NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Whelan TJ, O'Brien MA, Villasis-Keever M, et al. Impact of Cancer-Related Decision Aids. Rockville (MD): Agency for Healthcare Research and Quality (US); 2002 Jul. (Evidence Reports/Technology Assessments, No. 46.)

  • This publication is provided for historical reference only and the information may be out of date.

This publication is provided for historical reference only and the information may be out of date.

Cover of Impact of Cancer-Related Decision Aids

Impact of Cancer-Related Decision Aids.

Show details

4Development of Cancer-Related Decision Aids

Twenty-two studies that report the development of cancer-related decision aids are described in this chapter. These studies share one main characteristic: the development of a DA intervention that could be used in a real clinical situation. Seven of the 22 DAs have already been used to help people in the decisionmaking process,24-30 and the results regarding their effectiveness are described in Chapter 5. Of the remaining development studies, one of the studies reported that their DA intervention is currently being tested with patients at the point of decisionmaking.31 The aim of this chapter is to summarize and characterize the DAs' development process, in which the authors have tried to evaluate the psychometric characteristics of the instrument, as well as the participants' perceptions of using the DAs.

General Characteristics of Included Studies

Type of Decision Aid

Eight studies reported the development of decision board (DB) interventions.27-29,32-36 Seven studies assessed interactive computer programs.24,37-42 One study investigated the use of a computer program that generated patient-specific printed output.31 One study assessed the use of a videotape intervention,24 and one study explored the use of a tradeoff analysis supplemented with "visual aids."26 Three studies investigated the use of audiotape workbooks.43-45 One study30 reported the development of a complex DA intervention consisting of a DA videotape, DA brochure, and counseling.

Type of Decisions

Nineteen studies evaluated DAs in the context of treatment decisionmaking.24-29,31-34,36-40,42-44,46 Two studies involved screening decisions,35,41 and one study was about prevention.30

Type of Cancer

Of the 22 studies included, there were 14 studies on breast cancer24,27-31,35-40,43,44 and two each on prostate cancer,42,46 ovarian cancer,33 and lung cancer.25,26,34 There was one study each on colon cancer41 and leukemia.32

Populations

All the studies included adult populations. For the evaluation of the DA, the authors selected healthy volunteers in nine studies;27-29,32,33,35,36,41 in one study, primary care patients without cancer were included.46 In 12 studies, cancer patients participated; 1 study included previously treated cancer patients,26 8 were after the treatment decision was made,25,31,33,37-40,42 and 4 studies included cancer patients at the point of treatment decisionmaking.34,39,43,44 In one study, women suspected of having a genetic predisposition to breast cancer who were making preventive decisions were participants.30

Two studies were specifically performed to evaluate a DA in special populations. McTavish and colleagues39 assessed an interactive computer program, the Comprehensive Health Enhancement Support System (CHESS) in eight African-American breast cancer patients from impoverished neighborhoods.39 As well, Lawrence and colleagues developed a DB for breast cancer screening decisions among European-American and Mexican-American women in the United States.35

Type of Outcome Measures

Eight studies assessed the validity and reliability of the DA,26,27,29,30,32,33,35,36 and nine the feasibility of the DA.26,30,32-34,37-40 Participants in three studies evaluated the clarity of the DA,32,43,46 and also in three studies28,32,46 the comprehensibility of the DA. The acceptability of the DA was evaluated in 12 studies.25,29,31,33,34,37-43 Ten studies26,27,29,30,32-36,39,41 reported the decision made by the participants. Other outcomes can be seen in Table 18.

Table 18. Summary of included studies performed to develop decision aids.

Table

Table 18. Summary of included studies performed to develop decision aids.

Type of Study Design

Ten studies were case series,25,31,34,37-39,41-44 and six studies used a test/retest design.26,27,29,30,33,36 There was a combination of designs in two studies;32,35 a subgroup of participants in an RCT32 were enrolled in a test/retest study, and the authors of one case series35 used the same population to perform a test/retest study. There was one study each of the following study designs: survey,28 one-group pre/post intervention,46 RCT,40 and controlled trial.24

Quality Assessment

Since we were not aware of published quality assessment scales specifically developed to assess studies of the development of an intervention or tool, we elected to use the most appropriate items from the Downs and Blacks Checklist.22 The original Downs and Black Checklist contains 10 items for reporting, 3 for external validity, and 14 for the study of internal validity, with 27 as the maximum score. The internal validity is divided in two categories: bias and confounding, with seven items in each. Since not all the items were applicable to each study, the overall evaluation considered only the total of those items that could be reported; thus each study had a unique denominator. The reviewed studies could be evaluated on a score of 18 to 24 items, and each component is reported in Table 19. Considering percentages of the scores, among the 22 studies, 226,30 achieved the highest scores, 80 percent. Overall, according to each category, the highest scores were for the reporting (11 percent to 100 percent), followed by the internal validity (bias: 17 percent to 66 percent; confounding: 33 to 100 percent). The external validity had the lowest scores (0 percent to 33 percent).

Table 19. Quality assessment of development studies with the Downs and Black Checklist -- Overall evaluation.

Table

Table 19. Quality assessment of development studies with the Downs and Black Checklist -- Overall evaluation.

Description of the Studies According to the Type of Intervention

Videotape (Evidence Table 4.1)

Only one study described the development of a videotape intervention. Schapira and colleagues46 reported the development of a DA to assist patients in considering treatment options (radiation therapy, radical prostatectomy, watchful waiting) for localized prostate cancer.

Development Process of the DA

The videotape script content was based on a review of the literature about the morbidity and mortality associated with the three treatment options, and on input from a multidisciplinary team of experts (i.e., a radiation oncologist, a urologist). Subsequently, the relevance of the information was evaluated by two groups of patients who had been treated for clinically localized prostate cancer. These focus groups addressed the importance of information regarding treatment complications, such as impotence as well as the presentation of quantitative information rather than qualitative ('risk of impotence is 40 percent' instead of 'a small risk'). After gathering all the information, the final content of the script included: anatomy of the prostate gland, epidemiology of prostate cancer, treatment options and outcomes, efficacy of treatment and management of possible treatment side effects. A 20-minutes long videotape was designed following the principle of role modeling to encourage active involvement in decisionmaking. A professional filming company produced the videotape; and the edited version was piloted on a group of 10 previously treated prostate patients; men responded favorably to the DA with respect to the relevance of content, clarity and flow of information.

Evaluation of the DA

The final version was evaluated in a group of 32 men who did not have cancer and were recruited from a primary care outpatient clinic (Evidence Table 4.1a).

Outcomes

Validity. Not reported.

Reliability. Not reported.

Feasibility. Not reported.

Acceptability of DA

The authors reported that "97 percent of the participants felt that the information presented in the videotape was clear and easy to understand." Participants expressed satisfaction with the information in this format, and stated that the inclusion of patient interviews was helpful in making the information relevant.

Knowledge

The authors developed a 20-item questionnaire to evaluate knowledge about the videotape content. Participants answered the questionnaire before and after having viewed the videotape. After the intervention, there was a statistically significant increase in correct responses in 10 of the 14 questions (p < 0.05).

Other Outcomes

Using a qualitative analysis, the authors asked participants about their approach to decisionmaking by asking three open-ended questions, before and after the intervention. Prior to viewing the video, 28 percent of participants indicated that they would defer decisionmaking to the physician; after the intervention, 16 percent had the same response. For percentages of other factors influencing the approach to treatment decisionmaking, such as looking for a second opinion and joint decisionmaking discussions between the physician and patient, see Evidence Table 4.1b.

Interactive Computer Program (Evidence Tables 4.2 to 4.8)

Seven studies24,37-42 focused on the development of interactive computer programs (ICP). Four37-40 evaluated the use of the "Comprehensive Health Enhancement Support System" (CHESS) in breast cancer patients. Of the other three studies, one assessed an ICP for helping people choose among different alternatives for colon cancer screening.41 A second study was a preliminary evaluation of the videodisk, "A Shared Decision Program," for women with breast cancer.24 The third study assessed an ICP that was designed to assist prostate cancer patients in making treatment decisions.42

Since 1993, Gustafson and colleagues have been evaluating CHESS in different populations. Three of the reviewed studies were case series37-39 and one was an RCT.40 CHESS is a disease-specific computer-based system that is designed to meet information and support needs in a user-controlled and nonthreatening manner for people facing health problems. In the prototype, the visual interface was interactive, easy to use, and utilized color and graphics on a personal computer with a VGA monitor and a modem. This program contained modules for breast cancer, acquired immunodeficiency syndrome, substance abuse, sexual assault, and other conditions. The components of CHESS evolved from 7 in the first study37 to 11 in the last study.40 These components were grouped into three main areas: information, social support, and a problem-solving component. Of particular relevance was one module called, "Decision Aid," that used utility theory to help patients through difficult decisions such as what surgery to have and whether to take adjuvant chemotherapy or tamoxifen treatment. Users could read the description of each option or read an excerpt from a personal story of a woman who chose that option. They could read the criteria used by other women to make a choice and a research summary about their choice. Different treatment choices were included such as choice of surgery (mastectomy vs. lumpectomy), choice of adjuvant hormonal therapy or chemotherapy, or choice of joining a clinical trial. The authors reported that the "Decision Aid" section helped users work through difficult and specific decisions, but the module "does not tell users what they should do; it shows users how the computer used their input to predict the choice they might make." The "Action Plan" module helped users implement a decision by combining statistical decision theory and change theory. The module asked users their plan to implement a decision, predicted the likelihood of success, and suggested how they could strengthen their prospects. Through a modem, users of CHESS also had the opportunity to interchange information and experiences with other patients in an on-line support group that included trained facilitators to monitor the group.

Development Process of the DA

The selection of content for CHESS began with a needs assessment that used focus groups and survey research. A panel of breast cancer patients, their partners, and daughters; a surgeon and a nurse practitioner; and two researchers developed three information needs surveys for patients, partners, and adult children. The authors surveyed 400 patients and family members at diagnosis and 3 and 9 months later; all participants were recruited from breast cancer clinics. CHESS was planned for users' homes, workplaces, libraries, and other community sites; thus far, its evaluation has been conducted at patients' homes. The first version was designed at a 10th-grade reading level. The authors reported that they intended to simplify the reading level to grade 8 and to add sound and voice recognition, motion video, and touch screens; however, it is not clear whether or not these features have been incorporated. They also reported that CHESS is reviewed and updated biannually, and that "...this easy update feature allows users access to the most recent information and discoveries about breast cancer...."

Evaluation of the DA

The first two field studies37,38 included newly diagnosed breast cancer patients who had chosen surgical (mastectomy or lumpectomy) and adjuvant (radiation, chemotherapy, or tamoxifen) treatment. Ten breast cancer patients and five adult daughters were enrolled in the first pilot study, and 20 patients in the second. Women in the first pilot study suggested many enhancements; and after another year of development, the second pilot study was conducted (details not provided). In the third study, CHESS was evaluated among eight African-American women, with stage I or II breast cancer, from impoverished neighborhoods in Chicago. The final study was an RCT to assess the impact on quality of life and the costs of care. The authors studied 36 newly diagnosed stage 0-II breast cancer patients who received CHESS (intervention group, n = 19), or Dr. Susan Love's Breast Book (control group, n = 17). In this study, the authors focused their results on the patients' use and impressions of CHESS.

Outcomes

Validity. Not reported.

Reliability. Not reported.

Feasibility. Across the four studies, the authors quantified the time that each patient spent using each program component. Overall, the 10 women in the first study37 used CHESS 546 times during the 50-day trial, and the 20 women in the second study38 used it 1,131 times during the 70-day trial. In the first two studies, the Decision Aid and Action Plan were the least used components; however, the authors reported that the first CHESS versions only contained the surgery decision program, and most of the women had already made their decision before the studies had begun. African-American women in the third study39 used CHESS 886 times within the 15-week access period. Fifty five percent of the total time of use was spent in the social support sections, 41 percent in the information components, and only 4 percent in the Decision Aid and Conflict components. The authors reported that each of the 36 women in the RCT40 used an average of 62 "hits" during the 3 months of implementation. Initially, all participants used the system more frequently; its use decreased to an average of twice per week between the 3rd and 7th week, and once per week through to the end of the 3 months. The Discussion Group was the component most used, whereas the Decision Aid and Action Plan were less used.

Acceptability of DA

In the first two studies,37,38 participants rated the value of each CHESS component on a 1- to 5-point scale. In both studies, patients' mean scores were 3.0 or higher (see Evidence Tables 4.2b and 4.3b for each component). The African-American women39 utilized a 1- to 7-point scale for this outcome; their mean scores were 6.5 or higher.

Women in the first two studies were asked to appraise the perceived value of using CHESS for other patients during diagnosis, treatment, and after treatment. They assessed this item on a 1- to 5-point scale; participants' mean scores were 3.5 or higher. The value of CHESS during diagnosis and treatment was rated the highest.

In the second study,38 the authors reported that 6 of 17 patients preferred talking with a visiting nurse or with a patient survivor rather than using CHESS, because they wanted rapid and precise answers to their questions.

Decision

None of the studies were primarily focused on the actual decisions that were made. However, in the study of African-American women,39 the authors reported that one woman had avoided a surgical decision for months and, after receiving CHESS, she re-entered treatment.

Other Outcomes

Based on a 1- to 7-point scale, the authors asked women in the first three studies37-39 to rate their "emotions" while using CHESS. They considered two types of "emotions" -- positive (7 items) and negative feelings (11 items). Overall, participants' mean ratings were 4.0 or higher for the positive feelings and 4.0 or less for the negative feelings (see Evidence Tables 4.2b, 4.3b, and 4.4b).

In the RCT,40 the authors reported that CHESS was well received by the women. They found it easy to use and considered that having the software for 3 months was adequate. Fewer than half of all the CHESS users said they would be willing to purchase the tool, but 50 percent reported interest in borrowing computers or software.

Dolan41 described the results of using an analytic hierarchy process (AHP) in 20 volunteers who were offered five options regarding screening regimens for colon cancer. The authors reported that the AHP evaluated was a user-friendly technique that could provide a way to effectively engage patients in the process of medical decisionmaking because it enabled users to elicit subjective values and combine them in an explicit, unbiased manner.

Development Process of the DA

The author did not report how the AHP was developed. The decision aid used for the study was based on Eddy's "balance sheet" that described the pros and cons of alternative approaches to colon cancer screening for 50-year-old men who had a family history of colon cancer in a first-degree relative. At the top of the AHP model was the goal of the decision, to maintain health and well-being. The five decision criteria located in the middle level of the model represented the potential risks and benefits of screening (i.e., decrease the risk of colon cancer, avoid false positive tests, minimize costs). The five alternatives at the bottom level provided a spectrum of risk-benefit profiles varying from "careful followup," which was considered the least likely to decrease cancer risk but the best in terms of the other four criteria, to "colonoscopy every 5 years," which was described as the biggest reduction in cancer risk but the worst alternative in every other way.

The analysis consisted of pair-wise comparisons among the screening alternatives and pair-wise comparisons among the criteria relative to the goal. The author reported that "Expert Choice" was the software used for the AHP analysis, which was described as a standard AHP software package running on a laptop computer.

Evaluation of the DA

The intervention consisted of a structured interview where the participants imagined that they were 50 years old, had a relative with colon cancer, and were making a decision about a colon cancer screening program for the next 25 years. The author explained the different options for colon cancer screening: careful followup, sigmoidoscopy, annual stool test and sigmoidoscopy, barium enema, and colonoscopy. After the participant understood the problem, he/she completed a guided AHP analysis using "Expert Choice" software. Following the analysis, the results were discussed with the participants.

Outcomes

Validity. Not reported.

Reliability. Not reported.

Feasibility. The author reported this outcome as a cutoff time: the participants were judged to be capable of using the AHP if they completed the analysis in 45 minutes or less. The mean time required to complete the analysis was 35 minutes, and 19 participants finished the analysis in less than 45 minutes.

Acceptability of DA

Nineteen of the 20 participants agreed that if they had a close relative with colon cancer, they would prefer to go through this type of analysis before making a decision. Nineteen felt that they learned useful information about colon cancer screening. All the participants indicated that, if the scenario were real, they would choose the option that came out best on the analysis. Eighteen of 20 participants answered positively when they were asked about using this type of analysis for other health care decisions.

Other Outcomes

According to the AHP analysis, the best option was colonoscopy every 5 years for 10 participants, careful followup for 9, and barium enema every 5 years for 1. Minimizing the chances of developing cancer was ranked the most important priority by 15 participants; 4 considered the side effects, and 1 wished to avoid costs.

Chapman and colleagues24 reported the preliminary evaluation of the videodisk entitled "Treating Your Breast Cancer" produced by the Foundation for Informed Medical Decision Making. The Shared Decision Program (SDP) is an interactive videodisk containing two sections: a core program with relatively few interactive options and a more interactive "Learn More" section, where the viewer selects from a menu of topics.

Development Process of the DA

Not reported.

Evaluation of the DA

Participants were not exposed to the videodisk program. Instead, the authors used a 40-minute linear videotape version. Participants watched the core program, which was one of the two SDP sections. This program contained information about the breast and breast cancer and an overview of the surgical treatment options (breast-sparing with radiation treatment, mastectomy followed by breast reconstruction, and mastectomy followed by use of a breast prosthesis), including the advantages and disadvantages of each surgical procedure. The program included interviews with five breast cancer survivors who talked about the decision they made and how satisfied or unsatisfied they were after having made the decision and going through the treatment they chose. The authors compared the videotape with a decision aid brochure whose text was produced from the video script. The booklet contained photos selected to closely match those presented in the video, but did not contain the interviews and cartoon drawings. Nursing (n = 34) and psychology (n = 48) students participated in this controlled trial study; half of each student group viewed the video (n = 40), and the other half read the booklet (n = 42). The authors did not report how they assigned participants to the study groups.

Outcomes

Validity. Not reported.

Reliability. Not reported.

Feasibility. Not reported.

Acceptability of DA

Participants were asked to rate how much they liked the women in the interviews and the narrator on a 5-point scale. Two women in the videotape had the lowest ratings (p <0.0001); both were dissatisfied, one with her reconstruction and the other with her prosthesis (Evidence Table 4.7a).

Decision

The authors asked participants to imagine that a friend who had breast cancer had requested their advice about what surgical treatment to select. They expressed their recommendations by dividing 100 points among three options: breast-sparing surgery with radiation, mastectomy followed by breast reconstruction, and mastectomy followed by use of a breast prosthesis. The authors reported their results taking into account three different factors: intervention (video vs. booklet), sample (nursing vs. psychology students), and time (pre- vs. postintervention). After the intervention, participants in the videotape group showed a significant increase in their recommendations for breast-sparing surgery (p < 0.01); nursing students rated breast-sparing surgery higher than did psychology students (p < 0.01) (see Evidence Table 4.7b).

Knowledge

The authors reported the knowledge scores in the same order as described above. The authors reported that there was no difference between video and booklet groups (F < 1; p value was not specified). Nursing students pre- and posttest knowledge scores were significantly higher than those of the psychology students (p < 0.0001). Mean scores for knowledge were significantly higher in the posttest than in the pretest evaluation (p < 0.0001).

Jenkinson and colleagues42 developed "The Prostate Center," a hypermedia program integrating CD-ROM and Internet technology, to help patients with localized prostate cancer access detailed and current information about available treatment options (Evidence Table 4.8a).

Development Process of the DA

The first prototype was developed after interviewing 10 prostate cancer patients who provided their information needs and their opinions about three different interfaces. Subsequently, the authors incorporated information obtained from different sources, such as patient education materials and literature aimed at educating specialists. A group of experts reviewed the material to ensure accuracy and offered suggestions for improvement. The resulting interactive program contained information regarding the diagnosis (i.e., diagnostic test), treatment options (radical prostatectomy, radiation therapy, watchful waiting), followup (i.e., side effects of treatment), and role of support groups as well as links to the Internet, glossary, and bibliography. The prototype incorporated text, graphics, and a video interview with a patient. The patient discussed his reactions to a diagnosis of localized prostate cancer, his treatment choice, and his subsequent recovery from surgery.

Evaluation of the DA

The first prototype was evaluated in a qualitative pilot study with five patients.

Outcomes

Validity. Not reported.

Reliability. Not reported.

Feasibility. Not reported.

Acceptability of DA

Using a semistructured one-on-one interview, the authors reported that, after the session with the computer, participants had positive responses and described the prototype as relevant, useful, and navigable.

Computer Program With Printed Output (Evidence Table 4.9)

Ravdin and colleagues31 report the development of a computer program called "Adjuvant" that assisted women diagnosed with breast cancer in deciding whether to have adjuvant treatment (chemotherapy or endocrine therapy).

Development Process of the DA

The computer program produced prognosis estimates using survival outcomes from surveillance, epidemiology, and end-results data. The program used the results of the 1998 overviews of randomized trials of adjuvant therapy to provide estimates regarding the efficacy of adjuvant treatment.

Evaluation of the DA

The Adjuvant computer program was designed to allow an oncologist to enter patient-specific information (age, menopausal status, comorbidity estimate, tumor size, number of positive lymph nodes, and estrogen receptor status). The program then used actuarial analysis to project overall survival and disease-free survival with and without adjuvant therapy for an individual patient. For adjuvant chemotherapy, the computer program based its estimates on the efficiacy of polychemotherapy (cyclophosphamide/methotrexate/fluorouracil-like regimens, or anthracycline-based therapy, or therapy based on both an anthracycline and a taxane). For adjuvant endocrine therapy, the computer program used estimates of the efficacy of 5 years of tamoxifen treatment. The patient-specific survival estimates generated by the computer program were provided to the patient as a printed output in either a numerical or graphical format.

Outcomes

Validity. Not reported.

Reliability. Not reported.

Feasibility. Not reported.

Acceptability of DA

The authors assessed the acceptability of the printed output format with 24 breast cancer patients who were currently undergoing adjuvant treatment or followup care. The authors did not report the raw data; however, they did cite examples of the types of changes that were made to reflect patient responses. After the output format was revised, the authors then tested the acceptability of the new format with another group of breast cancer patients (n=25; no data provided).

Treatment-Tradeoff Interview Method (Evidence Table 4.10)

Brundage and colleagues26 evaluated a treatment-tradeoff interview for helping lung cancer patients make treatment decisions between high-dose radiotherapy versus low-dose radiotherapy, and between combination treatment with chemoradiotherapy versus high-dose radiotherapy.

Development Process of the DA

The content of the intervention was supported by the results of published clinical trials and from a chart review of patients who had attended the authors' institution.

Evaluation of the DA

The intervention consisted of a 1-hour structured interview. For each comparison, treatment description cards were laid out so that two treatments could be viewed side-by-side. The description components included details of the actual treatment regimen, side effects, and effects of the treatment on personal functioning, emotional state, social interaction, and disease symptoms. Survival data were displayed using a time line that indicated time from treatment and a sliding-bar instrument that indicated percentage alive and percentage dead at a given time. For both comparisons, participants stated their preference (participants could decline both treatments) after the two treatments were described. First, the authors displayed the 6-month survival percentages for the two treatments as being equal. Subsequently, the survival associated with the nonpreferred treatment was then raised systematically until the participant's preference shifted to the alternative treatment; the authors recorded the switch point of preference. This same procedure was performed for 1-year and 3-year survival.

The authors included 56 cancer patients who had completed treatment and 20 staff members. They were asked to imagine being a stage IIIB non-small-cell lung cancer patient and to state their treatment preferences for each comparison.

Outcomes

Validity

For both comparisons, the stated preferences were similar. The authors reported that most of the participants (cancer patients and staff) would have considered taking the more toxic treatment if it offered a certain survival advantage. With longer survival frames (from 6 months to 1 and 3 years), there was a shift in the treatment preference (see Evidence Table 4.10b).

The authors calculated the survival advantage threshold (SAT) for the six tradeoffs (desired survival percentage minus baseline survival percentage). There was a wide range of SATs for patient and staff, from -5 percent to 60 percent. The authors reported that cancer patient thresholds were generally lower than those declared by staff.

Reliability

Twenty participants were interviewed a second time to determine the consistency of their responses (preferences and SATs) 3 or more weeks after the first intervention. The authors reported that the preference consistencies ranged from 70 and 100 percent for the six tradeoffs. Percentages for SATs exact agreement ranged from 20 to 40 percent (median 32.5 percent) and agreements within 5 percent ranged from 55 to 85 percent. The lowest correlation coefficient was 0.7.

Feasibility

All but three participants completed the intervention; two lung cancer patients did not understand the method and could not state a preference. Another lung cancer patient became frustrated with the method, apparently because of the numbers and proportions; he could not participate in the second tradeoff. Three additional patients understood the method but were unable to provide responses for the 6-month and/or 1-year survival time points. The authors reported that the mean time to complete the interview was 60 minutes.

Other Outcomes

The authors also assessed participants' desired level of involvement in decisions; staff members were more likely than patients to prefer active roles (p < 0.01), and no staff members preferred a passive role in decisions about their own medical treatment.

Decision Board (Evidence Tables 4.11 to 4.18)

Whelan and colleagues47 described the DB as a visual aid that targeted the physician-patient interaction to improve communication by presenting information in a standardized manner, using spoken and written language supported by the use of visual aids and relying on repetition. The DBs permitted patients to state their preference for treatment or screening options. Eight studies27-29,32-36 about the development of the DB were reviewed; two of these studies were described in the same paper.33,34 All but one study focused on treatment decisions; Lawrence and colleagues35 designed a DB for screening decisions. Four DBs27-29,36 were developed for breast cancer patients, two for ovarian cancer,33,34 and one for leukemia.32

Development Process of the DA

Overall, the eight studies followed the same steps for developing the DB; a review of the literature about the different treatment or screening options was reported in five studies.27-29,32,35 In two studies,33,34 the authors based the DB content on the observation of four patient consultations with their oncologist and included information about the disease status, treatment options, and treatment protocol. Evaluation of the information contained in the DB by focus groups of patients and clinicians was described in six studies.27,29,33-36 Details of the DB clarity and comprehensibility were provided in all eight studies; these items were assessed only by clinicians,27,28,32,36 or by previously treated patients and clinicians.29,33-35

Evaluation of the DA

The following studies are reported based on the year of publication.

Breast cancer

Levine and colleagues27 enrolled 30 healthy volunteers who were asked to choose between adjuvant chemotherapy and no treatment after viewing a DB. Scenarios were initially created that described adjuvant chemotherapy options for patients with early stage breast cancer postsurgery. The scenarios provided information describing a patient's choice, a patient's quality of life based on her choice of treatment compared with no treatment, and the possible outcomes (recurrence compared with no recurrence). The authors presented the risk for treatment failure in terms of recurrence of breast cancer (negative framing). Morbidity was described to the patient in the chemotherapy scenario using probabilistic language "to capture the uncertainty of the potential treatment's side effects." The authors used a verbal description of the uncertainty (i.e., likely) rather than a numerical one (i.e., 0.8). The DB had three subtitles: "treatment choice," "chance of outcome," and "outcome," and, except for the titles, the board was empty when the interview began. The information cards were held by the patient while the interviewer read aloud; the plastic-laminated card was then attached to the DB with Velcro. By the end of the session, all the information was on the board. The board was described as "large enough to permit the patient to read the display, but not so large that it is cumbersome to store and carry."

Breast Cancer

Whelan and colleagues28 assessed, in 10 healthy female volunteers and 16 node-negative patients who had been treated with breast irradiation postlumpectomy, a DB developed for breast cancer patients to elicit their preference between radiation and no radiation after lumpectomy. The authors created an introduction and scenarios. The introduction provided the patient with some background information about the disease, prognosis, and the purpose of the board. The scenarios described the treatment options of radiation therapy or no radiation and the subsequent outcome of chance of recurrence of cancer in the breast for women with node-negative breast cancer who were treated by lumpectomy and axillary dissection. Descriptions of the risk of recurrence were developed using visual aids. The radiation therapy scenario described the treatment program as well as potential side effects. The probability rates were tailored to known patient prognostic factors of age and tumor size. The DB had two subtitles, "treatment choice" and "results of treatment choice," and the board was empty at the beginning of the interview. The patient and clinician (nurse or physician) read each information card and attached it to the board with Velcro. At the end of the discussion, all of the information cards were on the board.

Breast cancer

Whelan and colleagues29 developed a DB for primary treatment in breast cancer patients; 30 healthy female volunteers chose between lumpectomy plus radiation or mastectomy. Scenarios were developed to give information about the disease, the purpose of the decision instrument, the two treatment options (mastectomy and lumpectomy plus radiation), the acute and long-term side effects associated with each treatment, long-term survival, and quality of life. The DB was used by surgeons and had four titles: "treatment choice," "side effects," "results of treatment choice for the breast," and "results for treatment choice for survival." Below each heading were two information panels (one for mastectomy and one for lumpectomy plus radiation), resulting in eight separate panels. Initially, each panel was covered by a sliding door; the panel was then opened to reveal information in a sequential fashion. The patient and the surgeon read each panel together. During and after the presentation, the patient was encouraged to ask questions. At the end of the presentation, the patient was faced with a complete visual representation of the options and outcomes. The authors reported that an additional card with details regarding breast reconstruction was also available. After the DB presentation, the patient was given a takehome version of the DB.

Breast cancer

Carrère and colleagues36 reported the results of the initial evaluation of a DB in 40 healthy female volunteers; this treatment DB was developed for postmenopausal women with node-positive breast cancer and positive hormonal receptors. The postoperative treatment options were chemotherapy or no chemotherapy. The authors reported that the risks and benefits of the two treatment options were described in an unbiased manner by means of written material and visual aids. The written material included all the information to be described and served as the physician's guide for using the visual aids. As in other studies, the visual aid was empty at the beginning of the interview and, each time a piece of information was orally provided to the patient, a corresponding card with written information was attached to the board with Velcro. By the end of the interview, all the information cards were on the board. The authors presented the two options simultaneously in eight successive sections: results of cancer surgery; presentation of mandatory treatments (tamoxifen and radiation treatment); formulation of therapeutic choice; description of chemotherapy treatment; presentation of the risks of relapse associated with the two options (5-year relapse rates were 30 percent with chemotherapy vs. 40 percent without chemotherapy), information about overall survival; presentation of chemotherapy side-effects (hair loss, nausea, vomiting, tiredness, and infection); and treatment schedules. The treatment options were presented by colors (pale yellow for chemotherapy and pale orange for no chemotherapy). The information concerning relapse rates was provided using bar charts.

Outcomes

Validity

Levine and colleagues27 administered the DB using standard estimates of recurrence for a stage I patient (15 percent risk of recurrence within 5 years without treatment, which was reduced to 10 percent with chemotherapy); the participant was then asked to choose whether or not to be treated. Validity was evaluated by changing the information provided and determining whether a preference changed in a predictable manner. When women chose to receive chemotherapy, the absolute magnitude of benefit was reduced gradually and authors considered that the DB was valid if a patient changed her choice to no treatment when a very small or no gain was presented; the same change was predicted after increasing the therapy morbidity. The same validity test was done for those women who chose not to receive chemotherapy, but increasing the treatment potential benefit and reducing its toxicity. The authors reported that more than 82 percent of the women changed their preference according to the hypothesis (see Evidence Table 4.11b).

Whelan and colleagues29 assessed validity by changing the information provided as described above and determining whether preferences changed in a predictable manner; 89 percent of women who chose lumpectomy plus radiation switched their preference when survival was decreased, and 82 percent of women who chose mastectomy switched their preference when survival was reduced.

Carrère and colleagues36 hypothesized that modifying certain characteristics of the chemotherapy treatment (conditions of administration and side effects) would lead to predictable shifts in expressed choice. The authors reported that the choice of chemotherapy was not very sensitive to increasing side effects, and the choice of no chemotherapy was rather sensitive to change in the conditions of administration and side effects of chemotherapy (see Evidence Table 4.18b).

Whelan and colleagues29 did not report how validity was assessed.

Reliability

Levine and colleagues27 evaluated reliability in a second interview conducted 2 weeks after presentation of the first instrument. They used the same intervention described above; preference in both assessments was measured on a 6-point Likert scale. The authors reported that 28 of the 30 patients had the same preference in both assessments and that agreement between preference responses, expressed as a kappa statistic, was 0.86.

Whelan and colleagues29 reported that the DB was administered on two occasions by a skilled interviewer and that women's responses were stable over time. After 3 to 4 weeks, 28 of the 30 had the same preference (kappa statistic = 0.86)

Carrère and colleagues36 evaluated the across-time consistency of the preference. The authors measured a participant's choice and her strength of preference on a 0-to-10 visual analogue scale after the intervention and 2 weeks after the first intervention. The authors reported than none of the women shifted their choice between the two interviews (intraclass correlation coefficient = 0.97)

Whelan and colleagues28 did not report reliability data.

Feasibility

None of the four studies reported this outcome.

Acceptability of DA

Whelan and colleagues28 reported that participants found the instrument easy to read and that participants appeared to have a good understanding regarding most issues concerning breast irradiation. They had some difficulty understanding that breast irradiation was not associated with a survival benefit; however, the authors did not report raw data.

Whelan and colleagues28 reported that 80 percent of the women found the DB useful, 97 percent recommended its use for breast cancer patients, and 100 percent indicated that the DB was easy to understand.

Knowledge

Carrère and colleagues36 tested comprehension with a 13-item questionnaire; they reported that after the first intervention, the proportion of correct answers for each of the 13 questions was never lower than 87.5 percent, and "none of the questions were randomly answered (P < 5.10-6, one tailed X 2 test)."

Decision

Levine and colleagues27 reported that 17 of the 30 women chose chemotherapy, and 13 chose no chemotherapy. Carrère and colleagues36 reported that 18 of 40 participants chose chemotherapy, and 22 chose no chemotherapy.

For screening decisions, Lawrence and colleagues35 developed a mammography DA for European-American and Mexican-American women. They pretested a DB in 56 healthy female volunteers older than 49 years. Twenty-eight were mainly English speaking, and 28 mainly Spanish-speaking women. The authors reported that the DB provided evidence about the risks of screening and that it was comprehensible across a wide spectrum of educational levels. The authors described the DB as a visual aid that consisted of graphics and written material describing clinical alternatives, outcomes, and probabilities. As reported in previous studies, the DB was initially empty and then interactively assembled by a trained administrator with one or more participants. The authors reported that the interaction was a key feature of the decision aid and it included probes to ensure comprehension and encourage spontaneous questions. When completed, the DB visually summarized the choices, outcomes, and probabilities. The DB had four parts: (1) "introduction," which provided information about breast cancer and mammography; (2) "choice," which presented detailed information about logistics of screening (cost, procedure, description, discomfort, time needed), potential risk of false-positive and false-negative results, and other options (i.e., self-examination, examination by health professional, or no surveillance); (3) "chance of developing breast cancer," which provided age-based probabilities for developing breast cancer over 10 years; and (4) "outcomes of breast cancer," which provided the probabilities of dying from breast cancer and recurrence after 10 years with and without screening, the absolute risk reduction with mammography, and the possibilities of surgery, chemotherapy, and metastatic disease.

Outcomes

Validity

Lawrence and colleagues35 assessed validity with qualitative and quantitative testing. Comprehension of the information was reported as excellent, based on responses to standardized probes. For the quantitative testing of validity, the authors used a similar process to that previously reported. They hypothesized that, if women understood the information, they should reverse their decisions when risk or benefit information was reversed. Twenty-two of 28 women changed their preference as predicted with changed probabilities.

Reliability

The authors reported that all 28 women participated in a second intervention 1 to 2 weeks after the first DB presentation. They all made the same decision in the second session: 26 chose mammography and 2 did not choose mammography.

Feasibility

Not reported.

Leukemia

Sebban and colleagues32 evaluated a DB about leukemia treatment options; participants chose between bone marrow transplant (BMT) and chemotherapy. The DB included visual aids and written material. It provided detailed information on the patient's options, outcomes, probabilities of outcomes and their meaning, and quality of life associated with treatment choices and potential outcomes. The authors reported that they developed several scenarios. The sequence of scenarios was as follows: (1) a scenario introduced patients to the specific clinical problem and to the context of the choice, (2) how BMT would be performed, and (3) how the chemotherapy would be administered. There were four outcome scenarios for each treatment option (BMT vs. chemotherapy) describing potentially favorable (i.e., "you will feel better") and unfavorable (i.e., "your bone marrow might have difficulties to work properly") outcomes at different points of time (3 months, and 2 and 5 years). The scenarios emphasized the relevant components of the choice: severity of the disease and uncertainty of its evolution, potential toxicity of BMT, morbidity related to graft versus host disease, the intensive monitoring of BMT patients in contrast with the relatively easy monitoring of chemotherapy patients, and the inability of chemotherapy to cure the disease. The authors described the uncertainty about potential treatment side effects verbally rather than numerically (i.e., "likely" vs. 0.8). The administration of the DB was similar to those reported by others.27-29 The first evaluation of the leukemia DB was performed among 42 healthy volunteers (staff and individuals from a rural donor clinic). The authors reported that participants were randomly allocated to the final version of the DB or to the short version followed by the standard DB version. The short DB version only displayed the survival probabilities without details about the procedures and their outcomes. A test-retest study was completed with 16 of the 42 participants in the RCT.

Outcomes

Validity

The authors assessed validity using four different constructs: (1) systematic manipulations of the scenarios or of the survival probabilities to determine an subject's understanding of the information; (2) comparison of the two DB versions (short vs. standard DB) to evaluate the impact of the quantity of information on the expressed choice; (3) correlation of the preference with the participant's age (the authors hypothesized that older people made more conservative choices; i.e., chemotherapy); and (4) comparison of the satisfaction with the decision across the two DBs (short vs. standard DB). The authors hypothesized that participants using the standard DB would be more satisfied with their choices compared with participants using the short version. The first construct was tested in 16 participants; the authors reported that, in all cases, preferences shifted in the predicted direction. The 11 participants who initially chose BMT changed to chemotherapy when the survival probabilities for the BMT were reduced. All respondents shifted their preference when chemotherapy survival probabilities were increased. The same results were seen in the six participants who chose chemotherapy, but in the opposite direction.

The authors reported that, for the 20 individuals who were exposed to the short version of the DB first, the correlation between preference scores elicited using both versions was 0.7, "suggesting that choices elicited by the two instruments differ." They indicated that "respondents were inclined to choose BMT more often when presented first with the short version...." The authors did not find a statistically significant correlation between the respondent's age and treatment choice. Participants were significantly more satisfied with the standard DB version than with the short version (p < 0.01)

Reliability

Based on a 7-point Likert scale, the authors assessed the strength of preference (1 = "I definitely prefer BMT," to 7 = "I definitely prefer chemotherapy"); the intraclass correlation coefficient was 0.87.

Feasibility

The authors reported that all participants who were approached participated in the intervention. Each interview lasted from 20 to 30 minutes.

Acceptability of DA

Twenty six (61.9 percent) participants rated the DB as "very clear," and 16 (38.1 percent) rated it as "clear enough"; 32 (76.2 percent) rated the amount of information as "about right", and 10 (23.8 percent) felt the amount of information was not sufficient. The authors reported that 73.7 percent felt comfortable with the method of presentation and that three did not feel comfortable with the method.

Decision

The authors reported that 10 of 22 participants (45.4 percent) exposed to the standard DB chose BMT, 11 chose chemotherapy, and one did not choose either of the two options. Thirteen of the 20 (65 percent) exposed to the short version chose BMT, and the rest chose chemotherapy.

Other Results

The authors asked participants to indicate who should make decisions about treatment; 54.8 percent felt that it should be the patient and relatives, and the rest reported that it should be a shared decision between the patient (with or without the relatives) and the physician.

Ovarian Cancer

Elit and colleagues in two different studies33,34 evaluated the use of a DB for ovarian cancer patients. The authors asked participants to choose between two different treatments: cisplatin + cyclophosphamide (Plan A) vs. cisplatin + paclitaxel (Plan B). In the first study,33 37 healthy female volunteers and 11 women who had completed first-line chemotherapy for ovarian cancer were included. For the second study,34 13 newly diagnosed patients with stage III or IV epithelial ovarian cancer expressed their preference for one of the two treatments, but the patients did not undergo the treatment chosen.34 The authors reported that the DB was written in English, in a narrative format, and with a readability index at a grade 7 level. All outcome information was provided using a mixed frame (positive and negative), and the names of the chemotherapy agents in each treatment plan were not disclosed ("Plan A included two cancer-fighting drugs"). The DB was divided into five parts: "introduction," "treatment choices," "side effects," "outcome," and "description of the outcome." The introduction described information about the disease and the need for chemotherapy. The treatment option section described the administration of the two chemotherapy combinations; bar graphs were used to portray the rates of the most common side effects. The outcomes (cure, recurrence) were presented as probabilities on a wheel and described by narrative material. ("All tests and examinations over the next 5 years show that you have no cancer. ...Even though... you are cancer-free, from time to time you may worry about the cancer coming back.") The administration was similar to that described previously.27-29 The authors designed the DB for administration by a physician or a research assistant.

Outcomes

Validity

Validity was assessed in a similar manner to that reported in the other studies. The authors reported that the predicted shifts in treatment choice occurred 95 percent of the time, and that the strength of preference for Plan B was statistically higher than for Plan A (p < 0.001).

Reliability

The authors evaluated the reproducibility of the participant's choices in 10 volunteers; participants were interviewed on two occasions approximately 3 weeks apart. The authors reported that women's treatment choices showed 100 percent agreement on repeated testing (kappa = 1.0)

Feasibility

In the first study, the authors reported that respondents took approximately 20 to 30 minutes to review the DB. For the second study, the authors approached 13 ovarian cancer patients; only one was not interested in reviewing the survival information.

Decision

The authors reported the treatment choice of the 10 female volunteers involved in the test-retest study. One respondent chose Plan A, and nine chose Plan B. In the second study, the authors gave participants 24 hours to make a treatment decision in a hypothetical scenario; four chose Plan A, and eight chose Plan B.

Acceptability of DA

The authors reported that 96 percent of women in the first study group found the DB either easy or very easy to understand, and 89 percent said they would not change the information on the DB. Nine women in the second study recommended using the DB with other patients like themselves.

Other Outcomes

The authors used the State Trait Anxiety Inventory to assess anxiety in both studies. They reported that anxiety did not change appreciably before and after the DB application (p > 0.05), but they did not provide raw data. The authors reported that the 11-item comprehension questionnaire was answered correctly by 96 percent of healthy volunteers, 92 percent of ovarian cancer patients who had completed chemotherapy, and 86 percent of the newly diagnosed ovarian cancer patients.

Audiotape Workbooks (Evidence Tables 4.19 to 4.21)

Three studies reported the development of audiotape workbook interventions.25,43,44 Two of the studies were conducted by Sawka and colleagues,43,44 and the results were published in the same article. The authors developed and pilot tested a DA for women facing surgical treatment for breast cancer. The third study, by Fiset and colleagues,25 reported the development of an audiotape workbook to assist patients with non-small-cell stage IV lung cancer in deciding whether to have chemotherapy in addition to supportive care and radiation therapy.

Sawka and colleagues developed and conducted two pilot studies43,44 of a DA designed to help breast cancer patients decide between lumpectomy plus radiation therapy versus mastectomy. The DA consisted of an audiotape supplemented by a workbook. The intervention contained specific information regarding the two treatment options: the surgical procedures, a description of radiation, side effects of radiation, and appearance after mastectomy (refer to Evidence Tables 4.19a and 4.20a). The workbook was illustrated with color photographs of the treatment effects, figures providing quantitative information, and a value-clarification exercise in which the women could achieve a better understanding of their choices with similar hypothetical cases. The audiotape provided a verbal description of the workbook contents.

Development Process of the DA

Before the DA was tested, a steering committee (surgeons, a radiation oncologist, a medical oncologist, a breast cancer survivor, and two researchers) oversaw all aspects of development and evaluation of the DA. The content was based on a systematic literature review, the steering committee's input, a needs assessment based on focus groups of women previously diagnosed with breast cancer, and women who were at the point of making a decision about breast cancer. A draft DA was reviewed and critiqued by the steering committee, by nine women who had previously participated in the needs assessment, and by two radiation oncologists and four surgeons not associated with the study. The revised DA was also reviewed by two focus groups of oncology nurses.

Evaluation of the DA

The authors performed two pilot studies;43,44 both studies (n = 18 and 10, respectively) included women with stage I or II breast cancer who were making a decision about surgical management. With their input and the comments of four surgeons and seven oncology nurses, a preliminary version of the DA was developed. In the second study, the authors provided a final version to the patients. In both studies, the patients reviewed the DA after discussion with the surgeon about the diagnosis and treatment options. During the first pilot study, a research nurse answered questions.

Outcomes

Validity. Not reported.

Reliability. Not reported.

Feasibility. Not reported.

Clarity of DA

In the first study,43 10/18 patients did not find any sections of the DA unclear. Three reported that the discussion about adjuvant systemic treatment was unclear. This outcome was not addressed in the second study.

Acceptability of DA

In the first study,43 18 patients indicated that the DA helped them to clarify information given by the surgeon, 17 had positive reactions and were very satisfied with the DA, 6 women thought that the photos were useful in decisionmaking, and 5 felt that the DA raised new questions about what the surgeon first explained. Again, this was not addressed in the second study.

Knowledge

In both studies, the authors evaluated the women's knowledge about breast cancer after the intervention with a previously validated 18-item questionnaire. The authors reported the data as the number of incorrect responses per item; 12/18 patients in the first study answered the questionnaire, and all women answered it in the second study. The high number of incorrect responses in the first group led the authors to change the wording of the questionnaire. The results were presented as number of incorrect responses per group, and no comparisons between groups were reported.

Anxiety

Based on the women's self reports, the authors described that among patients in the first study, 13/18 found themselves "unaffected," and four "relieved" after using the DA. No information was given about anxiety in the second study.

Decisional Conflict

After the intervention, both groups completed the O'Connor's Decisional Conflict Scale.48 The mean score for the first group was 1.8; and for the second, 2.2. The authors reported that the results were consistent with respondents who make decisions after being informed.

Fiset and colleagues25 developed a DA to assist patients with non-small-cell stage IV lung cancer in deciding whether or not to have chemotherapy in addition to supportive care and radiation therapy. The intervention consisted of an audiotape and a companion workbook. The audiotape provided a description of lung cancer and its stages, the functional impact of the disease, and the treatment options available to the patients (i.e., chemotherapy, radiation therapy, and supportive care). The audiotape detailed the risks and benefits of chemotherapy and then discussed the steps in the decisionmaking process. The accompanying workbook consisted of a booklet containing information similar to the audiotape and a worksheet. The worksheet was used by patients to clarify personal values regarding the advantages and disadvantages of chemotherapy, identify any remaining questions, identify patient preference for participation in the decisionmaking process, and finally to determine patient treatment preference.

Development Process of the DA

The content of the audiotape workbook intervention was based on the Cancer Care Ontario practice guidelines for the treatment of stage IV non-small-cell lung cancer. The DA intervention incorporated the Ottawa Decision Support Framework : (1) to provide information on options and outcomes to improve patient knowledge, (2) to provide guidance and examples of the decisionmaking process, and (3) to help to clarify a patient's personal values.49 Before the DA was tested, a development panel (two medical oncologists, two decisionmaking researchers, and two oncology nurses) reviewed each draft of the DA intervention. The draft DA was then reviewed and critiqued by a practitioner panel (all the medical and radiation oncologists in the region who treated lung cancer patients).

Evaluation of the DA

The authors tested the acceptability of the audiotape workbook with stage IV non-small-cell lung cancer patients (n=6) who had already made a decision about chemotherapy (four accepted and two declined). The authors also conducted a cross-sectional mail survey of thoracic surgeons and respirologists currently involved in the treatment of lung cancer patients to investigate physician opinion of the DA intervention.

Outcomes

Validity. Not reported.

Reliability. Not reported.

Feasibility. Not reported.

Acceptability of DA

All of the patients (n=6) found the audiotape workbook acceptable in terms of amount, length, clarity, and appropriateness. Three of the patients found the absolute survival benefit of chemotherapy upsetting (all three were currently on chemotherapy). Of the 29 physicians who reviewed the DA and completed the survey, 25 agreed with the clinical practice guideline on which the intervention was based. Nineteen physicians responded that they would be comfortable giving the DA to their patients, six physicians were neutral, and four physicians stated that they would be uncomfortable giving the DA to their patients.

Other Outcomes

The authors also assessed the usefulness of the DA intervention. All of the patients (n=6) responded that the audiotape workbook would be useful for patients faced with this decision. Of the 28 physicians who answered this part of the survey, 17 felt that the DA would complement their usual approach; 14 stated that the DA would be easy to use in their practice; and 21 responded that they would be very likely, likely, or somewhat likely to use the audiotape workbook in the future.

Time-Tradeoff Method + Videotape + Brochure (Evidence Table 4.22)

Unic and colleagues30 assessed a time-tradeoff (TTO) method to elicit preferences for breast cancer prevention strategies, specifically for prophylactic mastectomy (PM).

Development Process of the DA

Not reported.

Evaluation of the DA

The authors evaluated the intervention in 44 women suspected of carrying the BRCA1 and BRCA2 mutation. The TTO was used as part of a "shared-decisionmaking program" that the authors developed. This program included three or four sessions. During the first session, participants received information (videotape and pamphlet) to be studied at home and the preference assessment was administered. In the 32-minute videotape called, "Breast Cancer in the Family," high-risk women, their partners, and relatives talked about the physical and psychological aspects of PM and breast cancer screening (BCS). The brochure contained information about breast cancer and risk of breast cancer for BRCA1/2-mutation carriers, BCS, PM, and breast reconstruction. In session two, the first replication of the preference assessment was done. In the third session, the second replication of the preference assessment was carried out; and in the forth session, the authors reviewed the decision analytic results and provided advice to the participants.

The TTO offered the women two alternatives: inferior health (x) for a certain number of life years (with PM or BCS), followed by death; and the best-ranked health state (t) for a number of life years, followed by death (with PM or BCS). The authors began by presenting x equal to t to check the understanding of the TTO questions. Next, x was set equal to zero. Then, "x was varied via bisection until the woman was indifferent between the alternatives." Four tradeoffs were done for different values of t, depending on the woman's age. The authors did not provide information about the sources they used to build the TTO.

Outcomes

Validity

The authors compared the women's preference for PM using the TTO with a rating scale (RS) to evaluate convergent validity. The authors reported that four women (proven carriers) chose PM as the most preferred health state; and 47 of 51 women ranked BCS as the most preferred health state. The comparison of the two instruments was done using results obtained in the second preference assessment. The correlation coefficient between the TTO values and the RS values was 0.5; RS values were significantly lower than TTO values (p < 0.001).

Reliability

This outcome was evaluated in 46 women. The authors reported that Pearson's correlation coefficient between the RS values for the two first sessions was 0.67; however, the TTO values coefficient was obtained with pooled data (the first two and last two sessions).

Feasibility

The authors reported that 42 women were able to complete the TTO and the RS on two occasions. Three women completed both preference assessments on only one occasion. Nine women completed the TTO a third time because of discrepancies between the TTO values obtained in the first and second sessions.

Acceptability of DA

The RS was judged as "rather easy" and the TTO as "rather difficult;" the mean score for this judgment was statistically significant (p < 0.0001). Both instruments were rated as "very sufficient" in their "ability to express own values" (p = 0.25). "Feeling relaxed during" the TTO and RS was rated "very sufficient" (p = 0.07). For more details, see Evidence Table 4.22b.

Summary

In this chapter, we reviewed 22 published studies that described the developmental process of DAs. Nineteen DAs were developed for treatment decisions, two for screening, and one for prevention. The type of DA most frequently described was the DB (eight studies), followed by interactive computer programs (seven studies: four CHESS and three other), and audiotapes complemented by workbooks (three studies: two with same DA and one other). Across all interventions, written or verbal information was provided with the aid of visual or graphic displays. Most DAs focused on breast cancer; followed by prostate, lung, and ovarian cancer. All study participants were adults from developed countries. Only two DAs were developed for members of special populations (Spanish-speaking Mexican-American women and low-income African-American women).

The rationale for the development of the DA was reported in 10 studies. In seven studies, the authors reported that the DA was developed to promote shared decisionmaking, although there was insufficient description of this process. In three studies, the DA was developed to support an "informed" process of decisionmaking.

In most studies, some psychometric properties, such as reliability, validity, feasibility and clarity of the DA, were measured. Reliability was determined by focusing on the stability of participants' decisions (eight studies). Construct validity was usually measured by altering the information presented to determine whether users' preferences shifted in the expected direction (seven studies). Twelve studies measured acceptability, and 10 studies assessed feasibility. In contrast, for the outcome measures (e.g., knowledge, satisfaction, or decision), there was an almost complete lack of reporting on why particular outcome measurement tools had been selected or their psychometric properties.

Overall, the studies in this chapter followed similar steps or phases in the process of developing a DA. In the initial phase, the investigators determined why an intervention was required and exactly what was needed. During the second stage, the authors gathered scientific and nonscientific information to be presented in the DA. Information contained in the DA was derived from a variety of sources: research evidence, input from multidisciplinary experts, and, less frequently, patient chart reviews. In general, information consisted of the nature and prognosis of the disease; description of the options for prevention, screening, or treatment; as well as the positive and negative (physical, social, emotional) outcomes associated with the options. Some DAs also included plans regarding implementation of the decision, managing negative outcomes, and links to support groups or other sources of information. Once the information was collected, most authors attempted to determine its relevance through surveys, interviews, or focus groups with experts (e.g., clinicians, epidemiologists, surgeons) and former patients. The fourth phase of the development process involved the actual construction of the DA. After constructing the DA, the final phase of the development process involved evaluation. Generally, in most studies, the newly constructed DA was first pilot-tested among different focus groups such as health care professionals, healthy volunteers, former cancer patients, or patients with other illnesses. Subsequently, DAs for treatment decisions were tested with actual cancer patients at the point of decisionmaking (two studies) or previous patients (three studies). For decisions related to prevention, the DA was tested in women at risk of cancer (one study). In the two studies about screening decisions, healthy volunteers were used.

In summary, most of the studies involved a thorough systematic process for the development of DAs. However, only half of the studies described the rationale for the development of the DA and, consequently, there was sizeable variation in the types of outcome measures selected. Surprisingly, given the amount of time and effort needed to develop a DA, few investigators to date have conducted followup studies to evaluate the effectiveness of the DA in controlled studies. In future studies, investigators should provide documentation of the process used to develop a DA, including the rationale for the choice of DA and outcome measures as well as the psychometric properties. In studies about the development of DAs for treatment, the aid should be field-tested with cancer patients. To date, few DAs have been developed for special populations, resulting in a gap in knowledge about their needs. Finally, as discussed in subsequent chapters, once a DA has been developed and tested, more studies need to be conducted to evaluate the effectiveness in actual clinical situations.

Views

  • PubReader
  • Print View
  • Cite this Page

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...