in increasing the delivery of cervical cancer screening (OR 1.15, 95 percent CI 0.89 to 1.49) when compared with results of six other forms of preventive care examined within the review (i.e., colorectal cancer screening, cardiovascular risk reduction, and breast cancer screening).

   Three other reviews specifically addressed the use of GP reminders/prompts to promote the delivery of Pap screening among their patients. In all three of the reviews, the use of reminders/prompts significantly increased cervical cancer screening practice. Pirkis et al.¹¹⁴ reported that the women whose GPs had been prompted to remind them to have a Pap test were significantly more likely to do so than were control women (typical risk difference [TRD] 6.6 percent, 95 percent CI 5.2 to 8.0). Balas et al.⁹⁴ also found a significant increase in preventive care performance when prompting of physicians was utilized. It was determined that prompting can significantly increase Pap smear delivery by up to 18.3 percent (95 percent CI 11.6 to 25.1); this too was echoed by Austin et al.¹¹³ who found the increase to be significant (OR 1.18, 95 percent CI 1.02 to 1.34). The review by Shekelle et al.¹⁰³ addressed office-system prompts as part of a larger review of screening services. Office-system prompts specific to cervical cancer screening were not presented separately; however, the authors reported that office-system prompts, in the form of physician reminders, were shown to be effective and calculated the OR to be 1.40 (95 percent CI 1.27 to 1.54).

3. Audit and feedback

   Three reviews were identified by our search that specifically addressed the use of audit and feedback to promote the delivery of cervical cancer screening by physicians.^{103, 105, 114} One of the reviews¹⁰⁵ only addressed audit and feedback as a part of a multicomponent intervention and is referred to in the multicomponent interventions section of this chapter. The only review identified that included information regarding the use of audit and feedback was conducted by Kupets and Covens.⁹⁷ The review included 14 RCTs and all were related to cervical cancer screening (also tagged with mammography). Of the 14 studies, two specifically addressed the use of audit and feedback to promote delivery of Pap smears. The results indicate that when comparing a control arm of no intervention with audit and feedback, neither study showed improvement in cervical cancer screening rates. The review by Shekelle et al.¹⁰³ addressed feedback as part of a larger review of screening services. The authors reported that feedback was shown to be effective and calculated the OR to be 1.12 (95 percent CI 0.97 to 1.30).

Individual-directed interventions

1. Invitations or reminders

   Four systematic reviews identified were conducted to determine the effectiveness of the use of invitations or reminders directed at patients or individuals to promote the uptake of cervical cancer screening^{96, 103, 114, 115} of which three incorporated meta-analysis.^{103, 114, 115}

   In the review conducted by Tseng et al.,¹¹⁵ a meta-analysis was performed on 10 articles that specifically investigated the use of mailed letter reminders on cervical cancer screening. They reported patient reminder letters in the form of mailed letters increased the rate of cervical cancer screening. The authors reported that most notable within the results was that those studies evaluating lower socioeconomic groups had a smaller response (OR 1.16, 95 percent CI 0.99 to 1.35) than those studies using mixed populations (OR 2.02, 95 percent CI 1.79 to 2.28).

   Similarly, Pirkis et al.¹¹⁴ found that the use of patient reminders would appear to be more effective than GP reminders when compared with normal care, in promoting cervical cancer screening. The authors reported that women whose GPs had been prompted to remind them to have a Pap test were more likely to do so than were control women (TRD 6.6 percent, 95 percent CI 5.2 to 8.0). They reported the estimate of the number of women needed to be involved in a GP reminder scheme in order to produce one additional screen (NNI) is 15.2 (95 percent CI 12.6 to 19.3). Sensitivity analysis revealed that one study stood out as exceptional, and when omitted, the TRD was 7.9 percent (95 percent CI 6.5 to 9.4). The TRD for the group of six patient reminder studies, after removal of one exceptional study to produce homogeneity, was 10.8 percent (95 percent CI 8.1 to 13.6). Jepson et al.⁹⁶ also found evidence that letters were effective in increasing uptake, with greater effects demonstrated for cervical cancer screening than mammography. However, the review stated that there was not enough evidence to detect whether GP letters suggesting a Pap smear were more effective than those from another source. In the Shekelle et al. review,¹⁰³ they reported patient reminders for cervical cancer screening as part of a larger review of screening. For cervical cancer screening, the authors reported an OR 1.84 (95 percent CI 1.67 to 2.02).

2. Healthcare provider advice

   Only one systematic review examined the use of health provider advice or counseling for promoting uptake of cervical cancer screening. The Jepson et al.⁹⁶ review identified five studies that evaluated the use of face-to-face counseling by a health professional in either the home or in a healthcare setting. The results indicated that there was no effect (numerical results not available) of the intervention for increasing the utilization of cervical cancer screening.

3. Educational materials

   The use of educational materials to promote the use of cervical cancer screening has not been widely investigated. Two systematic reviews were identified by our search which incorporated several interventions to promote the uptake of cervical cancer screening.^{96, 103}

   The review by Jepson et al.⁹⁶ addressed several different types of educational interventions, which included the use of printed educational materials versus controls, audio/visual (videos, tape-slide shows, and computers) and group teaching (classes and workshops). For the three RCTs for which RRs could be calculated, no effect of printed materials was found compared with a control arm. RRs were not calculated for seven other studies. For these remaining studies, they report that results varied from no effect (n=6) to moderately effective (n=1). The information available for the use of audio/visual equipment was limited in this review. Four studies were identified (two RCTs and two quasi-RCTs); however, the results were grouped with other prevention topics. There was a brief mention of one study which found that tape-slide programme playing in a clinic waiting room had no effect on the uptake of Pap smears when compared with controls (author calculated OR 0.97, 95 percent CI 0.63 to 1.49). The Shekelle et al. review¹⁰³ found that patient-education interventions were effective and reported an OR of 1.53 (95 percent CI 1.30 to 1.82). The descriptions of studies were reported narratively and overall conclusions were done across screening topics.

4. Other types of interventions

   The Shekelle et al.¹⁰³ review addressed organizational change as an intervention to promote screening mammography. They reported that organizational change was consistently one of the most (or the most) effective interventions at increasing the use of preventive services. The OR for improving mammography was 2.65 (95 percent CI 2.26 to 3.12. Organizational change was the most heterogeneous intervention and was often combined with reminder letters.

   Within the Jepson et al.⁹⁶ review, they identified only one study: an RCT that evaluated the effectiveness of an organized programme of prevention that included the use of a health-promotion nurse. The intervention was so effective (Pap smear RR 1.56, 95 percent CI 1.44 to 1.69) that the trial was discontinued after two years (instead of three years) because GPs were no longer willing to exclude half of the participants from accessing the health-promotion nurse. Information regarding the specific outline of the study (other than study design description) was not provided by the authors.

Access-enhancing interventions

1. Financial barriers

   The information for removal of financial barriers to promote the uptake of cervical cancer screening is inadequate. The use of removal of financial barriers was not addressed systematically; however, one study mentioned within a much larger review by Jepson et al.⁹⁶ was identified. Jepson et al. reported that the RCT compared using a voucher for free preventive visits with a control group and the results for cervical cancer screening were part of a larger study that included tuberculosis screening and preventive visits. This study concluded that older individuals will respond to these programs, and such services will result in modest health gains. The overview by Shekelle et al.¹⁰³ included meta-regression analysis for financial incentives as part of a much larger review of preventive services, and the authors report that this intervention was effective in promoting cervical cancer screening with an OR of 3.12 (95 percent CI 2.62 to 3.72).

2. Access barriers

   None of the included reviews addressed access barriers for cervical screening.

3. Media campaigns

   The use of media campaigns to promote the uptake of cervical cancer screening has been understudied. The use of media campaigns was not addressed systematically; however, one controlled trial mentioned within a much larger review by Jepson et al.⁹⁶ was identified. The controlled trial compared the use of a media campaign with control (communities) to increase cervical (and breast) cancer screening and found that the intervention had no effect on being up to date for any of the tests compared with control.

4. Multi-strategy interventions

   Several systematic reviews addressd the use of multi-strategy or multicomponent interventions for uptake of cervical screening.^{96, 97, 102, 105} The interventions included were physician-directed, patient-directed and a combination of both physician- and patient-directed interventions (i.e., such as audit and feedback combined with physician education and the use of a flow chart to enhance delivery of Pap smears).

   Snell and Buck¹⁰⁵ conducted a systematic review of interventions to increase cancer screening. Within this review cervical, breast, or colorectal cancer screening were included. The effect size for Pap smear (n=35 cases) was d +0.0083 (95 percent CI -0.0174 to +0.0340. The results were also presented generally pertaining to cancer screening and determined that screening behavior improved when the physicians were the target of more than one intervention, but not more than three (d [the average amount of change in standard deviation units achieved by individuals in a treated group vs. the change achieved by members of a control/comparison group for a particular study] +0.1360, d +0.2495, d +0.6829, d -0.0058).

   The Shea et al.¹⁰² review examined the use of computer versus manual reminders for improving preventive services and calculated the effect of the combined intervention versus control groups and found only a small benefit (OR 1.12, 95 percent CI 0.82 to 1.51). Similarly, Kupets and Covens⁹⁷ found that in the three studies identified by their search, one study indicated an additive effect of the combined intervention (patient letter combined with computer-generated reminder) of cervical cancer screening of 15 percent, while the two other studies (patient reminder letter combined with computer-generated reminder, and physician reminder combined with patient-carried health maintenance prompt card) did not show an additive effect.

   The Jepson et al.⁹⁶ review also addressed the use of combined interventions on the uptake of cervical cancer screening. Fifteen studies (10 RCTs, two quasi-RCTs, and three controlled trials) evaluated a combination of interventions to increase uptake of screening. The following combinations were shown to be effective at increasing uptake when compared with control: invitation letter from GP plus education, an invitation letter plus followup call from a health educator, and physician reminder combined with invitations to individuals. Finally, the Shekelle et al review¹⁰³ presents narratively some of the studies that look at multi-strategy interventions and based conclusions on multi-strategy interventions across several screening topics. For this reason, we could not extract enough data from the review to present results for cervical cancer screening specifically.

Inverventions to promote evidence of cancer pain

One systematic review examining the World Health Organization (WHO) Analgesic Ladder was excluded as the studies did not evaluate the effectiveness of the intervention.¹¹⁹ Two reports^{120, 121} met the inclusion criteria, but did not have extractable data and are included in the general table (Appendix G).

The objective of the review by Allard et al.¹²² was to identify and describe educational interventions conducted to improve pain control in patients with advanced cancer. Outcomes were participants' attitudes and knowledge, pain management, patients' quality of life, and pain levels. Thirty-three studies targeting either health professionals (HPs) (25 studies) or patients with advanced cancer (seven studies) or their family caregivers (one study) were included. Across all studies, the format and duration of the interventions varied (Summary Table 9). Of the 25 studies targeting HPs, the most frequently studied interventions were educational courses or workshops (16 studies). Some of these interventions targeted opinion leaders (one study) or other types of role models (five studies). Other interventions included the use of pain assessment tools (three studies), provision of guidelines or a treatment algorithm (three studies), provision of patient pain scores to clinicians (two studies), and a combination of education and the creation of a supportive care service (one study).

The methodological quality of the review was scored as “moderate” (Summary Table 10).

Of the eight studies targeting patients or families, all the interventions included education with or without the provision of a booklet or diary. Six studies were RCTs (four targeting HPs and two targeting patients/family caregivers) and three were quasi-experimental designs (QEDs) with a non-equivalent control group (two targeting HPs and one directed to patients/family caregivers). Of the nine RCTs and QEDs, only four studies enrolled over 100 patients.

Healthcare provider-directed interventions

Across all 25 studies, knowledge or attitude was improved to a variable degree in all 13 studies in which this outcome was measured. When only the studies with stronger designs (four RCTs and two QEDs) were considered, knowledge was increased in the only two studies in which it was measured. Pain management strategies such as compliance with guidelines were improved in seven of 10 studies. In the six studies with stronger designs, only two of four studies showed an improvement in adherence to guidelines or analgesic prescription. Patients' pain relief was measured in seven studies with at least some improvement noted in four. In the stronger studies, pain was assessed in all six studies and was improved in three.

Promising interventions in an outpatient setting appeared to be the transmission of patients' self-reported pain scales to oncologists and the use of treatment algorithms for improving prescribing and reducing pain. Interventions involving role modeling or nursing pain assessments and use of a flow sheet may reduce pain, but these studies were relatively small and employed less rigorous designs.

Individual-directed interventions

Patient knowledge or attitude was measured in six of eight studies and was improved in all six including those with stronger designs. In contrast, pain management was measured in only three of eight studies and was improved in all three including one of the strong studies. Pain relief was improved in five of eight studies, including three of the stronger studies. The authors of the review reported that promising interventions for pain control in ambulatory settings appeared to be brief nursing interventions to patients combined with a daily pain diary.

In summary, a nursing pain education program coupled with a daily pain diary can increase pain relief. A brief nursing counseling intervention combined with a pain management booklet is also promising. There are too few studies of interventions of the education of family caregivers to comment on their effectiveness.

Dissemination studies that targeted healthcare providers

1. Train-the-trainer

   Two studies evaluated train-the-trainer approaches.^{134, 136} Albright et al.¹³⁴ evaluated physician training in disseminating a clinical preventive medicine curriculum (CPM). The CPM included six topics: risk factors for coronary heart disease and cancer; smoking cessation, nutrition, weight control and exercise, pharmacological interventions for hyperlipidemia, and screening. Ten general internists from across the US were trained to deliver the CPM to local faculty (n=91). Outcomes included process measures such as the fidelity of the training given to local faculty; change in knowledge; attitudes, and clinical practice of participants and faculty; and the subsequent teaching of CPM topics by local faculty to house staff.

   General internal medicine faculty trained to deliver the CPM program reproducibly delivered this to local faculty.¹³⁴ Local faculty was more knowledgeable about smoking cessation. Using a self-rated scale, they were more likely to implement smoking cessation interventions following the seminars (p<0. 0001). There were significant increases in the proportion of patients with whom smoking cessation was usually or always discussed and in the utilization of specific behavior-change strategies for smoking cessation. Between 85 and 96 percent of faculty reported teaching CPM strategies to house staff. There were significant increases in self-reported efficacy of the 125 house staff to implement specific CPM strategies. However, significant changes in the frequency of advice to stop smoking were not observed. No information was provided on potential barriers to this approach.

   Muramato et al.¹³⁶ evaluated the Arizona Department of Health Services Tobacco Education and Prevention Program (AzTEPP) of certification in tobacco cessation skills. This is a three-tiered certification model to disseminate Public Health Service (PHS) clinical practice guidelines in smoking cessation. It is funded from revenue generated from tobacco tax excises. Personnel working in Arizona's tobacco control program local community-based projects were trained in basic tobacco cessation skills. A proportion of these individuals received additional training as tobacco cessation specialists (including training individuals in basic tobacco cessation skills), and tobacco treatment services managers. Only the first two tiers were evaluated. The evaluation included process measures such as satisfaction and knowledge, plus behavioral outcomes including smoking cessation-related activities. Data are being collected on intermediate and longer-term outcomes including quit attempts, quit methods, use of intensive services, and statewide tobacco use prevalence.

   This study also demonstrated positive outcomes in tobacco control activities from a train-the-trainer approach. Significant increases in knowledge were observed following basic and intensive training courses. Eighty-two individuals received specialist level certification. Nearly half of these individuals were delivering intensive smoking cessation interventions and 69.5 percent were teaching basic certification courses. Individuals who completed basic level certification also showed significant increases in baseline knowledge following the training, and 81 percent reported they had performed at least one brief intervention. No data were presented on smoking cessation outcomes.

2. Recruitment of professional organizations

   Epps et al.¹²⁹ reported on a strategy to recruit professional organizations to co-sponsor train-the-trainer sessions with the NCI. This study was a post-test only design. Data were reported on the recruitment of professional organizations, but not the outcomes of the train-the-trainer programs. Eight national organizations were initially identified and approached, of whom six agreed. Five additional organizations were subsequently involved. Cooperating organizations collaborated in developing training plans, setting dates and locations, and naming coordinators for all seminars. Over four years, 53 train-the-trainer seminars were conducted in 22 states, training 2,098 individuals. Recruitment of professional organizations appeared to be an effective dissemination strategy. However, no information is presented on the outcome of the train-the-trainer component.

3. Educational facilitators

   Three studies evaluated educational facilitators as a dissemination strategy.^{124, 143, 144} Two studies examined the use of educational facilitators to disseminate information about office systems to promote the delivery of preventive services in family practice,^{143, 144} and one examined the use of educational facilitators to disseminate information about smoking cessation kits.¹²⁴

   Cockburn et al.¹²⁴ compared three strategies to disseminate smoking cessation kits to family practitioners. This was the only trial comparing different dissemination strategies. General practitioners (GPs) were randomized to delivery by an educational facilitator, delivery by friendly courier, or postal delivery of smoking cessation kits. Academic detailing by an educational facilitator was considered a dissemination strategy in this case, as it was used to promote the delivery of evidence-based smoking cessation interventions by GPs. The main outcome assessed was use of the kit by the GPs. Process measures were also assessed including initial reaction to the kit, motivation engendered by mode of delivery, and overall acceptability of the kit.

   Delivery by an educational facilitator was rated as more motivating than personal delivery by a courier or postal delivery. GPs in the educational-facilitator group were significantly more likely to have seen the kit (facilitator 99 percent, courier 83 percent, mail 88 percent, p=0.003). They were more likely to believe the kit was less complicated and were more knowledgeable about use of the kit. GPs in the educational-facilitator group used significantly more contract cards than GPs in the courier or mail groups (mean 6.54 vs. 3.79 vs. 1.92, p=0.02). There were no differences in use of other interventions in the kit. No data were presented on smoking cessation rates.

   There were no significant differences between the groups in their perceptions of the overall acceptability of the kit. While some measures favored the educational- facilitator group, delivery by an educational facilitator was 24 times the cost of postal delivery (A$142 vs. A$6). In the absence of data on smoking cessation rates, the additional cost does not appear to be warranted.

   Lemelin et al.,¹⁴³ randomized health service organizations (community primary care practices that have a payment system primarily based on capitation) to an educational facilitator or no intervention. The intervention used seven strategies based on review of the literature. These included audit and feedback, consensus building, opinion leaders and networking, academic detailing and educational materials, reminder systems, patient mediated activities, and patient educational materials. The educational facilitators discussed strategies with the physicians and practice staff, worked with them to adopt the strategies, provided feedback about performance using mini audits, and provided management support to practices. The primary outcome was a preventive performance index. This was defined as the proportion of eligible patients who received recommended preventive maneuvers minus the proportion of eligible patients who received inappropriate preventive maneuvers (as defined by Canadian Task Force on Preventive Health Care). Preventive maneuvers were assessed by audit of 100 charts per practice.

   Forty six practices were randomized. All facilitator-group practices received preventive performance audit and feedback, achieved consensus on a plan for improvement, and implemented a reminder system. Ninety percent implemented a customized flow sheet, 10 percent used a computerized reminder system, 95 percent wanted critically appraised evidence for prevention, and 100 percent received patient educational materials. The overall preventive performance index improved 11.5 percent in favor of the intervention group (31.9 percent vs. 32.1 percent pre- intervention; 43.2 percent vs. 31.9 percent post-intervention, p<0.001). There was no significant difference in smoking cessation-counseling activities (37.6 percent vs. 40.5 percent pre-intervention; 41.2 percent vs. 38.7 percent post-intervention, p>0.05).

   Dietrich et al.,¹⁴⁴ used a factorial design to randomize family physicians and internists to an educational facilitator, a workshop, both, or neither. The workshop was a day long educational session, plus written syllabus. The educational facilitator provided consultation on design and implementation of an office system including a preventive care flow sheet in patients' records, identification of smokers, health education posters and brochures, and patient-held diaries. Outcome data were collected in two different cross-sectional surveys of 20 to 30 patients from each physician in the study (pre-study and 12 to 14 months later). A significantly higher proportion of patients in the educational-facilitator- only group than the control group reported their physician had advised them to quit smoking (0.84 vs. 0.67, p<0.05). No significant differences were observed between the educational-facilitator-plus-workshop group compared with the control group (0.80 vs. 0.67, p>0.05). The authors concluded that overall provision of cancer early detection and preventive services was improved through the use of an educational facilitator to establish office systems.

   The overall findings of these two studies suggest that the delivery of preventive services can be improved through the use of educational facilitators disseminating information about office systems. However, there is uncertainty whether this approach impacts on smoking cessation activities.

4. Postal delivery

   Three studies examined postal delivery as a dissemination strategy.^{125, 128, 133} Fowler et al.,¹³³ used a cross-sectional survey to examine the impact of postal delivery of a smoking cessation booklet to UK GPs. A random sample of, 5,000 GPs were sent a survey. The response rate to the original questionnaire was less than 50 percent. Therefore a shortened version of the questionnaire was sent out. An overall response rate of 75 percent was achieved. Only 50 percent of responding GPs recalled receiving the smoking cessation book and 28 percent read it. Among doctors who read the book, only 19 percent could recall one or two essential steps to advising their patients to stop smoking, and 12.5 percent recalled all three steps.

   One of the barriers to dissemination may have been the mode of delivery. Booklets were delivered with the British Medical Association News Review. Some GPs may discard this without opening it. Additionally, the booklet was distributed with minimal publicity. Nevertheless, simple mailing of smoking cessation booklets to GPs appears to be an ineffective dissemination strategy.

   Mullins et al.¹²⁵ used an interrupted time series to survey members of the general public regarding the extent to which they remembered GPs discussing smoking. The survey was conducted every two years between 1990 and 1996. The aim was to assess the impact on GPs' smoking cessation activities of the dissemination by mail of a self-help booklet on quitting smoking. The main outcome was individuals' recollections of GPs' discussions about smoking cessation. Over 95 percent of GPs recalled receiving a patient self-help book on smoking cessation delivered by post and 97 percent of GPs who recalled receiving the self-help books distributed them. There were no changes between 1990 and 1996 in the proportion of patients that smoked who reported being asked about smoking status by their GP (22.4 percent vs. 21.3 percent), or who recalled being advised to quit smoking by their GP (34.8 percent vs. 37.4 percent). However, significantly more smokers (10.7 percent vs. 20.6 percent, p<0.001) were given information or help to stop after dissemination of the self-help books.

   The third study evaluated postal delivery of a letter, recruiting pharmacists to distribute patient self-help booklets on smoking cessation.¹²⁸ The Manitoba Pharmaceutical Association supported this strategy. In addition, a media campaign (television, radio, and billboards) promoted the distribution of the booklets. Only process measures were reported (number of booklets distributed). The majority of pharmacies agreed to participate (87 percent), and 93.5 percent of 46,000 booklets were picked up. No additional information was provided.

   The findings from the last two of these studies suggest that mailed invitations or postal delivery are strategies to disseminate patient self-help materials. However, the strength of this evidence is weak. No information is provided in either study about behavioral outcomes such as quit rates.

5. Multiple dissemination strategies

   One additional study targeting healthcare professionals utilized multiple dissemination strategies as part of an approach to improve smoking cessation counseling practices of family physicians in Montreal.^{138, 142} The multicomponent approach included: postal delivery of smoking cessation guidelines (based on PHS clinical practice guidelines); educational materials to facilitate counseling; smoking cessation publications (including publication of guidelines); public awareness campaigns to encourage individuals to ask for help to stop smoking from their family practitioner; and desktop cards presenting an algorithm of smoking cessation interventions. The program was initiated in 1997. Mailed questionnaires were sent to Montreal family physicians in 1998 and 2000. The outcomes assessed were family physicians' self-reported attitudes and beliefs about smoking cessation counseling, perceived skills, and importance of perceived barriers to smoking cessation activities.

   There were no changes in process measures (attitudes, perceived abilities, or interest in updating smoking cessation counseling skills) among male physicians. Female physicians perceived ability to provide smoking cessation counseling improved over time. Female physicians showed more improvement in smoking cessation-counseling practices between the two evaluations. Lack of time was perceived as a significant barrier to smoking cessation counseling by both male and female physicians. No data on behavioral outcomes such as actual physician smoking cessation activities, or patient outcomes such as quit rates were assessed. It is therefore difficult to make any firm conclusions regarding the effectiveness of this multicomponent dissemination strategy.

Dissemination studies that targeted individuals

1. Media awareness campaigns

   Seven studies reported on the effectiveness of the media as a dissemination strategy for smoking cessation interventions.^{126, 127, 130, 131, 140, 141, 146} Two studies retrospectively evaluated the importance of different media sources as strategies for patient recruitment.^{127, 140} Elder et al.¹²⁷ evaluated the importance of different media sources for recruitment to a community smoking cessation contest, whereas Anderson et al.¹⁴⁰ surveyed callers to Cancer Information Service (CIS) telephone hotlines inquiring about smoking, diet and nutrition, pap smears, and breast self examination, to determine how individuals found out about the CIS.

   Elder et al.¹²⁷ found that television was the most important media source for people joining a Quit to Win community smoking cessation program.¹²⁷ The importance of television increased with decreasing income (71 percent for income <$20,000 v. 40.5 percent for income >$40,000, p<0.05). Similarly, Anderson et al.¹⁴⁰ reported that television was the most important source of learning about the CIS hotline. Men (72 percent) were more likely to cite television than women (61 percent). The likelihood of citing television as major source of learning about CIS increased with younger age (<20 years 81.7 percent vs, >60 years 39.6 percent) and lower education. Television was the first source of knowledge about the CIS for the majority of people inquiring about smoking (78.5 percent). Both of these studies are retrospective and provide weak evidence for the effectiveness of the media as a dissemination strategy.

   Pierce et al.¹⁴¹ evaluated the impact of public service announcements (PSAs) about smoking cessation on use of CIS using an interrupted time series. Three of 12 PSAs explicitly encouraged viewers to call the CIS. Monthly call volumes to the CIS were compared with the timing of PSAs from the Office of Smoking and Health. The three peak call periods to the CIS over a five-year period were observed following these three PSAs. Television promotion increased the proportion of callers who were male, younger, and less educated.

   Three prospective studies also observed that media dissemination (primarily via television) increased calls to a CIS.^{126, 131, 135, 145, 146} Cummings et al.¹³¹ prospectively evaluated the impact of a targeted media campaign (television and print media) on calls to CIS in seven media markets. Matched media markets were utilized as controls. Mothers with young children were specifically targeted and the primary outcome was a comparison of call rates between the groups during the media campaigns. Calls regarding smoking cessation to the CIS were increased 4.8 times in the targeted media audiences compared with the control areas. In the experimental group, 28.9 percent of calls were from the target group compared with 9.5 percent in control markets. Call volumes increased substantially during the television campaigns and then dropped off back to baseline soon after. In followup interviews with the targeted group (women with young children), more women in the experimental group attempted to quit (64 vs. 46 percent, p> 0.05). However, there was no difference in the number of women who reported they quit (13 vs. 15 percent).

   Pierce et al. ^{145, 146} evaluated media dissemination of a smoking cessation clinic (“Quit for Life” campaign). Calls to the “Quit Line” and enrolments at the “Quit Centre” were monitored following a campaign of three television commercials. Over 50,000 calls were made to the Quit Line over the first three months of the campaign compared with an expected 8,600 calls based on previous call rates. Enrolments in the Quit Centre smoking cessation programs were almost 3,000 for the year of the campaign, compared with 352 in the prior year. Cohorts of people in two cities (intervention and control cities) were surveyed prior to and one year after the start of the campaign. Smokers in the intervention city were significantly more likely to report quitting or cutting down the number of cigarettes smoked (35 vs. 18 percent, p<0.05). They were also more likely to have attempted to quit during the 12 months following the media campaign (66 vs. 60 percent, p<0.05).

   Boyd et al.¹²⁶ targeted African American populations in a randomized trial of a media campaign to increase use of CIS. Several strategies were used including television and radio commercials, along with a community outreach packet utilizing peer leaders. The majority of calls received by the CIS in both control and experimental communities during the study period were from African Americans (565 vs. 144). The number of calls from African Americans in the experimental communities was 80 times greater than that from control communities. Increased call volumes persisted for about eight weeks following the first media wave and four weeks following the second media wave. Callers in the experimental communities were more likely to cite radio (51.4 percent) than television (41.6 percent) as the way they heard about the CIS. Radio generated 8.89 calls per 10,000 African American smokers and television generated 6.89 calls per 10,000 African American smokers. All other sources combined accounted for 1.38 calls per 10,000 African American smokers.

   A series of reports by Marin et al.^{130, 132, 139} provided information about media dissemination of the Programa Latino Para Dejar de Fumar to Latino residents in San Francisco. Information about a self-help manual targeted to Latinos was disseminated via a variety of media sources. Cross-sectional surveys were conducted annually between 1986 and 1993 to assess the impact of the program. Awareness of the smoking cessation program increased from 18.5 to 41.5 percent at the end of the evaluation. The increase in awareness was similar in the less acculturated as well as highly acculturated population. Awareness of printed smoking cessation material increased over time. The largest increase was in less acculturated Hispanics. The proportion of respondents who had a copy of the self-help manual increased from 7.6 to 19.7 percent over time. These changes were associated with an overall decrease in prevalence of smoking from 24.5 percent in 1986 to 16.4 percent in 1991.

   The findings of these seven studies provide some consistent evidence that media awareness campaigns are important strategies to disseminate information about CIS help lines. Television was the more important source of information in all but one study. Several of the studies suggest that television is a more important source of media awareness for certain demographic groups, such as younger people and people from less educated and lower income groups. However, the studies generally do not provide information about the subsequent outcomes following the call to CIS. It would be important to know about the use of other smoking cessation interventions, along with the quit rates from these populations of individuals. This is particularly relevant given the uncertainty about the effectiveness of reactive telephone counseling in the findings for key question #1. A significant consideration of media awareness campaigns is cost. This becomes an important issue given that the above studies suggest the impact of any media awareness campaign disappears within one or two months after the campaign finishes.

Dissemination studies that targeted worksites

1. Passive dissemination

   One paper evaluated a passive dissemination strategy in the worksite.¹³⁷ Control worksites in the Working Well Trial were given the intervention materials at the completion of the trial. The way in which this information was given to control sites is not specified. This study reported on smoking cessation rates two years after completion of the trial. Dissemination of the program to control sites had little impact on the level of smoking activities in control worksites. This passive dissemination strategy was ineffective.

Dissemination studies that targeted healthcare providers

1. Train-the-trainer

   One “train-the-trainer” study aimed at disseminating preventive medicine education to physicians.¹³⁴ Faculty from general internal divisions across the US were invited to apply for a month-long Stanford Faculty Development Program; 10 were chosen and trained to be Clinical Preventive Medicine facilitators. They then went to their home institutions and trained other faculty at their home site. Fidelity checks concluded that facilitators adhered closely to the curriculum they had been taught. Those medical faculty educated by the facilitators had an increase in knowledge and self-efficacy to use behavior changes to promote healthy diets. Subsequently, house staff physicians interacting with faculty who had attended the facilitator-run sessions reported an increase in the degree of preventive medicine content in teaching interactions and an increase in their ratings of self-efficacy to implement preventive medicine strategies.¹³⁴ While the train-the-trainer model shows some promise, it needs to be evaluated with a more rigorous design; furthermore, many biases are likely to be inherent in the selection of internists who were able to leave their work situation for a month of training.

2. Academic detailing (educational facilitators)

   One RCT¹⁴⁴ targeted dissemination to healthcare providers using academic detailing. In this trial by Dietrich et al., primary care medical practices were randomized to one of four groups: facilitator only, facilitator-plus-workshop, workshop only, or a control group. Practices in the facilitator-only group (n=24) received three to four visits from a facilitator who provided detailed instruction and assistance in selecting and implementing non-computer-based office-system interventions. Practices in the facilitator-plus-workshop group (n=26), in addition to receiving visits from an educational facilitator, had a physician from the practice attend a one-day workshop. The workshop session reviewed NCI's prevention and screening recommendations, but did not provide information on the use of office-system interventions. Practices in the workshop-only group (n=24) attended the workshop. Practices in the control group (n=24) received no information.

   Cross-sectional patient surveys were conducted before randomization and again at 12-month followup. The study reported on two diet-related outcomes: (1)the number of patients reporting that their physician had advised them to reduce their fat intake and (2) the number of patients reporting their physician had advised them to increase their fiber consumption. At 12-month followup, significantly more eligible patients in the facilitator-only group reported their physician had advised them to reduce their fat intake compared with patients in the control group (0.56 vs. 0.47, p<0.05). There was no significant difference in the number of patients reporting advice to decrease fat intake between the facilitator-plus-workshop group and the control group at 12- month followup (0.51 vs. 0.47). There was no significant increase in the number of eligible patients in the facilitator-only or facilitator-plus-workshop groups reporting advice to increase fiber consumption compared with patients in the control group at 12-month followup (facilitator vs. control 0.48 vs. 0.38; facilitator-plus-workshop vs. control 0.41 vs. 0.38). The overall conclusion from this RCT was that the use of educational facilitators to disseminate and implement office-system interventions can improve the provision of prevention and early detection services in community practices.

   The use of educational facilitators (academic detailers) to disseminate office-system interventions appears to be a promising strategy. Further research in this area is needed.

3. Workshops

   The RCT Tziraki et al.¹⁵¹ assessed the effectiveness of two strategies for promoting the use of an NCI nutrition manual by primary care physicians and their office staff. The nutrition manual was modeled after the NCI publication “How to help your patients stop smoking”. Medical practices randomized to the workshop group (n=244) were invited to send one staff member to a three-hour training workshop on how to use the nutrition manual. Training was provided in four major components of the manual: (1) how to organize the office environment, (2) how to screen for patient adherence, (3) how to provide dietary advice, and (4) how to implement a patient followup system. Medical practices assigned to the postal-delivery group (n=256) received the nutrition manual in the mail with no further information. Medical practices in the control group (n=255) did not receive the nutrition manual.

   Followup interviews with medical staff and observational assessments were conducted at four to six months after dissemination of the manual. Adherence scores were calculated for four areas: office organization, nutrition screening, nutrition advice or referral, and patient followup. There was low attendance at the workshop session; less than 50 percent of assigned practices sent representatives (120 of 244). The authors of the trial used an “intent to treat” approach for the primary statistical analyses and included all practices in the workshop group regardless of attendance. The workshop group was significantly more adherent to the manual's recommendations for office organization at followup than either the postal-delivery group (28.5 vs. 24.7 percent, p<0.005) or the control group (28.5 vs. 23.0 percent, p<0.001). Of those practices who sent a representative to the workshop, 30.6 percent were adherent to the recommendations for office organization. There was no significant difference between the postal-delivery group and the control group for office organization (24.7 vs. 23.0 percent).

   The workshop group was also significantly more adherent to the manual's recommendation for nutrition screening than either the postal-delivery group (23.5 vs. 21 percent, p<0.05) or the control group (23.5 vs. 20.5 percent, p<0.05). Of those practices that sent a representative to the workshop, 25 percent were adherent to the nutrition screening recommendations. There was no significant difference between the postal-delivery group and the control group for nutrition screening (21 vs. 20.5 percent). There was no statistically significant difference between the three groups for providing nutrition advice (workshop 54.9 percent, postal delivery 53 percent, control 52.3 percent), nor for patient followup (workshop 14.6 percent, postal delivery 13.6 percent, control 13.6 percent). A secondary analysis showed that those practices who attended the workshop were significantly more likely than either the postal-delivery group (57 vs. 53 percent, p<0.05) or the control group (57 vs. 52.3 percent, p<0.05) to provide nutrition screening. There was no significant difference observed for patient followup on secondary analysis.

   Training workshops appear to hold some promise as a dissemination strategy; however, motivating medical professionals to attend these sessions may be a difficult barrier to overcome. Further research in this area is needed.

4. Postal delivery

   One RCT¹⁵¹ evaluated the effectiveness of postal delivery as a dissemination strategy. This trial compared the effectiveness of postal delivery with a training workshop to disseminate an NCI nutrition manual to primary care practices. Postal delivery was not found to be an effective method to disseminate the nutrition manual. Please refer to the section above on Workshops for the detailed results of this study.

Dissemination studies that targeted worksites

1. Passive dissemination

   The Working Well Trial^{92, 137} randomized 114 worksites of over 28,000 workers to test the effectiveness of health promotion activities that were planned and delivered with a high level of employee participation. The intervention phase lasted for two years, and then nutrition materials were disseminated to the control sites, followed by a further two-year assessment. The investigators were particularly interested to see if the control sites would utilize the materials. No information was given about the actual strategies used to get the nutrition intervention materials to the control group, nor was any report of measure of uptake given. No changes occurred in the level of nutrition activities in the control sites.

   An opinion leader strategy was tested using peer educators in the worksite intervention called “5-A-Day: Healthier Eating for the Overlooked Worker”. While rated methodologically weak, it holds promise as an area for further research. It was an RCTof 5-A-Day intervention to increase fruit and vegetable consumption in an ethnically mixed population of 2,091 lower socioeconomic and trade employees.^{147, 148} Both the intervention group and the control worksites received an 18-month intervention program of education materials through workplace mail, cafeteria promotions, and speakers. In the intervention group, naturally occurring work “cliques” were identified, and within those, ratings were given to each individual regarding their degree of “centrality” to communication ties and flow. Those rated highest in “centrality” became the peer educator for that clique, mimicking the “opinion leader” strategy.

   Peer educators attended a 16-hour training program where they were given information about health benefits of eating fruits and vegetables, cultural trends in dietary practices, peer educator's roles and responsibilities, and five persuasive communication strategies (foot-in-the-door, fear appeal, benefits, peer pressure, and questioning) and ways to initiate informal conversations about fruits and vegetables. They were instructed to engage in nutrition education of the co-workers for about two hours per week, on work time. They also distributed 5-A-Day materials produced specifically for this population: a nine-booklet resource guide, four issues of a newsletter, enabling gifts such as a recipe book, and vegetable seeds. The peer educator intervention lasted nine months, with consumption measured at the end of the intervention and six-month followup.

   The result was an increase in fruit and vegetable consumption of 0.77 total servings per day more in the intervention group compared with the controls (measured by recall, p<0.001) and an increase of 0.46 total daily servings (measured by food frequency, p<0.002).¹⁴⁷ The effect was maintained at six-month followup for intake recall (increase of 0.41 daily servings, p=0.034) but not for food frequency¹⁴⁷. In analysis of the frequency and duration of peer-education contact with co-workers, greater contact with the peer educators was related to larger immediate increases in fruit and vegetable intake, particularly vegetable intake, but was not related to total intake at six-month followup.¹⁴⁸ A qualitative design, used to study the educational strategies used by the peer educators in the intervention group,¹⁴⁹ found that these studies differed by gender and ethnicity.¹⁴⁹ Hispanic educators were more likely to use individual, rather than group, change strategies than non-Hispanic educators; men more frequently used strategies such as “mock competition”, “giving materials” and “encouragement”, while female peer educators more often used “creating context”, and “keeping 5-A-Day visible”.¹⁴⁹

   Few worksite dissemination strategies have been evaluated. In one, the dissemination strategy was not evaluated.⁹² The other study using an opinion leader strategy had at least a short-term impact on consumption.

Dissemination studies that targeted individuals

1. Media strategies

   Two studies evaluated multiple media channels (print, television, and radio) to assess the impact of the media campaigns on telephone calls to an information telephone line.^{140, 150} “Project Lean” (Low-Fat Eating for America Now) was a three-year initiative, begun in 1989, to reduce dietary fat consumption. The media campaign led to hotline access of 300,000 consumer calls in 18 months (25,000 to 28,000 calls/month), but the calls declined as publicity declined, and the line was terminated due to expense, estimated to be US $300,000 per year.¹⁵⁰ While these outcomes were not assessed in a direct comparison, some important lessons were learned in this study:¹⁵⁰ that well-placed advertising may be the most appropriate and effective communications strategy for a national nutrition social marketing campaign as it can, more easily than PSAs, be tailored to the particular audience; can communicate information more directly and can reduce the need for an information hotline or followup materials. Furthermore, building a network of state and local programs and partnerships with the food service industry allowed the campaign to reach a broader audience.¹⁵⁰

   A second primary study was identified which was an analysis of calls to the CIS hotline. Callers were asked, “How did you first find out about the CIS?” Records of a subsample of people (214,472) who inquired about smoking, nutrition, Pap smears, and breast self-evaluation were reviewed. Television was the most frequently reported source of learning about the information line, regardless of age, gender, or ethnic group (except callers of Asian or Pacific heritage, who reported publications as the more common source of information about the hotline).¹⁴⁰

   The media dissemination strategies, particularly television messages, can make people aware of information lines and prompt them to call. However, from these two studies, it appears that the lines are expensive to advertise and maintain.

Dissemination studies that targeted healthcare providers

1. Academic detailing (educational facilitators)

   Of the four studies^{143, 144, 153, 156} that targeted dissemination to healthcare providers using academic detailing, the Kinsinger et al.¹⁵³ RCT provides the strongest evidence. Medical practices within the experimental group (n=32) received detailed instruction and assistance on how to implement office systems tailored to increase breast cancer screening. Practices in the control group (n=31) received no academic detailing visits and were not provided with any information on the development of office systems. Five indicators of the use of an office system were assessed at baseline and again at followup in each of the practices¹⁵³. These indicators were: (1) ≥50 percent of patient records have an entry on a flow sheet; (2) ≥50 percent of physicians in the practice report having written preventive care policy; (3) ≥50 percent of the physicians in the practice report that nurses frequently or sometimes recommend mammograms to patients; (4) ≥50 percent of physician in the practice report that nurses identify patients who are due for mammograms; and (5) ≥50 percent of physicians in the practice report frequent use of flow sheets or computerized reminders to identify patients due for mammograms.

   The mean number of indicators increased significantly in practices in the experimental group (1.3 to 2.8) compared with control group practices (decrease from 1.5 to 1.4) (p<0.0003). There were significant increases in three of the five indicators in experimental group practices compared with practices in the control group (Indicators [1], [2], and [3] above; all p<0.05). At followup, however, no indicator was present in more than two-thirds of intervention practices, and only seven of the 32 practices in the experimental group reported a complete office system for breast cancer screening.

   Kinsinger et al.¹⁵³ also conducted a chart review on a random sample of eligible patient records at baseline and again at followup. There was no significant difference between experimental and control practices in the change in proportion of women's records with an actual mammogram report within the last year (increase of 4.7 vs. 3.4 percent). The study identified physician attitude regarding the value and feasibility of office systems for mammography screening as a barrier to successful implementation of complete office systems.

   Lemelin et al.¹⁴³ randomized health service organizations (community primary care practices that have a payment system primarily based on capitation) to an educational-facilitator group (n=22 practices) or to a control group (n=23 practices) that received no visits from an educational facilitator. Over an 18-month period, educational facilitators visited practices in the experimental group an average of 33 times with each visit lasting about one hour. The facilitators performed an initial audit and feedback of each practice's baseline preventive performance. The academic detailers then acted to facilitate the development of practice goals and policy for preventive care and assisted the practices in selecting and implementing interventions to improve preventive care. All of the practices in the experimental group implemented a reminder system. Ninety percent implemented a customized flow sheet, 10 percent used a computerized reminder system, 95 percent wanted critically appraised evidence for prevention, and 100 percent received patient educational materials. Ninety-five percent of the physicians in the experimental group reported that they were either satisfied or very satisfied with the educational-facilitator approach.

   The primary outcome measured was a preventive performance index. This was defined as the proportion of eligible patients who received recommended preventive maneuvers minus the proportion of eligible patients who received inappropriate preventive maneuvers (as defined by Canadian Task Force on Preventive Health Care). Preventive maneuvers were assessed by audit of 100 charts per practice. At baseline, the preventive performance index was not significantly different between the facilitator and control groups (31.9 vs. 32.1 percent, respectively). At followup, the corresponding values were 43.2 percent for the facilitator group and 31.9 percent for the control group. The absolute increase in the preventive performance index of 11.5 percent in the educational-facilitator group was statistically significant (p<0.001). At followup, there was no significant improvement in the number of eligible patients in the facilitator group having mammograms compared with the control group (67.5 vs. 58.7 percent, respectively). Refer to Evidence Table 8 for further detail. The overall conclusion of the study's authors was that use of this educational-facilitator approach resulted in significant improvements in preventive care performance.

   Dietrich et al.¹⁴⁴ conducted an RCT in which primary care medical practices were randomized to one of four groups: facilitator only, facilitator plus workshop, workshop only, or a control group. Practices in the facilitator-only group (n=24) received three to four visits from a facilitator who provided detailed instruction and assistance in selecting and implementing non-computer-based office-system interventions. Practices in the facilitator-plus-workshop group (n=26), in addition to receiving visits from an educational facilitator, had a physician from the practice attend a one-day workshop. The workshop session reviewed NCI's prevention and screening recommendations, but did not provide information on the use of office-system interventions. Practices in the workshop-only group (n=24) attended the workshop. Practices in the control group (n=24) received no information.

   Cross-sectional patient surveys were conducted before randomization and again at 12-month followup. The proportion of patients reporting having a mammogram at 12- month followup was significantly higher in each of the three experimental groups compared with the control group (facilitator plus workshop vs. controls 0.78 vs. 0.57, p<0.01; facilitator only vs. controls 0.77 vs. 0.57, p<0.01; workshop only vs. controls 0.71 vs. 0.57, p<0.01). There was no significant difference among the three experimental groups in the proportion of eligible patients reporting having had a mammogram. A chart review of a random sample of patient records confirmed the mammography-related findings from the cross-sectional patient surveys (see Evidence Table 8 for further detail). The overall conclusion from this RCT was that the use of educational facilitators to disseminate and implement office-system interventions can improve the provision of prevention and early detection services in community practices.

   Williams et al.¹⁵⁶ also evaluated the use of academic detailing as a dissemination strategy. In this one-group pre-post-test study, GPs were provided with information about the effectiveness of medical record prompts and recall systems, and the availability of ACS resources.¹⁵⁶ The academic detailers also provided the physicians with ACS patient educational materials and display racks. At baseline only one of the 10 practices used ACS patient educational materials. After academic detailing visits, all 10 practices used the ACS materials and nine displayed the information in the racks provided. In contrast, only minor changes to office systems were found at followup. Practices that had not used medical record prompts at baseline did not add them. However, practices that previously used chart summaries or prompts added items, typically mammography or Pap test notations (no further details provided). At baseline, only one practice had a recall system for scheduling mammograms. At followup, one practice with a Pap test recall system at baseline had added mammography recalls and one practice with no recall system at baseline implemented both Pap and mammography recalls. The principal barriers to intervention implementation identified were: time, lack of administrative process or infrastructure, and lack of third party reimbursement.

   The use of academic detailers (educational facilitators) was reported to be acceptable to healthcare providers in the three studies in which it was assessed. This strategy yielded mixed results for disseminating office-system interventions to increase mammography uptake.^{143, 144, 153, 156} This strategy may be a promising way to disseminate patient educational materials and information about local resources for healthcare providers.¹⁵⁶ Further research is needed in this area to clarify whether educational facilitators can be an effective strategy to disseminate office-system interventions or patient educational materials. Further research is also needed to elucidate variables that influence the effectiveness of this strategy.

2. Workshops

   Scott et al.¹⁵⁵ compared two strategies to disseminate a manual of evidence-based interventions promoting mammography uptake to managed care organizations. Half of the managed care organizations attended an intensive one-day workshop on how to use the intervention manual and received an accompanying user guide. In contrast, managed care organizations in the passive-dissemination group received only the intervention manual with no training workshop or user guide. This was predominantly a descriptive study in which in-depth interviews were conducted with quality-improvement personnel at each managed care organization to elucidate motivating factors and identify barriers to use of the manual or subsequent implementation of interventions from the manual.

   Little difference was found between managed care organizations that attended the workshop and those in the passive dissemination group in use of the manual or in interventions implemented (no statistical analysis reported). Overall, seven of the eight managed care organizations in the study used the manual and implemented more interventions to promote mammography uptake in the year after receiving the manual compared with the year prior to its dissemination (no statistical analysis reported).¹⁵⁵ Improvement was found in the type of interventions implemented (i.e., more were evidence-based). All seven plans implemented a physician-directed intervention. This was most frequently a performance feedback letter encouraging physicians to recommend mammograms to their patients who were due for them. Some of the managed care organizations also implemented interventions directed towards patients (e.g., reminder letters). The only managed care organization that did not use the manual had the point person for the organization change jobs during the study.

   Mammography rates in the year prior to dissemination of the manual were compared with the rates in the year after dissemination. In all seven of the managed care organizations that used the intervention manual, the mammography rates increased (range 0.22 to 4.0 percent). Mammography rates in the managed care organization that did not use the intervention manual decreased by 2.67 percent.

   In this study, the intensive one-day workshop strategy did not seem to confer an added advantage to successful dissemination of the manual, nor to the subsequent implementation of interventions from the manual compared with passive dissemination. Overall, the managed care organizations were receptive to the dissemination of cancer control interventions to promote mammography.

   In-depth interviews with key personnel at each of the managed care organizations identified local factors that influenced implementation of interventions from the manual. A key factor for implementation seemed to be the length of employment of the point person for the managed care organization. The two organizations that implemented the least intensive interventions had point people who had only been in their positions for a relatively short period of time. Other factors that facilitated use of the manual and implementation of interventions were: (1) the motivation of the point person to improve mammography rates, (2) support of senior management, (3) adequate resources (time, personnel, and funds), and (4) the organization and content of the intervention manual. The two major barriers to implementation of interventions from the manual were resources and data limitations. The resource limitations identified were finances, time, and programming. The data limitation concerned the ability to identify the baseline population of members to be targeted for the interventions.

Dissemination studies that targeted worksites

One study evaluated the dissemination of interventions that promote mammography uptake to worksites. In this self-selected, four-group post-test only study, Paskett et al.¹⁵⁴ evaluated whether approaching worksite management by introductory letters, followed by telephone contact, was an effective strategy to recruit worksites to sponsor breast cancer screening education programs for their employees¹⁵⁴. Worksite management was offered the choice of sponsoring three increasingly intensive interventions: (1) printed educational materials, (2) breast cancer screening educational sessions, or (3) training of worksite nurses in breast cancer education. Of the 102 worksites approached, 97 completed the baseline survey. Of these, 63 worksites accepted and offered a program to their employees. Fourteen worksites chose the intensive nurse training, 14 sponsored worksite classes, and 35 chose the educational display of brochures. Worksites that chose to sponsor one of the interventions were more likely to have sponsored breast cancer education programs before (p=0.027) or to have a medical department (p=0.006).

The intervention chosen was significantly associated with a history of sponsoring other health education programs (p<0.01). Worksites that had sponsored a similar program in the past were more likely to send a nurse to be trained. Of the 73 worksites that had never sponsored a breast cancer program, a majority (n=43) was responsive to this dissemination strategy and chose to sponsor one of the interventions. The least intensive intervention (educational displays with brochures) was selected by 29 of these worksites.

Responses from the 34 worksites that chose not to sponsor one of the interventions indicated the presence of several barriers to implementation. These barriers were: (1) finances, (2) geographic composition (several branches in different locations), (3) employee characteristics (majority of employees not women or not in target age range), and (4) presence of a corporate policy that prohibits offering a program to a subgroup of employees (while excluding others).

Approaching worksite management by introductory letters and telephone contact appears to be a promising strategy to disseminate interventions to promote mammography uptake. This strategy was even successful in recruiting worksites that had not previously sponsored programs for their employees to implement one of the interventions promoting mammography uptake. Further research in this area should definitely be undertaken.

Dissemination studies that targeted healthcare providers

1. Academic detailing (educational facilitators)

   Three studies^{143, 144, 156} assessed the effectiveness of disseminating interventions to healthcare professionals using educational facilitators (academic detailing). Lemelin et al.¹⁴³ randomized health service organizations (community primary care practices that have a payment system primarily based on capitation) to an educational-facilitator group (n=22 practices) or to a control group (n=23 practices) that received no visits from an educational facilitator. Over an 18-month period, educational facilitators visited practices in the experimental group an average of 33 times with each visit lasting about one hour. The facilitators performed an initial audit and feedback of each practice's baseline preventive performance. The academic detailers then acted to facilitate the development of practice goals and policy for preventive care and assisted the practices in selecting and implementing interventions to improve preventive care. All of the practices in the experimental group implemented a reminder system. Ninety percent implemented a customized flow sheet; 10 percent used a computerized reminder system; 95 percent wanted critically appraised evidence for prevention; and 100 percent received patient educational materials. Ninety-five percent of the physicians in the experimental group reported that they were either satisfied or very satisfied with the educational- facilitator approach.

   The primary outcome measured was a preventive performance index. This was defined as the proportion of eligible patients who received recommended preventive maneuvers minus the proportion of eligible patients who received inappropriate preventive maneuvers (as defined by Canadian Task Force on Preventive Health Care). Preventive maneuvers were assessed by audit of 100 charts per practice. At baseline, the preventive performance index was not significantly different between the facilitator and control groups (31.9 vs. 32.1 percent, respectively). At followup, the corresponding values were 43.2 percent for the facilitator group and 31.9 percent for the control group. The absolute increase in the preventive performance index of 11.5 percent in the educational-facilitator group was statistically significant (p<0.001). At baseline, Pap testing was performed on 60.8 percent of eligible patients in the facilitator group and on 57.9 percent of patients in the control group. At followup, there was no significant improvement in the number eligible patients in the facilitator group having a Pap test compared with the control group (66.2 vs. 59.1 percent, respectively). The overall conclusion of the study's authors was that use of this educational-facilitator approach resulted in significant improvements in preventive care performance.

   Dietrich et al.¹⁴⁴ conducted an RCT in which primary care medical practices were randomized to one of four groups: facilitator only, facilitator plus workshop, workshop only, or a control group. Practices in the facilitator-only group (n=24) received three to four visits from a facilitator who provided detailed instruction and assistance in selecting and implementing non-computer-based office-system interventions. Practices in the facilitator-plus-workshop group (n=26), in addition to receiving visits from an educational facilitator, had a physician from the practice attend a one-day workshop. The workshop session reviewed NCI's prevention and screening recommendations, but did not provide information on the use of office-system interventions. Practices in the workshop-only group (n=24) attended the workshop. Practices in the control group (n=24) received no information.

   Cross-sectional patient surveys were conducted before randomization and again at 12-month followup. There was no significant difference in the proportion of patients reporting having had a Pap test at 12-month followup in any of the three experimental groups compared with the control group (facilitator plus workshop vs. controls 0.65 vs. 0.61; facilitator only vs. controls: 0.71 vs. 0.61; workshop only vs. controls 0.63 vs. 0.61). The overall conclusion from this RCT was that the use of educational facilitators to disseminate and implement office system interventions can improve the provision of prevention and early detection services in community practices.

   The RCTs by Lemelin et al.¹⁴³ and Dietrich et al.¹⁴⁴ both found that using educational facilitators to disseminate office-system interventions resulted in significant improvements in overall indicators of preventive care. Neither study, however, found that the use of educational facilitators led to an increase in cervical cancer screening rates.

   Williams et al.¹⁵⁶ conducted a demonstration project using academic detailing to provide GPs with information about cancer-screening guidelines, the effectiveness of medical record prompts and recall systems, and the availability of ACS resources. This was a one-group pre-post-test study. Four education representative volunteers (ERVs) were recruited and trained to lead discussions and to involve participating staff and physicians. The academic detailers provided the primary care physicians with ACS patient educational materials and display racks. Of the 10 primary practices visited, only one of the practices used ACS patient educational materials pre-intervention. Post-intervention, all 10 practices used the ACS materials, and nine displayed the information in the racks that were provided. In contrast, only minor changes to office systems were found at followup. Practices that had not used medical record prompts at baseline did not add them. However, practices that previously used chart summaries and/or prompts added items (i.e., Pap notations). One practice, which was initially without a recall system, instituted one that included Pap test and mammography recalls. The principal barriers to delivering preventive care were: (1) time (which was reported by all physicians and staff as the major barrier to implementation), (2) lack of administrative process or infrastructure, and (3) third party reimbursement.

   The use of academic detailing for healthcare providers was reported as acceptable in the two studies in which it was measured.^{143, 156} While this strategy improved overall indicators of preventive care, it did not yield promising results for disseminating office systems to increase cervical cancer screening.^{143, 144, 156} Further research is needed in this area, in particular, to identify variables that may impact on the effectiveness of this strategy for cervical cancer screening.

Dissemination studies that targeted individuals

1. Media Awareness campaigns

   One study evaluated the importance of different media sources in disseminating information about the telephone-based CIS. In this descriptive study, Anderson et al.¹⁴⁰ examined inquiries received by the CIS to determine effects of different media in stimulating calls to the service, as well as demographic characteristics of callers in four cancer prevention and early detection topics: Pap smear screening, smoking cessation, nutrition, and breast self-examination.

   A retrospective analysis of five years of inquiries to one national and 26 local CIS offices in four topics provided data. A standardized call record form was completed for each call. The variable of interest was one of the questions posed to callers: “How did you first find out about CIS?” The local CIS office manager assigned one out of a possible 52 codes to their response. For the purpose of their analysis, the codes were collapsed into six categories: (1) television - including several prime and fringe time public service announcements; (2) publications - advertisements, magazines, newsletters, and health brochures/pamphlets; (3) radio - combining several PSAs; (4) healthcare providers - including health agencies such as the ACS, physicians, or members of their staff; (5) significant others - including friends, relatives, neighbors, and co-workers; and (6) telephone assistance - which included telephone book, directory assistance, and the 1-800- operator assistance.

   Telephone assistance (phone book and directory assistance) was the most frequently reported source of learning about the CIS by callers seeking Pap-smear screening information (27.7 percent). The second most cited source was healthcare providers (22.7 percent). Publications (including newspapers, magazines, pamphlets, and posters) were cited by 19.7 percent of Pap-smear- related callers. Television was cited by 17.7 percent of Pap-smear-related callers, 10.3 percent cited significant others, and two percent cited radio. Television was the primary source reported for callers for all education levels. In general, the lower the caller's level of education the more frequently television was cited as the information source.

   Four of the five ethnic groups identified across all preventive topic areas - Caucasians, African Americans, Hispanics, and Native Americans - identified television as the primary source. For callers of Asian or Pacific Island heritage, the most frequently cited source was publications (46.7 percent), followed by television (32.1 percent). Further analysis suggested that news publications, not health publications, were the greatest source of information for this ethnic group.

   This review provides “weak” evidence for the effectiveness of media dissemination strategies to raise awareness of CIS. Additionally, there is little evidence from systematic reviews regarding the effectiveness of reactive telephone counseling for cervical cancer screening.

   The use of physician-directed interventions, such as office systems in the form of medical record prompts and recall systems, has been shown to be somewhat effective in promoting the uptake of cervical cancer screening. The use of the CIS was not an intervention that was identified by our review of systematic reviews addressing cancer control interventions to promote the uptake of cervical cancer screening.

Dissemination studies targeted to healthcare providers

1. Application of an algorithm with community oncologists and nurses

   Du Pen et al.¹⁵⁸ randomized nine institutions (two managed care organizations, three small community hospitals, four large community/regional hospitals) to either training or no training group. Each group had 10 oncologists and 18 to 20 nurses. Patients were included if they were: English speaking, had diagnostic evidence of a locally invasive or metastatic solid tumor, were ambulatory, and had at least six-month life expectancy. The curriculum for the dissemination strategy was designed to address problems previously identified. It consisted of a one-time five-hour training session incorporating the physician/nurse teams as well as a workbook containing the algorithm flow chart, guiding principles of pharmacologic management, and a pain flow sheet. It was unclear if the actual participants were passively or actively involved in the training. Patients were not trained. Overall, there was a statistically significant improvement in overall provider adherence, but there were no significant differences in any of the subscales. A significant deterioration in training effect was found over the study. Patient adherence was a confounder in both groups and was highly correlated with usual and worst pain at the end of the study. The authors reported that other factors, such as a lack of time and resources, may have been factors in the inability to sustain the effect of the training.

2. Role-modeling

   In a pre-post design, Weisman and Dahl¹⁵⁹ evaluated the Wisconsin Cancer Pain Initiate Role Model program. This purpose of this one-day program was to train pairs of clinician educators and their clinical partners to be role models in their local communities. The authors reported significant improvement in knowledge. Sixty-four percent of participants reported that they had met their goal of conducting an in-service program or longer-term programs such as integrating pain assessment into clinical practice.

   Breitbart et al.¹⁵⁷ also used a pre-post design to evaluate the Observership Program, which was one of four components of the “The Network Project”, a project funded by the NCI. This strategy consisted of a two-week educational program targeted toward individuals who were likely to promote education in their own institution or community.

   The program consisted of attendance at rounds, meetings, and observing doctor-patient interactions. The authors reported increases in participant knowledge after training compared with baseline. Thirty-seven percent completed a followup questionnaire pertaining to their local educational and training activities. Of those responding to the survey, just under half reported that they had organized a pain management symposium or educational program.

Dissemination studies targeted to individuals

Although there is evidence that a nursing pain-education program coupled with a daily pain diary or a nursing counseling intervention combined with a pain management booklet can increase patient pain relief, no dissemination studies of these types of interventions were identified.