This publication is provided for historical reference only and the information may be out of date.
Summary of Main Findings
Key Question 1. What Definitions are Used for the Adult PHE in Studies of its Value?
Two central elements used to define the PHE were a) the clinical history and risk assessment of patients, and b) the performance of a physical examination. However, the specific composition of these central elements of the PHE varied among studies. For history and risk assessment, the most frequently cited types of history and risk assessment performed were assessment of dietary risk, alcohol and substance abuse risk, tobacco smoking risk, and physical activity. In most cases, the physical examination was referred to with no specific clarification of what components were included. When specific components of the physical examination were specified, the most frequently cited components were assessment of blood pressure, assessment of weight, assessment of height, breast examination, gynecological examination, and rectal examination.
Key Question 2. What is the Evidence that a PHE, Delivered at Different Patient Ages or Different Frequencies, is Associated with Benefits Compared to Care Without a PHE?
A summary of study designs assessing outcomes, the strength of the best available evidence assessing each outcomes and the direction of the evidence pertaining to each outcome is contained in Table 9.
Delivery/receipt of clinical preventive services. Among the best available evidence, the PHE consistently improved the delivery/receipt of the gynecological examination/Pap smear, cholesterol screening, and fecal occult blood testing. The strength and consistency of evidence for these outcomes ranged from “medium” (cholesterol screening) to “high” (gynecological examination/Pap smear and fecal occult blood testing). Effects of the PHE were mixed among studies assessing the delivery/receipt of preventive counseling, immunizations, and mammography). The strength and consistency of the evidence regarding these outcomes ranged from “low” (mammography and counseling) to “medium” (immunizations).
Proximal clinical outcomes. One study assessing patient attitudes reported the PHE had a positive effect on patient “worry.” The strength and consistency of the evidence from this study was graded as “medium.” Among the best available evidence, the PHE had mixed effects on disease detection, health habits, blood pressure, serum cholesterol, and BMI. The strength and consistency of the evidence assessing these outcomes ranged from “low” (serum cholesterol) to “medium” (disease detection, health habits, health status, blood pressure, and BMI).
Distal clinical and economic outcomes. Among the best available evidence, the PHE had mixed effects on costs, disability, hospitalization, and mortality. The strength and consistency of the evidence ranged from “medium” (costs, disability, mortality) to “high” (hospitalization).
Key Question 3. What is the Evidence That a PHE, Delivered at Different Patient Ages or Different Frequencies, is Associated With Harms Compared to Care Without a PHE?
We identified no studies focused on the delivery of non-recommended preventive services or the inducement of poor health outcomes as a result of the PHE.
Key Question 4. What System-based Interventions Improve the Receipt or Delivery of the PHE?
Among the best available evidence, two interventions (scheduling of appointment for the PHE and offering a free PHE) improved delivery of the PHE with medium to large positive effects. The strength and consistency of the evidence assessing this outcome was “medium.”
Limitations of the literature studied and this review deserve mention. First, we used comparative studies of the effect of the PHE on clinical outcomes to assess the ways in which the PHE is defined. Given that the studies did not set out to define the PHE themselves, this may represent a suboptimal approach. It is possible qualitative assessment of definitions of the PHE obtained through interviews of health care providers or patients with a vested interest in the PHE would reveal perceptions regarding the nature of the PHE that are different from our findings. Second, there were few large-scale randomized controlled trials assessing the effect of the PHE on the receipt of clinical preventive services and outcomes. The largest trials to directly assess the effect of the PHE on clinical outcomes were performed in Medicare demonstration projects in the late 1980's and 1990's, among Kaiser enrollees in the early 1960s, and among residents Southeast London in the late 1960s. Thus, inferences are limited not only to these select populations but are also limited by differences in the timeframe of the studies. Studies performed prior to the first USPSTF guidelines in 1989 were less likely to incorporate clinical preventive services that are most frequently used today and may have implemented clinical preventive services in a way that would be considered inappropriate today, further limiting the generalizability of their results. Despite this limitation, we included these studies in the review because we felt they could provide information regarding benefits of the PHE which might not be explicitly linked to the delivery of currently recommended clinical preventive services. Results of studies performed before 1980 largely mirrored results of more recent studies or yielded neutral results (in the case of long-term outcomes such as mortality). Thus, we do not feel their inclusion substantially altered our main conclusions. While we incorporated observational studies in our review in an attempt to observe effects of the PHE across a variety of clinical settings and in various patient populations as well as to include more recent studies, these studies were often limited by their design (many studies were not specifically designed to assess the effect of the PHE on the receipt of clinical preventive services or clinical outcomes) or their inability to completely account for potential confounding of results.
Heterogeneity in the definitions of the PHE incorporated by studies pose a particularly important limitation in this review. Although we developed a standard definition of the PHE for identification of the PHE in studies, we found substantial differences in the composition of the PHE across studies as well as substantial variation in the degree to which different studies also incorporated interventions to enhance the delivery of the PHE itself (such as patient reminders or physician prompts regarding PHE attendance). This heterogeneity could result in variation in the magnitude and direction of studies' results and hinders drawing broad conclusions regarding the effect of the PHE on a variety of outcomes. For instance, many studies (such as the Medicare demonstration trials) bundled the PHE with other forms of structured counseling (such as nurse-led educational classes). While we attributed changes in outcomes to the PHE delivered in different forms, it is possible changes in outcomes were related to the structured programs themselves and not the PHE. This concern may be particularly relevant when considering studies evaluating the effect of the PHE on patient behaviors, which may be greatly impacted by multifaceted interventions.48 It is possible findings of positive behavior change associated with the PHE could be attributed to interventions delivered in conjunction with the PHE and not the PHE itself. In addition, many studies contained incomplete descriptions of the PHE, making it difficult to ascertain which components of the evaluation contributed most to observed effects of the PHE. It is unclear how well the PHE employed in these studies reflects the PHE as practiced in real-world settings. The PHE was also delivered by various personnel in these studies, further complicating the interpretation of findings. Many studies identified the PHE as an intervention led by nurses or nurse-practitioners while other identified the PHE as involving physician interaction. In some cases, it was unclear if studies intended to assess the feasibility of performing the PHE without substantial physician involvement. If nurse and physician approaches to the PHE are different (particularly in their approaches to counseling or the performance of diagnostic testing), inferences regarding the effect of the PHE could be influenced by these differences. Finally, many studies included an invitation to the PHE as part of the intervention, however, adherence or uptake of the PHE among study subjects was variably achieved. In addition, people attending the PHE may be more healthy than non-participants. The power to detect differences between the intervention group and persons receiving usual care would be limited if studies failed to achieve a meaningful separation in rates of receipt of the PHE between study groups or if participants had low risk of developing outcomes (such as death). Most RCTs did report moderate to high rates of PHE attendance.
Outcomes in some categories were heterogeneous (e.g., the effect of the PHE on several types of counseling was reported across studies), limiting our ability to draw definitive conclusions regarding the effect of the PHE on many outcomes. In some cases, the assessment of outcomes could be biased by their measurement. For example, many studies assessing the effect of the PHE on behavior change assessed behaviors from patient-self report. Measurement of behavior change in this manner could be strongly biased by patient recall. Further, there was little evidence to address the effect of the PHE on many meaningful intermediate outcomes. For example, few studies assessed the effect of the PHE on blood glucose control, diabetes management, or control of other common risk factors. Similarly, while some studies reported on disability, few studies were performed to measure potential enhancements of worker productivity in association with receipt of the PHE. Evidence regarding the cost-effectiveness of the PHE was similarly sparse. As many studies captured direct costs of care associated with the PHE, few captured indirect costs, and we found only one study directly assessing both the costs and effectiveness of the PHE. In addition, many of our outcomes were reported among a few RCTs. The effect of an individual study's design on the direction of multiple outcomes measured within that study could be substantial. This is important, given the heterogeneity of interventions among our studies—it is possible the benefit of the PHE could be overestimated if multiple positive outcomes are reported among a select few studies. Studies reporting on multiple studies may also be limited by lack of power to assess some outcomes, potentially contributing to the reporting of neutral results.
Many studies described the PHE being compared to “usual care” with little or no description of the nature of usual care. This limitation reflects not only lack of specificity within the studies, but a lack of clarity in clinical practice regarding what constitutes “usual care.” Usual care could vary widely, depending on the system of care which is being examined, and could include the delivery of preventive services at specific intervals during short visits or systems which provide reminders to perform prevention at acute visits. Lack of specificity in identifying the components of usual care could significantly affect outcomes, particularly if some preventive services are delivered as a part of usual care.
Limitations in studies assessing the long-term outcomes associated with receipt of the PHE deserve special attention. While assessment of the PHE's effects on long-term outcomes such as hospitalization or death is desirable, the feasibility of isolating the effect of the PHE on these long-term outcomes is unclear, especially given the periodic nature of the PHE and given multiple other episodes of patient care that typically occur outside of the PHE. It is possible that, although patients receive a PHE at baseline, the effect of other episodes of care (such as management of chronic illnesses detected before or after the PHE) have a more powerful effect on long-term outcomes than the PHE itself. It is also possible that the receipt of more frequent PHEs results in improved outcomes over a single PHE, particularly for persons with chronic illnesses who might require more than one visit to adequately address their prevention needs. While many studies evaluated the institution of a PHE for one to two years, others evaluated the effect of a single PHE. It is possible differences in outcomes could be attributed to differences in the intensity of the PHE or the frequency with which patients received the PHE in different studies. It is also possible differences in outcomes could be related to differences in the burden of comorbid illnesses among participants of different studies.
Our review is also subject to potential publication bias, in that investigators may have been more likely to publish articles reporting the PHE improved outcomes. A lack of enough RCTs assessing the effect of the PHE on several outcomes prohibited a formal analysis of publication bias, however. In addition, all articles reported on benefits of the PHE and none specifically studied the inducement of harms associated with the PHE. Lack of evidence on harms may reflect not only difficulty in collecting this information for some outcomes but also a bias on the part of researchers toward publicizing the benefits of the PHE. While the inclusion of observational studies in this review allowed for the ascertainment of the effect of the PHE across a more broad group of populations than did the RCTs alone, these studies are more subject to residual confounding of results that were incompletely accounted for in analyses, potentially enhancing the probability of positive findings.
Finally, we assigned grades regarding the strength and consistency of the evidence pertaining to each outcome in an effort to provide readers with information regarding the confidence with which inferences regarding summary results can be drawn. However, one tenet of the GRADE framework we used to guide our assessments is that the RCT represents the highest level of evidence to assess any one outcome. While we agree the RCT represents the ‘gold standard’ approach to assessing the effect of interventions in while minimizing sources of bias and unobserved confounding, institution of the RCT to assess system-level interventions may not always be feasible. Thus, it is possible our grade of evidence pertaining to studies of system interventions to improve the receipt of the PHE (Key Question 4) is artificially low.
Recommendations for Future Research
While the available evidence reports on the effect of the PHE on the delivery/receipt of some clinical preventive services, it does not report on the effect of the PHE on the delivery of recommended versus non-recommended clinical services. Similarly, little evidence is available to discern the effect of the PHE on clinical harms (e.g., potential increase in patient complications from inappropriate testing). Studies specifically designed to assess whether the PHE could encourage delivery of inappropriate clinical preventive services or enhance the potential for harms inflicted on patients as a result of such inappropriate care could shed important light on ways in which the PHE should best be implemented.
Little evidence is available to ascertain whether the PHE improves intermediate clinical outcomes such as disease management (e.g., blood pressure or glucose control) or changes in worker productivity. The evidence is also sparse with regard to the PHE's effect on the incidence of clinical morbidity (e.g., cardiovascular disease, cancer). In addition, many studies evaluating proximal clinical outcomes followed patients for short time periods, which may not have provided ample enough opportunity to capture long-term changes in proximal clinical outcomes. While the best available evidence is largely neutral with regard to the effect of the PHE on mortality, it is possible the PHE could have an effect on more proximal outcomes, thus potentially leading to improvements in patients' quality of life. Work to elucidate the magnitude and duration of effects of the PHE on more proximal clinical outcomes, including potential enhancements in worker productivity may also help clarify the potential role of the PHE in affecting health care utilization and costs.
Studies reporting on the effect of the PHE on costs of health care reported primarily on direct costs of clinical care, with little focus on the effect of the PHE on indirect health care costs (e.g., potential cost savings associated with less time lost due to premature morbidity, mortality and illness) or the cost-effectiveness of the PHE. Work more fully elucidating the effect of the PHE on both direct and indirect costs may help health care practitioners and policy makers assess the economic value of the PHE more effectively. Cost effectiveness models are needed to more fully understand the complex interplay of induced costs associated with preventive services offered as a result of the PHE as well as reduced costs associated with potentially improved management of chronic illnesses and potential improvements in quality of life which could occur as a result of the PHE.
Although some studies reported on the effect of the PHE on patient health habits, we identified no studies reporting on whether the PHE could affect patients' motivations to change, self-efficacy, or adherence to continuous care. Work to elucidate the PHE's effect in these areas would help to clarify mechanisms through which the PHE could improve both proximal and distal clinical outcomes.
While some evidence is available regarding the effect of the PHE on patient attitudes, we found no evidence regarding the potential effect of the PHE on patient knowledge of clinical guidelines, health care system use, or the patient-physician relationship. As consumer-driven health care is increasingly touted as a mechanism through which health care costs could be contained and greater patient satisfaction could be achieved, research to identify the effects of the PHE on patient knowledge and health care system use could prove valuable.85, 86 In addition, the patient-physician relationship is increasingly reported as important in affecting patient satisfaction, adherence to clinical recommendations, and receipt of appropriate clinical care.87–89 Work to determine whether the PHE enhances or detracts from the quality of the patient-physician relationship could be very important in guiding future clinical practice.
The available evidence does not address whether the implementation of preventive services in the context of the PHE results in improved public health outcomes such as communicable disease containment or improvements in family health. Such outcomes represent the potential for broad societal benefit of the PHE's strong focus on risk assessment and disease prevention. While studies of these outcomes may be difficult to perform, work employing modeling techniques to estimate the potential benefits or harms of the PHE for society could prove fruitful for health care policy makers and public health practitioners.
In addition, the evidence did not address in a systematic way the frequency and intensity of the PHE required to achieve potential improvements in clinical outcomes, nor did it assess ways in which the content of the PHE should change for persons of different age groups. Work is needed to ascertain the effects of both the frequency of the PHE (as opposed to a single visit) on outcomes as well as whether tailoring the PHE for persons at different levels of risk would be beneficial. Few studies addressed the persistence of the effect of the PHE, which may be short-lived, particularly if it is delivered only once. It is also unclear if the effect of the PHE would change based on the type of clinician delivering the PHE (i.e., physicians versus nurses or physicians of different clinical specialties) and the resources available to clinicians implementing the PHE. The potential role of the electronic health record in enhancing the delivery of the PHE could provide insight to mechanisms through which the PHE might be delivered more efficiently.
Finally, a paucity of studies evaluated interventions to improve the receipt of the PHE. Performance of additional, well-designed studies is needed to strengthen the evidence for or against such interventions.
The best available evidence suggests delivery of recommended clinical preventive services, patient attitudes, and patient health status are improved by the PHE and may be more directly affected by the PHE than other proximal clinical outcomes or long-term financial and clinical outcomes. Given that it may be impossible to entirely isolate the effect of receipt of the PHE on intermediate clinical outcomes which require ongoing management such as blood pressure or long-term outcomes such as mortality, studies linking the PHE with improved delivery of recommended clinical services may provide the best evidence of its value. Since appropriate implementation of currently recommended clinical preventive services has been demonstrated to improve health in evidence which provides the basis for USPSTF recommendations, findings of increased delivery of preventive services in the setting of the PHE may provide adequate justification for implementation of the PHE. Indeed, if the PHE, instituted in some standard fashion, could be consistently demonstrated to improve the delivery of several recommended clinical preventive services across a variety of settings, the value of the PHE might be substantial. This hypothesis assumes, however, that combining multiple evidence-based preventive services in the context of the PHE has additive benefits and that delivery of the same preventive services during other types of office visits (e.g., visits for management of chronic illnesses) would not be as beneficial. While achieving consistency in the definition and delivery of the PHE stands as an important remaining challenge, efforts to clarify the underlying long term benefits (or harms) of receiving multiple clinical preventive services in the context of the PHE versus other types of ambulatory care visits are needed to fully clarify the PHE's value.
Mechanisms through which improvements in care attributed to the PHE occur are unclear, as studies were so heterogeneous in terms of the content of the PHE and their institution of additional interventions to enhance delivery of the PHE as to prohibit formal analysis in this regard. The PHE may provide clinicians, who are routinely pressured to deliver care in short intervals of time, time to consider preventive care more fully, thus leading to their institution of preventive measures more frequently. Given the heterogeneity of studies, it is unclear if differences in the effect of the PHE on the delivery of different preventive services represents differences in studies reporting on different preventive services, or if differences are related to the preventive services themselves. It is possible the PHE has a stronger effect in improving the delivery of preventive services which are performed by clinicians at the time of the office visit (such as gynecological examinations/Pap smears or fecal occult blood testing) when compared to preventive services which require patients to schedule appointments outside of the initial office visit for the PHE (such as mammography).
Improvements in patient worry (one study) and health status (one study) associated with the PHE may provide insight to reasons patients and clinicians have persisted in implementing the PHE despite evidence to conclusively support its use as well as why the PHE may be associated with enhanced delivery of clinical preventive services. Elimination of worry or concern regarding possibly undetected illnesses or prevention of illnesses which has not yet occurred may represent a powerful motivator for action on the part of patients. The PHE, in providing an opportunity for both patients and physicians to contemplate potential risks, may provide a vehicle through which worries can be more fully elucidated from patients and addressed through completion of the evaluation. Evidence reflecting improvement in self reported health status may reflect the provision of time for physicians to consider patients' needs in greater entirety and may allow physicians to address less frequently assessed aspects of health (e.g., depression and functional status).
Several unanswered questions remain regarding the circumstances under which the PHE may provide the most benefit. Studies are needed to ascertain the frequency and intensity of the PHE needed to consistently improve outcomes (with study of precisely which components of the PHE are necessary), the patient populations that could benefit most from the PHE, and systems of care in which the PHE might be best delivered. Work is also needed to more adequately assess the potential benefit of the PHE on patient attitudes and patient health status as well as to assess whether the PHE could encourage the delivery of inappropriate clinical services or inflict harm on patients. Work to ascertain mechanisms for differential effects of the PHE on delivery of different clinical preventive services, to identify whether the PHE consistently improves intermediate clinical outcomes, to characterize the effect of the PHE on the patient-physician relationship, and to assess the effect of the PHE on broad societal outcomes such as disease containment will contribute greatly to knowledge regarding the value of the PHE.
The design of future studies to more completely assess the value of the PHE as it is currently delivered will require careful attention. While observational studies leave open the possibility for inadequate adjustment for potential confounding or bias in findings, larger randomized controlled trials should incorporate study populations which are generalizable to the majority of patients seeking health care in the U.S., including persons of a variety of ages, women, persons of diverse ethnicity and race, and persons utilizing different health plans. In addition, such studies should seek to carefully and clearly define systems of “usual care” with which the PHE is to be compared, to measure the degree to which both intervention and comparison groups comply with assignments to receive the PHE, and to capture outcomes in a standardized way. Large scale trials could be costly and may be unable to adequately capture long-term effects of the PHE on outcomes such as costs and mortality, as these outcomes could be influenced by multiple factors, including the degree to which individuals seek health care for other reasons such as the management of chronic illnesses. For this reason, the development of computerized models (incorporating evidence identified in this review, evidence from future studies, and existing evidence regarding the long-term value of preventive services delivered in the context of the PHE) to simulate trajectories of quality of life, the development of morbidity and mortality as well as direct and indirect costs incurred or saved as a result of the PHE could be most helpful in clarifying the value of the PHE.
Agency for Healthcare Research and Quality (US), Rockville (MD)
Boulware LE, Barnes GJ II, Wilson RF, et al. Value of the Periodic Health Evaluation. Rockville (MD): Agency for Healthcare Research and Quality (US); 2006 Apr. (Evidence Reports/Technology Assessments, No. 136.) 4, Discussion.