Chapter  136:  Value of the Periodic Health Evaluation

Introduction

The periodic health evaluation (PHE) consists of one or more visits with a health care provider to assess patients' overall health and risk factors for preventable disease, and it is distinguished from the annual physical exam by its incorporation of tailored clinical preventive services and laboratory testing as part of health risk assessment. By promoting prevention, management of chronic conditions, and enhancing the patient-provider relationship, the PHE may improve patient outcomes and the public's health. However, it could also induce unnecessary costs and patient harms by promoting the use of non-recommended services. Early studies of the PHE, performed before the adoption of current preventive services guidelines, were costly and demonstrated minimal improvement in clinical outcomes, leading to concern regarding the PHE's value and to the promotion of episodic, targeted delivery of preventive services in the context of ongoing clinical care. More recent clinical trials have reported scattered benefits of the PHE. Thus, despite its continued practice, the value of PHE in improving health and healthcare costs has been largely unclear.

Private and public health insurance coverage for preventive services in the U.S. has gradually increased over time. However, increases are typically for one recommended service at a time, rather than a comprehensive set of preventive services. Recent legislation will provide coverage for a “Welcome to Medicare Visit” for new enrollees, incorporating a range of diagnostic and screening tests. Lack of clear evidence to support or refute its use, and recent legislation to cover preventive services on a wide scale provide the basis for this systematic review of the evidence to elucidate the value of the PHE.

Methods

The American College of Physicians posed preliminary questions regarding the PHE. We convened a panel of three internal and eight external technical experts to provide input into the refinement of questions to be addressed. We also recruited peer reviewers representing stakeholder organizations to give feedback on the draft report.

We address the following Key Questions concerning the value of the PHE for adults:

• 1

  What definitions are used for the adult PHE in studies of its value?

• 2

  What is the evidence that a PHE, delivered at different patient ages or different frequencies, is associated with benefits (i.e., improved outcomes) compared to care without a PHE (e.g., usual care)?

  Outcomes include:

  • a

    Delivery of recommended clinical preventive services.

  • b

    Patient attitudes/perceptions (e.g., knowledge, satisfaction).

  • c

    Behavioral outcomes (e.g., tobacco cessation, adherence).

  • d

    Proximal/intermediate clinical outcomes (e.g., cholesterol lowering, disease management).

  • e

    Distal clinical outcomes (e.g., death, or myocardial infarction).

  • f

    Economic outcomes (e.g., cost savings, health care utilization).

  • g

    Public health (e.g., improvements in family and community health).

• 3

  What is the evidence that a PHE, delivered at different patient ages or different frequencies, is associated with harms (i.e., worse outcomes) compared to care without a PHE (e.g., usual care)?

  Outcomes include:

  • a

    Delivery of non-recommended clinical preventive services.

  • b

    Patient attitudes/perceptions (e.g., worry).

  • c

    Behavioral outcomes (e.g., continuation of risky behaviors).

  • d

    Proximal/intermediate clinical outcomes (e.g., complications from testing).

  • e

    Distal clinical outcomes (e.g., events such as death).

  • f

    Economic outcomes (e.g., induced costs, increased health care utilization).

  • g

    Public health (e.g., declines in family and community health).

• 4

  What system-based interventions improve the receipt or delivery of the PHE (e.g., insurance premium reductions or provider reminders)?

We searched MEDLINE^®, the Cochrane databases, and the Cumulative Index to Nursing and Allied Health Literature (CINAHL), and we hand-searched 24 journals and bibliographies from pertinent articles through February, 2006. We used pre-specified, standard criteria to select studies. Pairs of reviewers screened the literature for relevant article titles. For articles promoted to abstract review, two investigators independently reviewed abstracts and excluded them if they: 1) had no useful information applying to the Key Questions; 2) were not written in English; 3) included subjects only 18 years or younger ; 4) contained no original data; or 5) had no comparison group. Titles and abstracts were promoted to further review if either of two reviewers did not exclude them. For articles promoted to final review, two reviewers sequentially performed full data abstraction for each article, including information on study design, location and setting, dates of performance, follow up length, enrollment, eligibility criteria, participant characteristics, components of the PHE, interventions, and outcomes.

Two reviewers independently judged individual studies' quality on several aspects of external and internal validity, including descriptions of: inclusion/exclusion criteria; subjects' baseline characteristics; handling of withdrawals; the intervention; adequacy of length of study follow up; outcomes; randomization and blinding (for RCTs); and the statistical analysis. At the completion of the article review, we summarized the magnitude of effects in RCTs by reporting Cohen's d (95% confidence interval (CI)) for mean differences and proportions. We considered effect sizes ranging from 0 to 0.25 to represent “small” effects, from 0.26 to 0.8 to represent “medium” sized effects, and greater than 0.8 to represent “large” effects. We also graded the quantity, quality and consistency of the “best available evidence” (comprised of studies least likely to present biased findings) addressing Key Questions 2 through 4 by adapting an evidence-grading scheme recommended by the GRADE Working Group (classifying bodies of evidence pertaining to each outcome as “high”, “medium,” “low,” or “very low” grade). Evidence grading incorporated assessments of studies' quality, consistency in the direction of reported results for an outcome, sparseness of data, probability of bias, and reported strength of association between the PHE and outcomes.

Results

We screened 6523 articles for eligibility at the title review level and reviewed 2021 at the abstract level, and 819 at the article inclusion/exclusion level. Of these, 54 articles were promoted for full review, representing 36 studies reporting multiple outcomes or follow up dates. All studies addressed Key Question 1, 36 studies addressed Key Question 2, no studies addressed Key Question 3, and five studies addressed Key Question 4.

Identified Studies

Table 1

Characteristics of Studies Eligible for Inclusion in (more...)

Table 1

Characteristics of Studies Eligible for Inclusion in the Review (N = 36)

Characteristics		n (%)
Study design	Randomized controlled trial	11 (31)
	Non-randomized controlled trial	1 (3)
	Prospective cohort	2 (5)
	Retrospective cohort	5 (14)
	Cross-sectional	14 (39)
	Pre-post comparison	3 (8)
First year study conducted	1940-1969	4 (11)
	1970-1989	9 (25)
	1990-2005	19 (53)
	Not specified	4 (11)
Country where study conducted	United States	25 (69)
	United Kingdom	3 (8)
	Canada	3 (8)
	Japan	2 (5)
	Taiwan	1 (3)
	Denmark	1 (3)
	Sweden	1 (3)
Study setting	Urban	17 (47)
	Suburban	8 (22)
	Rural	5 (14)
	Unclear	18 (50)
Delivery site for PHE	Ambulatory practice office	16 (44)
	Academic practice	7 (19)
	Resident/housestaff clinic	5 (14)
	Hospital outpatient clinic	4 (11)
	Community health center	1 (3)
	Health checkup/physical exam clinic	3 (8)
	Military (not otherwise specified)	1 (3)
	Employer health clinic	1 (3)
	Community center	1 (3)
	Not applicable (observational study with survey design)	8 (22)
	Not specified	4 (11)
Physician specialty delivering PHE	Family medicine practice	10 (28)
	Internal medicine practice	7 (19)
	General Practice	6 (17)
	Multispecialty	1 (3)
	Not specified	4 (11)
	Not applicable (observational study with survey design)	8 (22)
Health plan under which PHE delivered	National health system (non-U.S. studies)	10 (28)
	Medicare	4 (11)
	Employer health plan	4 (11)
	Staff model health maintenance organization	2 (5)
	Other managed care plan	2 (5)
	Veterans Affairs or other U.S. Department of Defense health plan	1 (3)
	Not specified or mixed (surveys, or NOS)	15
Subject of study	Patient
	Not otherwise specified	31 (86)
	Employee or executive	8 (22)
	Medical Providers
	Family practitioners	8 (22)
	General internists	5 (14)
	General practitioners	4 (11)
	Medical trainees (fellows, residents/house staff, medical students)	3 (8)
	Nurses or nurse practitioners	3 (8)
	Internists	2 (5)
	Physicians' assistants	1 (3)
	Health providers, not otherwise specified	1 (3)
	Obstetricians/gynecologists	0 (0)
	Medical specialist/subspecialist physicians	0 (0)
	Physicians, not otherwise specified	0 (0)
	Office staff	0 (0)

PHE = periodic health evaluation.

Percents may not add to 100. Reviewers were able to give multiple answers to many of the questions.

We identified a total of 36 studies containing information applicable to the Key Questions. A description of study characteristics is listed in Table 1. The most common study design was cross-sectional (14 studies), followed by randomized controlled trials (RCTs) (11) and cohort studies (7). Overall, the literature was characterized by complexity and heterogeneity in several dimensions. Studies were conducted over a period of several decades (19 from 1990 and later, 9 between 1970-1989, and 4 before 1970) (Table 1). Practice settings for the studies were also diverse, with 16 studies taking place in private offices, four in hospital outpatient clinics, and seven in academic practices. Studies reflected a range of health plans as well, with four studies in Medicare or Medicaid populations, 10 in non-U.S. national health plans, four in employer health plans, and two in staff-model HMOs. While 25 studies were performed in the U.S., we also identified relevant studies from the United Kingdom, Canada, Taiwan, Japan, Denmark and Sweden.

Limitations

The PHE was described with great heterogeneity, limiting inferences regarding which aspects of the PHE are most influential on outcomes. Few large-scale RCTs assessed the effect of the PHE, with some of the largest trials performed among select populations prior to guidelines of the U.S. Preventive Services Task Force (USPSTF) in 1989, limiting their generalizability. Outcomes in some categories (e.g., counseling) were heterogeneous, limiting inferences regarding these outcomes. Little evidence addressed the PHE's effect on intermediate outcomes (e.g. blood glucose control, diabetes management). Evidence regarding the cost-effectiveness of the PHE was similarly sparse. Many outcomes were reported among a few RCTs, leaving open the possibility that individual study designs heavily influenced the direction of multiple outcomes. The feasibility of isolating the effect of the PHE on long-term outcomes is unclear given the periodic (or one-time) delivery of the PHE in studies and given multiple other episodes of patient care that typically occur outside of the PHE.

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. In addition, observational studies in this review are subject to unaddressed residual confounding of results.

Recommendations for Future Research

Studies are needed to assess whether the PHE could encourage delivery of inappropriate preventive services or inflict harms on patients and to clarify the effect of the PHE on health habits, patient attitudes, health status, other intermediate outcomes such as clinical morbidity or worker productivity, and broad public health outcomes such as communicable disease containment or improvements in family health. Work to elucidate the magnitude and duration of effects of the PHE on outcomes is also needed. Studies elucidating the PHE's effect on both direct and indirect costs, long-term changes in quality of life, and clinical morbidity can be used in cost-effectiveness models, which are needed to more fully integrate findings regarding potential harms and benefits of the PHE.

Studies are needed of the frequency and intensity of the PHE required to achieve clinical improvements (or to induce harms) studies also are needed to assess differences in the PHE's effect when delivered in different health care systems or by different providers. Such studies would enhance knowledge regarding mechanisms through which the PHE can be delivered most efficiently.

Additional, well-designed studies are needed to strengthen the evidence for or against system-level interventions to enhance receipt of the PHE.

Conclusions

The best available evidence suggests delivery of some recommended preventive services are improved by the PHE and may be more directly affected by the PHE than proximal or long-term clinical outcomes and costs. It may be difficult 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. Thus, studies linking the PHE with improved delivery of preventive services may provide the best evidence of its value. Since appropriate implementation of 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. While achieving consistency in the definition and delivery of the PHE stands as an important remaining challenge, efforts to clarify the presumed long-term benefits of receiving multiple preventive services in the context of the PHE versus other types of ambulatory visits are needed to fully elucidate the value of the PHE.

Mechanisms through which improvements in care attributed to the PHE occur are unclear. The PHE may have a stronger effect on the delivery of preventive services which are performed by clinicians at the time of the office visit (e.g., fecal occult blood testing) versus preventive services requiring patients to schedule appointments outside of the office (e.g., mammography).

Future studies assessing the value of the PHE should incorporate diverse study populations and should seek to carefully define systems of “usual care” with which the PHE is to be compared, capture outcomes in a standardized fashion, and more clearly assess the PHE's cost-effectiveness. The development of computerized models may be most helpful in assessing the long-term value of the PHE.

The Periodic Health Evaluation

The periodic health evaluation (PHE) consists of one or more visits with a health care provider for the purpose of assessing patients' overall health and risk factors for preventable diseases. The PHE is distinguished from the complete physical examination by its incorporation of tailored clinical preventive services and laboratory testing as a part of health risk assessment. During the PHE, health care providers perform a history and risk assessment in addition to a physical exam. Based on the information gathered by providers, patients may receive counseling, immunizations, lab testing, or arrangements for other preventive health services as part of the evaluation. By promoting appropriate clinical management of chronic conditions, providing patient education, and the patient-provider relationship, the PHE has been hypothesized to improve intermediate and long-term patient outcomes as well as the public's health. Because of its focus on prevention and recommendations for chronic disease management, the PHE has potential to affect patient health and health care cost for the individual, the health care industry, and society as a whole.

Historical Changes in Conceptualization of the Content and Value of the PHE

Since the late nineteenth century, ideas regarding the content and the value of the PHE have continually evolved, reflecting changing views of the medical community and the public toward the role of prevention in health care. In 1861, Dr. Horace Dobell, considered the father of mass screening in the United Kingdom (U.K.) and a physician at the Royal Chest Hospital in England, outlined his basic belief that discovery of a pre-existent disease state could offer a chance for treatment and cure through the detailed examination of the individual. Others supported this notion such as Dr. George Gould, a prominent Philadelphia physician, who offered the “periodic examinations of patients” as an important mechanism through which future illness could be prevented and quality of life could be enhanced.¹

In the early 1900s, motivations for assessing and enhancing individuals' health through the PHE were often financial in nature. The life insurance industry advocated the use of medical histories and periodic physical exams to risk stratify patients for coverage decisions. Studies at the turn of the century, such as those performed by Dr. A.S. Knight of the Metropolitan Life Insurance Company and Eugene Lyman Fisk of the Life Extension Institute, both reported that policyholders undergoing an annual physical exam had lower mortality than would be expected based on actuarial data.¹ Similarly, private industry advocated for comprehensive laboratory and physical exams to insure the health of workers and contribute to productivity, morale, and operating efficiency of the work place.² The physical exam was also often required to attend school, enlist in the military forces, gain employment, and note the early signs of potentially serious diseases.³ However, the central role of physicians in performing the physical exam prompted organized medical agencies to endorse the PHE as an opportunity to establish the physician-patient relationship. Giving his address to Harvard medical students in 1925, Dr. Francis W. Peabody emphasized this relationship by noting, “One of the essential qualities of the clinician is interest in humanity, for the secret of care of the patient is in caring for the patient.”⁴

Despite the potential virtues of the PHE from both a financial standpoint and from the standpoint of the patient-physician relationship, the PHE was not considered standard medical care and lost momentum as a public interest during the 1930s and the Great Depression. Renewed interest in the PHE arose in the mid-1940s with the development of “multiphasic screening” whereby the PHE assumed the objective of mass screening. This approach only minimally involved physicians, and used technology for the detection of unrecognized diseases or defects.¹ While the medical literature at that time continued to emphasize the important role of the periodic health examination, there were few examples of studies objectively supporting the ability of the PHE to promote health and longevity.²

Even with growing popularity of the PHE as standard clinical practice, the emergence of evidence-based medicine in the 1960s raised serious questions concerning the value of the PHE within the medical establishment.⁵ During this period, several studies were conducted to assess the value of the PHE. In the U.S., Collen undertook a study at Kaiser, which followed 10,000 adults through 7 and 16 year follow ups, collecting data on morbidity and mortality. In this study, the authors concluded that periodic health examinations were associated with lower death rates from potentially postponable causes.⁶ In the U.K., two group general practices in South London and the Department of Community Medicine at St. Thomas's Hospitals embarked on an ambitious study to assess the value of introducing a general-practice-based screening service for persons 40 to 64 year old as an extension of the National Health Service.⁷ This study collected data on morbidity, hospital admission rates, certified sickness absence from work, and mortality - ultimately concluding little difference between the screened and unscreened groups. Costs of the screening were also calculated, and the results of this large study ultimately influenced British policy makers against investing in publicly supported multiphasic screening at that time.^{8, 9}

In the 1970s, health care providers moved toward individualizing the PHE. Rather than a single annual exam during which a universal battery of screening and assessment maneuvers were undertaken, the PHE began to be conceptualized as an amalgam of preventive services tailored to individuals' risk profiles. In this manner, the annual physical exam became defined as the comprehensive physical examination which included an extensive history, physical, counseling, and diagnostic testing which was used to determine the patient's baseline health status. In contrast, the periodic health evaluation evolved into one or more visits with the primary emphasis of evaluating and offering preventive health services based the patient's age, gender and risk profile for recognizable and treatable conditions. Redefinition of the PHE in this way prompted another wave of important studies assessing the value of preventive services. Frame and Carlson in 1975 assessed the accuracy of screening measures and their impact on altering disease progression and mortality with regard to 36 major medical conditions.¹⁰ In 1979 the Canadian Task Force on the Periodic Health Examination (now known as the Canadian Task Force on Preventive Health Care (CTF) looked at 78 medical conditions and also assessed the strength of evidence behind screening measures to reduce morbidity and mortality.¹¹ In 1984, the newly established U.S. Preventive Services Task Force (USPSTF) began to evaluate specific preventive interventions and their impact on morbidity and mortality in 60 medical conditions.¹²

These efforts prompted conceptualization of the PHE as an assessment targeted only to the preventive service elements demonstrating an impact on morbidity and mortality. This minimalist approach was endorsed by the American College of Physicians (ACP) and the American Medical Association (AMA) with the caveat that the absence of evidence should not be equated with the ineffectiveness of an omitted screening intervention. These groups also advocated for increased breadth within the PHE supporting the inclusion of counseling and immunization.¹³

National Task Forces to Evaluate Preventive Care and the PHE

Continued Use of the PHE Despite Recommendations

Continued implementation of the PHE, despite the CTF's guidelines eliminating the “routine checkup,” may reflect the significant influence of patient and provider expectations regarding the PHE in clinical practice. According to a study assessing patients' expectations of the PHE, over 90% desired such examinations, most often on an annual basis. Accordingly, patients desired extensive examinations which included laboratory and other procedures which were in excess of CTF guideline recommendations.¹⁹

In a qualitative assessments of healthcare providers' perspectives on the integration of preventive practices during clinical visits, several barriers to delivering care were identified. Barriers within the physician-patient relationship include lack of patient compliance with preventive recommendations, lack of continuity of care, and discordant expectations of patients and providers within the clinical encounter. Studies seem to suggest patients place greater reliance on diagnostic labs and tests than do providers who often use the clinical history and physical to guide their recommendations during periodic health visit.^{20, 21} Health systems barriers which providers believed affected the integration of preventive measures into the clinical visit included lack of time, remuneration, and lack of provider reminders or tools to aid in the receipt of care.

Private Insurance Coverage for the PHE

Numerous studies have demonstrated a positive association between health insurance coverage and the receipt of preventive services. Coverage for preventive services in the U.S. by both private and public payers has gradually increased over time. However, this increase in coverage is usually for one recommended service at a time, rather than a comprehensive set of preventive services.²² In one recent study of employer-sponsored insurance plans by Partnership for Prevention, roughly 80% of plans reimbursed for a general physical examination with health maintenance organizations being slightly more likely to cover this service when compared to preferred provider organizations or point of service health plans.²³ According to the National Health Policy Forum at George Washington University, states rarely mandate insurance coverage for preventive services based on USPSTF guidelines. However, of those preventive services which are mandated, large employer-based health plans remain exempt from such requirements through the Employee Retirement Income Security Act. Thus, there is no consistent policy regarding coverage for preventive health services, including the periodic health examination.

Centers for Medicare and Medicaid Service (CMS) Legislation and the PHE

Under the Medicare Prescription Drug Improvement and Modernization Act (MMA) of 2003, the PHE will be covered for some individuals for the first time in Medicare's history. The Act provides Medicare reimbursement for an initial preventive visit at enrollment into Medicare, as long as the enrollee completes the examination within six months of enrollment.²⁴ The examination will cover a wide range of services, including: medical history; physical examination; counseling; laboratory tests; radiological interventions; and electrocardiograms. The Secretary of Health and Human Services has been granted permission to make some coverage decisions, but many interventions such as bone mass measurement, cancer screening and immunizations are specifically provided for in the MMA. This new legislation reflects intentions and efforts of the U.S. Department of Health and Human Services to combat rising health care costs and improve patient outcomes through the promotion of preventive measures.²⁵

Need for Review of the Evidence on the Value of the PHE

Historical changes in the conceptualization and implementation of the PHE reflect not only the complex and multidimensional nature of the PHE but also the lack of clear evidence to support or refute its continued use. While the PHE could be seen as an explicit opportunity outside of compressed symptom-based ambulatory visits for clinicians to implement recommended clinical preventive services (particularly for preventive interventions which require more time to perform or advance planning), it is unclear whether any improvements that could be gained from the implementation of the PHE would be justified by increased costs or harms which may be associated with its implementation. At the same time, it is unclear if symptom-based visits allow adequate time for clinicians to address all age-specific recommendations for prevention or behavioral issues such as smoking and diet (which are substantial contributors to the development of many of the most costly chronic illnesses burdening patients today).^{26, 27} We therefore performed this comprehensive review of the extensive literature to elucidate the value of the PHE and ways in which the PHE could be improved. The review was intended to provide an evidence basis which patients, health care providers and health policy makers can use to guide future clinical practice.

Recruitment of Technical Experts and Peer Reviewers

At the beginning of the project, we recruited a panel of internal and external technical experts to give us input on key steps including the selection and refinement of the questions to be examined. The panel included three internal technical experts from the Johns Hopkins University who had expertise in various aspects of the periodic health exam and eight external experts who had interests in the periodic health exam (see Appendix A ^a). In addition to this panel of technical experts, we recruited a group of peer reviewers to examine a draft of the evidence report, as described further in the section on Peer Review. This group included representatives of organizations or agencies having different perspectives on the topic. We also sought input throughout the project from representatives of the ACP.

Key Questions

We worked with the ACP and technical experts from the CMS, the Agency for Healthcare Research and Quality (AHRQ), and academic and clinical centers (including practicing internists and family physicians) to develop the Key Questions contained in this report. The ACP posed an initial set of questions designed to help its internal medicine physician members gain more insight into the value of the PHE for their adult patients. After consulting with representatives of the ACP and technical experts, we expanded initial questions to incorporate an assessment of the definition of the periodic health evaluation and to identify interventions which might improve the delivery of the PHE. During this process, we developed a conceptual framework, which we used to help with refinement of the initial Key Questions, to help standardize the research team's conceptualization of the PHE, and to help guide the literature search and review.

We asked the following Key Questions concerning the value of the PHE for adults:

• 1

  What definitions are used for the adult PHE in studies of its value?

• 2

  What is the evidence that a PHE, delivered at different patient ages or different frequencies, is associated with benefits (i.e., improved outcomes) compared to care without a PHE (e.g., usual care or opportunistic delivery of clinical preventive services)? Outcomes include:

  • a

    Delivery of recommended clinical preventive services.

  • b

    Patient attitudes/perceptions (e.g., knowledge, satisfaction, trust, respect).

  • c

    Behavioral outcomes (e.g., tobacco cessation, adherence).

  • d

    Proximal/intermediate clinical outcomes (e.g., cholesterol lowering, disease management).

  • e

    Distal clinical outcomes (e.g., measurable clinical events such as death, or myocardial infarction).

  • f

    Economic outcomes (e.g., cost savings, improved health care utilization).

  • g

    Public health (e.g., improvements in family and community health, communicable disease containment).

• 3

  What is the evidence that a PHE, delivered at different patient ages or different frequencies, is associated with harms (i.e., worse outcomes) compared to care without a PHE (e.g., usual care or opportunistic delivery of clinical preventive services)? Outcomes include:

  • a

    Delivery of non-recommended clinical preventive services.

  • b

    Patient attitudes/perceptions (e.g., worry/anxiety).

  • c

    Behavioral outcomes (e.g., continuation of risky behaviors).

  • d

    Proximal/intermediate clinical outcomes (e.g., complications from testing).

  • e

    Distal clinical outcomes (e.g., measurable clinical events such as death).

  • f

    Economic outcomes (e.g., induced costs, less efficient health care utilization).

  • g

    Public health (e.g., declines in family and community health).

• 4

  What system-based interventions improve the receipt or delivery of the PHE (e.g., insurance premium reductions or provider reminders)?

We selected these questions as the final questions for study after assessing the feasibility of addressing these questions in the literature (including brief preliminary reviews of electronic databases for the presence of evidence to address the questions) as well as on the perceived applicability and importance of the questions to current clinical practice.

Conceptual Framework

Figure 1. Conceptual Framework

   Figure 1. Conceptual Framework

Figure 1

In our model, we defined the goals and expectations of patients, providers and society, which provide the impetus for institution of the PHE. Performance of the PHE, which consists (at minimum) of a risk assessment, including personal and family history taking and a core physical examination, would be followed by delivery of tailored (to patients' age, gender and clinical risk factors) clinical preventive services. Same day clinical preventive services could be delivered either in the context of a more detailed physical examination (e.g., the gynecological examination/Pap smear for an appropriately aged female) or in the context of the provision of preventive counseling, immunizations or laboratory testing. We defined follow up clinical preventive services as services occurring outside of the initial visit for the PHE. However, both same-day and follow up clinical preventive services could be considered potential outcomes of receiving the PHE. The PHE could occur once or over repeated intervals of time. In addition to the receipt of clinical preventive services, we defined other potential benefits and harms of the PHE, including changes in patient attitudes (e.g., trust or worry), changes in patient behaviors (e.g., adherence or continuity with care), proximal (e.g., blood pressure control) or distal (e.g., death) clinical outcomes), resource use and costs (e.g., costs associated with hospitalizations), and outcomes related to public health in general (e.g., communicable disease containment).

To standardize the investigative team's conceptualization of the PHE, we summarized our conceptual framework using the following statement, which all investigators were encouraged to refer to when reviewing the literature at all stages of the study:

“The PHE consists of one or more visits with a health care provider for the primary purpose of assessing a patient's overall health and risk factors for disease which may be prevented by early intervention. During the PHE, health care providers typically perform a history and risk assessment, followed by a tailored physical exam. Based on the information gathered, patients may receive counseling, immunizations, lab testing or arrangements for other tailored preventive health services during the evaluation. The goal of the PHE is to improve intermediate and long-term patient outcomes and ultimately the public's health by appropriate clinical management of chronic conditions, patient education, and fostering the patient-provider relationship. The PHE has the potential to affect patient health and health care cost for the individual, the health care industry, and society as a whole.”

Literature Search Methods

Organization and Tracking of Literature Search

Figure 2. Summary of literature search and review process (more...)

   Figure 2. Summary of literature search and review process (number of articles)

The results of the searches were downloaded and imported into ProCite® version 5 (ISI ResearchSoft, Carlsbad, CA). From ProCite, the articles were uploaded to SRS 3.0. We used the duplication check feature in SRS 3.0. This feature allowed us to scan for exact article duplicates, author/title duplicates, and title duplicates. Additionally, this database was used to store citations in PDF (portable document format) and to track the search results at title review, abstract review, article inclusion/exclusion, and data abstraction levels (Figure 2).

Title Review

After the electronic databases were searched, citations were downloaded into ProCite, and uploaded to the SRS 3.0 tracking system. The study team scanned all titles. Title scans were conducted in a parallel fashion by two independent reviewers. For a title to be eliminated at this level, both reviewers had to indicate that it was ineligible. If the two reviewers did not agree on the eligibility of an article, it was automatically promoted to the next level (see Appendix D ^a, Title Review Form). The title review phase was designed to capture as many studies as possible reporting on the PHE. All titles related to the delivery of clinical preventive services or the PHE itself were included in the initial search and promoted to the abstract review level.

Abstract Review

Article Inclusion/Exclusion

Because of the broad array of potentially eligible articles obtained at the abstract review phase, full articles initially selected for final review underwent another independent parallel review by investigators to determine if they should be included for full data abstraction. At this phase of review, investigators determined which of the Key Questions (2–4) each article addressed (see Appendix D, Article Inclusion/Exclusion Form). For Key Questions 2 and 3, randomized controlled trials were deemed to have applicable information if they contained a group receiving the PHE compared to a group receiving usual care. Observational studies were deemed to have information applicable to assessing Key Questions 2 and 3 if they compared one group who had received the PHE to a group of persons not receiving the PHE (e.g., if persons reported receipt of a PHE in a cross-sectional survey) or if the compared persons before and after receipt of a PHE (in pre-post study designs). Studies were considered eligible if they focused on adults, and not children. All definitions of the PHE were included without regard to the targeting of adults of specific age groups. Although our uniform conceptualization of the PHE prior to the search stated health care providers “typically perform a history and risk assessment followed by a physical exam” in the PHE, we included articles even if they did not explicitly state which components of the PHE were included. Randomized controlled trials were deemed to be not applicable if they contained two groups both receiving the PHE or if they compared groups receiving different forms of the PHE (e.g., a PHE delivered by a nurse versus a physician). For Key Question 4, studies were deemed to have applicable information if they featured interventions designed to enhance patient attendance at the PHE. This could include randomized controlled trials randomizing certain practices or communities to interventions to enhance delivery of the PHE. Articles still deemed to have applicable information at this stage were included in the final article review. All articles deemed to apply to Key Questions 2–4 were used to answer Key Question one (assessing definitions of the PHE in studies). Differences in opinions regarding article inclusion or exclusion were resolved through consensus adjudication.

Article Review

The purpose of the article review was to confirm the relevance of each article to the Key Questions, to determine methodological characteristics pertaining to study quality, and to collect evidence that addressed the Key Questions. Articles eligible for full review could address one or more of the Key Questions. If reviewers felt an article addressed more than one question, multiple data abstraction forms were used.

Two investigators reviewed each study for assessment of study quality and full data abstraction. Each reviewer independently judged study quality and rated items on standard quality assessment forms. For all data abstracted from studies, we used a sequential review process. In this process, all data abstraction forms were completed by the primary reviewer. The second reviewer confirmed the first reviewer's data abstraction forms for completeness and accuracy. Reviewer pairs were formed to include personnel with both clinical and methodological expertise. A third reviewer re-reviewed all articles that were marked as “ineligible” by the first two reviewers to ensure consistency in the classification of the articles. Reviewers were not masked to the articles' authors, institution, or journal. In most instances, data were directly abstracted from the article. If possible, relevant data were also abstracted from figures. Differences in opinion were resolved through consensus adjudication.

For each article, data abstracted included: 1) study design; 2) study location (including country of study); 3) dates the study was performed and length of follow up; 4) study setting (geographic setting as well as health care delivery structure); 5) numbers of study subjects enrolled; 6) study eligibility criteria for patients and providers; 7) descriptive characteristics of study patients (including race, gender, education, and income) and providers (including clinical specialty and practice setting); 8) components of the PHE in each study; 9) descriptive information about study interventions; and 10) study outcomes (including baseline and follow up rates of delivery of recommended preventive services, proximal clinical outcomes, distal clinical and economic outcomes as well as improvements in the delivery of the PHE) (see Appendix D, Data Abstraction Review Forms).

Data Abstraction

Table 9

Summary of Results from Best Available Evidence to (more...)

Table 9

Summary of Results from Best Available Evidence to Assess Each Outcome

Outcome	Type* of Evidence Assessing Considered (number of studies)	Strength and Consistency of Evidence	Range of Magnitude and Direction of Effects of PHE on Outcome**
Delivery of Clinical Preventive Services
Gynecological examination/ Pap smear	RCTs (2)	High	Small to Large Positive
Counseling	RCTs (1)	Low	Mixed
	Observational (6)
Immunizations	RCTs (3)	Medium	Mixed
Cholesterol Screening	RCTs (1)	Medium	Small to Large Positive
	Observational (4)
Colon Cancer Screening (Fecal Occult Blood Testing)	RCTs (2)	High	Large Positive
Mammography	RCTs (1)	Low	Mixed
	Observational (1)
Proximal Clinical Outcomes
Disease Detection	RCTs (2)	Medium	Mixed
Health Habits	RCTs (5)	Medium	Mixed
Patient Attitudes (Worry)	RCTs (1)	Medium	Positive†
Health Status	RCTs (2)	Medium	Mixed†
Blood Pressure	RCTs (2)	High	Mixed
Serum Cholesterol	RCTs (1)	Low	Mixed
	Observational (1)
Body Mass Index	RCTs (3)	Medium	Mixed
Distal Clinical and Economic Outcomes
Costs	RCTs (4)	Medium	Mixed
Disability	RCTs (2)	Medium	Mixed
Hospitalization	RCTs (3)	High	Mixed
Mortality	RCTs (5)	Medium	Mixed
Improvement in Receipt of PHE	RCTs (1)	Medium	Medium to Large Positive
	Non-RCTs (1)

*

RCT=Randomized controlled trial; Observational=Studies with observational design; non-RCT=non-randomized controlled trials

**

Magnitude and direction of effect of receipt of PHE on outcome, based on standardized effect sizes calculated using Cohen's d. We considered effect sizes ranging from 0 to 0.25 to represent “small” effects, ranging from 0.25 to 0.8 to represent “medium” sized effects, and effect sizes greater than 0.8 to represent “large” effects. Effect sizes can be thought of as the average percentile standing of the average treated (or experimental) participant relative to the average untreated (or control) participant. An ES of 0.0 indicates that the mean of the treated group is at the 50th percentile of the untreated group. An ES of 0.25 indicates that the mean of the treated group is at the 58th percentile of the untreated group. An ES of 0.8 indicates that the mean of the treated group is at the 79th percentile of the untreated group.

†

Standardized effect size could not be calculated for the study or studies assessing this outcome.

All information from the article review process was entered in a relational database (Recruitment Evidence Database). The database was used to maintain and clean the data, as well as to create detailed evidence tables and summary tables (see Appendix G and Tables 1 through 9).

Data abstracted to assess the definition of the PHE (Key Question 1). Data were abstracted on the components of the PHE in each study. Components of the PHE could include: 1) the history and risk assessment of patients (including collection of a history on patients' diet, alcohol/substance abuse, injuries, sexual practices, tobacco use, calcium and folic acid intake, sun exposure, or poly-pharmacy); 2) physical examination of patients (including assessment of blood pressure, height, weight, pulse, and examination of breasts, cardiovascular system, pulmonary system, abdominal region, neurological system, gynecological or urological systems, and extremities); 3) counseling provided to patients (including counseling regarding diet, physical activity, alcohol/substance abuse, injury prevention, safe sexual practices, tobacco use, use of folic acid, sun exposure, oral health, poly-pharmacy); 4) delivery of immunizations during the PHE; and 5) delivery of clinical preventive services during the PHE (including Pap smears, gonorrhea/chlamydia screening, audiometry, vision screening, electrocardiograms, chest x-rays, mammography, sigmoidoscopy, colonoscopy, fecal occult blood tests, bone mineral density tests, serum glucose, lipids, hemoglobin A1c, blood counts, chemistries, prostate specific antigen, urinalysis, and purified protein derivative skin test (PPD) screening for tuberculosis exposure). When the PHE included other components (not listed), they were abstracted for later categorization.

Data abstracted to assess outcomes of the PHE (Key Questions 2 through 4). For studies assessing the benefits and harms associated with the PHE, data were abstracted to capture changes in the delivery (by health care providers) or receipt (by patients) of recommended clinical preventive services which were delivered as a result of the PHE, including the delivery of recommended aspects of the physical examination (e.g., blood pressure measurement, gynecological examination), counseling (e.g., substance abuse counseling), immunizations (e.g., influenza vaccination), and clinical screening tests (e.g., cholesterol testing). Data were also abstracted regarding changes in patient attitudes/perceptions as a result of the PHE (e.g., knowledge, satisfaction), changes in patient behavioral outcomes as a result of the PHE (e.g., rates of tobacco cessation), proximal/intermediate clinical outcomes (e.g., cholesterol lowering, disease detection), distal clinical outcomes (e.g., death), economic outcomes (e.g., cost, health care utilization), and public health outcomes (e.g., communicable disease containment). In studies of system-level interventions, data were abstracted on interventions associated with receipt of the PHE. For randomized controlled trials, non-randomized controlled trials, and comparative observational studies, baseline and follow up data from both intervention and control groups were abstracted for comparison. For observational studies with a pre-post design, outcomes were abstracted at baseline and follow up for the single group under observation.

Article Quality Assessment

Two reviewers independently judged articles on several aspects of study external and internal validity, including: 1) description of inclusion and exclusion criteria for study subjects (best scores assigned for explicit reporting of criteria); 2) description of study subjects' baseline characteristics (best scores assigned for reporting of all important characteristics including age, gender, race, socioeconomic status, and comorbidities); 3) description of study non-enrollees (best scores assigned for description of differences in sociodemographic or clinical characteristics between study groups); 4) description of handling of study withdrawals (best scores assigned for use of intention to treat analyses with sensitivity analyses to examine differences between as-treated and intention-to-treat analyses); 5) description of the study intervention (best scores assigned for studies in which reviewers judged the intervention could be replicated with the completeness and detail included in the description); 6) adequacy of length of study follow up (best scores assigned when the length of follow up was appropriate for fully capturing outcomes); 7) study subject attrition (best scores assigned when the percentage of subjects remaining study was ≥85%); 8) description of study outcomes (best scores assigned for studies clearly describing outcomes so they could be understood easily); 9) relevancy and appropriateness of outcomes (best scores assigned for studies in which outcomes were deemed to be relevant and appropriate for the study as well as feasibly measured); 10) quality of outcomes assessment (best scores assigned with assessment of outcomes was both standardized and valid); 11) quality of randomization for RCTs, (best scores assigned for reporting of centralized randomization scheme and the presence of sufficient documentation regarding randomization); 12) quality of blinding for RCTs (best scores assigned for studies documenting adequate blinding of patients, providers, and outcomes assessors when appropriate); 13) comparable treatment of treatment groups for RCTs (best scores assigned for studies reporting comparable treatment of study groups with the exception of the intervention); 14) comparable characteristics of enrolled subjects for control and treatment groups for RCTs (best scores assigned when studies reported no significant difference in any characteristic likely to affect the success of the intervention or any outcome); and 15) statistical analysis. Assessments of quality of statistical analyses included assessment of : 1) study power to assess study outcomes (best scores assigned when a priori estimates of the statistical analysis were reported); 2) study investigator choice for statistical tests (best scores assigned when appropriate choice of statistical tests were made); 3) the presentation of statistical significance (best scores assigned when studies reported statistical significance in the form of confidence intervals or p-values); 4) the assessment and adjustment for potential confounding, when present (best scores assigned when multivariable analyses adequately accounted for potential confounding); and 5) potential problems with unit of analyses (best scores assigned for studies with no potential problems or for studies in which potential problems existed but were appropriately addressed). For both experimental and observational studies, we applied a total quality score, based on Chalmers et al, in which items assessing the external validity of studies received 35% of the score, items assessing the internal validity of studies received 35% of the total score, and items assessing the quality of the statistical analysis received 30% of the total score (see Appendix D, Quality Review Form).⁴⁸ In developing overall quality scores for individual studies, scores for each item were averaged between two reviewers. Total quality scores for each study could range from 0 (worst quality) to 100 (best quality). In the absence of universal standards for recognizing studies of high or low quality, we classified studies according to their score relative to the distribution of all other study scores, defined by tertiles of the distribution of all scores. Studies with quality scores falling within the top 33% of all study quality scores were deemed to have “high” scores, studies with quality scores falling within the middle 33% of all study quality scores were deemed to have “medium” scores, and studies with quality scores falling within the lowest 33% of all scores were deemed to have “low” scores. Because trials were judged on slightly different criteria than observational studies, trials were rated in relation to the scores of all other trials, and observational studies were rated in relation to the scores of all other observational studies.

Data Entry and Quality Control

Initial data were abstracted by investigators and entered directly into Web-based data collection forms; SRS^® 3.0 (TrialStat! Corporation, Ottawa, Ontario, Canada) (Appendix D). After data were reviewed by a second author, adjudicated data were re-entered into Web-based data collection forms by trained research assistants. A standard process for data quality checks was instituted in which research assistants individually inspected all data entries. In addition, research assistants used a redundant system of random data checks to assure data quality.

Grading of the Evidence

At the completion of our review, we graded the quantity, quality and consistency of the best available evidence addressing Key Questions 2–4 by adapting an evidence-grading scheme approach recommended by the Grading of Recommendations Assessment, Development and Evaluation (GRADE) Working Group.⁴⁹ GRADE is a systematic approach to grading the strength of the total body of evidence that is available to support recommendations on a specific clinical management issue. In applying the GRADE system to the evidence, we incorporated assessments of studies' design, studies' quality, consistency of findings, and magnitude of findings.

Process for Assigning Evidence Grades

First, we assessed study designs of the best available evidence to assess an individual outcome. We used the term “best available evidence” to indicate studies assessing the effect of the PHE on outcomes of interest in the least biased manner. We considered randomized controlled trials to represent the best study design to address Key Questions 2–4. Thus, when RCTs assessed outcomes, we considered the best available evidence assessing the outcome to be comprised of two or more RCTs. If an outcome was evaluated by at least two RCTs as well as observational studies, our evidence grade was based only on the RCTs and observational studies were ignored. If an outcome was evaluated by one or no RCTs, our evidence grade was based on the single randomized controlled trial in addition to the best available non-randomized controlled trial or the best available observational studies (cohort studies (considered best), followed by cross-sectional studies and studies with pre-post observational design (considered worst)). We reported the number of studies within the category of best available evidence to assess the quantity of evidence.

Based on the design of at least two studies comprising the best available evidence assessing specific outcomes, we designated a starting numeric value between one and four for the evidence applying to each outcome. Accordingly, we assigned a value of 4 (highest value) if the body of evidence for the outcome included two or more RCTs; a value of 3 if there was one RCT with or wihtout at least one non-randomized controlled trial, one RCT with or without at least one cohort study (prospective or retrospective), or one RCT and one pre-post study; a value of 2 if there were cohort studies only (prospective or retrospective) or if there was one controlled trial and two cross-sectional studies. All other study designs started with a value of one (lowest value). Next, we assessed the quality of the individual studies providing the evidence on specific outcomes. We used the standard assessment of individual study quality (described above and completed prior to the evidence grading process) to guide our evaluation of the overall quality of evidence assessing the outcome, including variations in studies' external validity, internal validity, and approach to statistical analysis. We evaluated the consistency of the direction of results reported in the evidence by evaluating the number and type of studies reporting the PHE had positive, negative or no effects on specific outcomes. Bodies of evidence in which results from individual studies were consistent in direction for a specific outcome received no point deduction for inconsistency. Bodies of evidence in which some studies reported results in one direction (either positive or negative) but some studies reported neutral effects of the PHE received a 0.5 point deduction for inconsistency. Bodies of evidence in which studies reported both positive and negative results received a full one point deduction for inconsistency. Finally, we evaluated the directness of evidence by considering how individual studies handled plausible confounders, and we evaluated the strength of the associations between the PHE and outcomes based on the magnitude of effect sizes indicating clinically significant differences in outcomes between groups receiving the PHE and groups not receiving the PHE.

Figure 3. Explanation of GRADE Using Colon Cancer as (more...)

   Figure 3. Explanation of GRADE Using Colon Cancer as an Example

We based the overall grade of evidence on these four key elements for each outcome, categorized as “high” grade (score of 3.0 to 4.0), 2) “medium” grade (score of 2.0 to 2.9), 3) “low” grade (score of 1.0 to 1.9), and 4) “very low” grade (score less than 1.0). A grade of “high” signifies that further research would be unlikely to alter observed effects, a grade of “medium” signifies that further research could alter the observed effects, a grade of “low” signifies that further research would be very likely to alter the observed effects in the abstracted literature, and a grade of “very low” signifies that any estimate of effect is very uncertain. In Figure 3, utilizing colon cancer as an example outcome, we have provided an example of our approach to grading the evidence for each outcome.

Estimating the Magnitude of Effect of the PHE on Outcomes in RCTs

In an effort to provide standard estimates of the effect of the PHE across outcomes, we calculated effect sizes (using Cohen's d Effect Size Estimate for mean differences and differences in proportions) for comparative studies evaluating the effect of the PHE on outcomes where possible.^{50, 51} We considered effect sizes ranging from 0 to 0.25 to represent “small” effects, ranging from 0.26 to 0.8 to represent “medium” sized effects, and effect sizes greater than 0.8 to represent “large” effects.⁵⁰ Effect sizes can be thought of as the average percentile standing of the average participant receiving the PHE relative to the average participant not receiving the PHE. An effect size of 0.0 indicates that the mean of the group receiving the PHE is at the 50th percentile of group not receiving the PHE. An effect size of 0.25 indicates that the mean of the group receiving the PHE is at the 58th percentile of the group not receiving the PHE. An effect size of 0.8 indicates that the mean of the group receiving the PHE is at the 79th percentile of the group not receiving the PHE. Thus, larger effect sizes represent greater separation of findings between treatment and control groups.⁵² We also noted the direction of effects. We considered evidence neutral when the 95% CI of the estimate of effect included zero. When enough data were not presented in articles to present effect sizes (e.g., no information reported regarding the variance of reported means), we presented other standard estimates of effect (e.g., rate ratio) or estimated the direction and clinical significance of reported results.

Peer Review

Throughout the project, we sought feedback from the technical experts through ad hoc and formal requests for guidance. A draft of the completed report was sent to the technical experts and peer reviewers, as well as to the representatives of the ACP and AHRQ. The range of reviewers included a representative of the sponsor of the Key Questions (ACP), academic experts in the assessment of clinical preventive services and primary care, patient stakeholder organizations (American Association of Retired Persons, American Cancer Society), private and public health insurance stakeholders (American Health Insurance Plans and CMS), and experts from the AHRQ. Substantive comments were entered into a database, and revisions to the draft report addressed reviewer comments. The disposition of all comments was submitted to the AHRQ with the final report.

Results of Literature Search and Abstract Review Process

Figure 2

In addition to the 7003 citations retrieved by the search methods, we retrieved 64 citations through hand searching. Using the duplicate removal feature of SRS 3.0, and reviewer observation, we identified 544 duplicates, leaving 6523 for title review. Of these, we reviewed 2021 at the abstract level. We included 819 articles in the full article inclusion/exclusion portion of the review. Of these, 54 articles were promoted for full data abstraction and quality assessment. These 54 articles represented 36 studies that reported multiple outcomes and/or multiple follow ups. Full data abstraction was completed only on the 36 studies integrating data from all 54 articles.

Because many articles had more than one reason to be excluded the abstract reviewers did not need to agree on the main reason for exclusion applied at the abstract level. The two most frequent reasons for exclusion were that the article did not include any useful information for this review (762 abstracts), and no original data was presented (either a review or an opinion piece) (523 abstracts). The remaining reasons for exclusion were: study included only subjects less than 18 years old (75 abstracts), and not an English language study (4 abstracts). Articles could be excluded for more than one reason at this level.

Results of Article Inclusion/Exclusion Process

From the abstract review process, 819 citations were identified for the article inclusion/exclusion phase. At this level 762 articles (93%) were excluded, and 3 were not retrievable. The most frequent reasons for exclusion were that the article did not include any original data (390 articles), the article did not apply to any of the Key Questions (372 articles), the exposure in the study was not a PHE (310 articles), and the article focused on specific preventive service delivery (215 articles). Of the 54 articles (36 studies) included in this report, Key Question 1 was addressed by all studies, Key Questions 2 was addressed by 36 studies, and Key Question 4 was addressed by 5 studies. Articles could be excluded for more than one reason at this level. A listing of the included articles and the excluded articles with the reasons for exclusion is included in this report (Appendix E ^a

General Study Characteristics

Table 2

Summary of Study Results

Table 2

Summary of Study Results

Author, year	Year Study Began	Years (months of follow up)	Study population	Intervention Group	Comparison Group	Interventions outside of the PHE	Outcome(s) Assessed	Definition of PHE in this study	Study Limitations
Randomized Controlled Trials
Patrick, 199953	1993	48 months (2 years after intervention)	The study population consisted of 2,558 HMO enrollees in Seattle, WA. Medicare demonstration project.	The intervention group included Group Health Comparative of Puget Sound Medicare enrollees invited to receive preventive services benefits package for two years.	The comparison group consisted of Medicare enrollees receiving usual care.		1. Immunization - influenza	The PHE was described as a preventive service package that include - 1) health-risk assessment (telephone interview); 2) health-promotion visit (90 minute nurse visit with health risk appraisals, positive behavior reinforcement and referrals for interventions where appropriate); 3) disease-prevention visit (by nurse and physician who conducted history and physical examination and reviewed patients' health risks) and 4) follow-up classes (exercise, planning ahead). Counseling on exercise, high fiber/low fat diet and advance directives offered to all.	Internal Validity Concerns:
				Uptake of PHE: 90% attended health promotion and disease prevention visits year 1; 83% attended health promotion and disease prevention visits in year 2; 78% had visits in both years,; 9% attended none in any year; 24% attended any classes.			2. Health habits - improvement in:	Counseling on exercise, high fiber/low fat diet and advance directives offered to all.	1. Reporting on blinding
							- Physical activity	Health promotion visit and disease prevention visits and group exercise were offered in both years.	External Validity Concerns:
							- Diet (fat and fiber)		1. Description of outcomes not detailed
							- Advance directives		2. Results potentially not generalizable beyond Medicare recipient population
							- Breast self-exam		Statistical Validity Concerns:
							- Smoking		1. Potentially inadequate adjustment for residual confounding
							- Alcohol
							- Seat belt use
							3. Patient attitudes - mean score health worry‡
							4. Body mass index - at risk for obesity, 24-month F/U
							5. Costs - average total cost per participant
							6. Mortality
							- Mortality at 24 months
							- 48 months
Elder, 199554; Mayer, 199490	1992	48 months (2 years after intervention completion)	The study population consisted of 1,203 HMO enrollees in San Diego, CA. Medicare demonstration project.	The two-year intervention consisted of Medicare beneficiaries receiving a health risk appraisal with individual counseling and health promotion workshops.	The comparison group received usual care.		1. Health habits	The PHE was described as preventive services through a health risk appraisal with individual counseling, selected clinical tests and immunizations, and a series of 8 weekly group health promotion sessions (memory, mental alertness, coping with loss, choices for independent living, self-enhancement, exercise, nutrition, relaxation, self-care).	Internal Validity Concerns:
				Uptake of PHE: 96% completed health risk appraisal and individual counseling; 87% attended at least one group session, 72% attended at least 4, 59% attended at least 6.			- Fiber servings per day	Individual counseling was continued during year 2.	1. Reporting on differences between enrollees and non-enrollees
							- Fat servings per week		2. Reporting on blinding
							- Salt use		3. Reporting on withdrawals or crossovers
							- Caffeine drinks per day		External Validity Concerns:
							- Stretching minutes per week		1. Results potentially not generalizable beyond Medicare recipient population
							- Consumption of cruciferous foods
							2. Blood pressure
							- Mean systolic blood pressure at 12 months
							- Mean diastolic blood pressure at 12 months
							3. Body mass index
							- Mean BMI at 24 months (end of intervention period)
							- Mean BMI at 48 months (end of F/U)
Morrissey, 199555	1988	12–26 months after beginning of intervention (for cost outcomes, 3 years after beginning of intervention - 1 year after intervention completion)	The study population consisted of 1914 patients from 10 primary-care medical practices in central North Carolina. Medicare demonstration project.	The intervention group received full Medicare reimbursement to physicians for preventive care and health promotion packages, regular reminding of physicians to routinely schedule preventive care visits, a new office system in which nurse carried out many preventive procedures, and a form for charting preventive care. Duration of intervention was two years.	The comparison group received usual care.	Intervention Group:	1. Pap smear	The PHE was described as a preventive care visit offered once a year including a breast exam, eye exam, pap smear, hearing test, depression test, influenza & pneumovax immunization, cholesterol test, fecal occult blood test, urinalysis and a urinary incontinence test. One hour health promotion sessions were conducted every 6 months for physical activity, nutrition, stress management/problem solving and others based on risk.	External Validity Concerns:
				Uptake of the PHE: 88% received at least one clinical screening; 87% received at least one health promotion service.		Patients:	2. Immunization - influenza		1. Results potentially not generalizable beyond Medicare recipient population
						1. written materials	3. Cholesterol screening
						2. phone call	4. Colon cancer screening - fecal occult blood testing
						Providers:	4. Mammogram
						1. annual capitated payments for preventive care and health promotion visits	5. Costs
						2. prompting to schedule preventive care visits	- 3-year post-intervention cumulative Medicare charges.
						3. office system change for nurse delivery of preventive services	- 3-year post-intervention cumulative Medicare reimbursement
						4. form for charting preventive care	6. Hospitalization
							- Utilization data: hospital days per enrollee over two years of intervention and one year post-intervention
							- Admissions per enrollee over two years of intervention and one year post-intervention
Burton, 199556; German, 199579; Burton, 199780; Burton, 199578 Burton, 1995; German, 1995; Burton, 1997;	1989	24 months after beginning of intervention and for some outcomes 48 months after beginning of intervention (2 years after end of intervention)	The study population consisted of 4,195 older, community-dwelling Medicare recipients in Baltimore. Medicare demonstration project.	The intervention group received coverage for an annual preventive visit and tests (Medicare vouchers for 2 yearly preventive visits and optional counseling visits).	The comparison group received no coverage for an annual preventive visit and tests.	Intervention Group:	1. Pap smear	The PHE was described as a physical examination. The examination included a breast, pelvic (including Pap smear), and digital rectal exam, fecal occult blood testing, cholesterol testing, immunizations, counseling for health risks, and a complete history including vision, hearing, and dentition.	Internal Validity Concerns:
				Uptake of the PHE: 63% made preventive visit year 1; 52% counseling visit year 1; 52% preventive visit year 2; 33% counseling visit year 2.		Patients:	2. Health habits		1. Reporting on blinding
						1. Written material	- Smoking		External Validity Concerns:
							- Problem alcohol drinking		1. Results potentially not generalizable beyond Medicare recipient population
							3. Health status - change in health status of intervention and control groups from baseline to 2 years
							4. Costs
							- Total health care charges, Year 1.
							- Total health care charges, Year 2.
							- Mean monthly Medicare Part A charges, Year 1
							- Mean monthly Medicare Part A, charges Year 2
							- Mean monthly Medicare Part A charges Year 3 (1 year post-intervention)
							- Mean monthly Medicare Part A charges Year 4 (2 years post-intervention)
							5. Hospitalization
							- Mean inpatient days for the intervention and control groups who had a hospital discharge in that year (Year 1)
							- Mean inpatient days Year 2
							- Hospital discharges per 1000 Year 1
							- Hospital discharges per 1000 Year 2
							6. Mortality
Norman, 199260	1992*		The study population consisted of 818 patients from one general practice in Norfolk, England aged 30 to 41.	The intervention group consisted of patients who received an invitation letter with an appointment for a health check.		Intervention Group:	1. Receipt of PHE - attendance of PHE	The PHE was described as a health check that included the assessment of smoking behavior, alcohol consumption, diet and exercise levels, blood pressure and weight.	External Validity Concerns:
				The intervention group also included patients who received an open invitation letter to health		Patients:			1. Description of study population characteristics not detailed
						1. Written material			Statistical Validity Concerns:
						Control Group:			1. Potentially inadequate adjustment for residual confounding
						Patients:
						1. Written material
Belcher, 199044	1981	60 months	The study population consisted of 1,224 male patients who attended the Seattle Veterans Affairs Medical Center during October to December 1980. (We included 674 patients in the study who either attended a health promotion clinic or received usual care. Other groups received other interventions)	The intervention group was offered self-referral to a health promotion clinic.	The comparison group received usual care.	Intervention Group:	1. Counseling	The PHE was described as a physical examination similar to USPSTF recommended activities. Included history and physical examination items (alcoholism screen, smoking assessment, blood pressure check, breast examination); laboratory testing (fecal occult blood, cholesterol, tuberculin skin test, VDRL, Pap smear and mammography); tetanus/diphtheria and influenza vaccination, and counseling on breast self-examination and alcoholism and smoking cessation.	Internal Validity Concerns:
				Uptake of the PHE: 71% participated in health promotion clinic in year 1; 78% participated in health promotion clinic in year 1 or year 2; 90% attending in year 1 returned for year 2		Patients:	- Alcohol abuse		1. Reporting on blinding
						1. Written material	- Smoking cessation		External Validity Concerns:
						2. Phone call	2. Immunization - influenza		1. Description of study population not detailed
							3. Colon cancer screening - fecal occult blood testing		2. Results potentially not generalizable beyond Veterans Affairs (male) population
Stone, 198157; South-east London, 19777; South-east London, 200176; Stone, 1978a91; Stone, 1978b92; Trevelyan, 197393	1967	60–108 months	The study population consisted of 7229 patients in South-east London aged 40 to 64 years in 1967 receiving care in primary care group practices.	The intervention group were South London patients aged 40 to 64 years in specific group practices; received 2 multiphasic screenings 2 years apart.	The comparison group consisted of South London patients aged 40 to 64 years in specific group practices; received usual care.		1. Disease detection	The PHE was described as multiphasic screening. Screening for ischemic heart disease, elevated blood pressure, chronic bronchitis, diabetes, thyroid imbalance, arthritis, obesity, venous varicosities, hearing and visual defects. (PFTs, ECG, blood pressure, serum cholesterol, uric acid, fecal occult blood), abdominal exam, leg exam, breast and pelvic exam, chest x-ray, height, weight and skin fold, vision and audiometry testing, skin, mouth, teeth and joint exams	Internal Validity Concerns:
				Uptake of the PHE: 73% participated in first year screening; 99% of these had both physical exam and clinic tests			- Angina		1. Reporting on blinding
							- High diastolic blood pressure		External Validity Concerns:
							- Ischemia on electrocardiogram		1. Study performed before USPSTF or similar contemporary preventive services guidelines in effect
							- Bronchitis symptoms		Statistical Validity Concerns:
							2. Health habits - percentage still smoking		1. Differences in control and treatment group at baseline not specifically accounted for in analysis
							3. Disability - major disability (e.g., inability to dress or undress themselves)		2. Incomplete presentation of statistical significance
							4. Hospitalization - hospitalizations/ 1000 person years at risk (1976)
							5. Mortality mortality rate per 1000 person-years at risk:
							- All cause death
							- Neoplasm
							- Central nervous system
							- Cardiovascular disease
							- Respiratory disease
							- All other causes
Fletcher, 197745	1974	Outcomes (disease detection) measured directly after multiphasic screening visit.	The study population consisted of 112 patients 40 to 65 years old seen by 112 physicians. Physicians randomized to receiving results of multiphasic screening program versus reviewing prepared chart abstract versus reviewing actual chart	The intervention group consisted of patients who received multiphasic screening.	The first comparison group was the medical chart abstraction group (physicians given abstracted information about patients from chart).		1. Disease detection	The PHE was described as multiphasic screening that included a standard questionnaire, measurement of blood pressure, height, weight, visual acuity, tonometry, audiometry, blood leukocyte count, hematocrit and hemoglobin levels, serologic levels for syphilis, biochemistry of random blood specimen, urinalysis, ECG, and chest roentgenogram. The pHE also included clinical breast examination, pap smear and vital capacity.	Internal Validity Concerns:
		Charts were reviewed at 12 months to assess follow-up of new problems		Uptake of the PHE: Not mentioned, but by design, all intervention group participants would have received the multiphasic screening.	In the second comparison group, physicians reviewed the patient's actual chart.		- Disease detection of ALL problems before and after intervention (number of new medical problems detected at F/U)		1. Reporting on blinding
							- Disease detection of important problems before and after intervention		External Validity Concerns:
									1. Study performed before USPSTF or similar contemporary preventive services guidelines in effect
									2. Description of study population not detailed
									Statistical Validity Concerns:
									1. Potentially inadequate adjustment for residual confounding
Cutler, 197341; Collen, 197347; Dales, 197394; Ramcharan, 197395; Friedman, 19869; Dales, 197996; Norinder, 200297	1964	84–192 months	The study population consisted of 10,713 randomly selected California Kaiser Health Plan members in 1964, age 35–54.	The intervention group consisted of California Kaiser Health Plan members aged 35–54 encouraged to have an annual multiphasic check-up for 11 years.	The comparison group consisted of California Kaiser Health Plan members aged 35–54 who received usual care.	Intervention Group:	1. Costs	The PHE was described as an annual multiphasic health check-up (MHC) consisting of a multiphasic exam (which is a series of tests performed in the automated multi-test lab), and a follow-up evaluation of from multiphasic exam. The multiphasic exam included a standard questionnaire including history and present symptoms questions, measurement of blood pressure, visual acuity, tonometry, audiometry, urinalysis, ECG, and chest & breast x-rays,. The MHC also included anthropometry, spirometry, and a serum chemistry panel.	Internal Validity Concerns:
				Uptake of the PHE: 54% of intervention group received at least 4 PHEs over 7 years, 83% received at least one PHE over 7 years	Uptake of the PHE: 13% of intervention group received at least 4 PHEs over 7 years, 53% received at least one PHE over 7 years	Patients:	- Average annual cost for physician visit per participant at 7 years (men, aged 45–54 years at baseline)		1. Reporting on blinding
						1. Written Material	- Average annual cost for physician visit per participant at 11 years (men, aged 45–54 years at baseline)		External Validity Concerns:
						2. Reminder	- Average annual expense per participant in multiphasic health check-up expense at 7 years. (men, aged 45–54 years at baseline)		1. Study performed before USPSTF or similar contemporary preventive services guidelines in effect
						3. Phone calls	- Average annual expense per participant in multiphasic health check-up expense at 11 years. (men, aged 45–54 years at baseline)		Statistical Validity Concerns:
							2. Disability		1. Potentially inadequate adjustment for residual confounding
							- Disability at 7 years
							- Disability at 11 years among men aged 45–54
							3. Mortality
							Deaths, rate per 1000 persons
							- All cause deaths
							- Death from potentially postponable causes¶
							- Death from colorectal cancer
							- Death from breast cancer (women only)
							- Death from cervical/uterine cancer (women only)
							- Death from prostate cancer (men only)
							- Death from hypertension-associated causes
							- Death from ischemic heart disease
							- Death from respiratory system disease.
							- Death from musculoskeletal disease
							- Death from mental, nervous, or sensory organ disease
							- Death from endocrine, nutritional, and metabolic disease
							- Death from suicide
							- Death from lymphohematopoetic cancer
Theobald, 199858	1969	20 years	The study population consisted of 32, 186 patients aged 18–65.	The intervention group consisted of Stockholm residents aged 18–65 who were offered a general health examination.	The comparison group were Stockholm residents aged 18–65 who received usual care.		1. Mortality	The PHE was described as a general health examination that included social, psychiatric, and medical interviews and exams to determine social and medial needs. Also, blood tests, physical examinations, ECGs, exercise tests, psychological tests and eye and dental examinations.	Internal Validity Concerns:
				Uptake of PHE: 2578/3064 (84%) of those offered the PHE were examined.			- All cause mortality		1. Reporting on differences between participants and non-participants
							- Cardiovascular disease mortality		2. Reporting on blinding
							- Cancer mortality		3. Description of PHE is not detailed
							- Accidents and intoxication mortality		External Validity Concerns:
									1. Description of study population characteristics not detailed
									2. Study performed before USPSTF or similar contemporary preventive services guidelines in effect
									Statistical Validity Concerns:
									1. Potentially inadequate adjustment for residual confounding
OXCHECK, 199559; Langham, 199681]	1989	36 months	The study population consisted of 11,090 patients aged 35 to 64 from 5 general practices in Bedfordshire, England.	The intervention group received a health check at baseline (year 1) and in year 4.	The comparison group received no health check at baseline but received a health check in year 4.	Intervention Group:	1. Health habits	The PHE was described as a health check and consisted of medical history, lifestyle questionnaire, measurements of height, weight, blood pressure, and serum cholesterol levels, and post-visit counseling.	Internal Validity Concerns:
			4,908 patients receiving PHE n year 1, year 4 (intervention group) and in year 4 (control group) comprise the study population for this analysis.	Uptake of the PHE: Of the 2205 receiving PHE in year 1, 75% received PHE in year 4		Patients:	- Smoking		1. Reporting on blinding
						1. Written Material	- Alcohol use		Statistical Validity Concerns:
						2. Reminder	- Exercise less than once per month		1. Potentially inadequate adjustment for residual confounding
						3. Phone call	- Use full cream milk
						4. Encouragement to make a visit during other healthcare visits	- Use butter or hard margarine
						Control Group:	2. Blood pressure
						Patients:	- Systolic blood pressure at 3-year follow up
						1. Written Material	- Diastolic blood pressure at 3-year follow up
						2. Reminder	- Proportion of high risk diastolic pressure (≥100mm Hg) from 3 year F/U when compared to control
						3. Phone call	3. Changes in serum cholesterol
						4. Encouragement to make a visit during other healthcare visits	- Mean total cholesterol at 3-year F/U
							- Proportion of high risk cholesterol (≥8mmol/l) at 3 year F/U
							4. Body mass index
							- Mean BMI at 3-year F/U
							- Percentage of participants with BMI ≥ 30
							5. Cost-effectiveness
Non-randomized Controlled Trials
Christensen, 199584	1990		The study population consisted of 2,452 patients of 65 general practitioners.	The intervention group consisted of Denmark patients of general practitioners who received mailing that a preventive health examination was free.	The comparison group were Denmark patients of general practitioners who received mailing that a preventive health examination was 40 Danish Krone.	Intervention Group:	1. Receipt of PHE - attendance at PHE	The PHE in this study was described as a preventive health examination.	External Validity Concerns:
						Patients:			1. Description of PHE is not detailed
						1. Written material
						2. Financial disincentive (charge)
						Control Group:
						Patients:
						1. Written material
						2. Financial disincentive (charge)
Prospective Cohorts
Roberts, 196962	1950	180 months	The study population consisted of 20,648 male patients who had employee-sponsored periodic health exams.	The intervention group consisted of U.S. employed men receiving a employer-sponsored periodic health examination.	The comparison group consisted of a historical comparison of U.S. white men.		1. Mortality - actual/expected deaths	The PHE was described as a periodic health exam that included a health history, and a thorough physical examination supplemented by laboratory, x-ray, and ECG studies.	External Validity Concerns:
									1. Study performed before USPSTF or similar contemporary preventive services guidelines in effect
									2. Results potentially not generalizable beyond persons who are non-White men
Chiou, 200283	1993	6 years	The study population consisted of a sample of 1,193 elderly people in each of the 11 districts in Kaohsiung City, Taiwan.	The intervention group consisted of Taiwanese adults aged 65 years and older reporting receiving a physical examination in past year.	The comparison group were Taiwanese adults aged 65 and older not receiving physical examination in past year		1. Mortality - Relative risk of mortality	The PHE was defined as an annual physical exam that included measurements of weight, height, blood pressure, pulse, visual acuity, oral health, and hearing. A PHE also included urine, fecal occult blood, fasting blood lipids, and glucose laboratory tests.	External Validity Concerns:
									1. Results potentially not generalizable beyond elderly Taiwanese.
Retrospective Cohorts
Burton, 200229	1989	3 years	The study population consisted of 1,773 Bank One executives who were enrolled in the Bank Medical Plan, or a preferred provider plan.	The intervention group consisted of executives that were eligible for and receiving the periodic health examination.	The comparison group consisted of executives eligible for but not participating in the periodic health examination.		1. Costs - average cost in medical claims paid per employee	The PHE was described as a complete history and physical examination, fasting laboratory tests including multiphasic chemistries (blood count etc), lipid profile, total cholesterol, HDL-cholesterol, calculated LDL-cholesterol, dipstick urinalysis, resting 12 lead-electrocardiogram, pulmonary function testing and vision and glaucoma screening.	Internal Validity Concerns:
							2. Disability		1. Use of claims data not specified for research purposes
							- Average number of short-term disability days per employee		External Validity Concerns:
							- Total short-term disability days in 3 years		1. Results potentially not generalizable beyond executive employees
							- Any short-term disability days (%)
Hama, 200177	1999	12 months	The study population consisted of 240 employees of the Japan Maritime Self-Defense Force working on the Iwo Jima military defense base.	The patients in the intervention group received a pre-assignment medical exam 1 year before their assignment.	The patients in the comparison group didn't receive a pre-assignment medical exam 1 year before their assignment.		1. Disease detection	The PHE was defined as an annual health examination including analysis of height, body weight, blood pressure, BMI, chest radiography, electrocardiography, vital capacity, serum chemistry, stool samples, and urine samples.	External Validity Concerns:
							- Cardiac arrhythmia		1. Results potentially not generalizable beyond Japanese males.
							- Neurological problems		Statistical Validity Concerns:
							Hyperlipidemia		1. Potentially inadequate adjustment for residual confounding
							- GI ulcers
							Hypertension
							- Severe obesity (BMI >28.6 kg/m2)
							- Proteinuria
							2. Blood pressure
							- Mean systolic blood pressure
							- Mean diastolic blood pressure
							- Proportion of hypertension
							3. Changes in serum cholesterol levels
							- Mean total cholesterol
							- Mean LDL cholesterol
							- Mean triglycerides
							- Mean HDL cholesterol
							- Proportion of hyperlipidemia
							4. Body mass index
							- Mean BMI;
							- Proportion of severe obesity (BMI ≥ 28.6)
Freedman, 200016	1995	18 months	The study population consisted of 136 community-dwelling patients aged 70 and older.	The intervention group included patients who received a periodic health examination.	The comparison group received no periodic health examination and attended clinic 3 or more times.	Intervention Group:	1. Immunization	The PHE in this study included screening for smoking, alcohol, influenza vaccination, tetanus vaccination, exercise, nutrition, blood pressure, hearing, and vision.	Internal Validity Concerns:
						Providers:	- Received influenza vaccine		1. Reporting on differences between enrollees and non-enrollees
						1. Chart-based reminders	- Received tetanus vaccine		External Validity Concerns:
									1.Description of study inclusion/exclusion criteria not detailed
									2. Description of study population characteristics not detailed
									3. Description of PHE is not detailed
									Statistical Validity Concerns:
									1. Potentially inadequate adjustment for residual confounding
									2. Incomplete presentation of statistical significance
Williams,199865	1998*	12 months	The study population consisted of 50 adult patient's medical records before and 50 after intervention in each of 60 primary care practices.	The intervention group consisted of patients who received a Health Maintenance Exam (HME) and interacted with a touch-sensitive computer system (TSCS), which provided patient-specific preventive service recommendations.	The comparison group consisted of patients who had an HME and did not use a TSCS.	Intervention Group:	1. Pap smear	The PHE, called a HME in this study, was defined as an office visit specifically for a physical exam, breast examination, pap smear, and pelvic examination, or annual check-up.	Internal Validity Concerns:
			The study population also consisted of 507 touch-sensitive computer system users.			Patients:	2. Colon cancer screening		1. Reporting on differences between enrollees and non-enrollees
						1. touch-sensitive computer system	- Flexible sigmoidoscopy		External Validity Concerns:
						Providers:	- Fecal occult blood test		2. Description of study population characteristics not detailed
						1. touch-sensitive computer system	3. Mammogram		3. Description of PHE is not detailed
						Control Group:			Statistical Validity Concerns:
						Patients:			1. Potentially inadequate adjustment for residual confounding
						1. touch-sensitive computer system			2. Incomplete presentation of statistical significance
						Providers:
						1. touch-sensitive computer system
Bernacki, 198828	1983	36 months	The study population consisted of 710 male executives of a multinational US corporation.	The intervention group consisted of corporation executives that were eligible for periodic physical exam (PPE) and receiving 3 PPEs during 3-year study period.	The comparison group consisted of corporation executives eligible for PPE and not receiving a PPE during the 3-year study period.		1. Costs - health care claims cost per capita in Year 3	The PHE was described as a periodic physical examination that included a medical history, physical examination, visual acuity testing, resting electrocardiogram, multiple lab studies, audiometry, cervical cytology, chest radiograph, proctosig-moidoscopy, tonometry, pulmonary function test, maximal exercise electrocardiography, and a barium enema.	Internal Validity Concerns:
				The second intervention group were corporation executives eligible for PPE and receiving 1 or 2 PPEs during 3-year study period					1. Reporting on withdrawals
									2. Standard/valid reporting of outcomes
									External Validity Concerns:
									1. Description of study population characteristics not detailed
									Statistical Validity Concerns:
									1. Potentially inadequate adjustment for residual confounding
Grimaldi, 196561	1956	96 months	The study population consisted of 194 employees.	The intervention group consisted of corporation middle management employees opting to participate in a periodic physical examination (PPE).	The first comparison group consisted of corporation middle management employees opting not to participate in PPE.		1. Costs - mean medical expense per claim	The PHE was described as a preventive health examination that included a thorough self-administered health inventory questionnaire, a physical examination, a 14 × 17 x-ray film of the chest, audiometric testing, visual acuity, tonometry, 12-lead ECG, urinalysis for albumin and sugar, hematocrit and microscopic study of the blood smear, blood sugar determination, and a protoscopic examination when indicated.	External Validity Concerns:
					The second comparison group were employees from another site not offered the PPE.				1. Study performed before USPSTF or similar contemporary preventive services guidelines in effect
									2. Description of study population characteristics not detailed
									Statistical Validity Concerns:
									1. Potentially inadequate adjustment for residual confounding
									2. Incomplete presentation of statistical significance
Cross-sectional Studies
Lin, 200473	1997		The study population consisted of 21,025 patients who visited the outpatient department from 1997 through 2000. (NHAMCS data)	Patients receiving an outpatient department visit including a nurse practitioner defined the Intervention group.	Patients receiving an outpatient visit not including a nurse practitioner defined the comparison group.		1. Counseling	The PHE was defined as a non-illness care visit to the outpatient department.	External Validity Concerns:
							- Diet counseling		1. Data from pediatric outpatient clinics included
							- Injury prevention counseling		2. Study not specifically designed to address Key Question
							- Physical activity counseling		Statistical Validity Concerns:
							- Safe sexual practices counseling		1. Potentially inadequate adjustment for residual confounding
							- Tobacco use counseling
Flocke, 200474; Eaton, 200298	1994		The study population consisted of 2,670 adult outpatients, visiting 138 family physicians in 84 practices from October 1994 through August 1995 in Northeast Ohio.	The intervention group included patients who were seen by a health care professional for well care.	The first comparison group included patients who were seen for chronic illness.		1. Counseling	The PHE was defined as a well care visit with a health care professional.	Internal Validity Concerns:
					The second comparison group included patients who were seen for acute illness. The last comparison group included patients who were seen for things other than chronic illness, acute illness or well care.		- Patient diet advice recall		1. Reporting on differences between enrollees and non-enrollees
							- Patient smoking counseling recall		External Validity Concerns:
							- Physical activity patient recall		1. Study not specifically designed to address Key Question
							- Nutritional counseling—univariate analysis		Statistical Validity Concerns:
							total n = 3475		1. Potentially inadequate adjustment for residual confounding
							- Nutritional counseling—multivariate analysis
							total n = 3475
Finkelstein, 200267	1994		The study population consisted of 2,232 women aged 20 and older who were residents in Ontario, Canada that completed the National Population Health Survey reporting their use of annual examinations with answers linked to their use of services in a national health insurance plan	The intervention group received an annual health examination/ preventive screening.	The comparison group received no annual health examination/preventive screening.		1. Pap smear	The PHE was defined as an annual or periodic health examination by a healthcare professional.	Internal Validity Concerns:
							2. Cholesterol screening		1. Data obtained from questionnaire, results subject to recall bias
							3. Mammogram		External Validity Concerns:
									1. Description of study population characteristics not detailed
									2. Results potentially not generalizable beyond female population
Hahn, 199966	1995		The study population consisted of an audit of the computerized billing data of 75,621 cross-sectional audit of outpatient billing claims for adults seen at least once by a primary care provider in 1995, classified by visit type (visits for preventive care vs. acute care).	The first intervention group received preventive services with only HMO insurance	The first comparison group did not receive preventive services with HMO insurance.		1. Pap smear	The PHE was defined as a physical exam or preventive services.	External Validity Concerns:
				The other intervention group received preventive services with only FFS insurance.	The second comparison group did not receive preventive services with FFS insurance.		2. Immunization - tetanus		Description of study population characteristics not detailed
							3. Cholesterol screening		Study did not provide information on level of insurance for the PHE and screening tests in HMO and FFS plans
							4. Colon cancer screening		Statistical Validity Concerns:
							- Sigmoidoscopy		Potentially inadequate adjustment for residual confounding
							- Fecal occult blood testing
							5. Mammogram
Tao, 200164	1997		The study population consisted of data on women age >18 years from the 1997 National Ambulatory Medical Care and National Hospital Ambulatory Medical Care Surveys in which physicians completed forms describing reasons for ambulatory visits (including general medical visits or gynecological) and the receipt of preventive services	The intervention group received general medical or gynecologic exam as defined by either physician or patient.	The comparison group received non-general medical or gynecologic exam visits as defined by both patient and physician.		1. Pap smear	The PHE was defined as a general medical examination, gynecologic exam, or periodic health examination.	Internal Validity Concerns:
							2. Counseling - family planning or contraceptive given		1. Data obtained from questionnaire, results subject to recall bias
							3. Mammogram		External Validity Concerns:
									1. Study not specifically designed to address Key Question
									Statistical Validity Concerns:
									1. Potentially inadequate adjustment for residual confounding
Parchman,200169	1996		The study population consisted of 1,409 Mexican American El Paso County, Texas residents, aged 18 to 64 years of age participating in a telephone and door-to door survey designed to assess access to and use of ambulatory health care	The intervention group reported they had received a check-up in the past year.	The comparison group reported they hadn't received a check-up in the past year.		1. Pap smear	The PHE was defined as a check-up or visit to a healthcare professional.	Internal Validity Concerns:
							2. Cholesterol screening in past 5 years		1. Data obtained from questionnaire, results subject to recall bias
							3. Mammogram		External Validity Concerns:
									1. Study not specifically designed to address Key Question
									Statistical Validity Concerns:
									1. Potentially inadequate adjustment for residual confounding
Nutting, 200175	1991		The study population consisted of 1,138 patient-visits by 93 physicians in the Ambulatory Sentinel Practice Network, in 50 community-based practices. Physicians were surveyed to recall the content of nonacute care visits with women age 40–75 years seen in their practices.	The intervention group had an annual examination.	The comparison group had a routine chronic are visit.		1. Mammogram	The PHE was described as a routine annual examination that didn't included visits for chronic care, intercurrent illness, emergent conditions, or injuries.	Internal Validity Concerns:
									1. Data obtained from questionnaire, results subject to recall bias
									External Validity Concerns:
									1. Description of PHE is not detailed
									Statistical Validity Concerns:
									1. Potentially inadequate adjustment for residual confounding
Stange, 200042	1994		The study population consisted of 4,049 patient-visits in the offices of 138 family physicians in North-east Ohio.	The intervention group was described by well care visits.	The comparison group was described by illness visits.		1. Counseling	The PHE was described as preventive services that consisted of screening, health habit counseling, and immunization services.	External Validity Concerns:
							- Mean % (SD) receipt of USPTF health habits counseling		1. Description of study inclusion/exclusion criteria not detailed
							- Mean % (SD) receipt of cancer-related health habits counseling		Statistical Validity Concerns:
							2. Immunization		1. Potentially inadequate adjustment for residual confounding
							- Mean % receiving USPSTF recommended vaccinations		2. Incomplete presentation of statistical significance
Faulkner, 199722	1991		The study population consisted of 34,236 adults aged 18 to 64 from the Centers for Disease Control's 1991 Behavioral Risk Factor Surveillance System studied to assess the association of health insurance coverage with the receipt of preventive services.	The intervention consisted of a Behavioral Risk Factor Surveillance Survey 1991: All preventive services covered by health plan.	The first comparison group consisted of patients having most preventive services covered by a health plan.		1. Receipt of PHE odds of receiving check-up according to level of health insurance compared to no health insurance coverage	The PHE was defined as a period health exam where a patient has receipt of recommended services within the periodicity schedules recommended for specific age/gender groups.	Internal Validity Concerns:
					The second comparison group consisted of patients that had some preventive services covered by a health plan.		- Men, aged 18–39 years		1. Reporting on differences between enrollees and non-enrollees
							- Men, aged 40–64 years		External Validity Concerns:
							- Women, aged 18–39 years		1. Description of study inclusion/exclusion criteria not detailed
							- Women, aged 40–64 years		2. Description of study population characteristics not detailed
									3. Description of PHE is not detailed
									Statistical Validity Concerns:
									1. Potentially inadequate adjustment for residual confounding
									2. Incomplete presentation of statistical significance
Kottke, 199731	1994		The study consisted of 6,830 randomly selected patients from 44 primary-care clinics who completed a mail survey to ascertain their receipt of preventive services in the context of “check-up physical examinations” versus other types of visits	The intervention group consisted of patients who had a visit (reason for visit as declared by patient) for a health risk appraisal.	The first comparison group included patients with visits (reason for visit as declared by patient) for urgent problems.	Intervention Group:	1. Pap smear	The PHE was defined as a physical examination or check up that consisted of a breast exam, blood pressure measurement, pap smear, smoking cessation counseling, influenza & pneumovax immunization, cholesterol screening, and mammogram.	Internal Validity Concerns:
					The second comparison group included patients with visits (reason for visit as declared by patient) for continuing condition.	Providers:	2. Counseling - smoking cessation		1. Reporting on differences between enrollees and non-enrollees
					The third comparison group included patients with visits (reason for visit as declared by patient) for a follow-up.	1. continues quality improvement initiative	3. Immunization		External Validity Concerns:
					The fourth comparison group included patients with visits (reason for visit as declared by patient) other than for follow-up, continuing education, urgent problems, or health risk appraisal.		- Rate** influenza vaccine offered by providers		1. Description of study population characteristics not detailed
							- Rate** pneumococcal vaccine offered by providers		2. Description of PHE is not detailed
							4. Cholesterol screening**		Statistical Validity Concerns:
							5. Mammogram		1. Potentially inadequate adjustment for residual confounding
									2. Incomplete presentation of statistical significance
Sox, 199770	1992		The study population consisted of 2,775 patients of family physicians and general internists in Vermont, age 42 years and older, with no life-threatening illness, and recently visited a physician completing a questionnaire and agreeing to review of medical records to assess their receipt of a “periodic health examination” and their receipt of recommended clinical preventive services	The intervention group received a periodic health examination.	The comparison group received cancer-specific, age-appropriate and sex-appropriate exams during usual care.		1. Pap smear	The PHE is this study was described as a routine physical examination that wasn't for a particular illness, but for a general check-up.	Internal Validity Concerns:
							2. Counseling		1. Reporting on differences between enrollees and non-enrollees
							- dietary		External Validity Concerns:
							3. Colon cancer screening		1. Description of study inclusion/exclusion criteria not detailed
							- Mean proportion of persons in each practice receiving sigmoidoscopy		2. Description of study population characteristics not detailed
							- Mean proportion of persons in each practice receiving fecal occult blood testing		3. Description of PHE is not detailed
							4. Mammogram		Statistical Validity Concerns:
									1. Concern regarding unit of analysis employed in presentation of results
									2. Incomplete presentation of statistical significance
Slesinger, 197663	1973		The study population consisted of 989 employed individuals who had health insurance responding to household survey regarding their receipt of clinical preventive services in the past year	The intervention group consisted of employees who chose a prepaid group insurance plan.	The comparison group consisted of a random sampling of employees who chose the traditional Blue Cross/Blue Shield plan.	Intervention Group:	1. Pap smear	The PHE was described as a general physical check-up or Physical examination.	External Validity Concerns:
						Patients:	2. Receipt of PHE - receipt of general check-up in the past year		1. Study performed before USPSTF or similar contemporary preventive services guidelines in effect
						1. Comprehensive benefit package on a pre-payment basis			Statistical Validity Concerns:
						Control Group:			1. Potentially inadequate adjustment for residual confounding
						Patients:
						1. Did not offer prepaid comprehensive benefits package (no reimbursement for MD office visits or physical exams)
Nakanishi, 199682	1992		The study population consisted of 227,581 inpatient and outpatient claims of residents aged 40 and over in 9 cities in Japan.	The intervention group consisted of Japanese adults aged 40 years and older in the National Health Insurance program.			1. Costs	The PHE was described as a health examination that included (1) health check-ups as basic health examination (interview, body measurement, physical tests, blood pressure measurement, urinalysis and blood test) with special examinations when indicated, and screenings for stomach cancer (stomach radiography), uterus cancer (visual examination, cytodiagnosis and internal examination as cervical cancer screening and cytodiagnosis as uterine body screening), lung cancer (chest radiography and phlegm cellular test), breast cancer (visual examination and palpation), and colon cancer (occult blood test, starting in 1992) (for uterus cancer screening and breast cancer screening, women aged 30 or more are eligible); (2) issuance of a health notebook (recording health examinations, providing the eligibility of patients to receive medical care and maintaining medical care records); (3) health education (health classes using brochures, posters, cable broadcasting, etc.); (4) individual health counseling; (5) rehabilitation programs; and (6) home-visit guidance (nursing techniques, treatment methods, training for activities fordaily living at home, etc.).	Internal Validity Concerns:
							- Inpatient cost per insured person (yen) correlated with rate of use of health check-ups		1. Reporting on differences between enrollees and non-enrollees
							- High inpatient cost (600,000 yen or more) correlated with rate of use of health check-ups		External Validity Concerns:
							- Outpatient cost per insured person correlated with rate of use of health check-ups		1. Results potentially not generalizable beyond Japanese population
							2. Hospitalization
							- Hospital admission rate per 1000 insured persons correlation with rate of use of health checkups
							- Length of stay of 180 days or more per 1000 insured persons correlated with rate of use of health checkups
Somkin, 200468	1999		The study population consisted of, 463 subjects aged 40 to 74 residing in Alameda County, California respondents to a telephone survey assessing their access to and satisfaction with preventive services	The intervention group included persons reporting they received a check-up in the last 12 months.	The comparison group included persons reporting they had not received a check-up in the last 12 months.		1. Pap smear	The PHE was defined as a check-up in the last 12 months.	External Validity Concerns:
							2. Mammogram		1. Study not specifically designed to address Key Question
Pre-post Studies
Schneider, 200371	1999	2 months	The study population consisted of 319 adult patients in an ambulatory family practice residency clinic in which physicians participated in a quality improvement program to enhance the delivery of the “health maintenance examination” and clinical preventive services	Patients received written material, and a reminder phone call.		Intervention Group:	1. Pap smear	The PHE was defined as a health maintenance examination.	External Validity Concerns:
				Providers received education on prevention measures.		Patients:	2. Counseling		1. Study not specifically designed to address Key Question
						1. Patient-held medical records	- Exercise counseling		Statistical Validity Concerns:
						Providers:	- Diet counseling		1. Potentially inadequate adjustment for residual confounding
						1. Chart-based reminders	- Alcohol counseling
						2. Educational sessions on preventive measures	- Substance abuse counseling
						Control Group:	- Tobacco cessation counseling
						Patients:	3. Cholesterol screening
						1. Patient-held medical records	4. Colon cancer screening
						Providers:	- Percentage receiving fecal occult test.
						1. Chart-based reminders	- Percentage receiving sigmoidoscopy
						2. Educational sessions on preventive measures	5. Mammogram
							6. Receipt of PHE
Geiger, 199372	1990	5 months	The study population consisted of 23 family practice residents and faculty physicians who provided for 3,300 patients, using a practice-based teaching model to increase resident compliance with USPSTF guidelines	The intervention consisted of two sequential phases. In phase one, physicians were educated about providing preventive services in accordance with USPSTF guidelines in the context of a “health check.” In phase two, physicians were monitored for their delivery of a recommended preventive services during scheduled health checks.		Intervention Group:	1. Pap smear	The PHE was described as preventive services (health check; physical examination) that assessed blood pressure, breast exam, pap smear, height, weight, visual acuity, hearing, substance abuse activity, diet activity, injury prevention activity, oral health, and physical activity use. In addition, Influenza, pneumovax, & tetanus immunizations, cholesterol screening, mammography, urinalysis, and thyroid function.	Internal Validity Concerns:
						Patients:	2. Counseling		1. Standard/valid reporting of outcomes
						1. Written Material	- Substance abuse counseling		Statistical Validity Concerns:
						Control Group:	- Diet counseling		1. Potentially inadequate adjustment for residual confounding
						Patients:	- Oral health counseling (dental care)
						1. Written Material	- Physical activity counseling
							3. Cholesterol screening
							4. Mammogram

‡

Larger values indicate worse health.

¶

Colon/rectum, breast, cervix/uterine, prostate, and kidney cancer, hypertension, hypertensive cardiovascular disease, hemorrhagic cerebrovascular disease.

*

Date published; unspecified

**

Clinic weighted rate across 44 primary care clinics.

Table 3

Type and Number of Outcomes Reported in Studies

Table 3

Type and Number of Outcomes Reported in Studies

Type of outcomes and number reported in studies	n(%) studies
Delivery of Clinical Preventive Services
1 outcome	7(19)16,53,56,63,73–75
2 outcomes	2(6)42,68
3 or more outcomes	11(31)31,44,55,64–67,69–72
Proximal Clinical Outcomes
1 outcome	1(3)45
2 outcomes	2(6)56,57
3 or more outcomes	3(8)53,54,59,77
Distal Clinical and Economic
1 outcome	5(14)28,58,61,62,83
2 outcomes	4(11)29,53,55,82
3 or more outcomes	3(8)41,56,57
Interventions to improve receipt of PHE
1 outcome	5(14)22,60,63,71,84
All Outcomes (regardless of type)
1 outcome	13(36)45,58,60,62,83,8416,22,28,61,73–75
2 outcomes	5(14)29,42,63,68,82
3 or more outcomes	18(50)44,53–5731,41,59,64–67,69–72,77

Article Quality

Table 4

Quality of Identified Studies on the Value of the Periodic (more...)

Table 4

Quality of Identified Studies on the Value of the Periodic Health Evaluation

Author, year	External validity*	Internal validity†	Statistical Analysis‡	Total score§
Lin, 2004	low	high	high	high
Somkin, 2004	high	medium	high	high
Flocke, 2004; Eaton, 2002	medium	high	high	high
Schneider, 2003	medium	medium	medium	medium
Finkelstein, 2002	low	medium	high	medium
Hahn, 1999	medium	low	low	low
Chiou, 2002	low	low	high	medium
Burton, 2002	high	low	low	low
Tao, 2001	low	low	low	low
Parchman, 2001	low	medium	low	low
Nutting, 2001	low	low	medium	low
Hama, 2001	medium	high	medium	high
Patrick, 1999	medium	medium	low	low
Stange, 2000	high	high	medium	high
Freedman, 2000	high	high	medium	high
Williams, 1998	high	low	high	medium
Faulkner, 1997	low	high	medium	medium
Kottke, 1997	medium	low	high	medium
Sox, 1997	low	medium	medium	medium
Elder, 1995; Mayer, 1994	high	medium	high	high
Christensen, 1995	high	medium	high	high
Morrissey, 1995	high	high	high	high
Burton, 1995; German, 1995; Burton, 1997; Burton, 1995	high	high	medium	high
Morrissey, 1995	medium	high	medium	high
Norman, 1992	medium	high	low	medium
Belcher, 1990	medium	medium	medium	low
Bernacki, 1988	low	low	low	low
Stone, 1981; Stone, 1978; South-east London, 1977; Trevelyan, 1973; South-east London, 2001	high	medium	low	medium
Fletcher, 1977	high	low	medium	medium
Slesinger, 1976	low	low	low	low
Cutler, 1973; Collen, 1973; Dales, 1973; Ramcharan, 1973; Friedman, 1986; Dales, 1979; Norinder, 2002	high	medium	low	low
Robert, 1969	low	low	low	low
Grimaldi, 1965	low	low	low	low
Theobald, 1998	medium	low	medium	low
OXCHECK, 1995a; OXCHECK, 1995b	medium	medium	low	medium
Nakanishi, 1996	medium	high	medium	high

*

External validity includes quality of reporting on study inclusion/exclusion criteria, characteristics of study population, description of periodic health evaluation (PHE) or interventions to change the delivery of the PHE, and description of outcomes. See text under “Article summary quality” for more detail regarding assessment of quality.

†

Internal validity includes assessment of randomization scheme (for trials), appropriateness of control group (for trials), assessment of those who enrolled versus those who did not enroll, assessment of withdrawals, blinding of intervention assignment and outcome assessment (for trials), and adequacy of outcome measurement. See text under “Article summary quality” for more detail regarding assessment of quality.

‡

Statistical analysis quality includes reporting on sample size calculations, presentation of statistical significance, and appropriateness of statistical methods. See text under “Article summary quality” for more detail regarding assessment of quality.

§

Scores of high, medium, or low indicate that the article scored in the highest, middle, or lowest tertile of scores.

The quality of reporting on studies varied, with only one study receiving quality scores in the highest tertile for external validity, internal validity, and quality of statistical analysis.⁵⁵ (Table 4) The majority of studies received varied ratings for external validity and internal validity, while five studies received scores in the lowest tertile for external validity, internal validity, and quality of statistical analysis (Table 4).^{28, 61}–⁶⁴ Total quality scores for experimental trials were generally high (median score of 68 (range 56 to 87) out of 100 total possible points). Total quality scores for observational studies were also generally high, but with more variability in range (median score of 63 (range 37 to 77) out of 100 total possible points). Total quality scores are included in the evidence tables Appendix G).

Key Question 1: What Definitions are Used for the Adult PHE in Studies of its Value?

Summary of findings. Definitions of the PHE were heterogeneous. While central elements used to define the PHE included the clinical history and risk assessment of patients and a physical examination, the specific composition of these central elements varied among studies. The most frequently cited types of history and risk assessment performed were assessment of dietary, alcohol/substance abuse, and tobacco smoking risks; the least frequently cited types of risk assessment included assessment of calcium and folic acid intake. In many cases, the physical examination was referred to with no specific clarification of what components were included. When specified, the most frequently cited components of the examination were assessment of blood pressure, weight and height, breast examination, gynecological examination, and rectal examination; the least frequently cited components included neurological and foot examinations.

Table 5

Components Which were Included (or may have been included) (more...)

Table 5

Components Which were Included (or may have been included) in Studies on the Periodic Health Evaluation (N=36)

Component of PHE	n(%)
History and Risk Assessment
Tobacco smoking	14(39)
Alcohol/ substance abuse	13(36)
Dietary risk factors	12(33)
Physical Activity	10(27)
Injury prevention	6(17)
Safe sexual practices	6(17)
Sun exposure	4(11)
Oral health	4(11)
Medications/ Poly-pharmacy	4(11)
Calcium intake	2(6)
Folic acid intake	2(6)
Physical Examination
Blood pressure assessment	18(50)
Examination (not otherwise specified)	14(39)
Breast examination	12(33)
Weight	12(33)
Height	10(28)
Gynecological examination	10(28)
Cardiovascular examination	5(14)
Pulmonary examination	5(14)
Eye* examination	5(14)
Pulse	4(11)
Rectal examination	4(11)
Prostate examination	4(11)
Abdominal examination	4(11)
Neurological examination	3(8)
Foot examination	2(6)
Other†	13(36)

*

fundoscopic

†

vision testing, tonometry, audiometry

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?

Studies addressing Key Question 2 included studies reporting on the association of receipt of the PHE with: a) the delivery/receipt of seven clinical preventive services (gynecological examination/Pap smear, counseling, immunizations, cholesterol screening, colon cancer screening, and mammography); b) seven proximal clinical outcomes (disease detection, patient health habits, patient attitudes, health status, blood pressure, serum cholesterol, and BMI); c) three distal clinical outcomes (disability, hospitalization, and mortality); and d) economic outcomes (costs and cost-effectiveness).

Delivery of Clinical Preventive Services

Gynecological Examination/Pap Smear

Summary of findings. Thirteen studies (including two RCTs and eleven observational studies) evaluated the association of receiving the PHE with delivery/receipt of the gynecological examination/Pap smear. The best available evidence to assess this outcome was comprised of two large RCTs, performed in the late 1980's, and it was deemed to be of “high” grade based on standard criteria. In these studies, the PHE had small to large positive effects on the receipt of the gynecological examination/Pap smear (see below for details). While these RCTs were specifically designed to assess the effect of the PHE on this outcome, they focused on Medicare recipients, and thus may be limited in their generalizability to other populations. Observational studies of the association of receipt of the PHE with receipt of the gynecological examination/Pap smear revealed both positive and mixed results. Observational studies had a variety of limitations, including potential confounding of results not accounted for and use of data subject to recall bias.

Findings

Table 6

Grading of the Overall Strength of Evidence on the (more...)

Table 6

Grading of the Overall Strength of Evidence on the Value of the Periodic Health Evaluation

Outcome	Pap smear	Counseling	Immunizations	Choles-terol	Colon cancer	Mammogram	Disease detection	Health habits	Patient attitudes	Health Status	Blood Pressure	Serum Cholesterol	Body Mass Index	Costs	Disability	Hospitalization	Mortality	Receipt of PHE
Number of studies§	2 (2)	7 (1)	3 (3)	5(1)	2 (2)	2 (1)	2 (2)	5 (5)	1 (1)	2 (2)	2 (2)	2 (1)	3 (3)	4(4) ‡	2 (2)	3 (3)	5 (5)	2 (1)
Strength of study design*	4	3	4	3	4	3	4	4	4	4	4	3	4	4	4	4	4	3
Did the studies have serious (-1) or very serious (-2) limitations in quality? (Enter 0 if none)	-0.5	-1	-1	-1	-1	-1	-1	-1	-1	-1	-0.5	-1	-1	-1	-1	-0.5	-1	-0.5
Did the studies have important inconsistency? (-1)	0	-0.5	-1	0	0	-0.5	-1	-1	0	-0.5	-0.5	-0.5	-1	-1	-1	-0.5	-1	0
Were data imprecise or sparse? (-1)	0	0	0	0	0	0	0	0	-1	0	0	0	0	0	0	0	0	0
Did the studies have high probability of reporting bias? (-1)	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
Did the studies show strong evidence of association between intervention and recruitment outcome?†	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	0	0	0
Overall grade of evidence (high, medium, low, very low)	3.5 High	1.5 Low	2 Med	2 Med	4 High	1.5 Low	2 Med	2 Med	2 Med	2.5 Medium	3 High	1.5 Low	2 Med	2 Med	2 Med	3 High	2 Med	2.5 Med

*

Were study designs randomized trials (high quality), non-randomized controlled trials (medium quality), or observational studies (low quality)?

†

Evidence was deemed “strong” if significant relative risk or odds ratio >2 (or Cohen's d ≥0.8) based on consistent evidence from 2 or more studies with no plausible confounders (+1); “very strong” if significant relative risk or odds ratio >5 based on direct evidence with no major threats to validity (+2).

PHE = periodic health evaluation

§

Parentheses contain number of randomized controlled trials considered among the best available evidence.

‡

One study evaluating cost-effectiveness as cost per percent coronary risk reduction not included⁸¹ due to inability to assess direction of results.

Table 7

Comparison of Effect Sizes in Randomized Controlled (more...)

Table 7

Comparison of Effect Sizes in Randomized Controlled Trials

	Effect size (95% CI)			Not able to calculate ES
	Outcomes in studies with positive effect of PHE	Outcomes in studies with negative effect of PHE	Confidence interval crosses 0
Outcomes
Receipt of Pap smear	1.71 (1.69, 1.73)55
	0.07 (0.07, 0.07)80
Preventive counseling	1.09 (1.08,1.11)44c
	1.19 (1.17, 1.21)44b
Immunizations	0.35 (0.33, 0.36)55a	-0.22 (-0.24, -0.20)44a
	0.10 (0.10, 0.10)53a
Cholesterol screening	0.02 (0.00, 0.04)55
Colon cancer screening (fecal occult blood testing)	1.19 (1.17, 1.21)55
	1.07 (1.05, 1.08)44
Mammography	0.14 (0.12, 0.16)55
Disease detection	0.03 (0.02, 0.03)76f	-0.01 (-0.01, -0.01)76d	-0.01 (-0.01, 0.00)76e
	0.96 (0.84, 1.08)45a	-0.03 (-0.03, -0.03)76g
	0.53 (0.41, 0.64)45b
Health habits	0.28 (0.14, 0.42)54a	-0.040 (-0.043, -0.037)53h	0.000 (-0.14, 0.14)54b
	0.120 (0.117, 0.123)53b	-0.014 (-0.016, -0.012)76c	0.01 (-0.13, 0.15)54c
	0.040 (0.037, 0.043)53c	-.02 (-.03, -.02)78b	0.02 (-0.12, 0.16)54d
	0.345 (0.342, 0.348)53d		0.05 (-0.09, 0.19)54e
	0.080 (0.077, 0.083)53e		0.01 (-0.13, 0.15)54f
	0.020 (0.017, 0.023)53f
	0.020 (0.017, 0.023 )53g
	0.100 (0.098, 0.102)59a
	0.032 (0.030, 0.034)59b
	0.088 (0.086, 0.090)59c
	0.244 (0.242, 0.246)59d
	0.250 (0.248, 0.252)59e
	0.13 (0.11,0.14)78a
Patient attitudes				53
Health Status				56
Blood Pressure	0.12 (0.02,0.21)54g		0.03 (-0.06, 0.13)54h
	0.11(0.04, 0.18)59f
	0.13 (0.06, 0.19)59g
	0.022 (0.019, 0.024)59h
Changes in serum cholesterol levels	0.22 (0.16, 0.29)59k
	0.09 (0.09, 0.10)59l
Body Mass Index	0.087 (0.022, 0.153)59i	-0.020 (-0.023,-0.017)53i	-0.031 (-0.170, 0.108)54i
	0.032 (0.030, 0.034)59j		-0.036 (-0.174, 0.103)54j
Reduction in health care costs			0.06 (-0.03, 0.15)55d	47,53,56
			0.05(-0.04, 0.14)55e
Reduction in disability	0.060 (0.054, 0.066)96a	-0.014 (-0.016, -0.012)76
Reduction in hospitalizations	0.01 (0.00, 0.01)76a		0.02 (-0.07, 0.11)55b	56b,c,d,e
			-0.04 (-0.13, 0.05)55c
Reduction in all-cause mortality	0.06 (0.05, 0.06)56a	-0.03 (-0.04, -0.03)53a	Rate ratio: 1.03 (0.94,1.14)58
	0.004 (0.004, 0.005)9a	-0.002 (-0.003, -0.0003)76b
Receipt of PHE (Question 4)	0.69 (0.68,0.70)60

ES = effect size; CI = confidence interval

Citation(⁵⁵) a: influenza vaccination, b: hospital days per enrollee, c: Admissions per enrollee, d: 3-year post-intervention cumulative Medicare charges; e: 3-year post-intervention cumulative Medicare reimbursement

Citation(⁴⁵) a: disease detection of ALL problems before and after intervention, b: disease detection of important problems before and after intervention;

Citation(⁵³) a: Influenza vaccination, b: Physical activity, c: Diet (fat and fiber), d: Advance directives, e: Breast self-exam, f: Smoking, g: Alcohol, h: Seat belt use; I: at risk for obesity, 24-month F/U

Citation(⁴⁴) a: influenza vaccination, b: alcohol abuse, c: smoking cessation

Citation(⁷⁶) a: Hospitalizations, b: Mortality rate per 1000 person-years at risk: all cause death, c: percentage still smoking, d: angina, e: high diastolic blood pressure, f: ischemia on electrocardiogram, g: bronchitis symptoms;

Citation(⁵⁶) a: Death; b: Mean inpatient days for the intervention and control groups who had a hospital discharge in that year (Year 1), c: Mean inpatient days Year 2, d: Hospital discharges per 1000 Year 1, e: Hospital discharges per 1000 Year 2;

Citation(⁵³) a: Mortality at 48 months;

Citation(⁹) a: Deaths, rate per 1000 persons 16 years;

Citation(⁹⁶) a: Disability at 11 years

Citation(⁵⁴) a: fiber servings per day, b: fat servings per week, c: salt use, d: caffeine drinks per day, e: stretching minutes per week, f: consumption of cruciferous foods; g: mean systolic blood pressure at 12 months; h: mean diastolic blood pressure at 12 months; i: mean BMI at 24 months (end of intervention period); j: mean BMI at 48 months (end of F/U)

Citation(⁵⁹) a: smoking, b: alcohol use, c: exercise less than once per month, d: use full cream milk, e: use butter or hard margarine; f: systolic blood pressure at 3-year follow up; g: diastolic blood pressure at 3-year follow up; h: proportion of high risk diastolic pressure (≥100mm Hg) from 3 year F/U when compared to control; i: mean BMI at 3-year F/U; j: percentage of participants with BMI ≥ 30; k: mean total cholesterol at 3-year F/U; l: proportion of high risk cholesterol (≥8mmol/l) at 3 year F/U

Citation(⁷⁸) a: smoking, b: problem alcohol drinking

Strength and limitations of the evidence. The best available evidence assessing this outcome, emanating from two large RCTs, received an overall grade of “high.” In grading the evidence, these studies were found to have few serious limitations in quality, no important inconsistency with regard to the direction of effects, adequate data, and a low probability of reporting bias (Table 6). These two studies evaluated the effect of the PHE on the delivery/receipt of the gynecological examination/Pap smear among community dwelling Medicare recipients who received reimbursement for the PHE compared to Medicare recipients receiving usual care. While these studies were limited in their generalizability to other ambulatory populations, they were specifically designed to evaluate the effect of the PHE on delivery/receipt of clinical preventive services and therefore directly addressed Key Question 2 (Tables 2, 6 and 7).

Randomized controlled trials. Two randomized controlled trials performed in 1988 and 1989 studied Medicare recipients (over 5000 patient combined total).^{55, 56} Follow up of patients ranged from twelve to 26 months. In one study, funded as a Medicare demonstration project, the PHE was delivered in the context of a “preventive care package” in which patients received clinical preventive services including annual history and physical on at least an annual basis, delivered by both a nurse and a physician. The comparison group received clinical preventive services as customarily offered in their physicians' practices.⁵⁵ In the other study, also funded as a Medicare demonstration project in which participants in the intervention group received vouchers for free preventive visits to be delivered by participants' primary care physicians, the PHE was described as consisting of a history and physical examination followed by the provision of USPSTF recommended clinical preventive services. The comparison group received no coverage for annual preventive visits or tests.⁵⁶ The PHE had a small positive effect (Cohen's d (95% confidence interval (CI)): 0.07 (0.07,0.07)) to a large positive effect (Cohen's d (95% CI):1.71 (1.69,1.73)) on delivery/receipt of the gynecological examination/Pap smear (Table 7, Evidence Table 1a). Limitations of these studies include their potential limited generalizability to non-Medicare populations (Table 2).

Table 8

Number of Studies (presented in cells) Reporting Outcomes (more...)

Table 8

Number of Studies (presented in cells) Reporting Outcomes According to Study Design and Direction of Results Reported

	Number of Studies According to Study Design and Direction‡ of Results															Totals
	Experimental						Observational
	RCT* (N†= 11)			Non-RCT* (N†= 1)			Cohort* (N†= 7)			Cross-Sectional* (N†= 14)			Pre-Post* (N†= 3)			Experimental			Observational
Examined Outcome	+PHE	øPHE	-PHE	+PHE	øPHE	-PHE	+PHE	øPHE	-PHE	+PHE	øPHE	-PHE	+PHE	øPHE	-PHE	+PHE	øPHE	-PHE	+PHE	øPHE	-PHE
Delivery of Clinical Preventive Services
Physical Examination
Pap Smear	2							1		7	1		1	1		2			8	3
Preventive Counseling
Counseling	1									5		1	1	1		1			6	1	1
Preventive Immunizations
Immunizations	2		1				1			3			2			2		1	6
Laboratory, Radiological Testing
Cholesterol Screening	1									4				2		1			4	2
Colon Cancer Screening	2							1		2			1			2			3	1
Mammography	1							1		7		1		2		1			7	3	1
Proximal Clinical Outcomes
Disease Detection	1	1					1									1	1		1
Health Habits	1	3	1													1	3	1
Patient Attitudes	1															1
Health Status	1	1														1	1
Blood Pressure	1	1						1								1	1			1
Serum Cholesterol	1							1								1				1
Body Mass Index	1	1	1					1								1	1	1		1
Distal Clinical and Economic Outcomes
Costs£	1	2	1				2		1	1						1	2	1	3		1
Disability	1		1				1									1		1	1
Hospitalization	1	2								1						1	2		1
Mortality	2	2	1				2									2	2	1	2
Interventions to improve receipt of PHE	1			1						1	1		2			2			2	1

*

Study design definitions: RCT=randomized controlled trial (study of two groups randomly assigned to intervention (versus control)); non-RCT=non-randomized controlled trial (study of two groups randomly assigned to intervention (versus control)—intervention assignment not random); Cohort (study with prospective or retrospective longitudinal observation of study population (no intervention assignment)); Cross-sectional (study population observed at one point in time (no intervention assignment, no prospective or retrospective observation); Pre-Post=Pre-post observational design (one study group in which baseline measurements are taken (pre-intervention phase). These measurements are repeated on the same study group following the implementation of an intervention (post-intervention phase).

†

N represents total number for entire review.

‡

Direction of results: +PHE = Articles reporting the PHE improves delivery (or is associated with improved delivery) of clinical preventive services, proximal clinical outcomes, or distal and economic outcomes; -PHE = Articles reporting the PHE worsens delivery (or is associated with worse delivery) of clinical preventive services, proximal clinical outcomes, or distal and economic outcomes; ø PHE = Articles reporting mixed results (positive, negative, or neutral) with regard to the association of receipt of the PHE with clinical outcomes.

£

One RCT examining cost-effectiveness is not included because of the inability to assess direction of results.⁸¹

Observational studies. Observational studies evaluating the association of receipt of the PHE with delivery/receipt of the gynecological examination/Pap smear included one retrospective cohort study, eight cross-sectional studies, and two observational studies with pre-post design performed from 1976 to 2004. Study populations for these studies included patients seen in community practices who interacted (or did not interact) with a touch-sensitive computer system placed in primary care practices to promote the delivery of preventive services,⁶⁵ a cross-sectional audit of outpatient billing claims for adults seen at least once by a primary care provider classified by visit type (visits for preventive care vs. acute care),⁶⁶ female residents in Ontario, Canada who completed the National Population Health Survey reporting their use of annual examinations with answers linked to their use of services in a national health insurance plan,⁶⁷ respondents to a California telephone survey who were contacted to assess their access to preventive services and satisfaction with preventive services,⁶⁸ data from the National Ambulatory Medical Care and National Hospital Ambulatory Medical Care Surveys in which physicians completed forms describing reasons for ambulatory visits (including general medical visits or gynecological) and the receipt of preventive services,⁶⁴ Mexican-American participants in a telephone and door-to-door survey designed to assess access to and use of ambulatory health care,⁶⁹ patients randomly selected from 44 ambulatory outpatient clinics who completed a survey to ascertain their receipt of preventive services in the context of “checkup physical examinations” versus other types of visits,³¹ patients from randomly selected community practices agreeing to complete a questionnaire and medical record review to assess their receipt of a “periodic health examination” and their receipt of recommended clinical preventive services,⁷⁰ employed health insurance enrollees responding to a questionnaire regarding the receipt of clinical preventive services in the past year,⁶³ patients in an ambulatory family practice residency clinic in which physicians participated in a quality improvement program to enhance the delivery of the “health maintenance examination” and clinical preventive services,⁷¹ and family practice residents and faculty physicians using a practice-based teaching model to increase resident compliance with USPSTF guidelines.⁷² Eight of these observational studies reported a positive association between receipt of the PHE and delivery/receipt of the gynecological examination/Pap smear, while three of these studies reported mixed results (Table 8, Evidence Tables 1b–d). Several limitations were noted among these observational studies including inability to completely control for potential confounding in several of the studies, lack of detail in studies' descriptions of the PHE, and the potential for recall bias in studies based on participant interviews/questionnaire responses (Table 2).

Preventive Counseling

Summary of findings. Thirteen studies (including one RCT and eight observational studies) evaluated the association of receiving the PHE with delivery/receipt of preventive counseling. A variety of types of counseling were examined within studies, including counseling regarding diet (6 studies), regarding physical activity (9 studies), smoking cessation (9 studies), alcohol/substance abuse (8 studies), injury prevention (3 studies), safe sexual practices (3 studies), calcium intake (one study), oral health (one study), sun exposure (one study), and general counseling (not otherwise specified) (one study). Four studies reported on other types of counseling. The delivery of all types of preventive counseling among studies was treated as a single outcome. The best available evidence to assess this outcome emanated from one RCT and six cross-sectional observational studies performed from1981 to 2004, and it was deemed to be of “low” grade based on standard criteria. Most studies reported a positive association of receiving a PHE with the delivery/receipt of preventive counseling with a strongly positive effect rendered by the PHE on delivery/receipt of smoking cessation counseling and alcohol abuse counseling in the RCT. Five observational studies reporting moderate to large positive associations of receipt of the PHE with receipt of counseling, while one observational study reported a negative association. The RCT was noted to have poor description of the study population and the PHE itself as well as its potentially limited generalizability to persons receiving care in the Veterans Affairs setting. However, this study did directly address Key Question 2. The seven cross-sectional studies were noted to have several limitations, including not directly addressing Key question 2, inability to completely control for potential confounding in several of the studies, lack of detail in studies' descriptions of the PHE, and the potential for recall bias in studies based on participant interviews/questionnaire responses.

Findings

Strength and limitations of the evidence. The best available evidence assessing this outcome, emanating from one RCT and six cross-sectional observational studies received an overall grade of “low.” In grading the evidence, these studies were found to have serious limitations in quality, moderate inconsistency in the direction of results (one observational study reporting a negative association with remaining studies reporting positive associations) but adequate data and low probability of reporting bias (Table 6). Although the RCT was designed to assess this outcome directly, its limitations included poor description of the study population and the PHE itself as well as its potentially limited generalizability to persons receiving care in the VA setting.⁴⁴ The seven cross-sectional studies were noted to have several limitations, including not being designed to specifically assess this outcome, inability to completely control for potential confounding in several of the studies, lack of detail in studies' descriptions of the PHE, and the potential for recall bias in studies based on participant interviews/questionnaire responses (Tables 2, 6 and 7).

Randomized controlled trial. In this study, patients attending a VA medical center were randomized to receive the PHE in the context of a “health promotion clinic” versus usual care.⁴⁴The study measured the delivery of both alcohol abuse counseling and smoking cessation counseling.⁴⁴ The study began in 1981 with follow up for five years. Limitations of this study included lack of detail provided on the study population or the content of the PHE and limited generalizability. Delivery/receipt of alcohol abuse and smoking counseling were improved by randomization to the health promotion clinic in this study (Cohen's d (95% (CI)): 1.18 (1.17,1.21) and 1.09 (1.08,1.11), respectively) (Tables 2 and 7, Evidence Table 2a).

Observational studies. Observational studies evaluating the delivery/receipt of preventive counseling included six cross-sectional studies and two observational studies with pre-post design performed from 1993 to 2004. Study populations included reports from ambulatory patients across the U.S.,⁷³ data from the National Ambulatory Medical Care and National Hospital Ambulatory Medical Care Surveys in which physicians completed forms describing reasons for ambulatory visits (including general medical visits or gynecological) and the receipt of preventive services,⁶⁴ outpatients seen in family practices in Ohio,^{42, 74} patients randomly selected from 44 ambulatory outpatient clinics who completed a survey to ascertain their receipt of preventive services in the context of “checkup physical examinations” versus other types of visits,³¹ patients from randomly selected community practices agreeing to complete a questionnaire and medical record review to assess their receipt of a “periodic health examination” and their receipt of recommended clinical preventive services,⁷⁰ patients in an ambulatory family practice residency clinic in which physicians participated in a quality improvement program to enhance the delivery of the “health maintenance examination” and clinical preventive services,⁷¹ and family practice residents and faculty physicians using a practice-based teaching model to increase resident compliance with USPSTF guidelines.⁷² Six of these observational studies reported positive associations of receipt of the PHE with receipt of counseling, while one study reported a negative association and one reported mixed results (Table 8, Evidence Tables 2b–c). Several limitations were noted among these observational studies including inability to completely control for potential confounding in several of the studies, lack of detail in studies' descriptions of the PHE, studies not specifically designed to examine Key Question 2, and the potential for recall bias in studies based on participant interviews/questionnaire responses (Table 2).

Preventive Immunizations

Summary of findings. Nine studies (including three randomized controlled trials and six observational studies) evaluated the association of receiving the PHE with delivery/receipt of preventive immunizations. The association of receiving the PHE with delivery of a variety of immunizations was examined within studies, including the delivery of influenza (7 studies), tetanus (6 studies) and pneumonia (4 studies) vaccinations. Two studies reported on the delivery of other immunizations. The delivery of all types of preventive immunization among studies was treated as a single outcome. The best available evidence assessing this outcome, emanating from three RCTs performed from 1981 to 1999, and it was deemed to be of “medium” grade based on standard criteria. Results in these three RCTs were mixed with two studies reporting small to medium sized positive effects (two studies) and one study reporting a small negative effect of the PHE on delivery/receipt of preventive immunization. While these RCTs were specifically designed to assess the effect of the PHE on this outcome, they focused on Medicare recipients and patients of a Veterans Affairs medical center and thus may be limited in their generalizability to other populations. Six observational studies reported the PHE improved the delivery/receipt of preventive immunizations. Several limitations were noted among these observational studies including inability to completely control for potential confounding in several of the studies, lack of detail in studies' descriptions of the PHE or study populations, and studies not specifically designed to assess this outcome.

Findings

Strength and limitations of the evidence. The best available evidence assessing this outcome, emanating from three RCTs, received an overall grade of “medium.” In grading the evidence, these studies were felt to have some serious limitations in quality, and important inconsistency in the direction of results (Table 6). These three studies evaluated the effect of the PHE on the delivery/receipt of preventive immunizations among Medicare enrollees and patients of a Veterans Affairs medical center compared to similar patients receiving usual care. While these studies were limited in their generalizability to other ambulatory populations, they were specifically designed to evaluate the effect of the PHE on delivery/receipt of clinical preventive services and therefore directly addressed Key Question 2 (Tables 2, 6 and 7).

Randomized controlled trials. Three randomized controlled trials performed from 1981 to 1999 studied Medicare recipients and patients attending a VA medical center (over 5000 patients combined total).^{44, 53, 55} Follow up of patients ranged from twelve months to five years. In one study, a Medicare demonstration project performed in 1993, the PHE was delivered in the context of a “preventive care package” in which patients received clinical preventive services including a health risk assessment, a health promotion visit (including health risk appraisals, positive behavior reinforcement and referrals for interventions where appropriate), disease prevention visit (visit with nurses and physicians who conducted history and physical examinations and reviewed patients' health risks), and follow up educational classes. The comparison group received clinical preventive services as customarily offered in their physicians practices.⁵³ In the second study, also a Medicare demonstration project, the PHE was delivered in the context of a “preventive care package” in which patients received clinical preventive services including annual history and physical on at least an annual basis, delivered by both a nurse and a physician. The comparison group received clinical preventive services as customarily offered in their physicians practices.⁵⁵ In the third study, patients attending a VA medical center were randomized to receive the PHE in the context of a “health promotion clinic” versus usual care.⁴⁴ The PHE improved delivery of preventive immunizations improved statistically significantly in two studies with effect sizes ranging from small to medium positive effects (Cohen's d (95% CI): 0.10 (0.10,0.10) and 0.35 (0.33,0.36), respectively).^{53 55} The PHE worsened delivery of preventive immunizations worsened in the group of VA patients attending a health promotion clinic, with a small magnitude of negative effect (Cohen's d (95% CI): -0.22(-0.20,-0.24))⁴⁴ (Table 7, Evidence Table 3a). These studies have potentially limited generalizability to non-Medicare or VA populations (Table 2).

Observational studies. Observational studies evaluating the association of receipt of the PHE with delivery/receipt of preventive immunizations included one retrospective cohort study, three cross-sectional studies, and two observational studies with pre-post design performed from 1993 to 2003. Study populations included community-dwelling patients aged 70 and older,¹⁶ a cross-sectional audit of outpatient billing claims for adults seen at least once by a primary care provider classified by visit type (visits for preventive care vs. acute care),⁶⁶ outpatients seen in family practices in Ohio,⁴² patients randomly selected from 44 ambulatory outpatient clinics who completed a survey to ascertain their receipt of preventive services in the context of “checkup physical examinations” versus other types of visits,³¹ patients in an ambulatory family practice residency clinic in which physicians participated in a quality improvement program to enhance the delivery of the “health maintenance examination” and clinical preventive services,⁷¹ and family practice residents and faculty physicians using a practice-based teaching model to increase resident compliance with USPSTF guidelines.⁷² All six of these observational studies reported a positive association between receipt of the PHE and the delivery of immunizations (Table 8, Evidence Tables 3b–c). Several limitations were noted among these observational studies including inability to completely control for potential confounding in several of the studies, lack of detail in studies' descriptions of the PHE or study populations, and studies not specifically designed to examine Key Question 2 (Table 2).

Cholesterol Screening

Summary of findings. Seven studies (including one RCT and six observational studies) evaluated the association of receiving the PHE with delivery/receipt of cholesterol screening. The best available evidence to assess this outcome was comprised of one RCT and four cross-sectional observational studies, performed from 1995 to 2003, and it was deemed to be of “medium” grade based on standard criteria. These studies demonstrated receiving the PHE was positively associated with receipt of cholesterol screening (small to large positive effect sizes). While the RCT was specifically designed to assess this outcome, it was limited to Medicare recipients and thus may be limited in its generalizability to other populations. The four cross-sectional observational studies had a variety of limitations, including the potential for recall bias in studies based on participant interviews/questionnaire responses, inability to completely control for potential confounding in several of the studies, and lack of detail in studies' descriptions of the PHE or study populations.

Findings

Strength and limitations of the evidence. The best available evidence assessing this outcome, emanating from one RCT and four cross-sectional observational studies, received an overall grade of “medium.” In grading the evidence, these studies were felt to have some serious limitations in quality, but they were consistent in the direction of findings, had adequate data to assess the outcome, and they had low probability of reporting bias (Table 6). The RCT evaluated the effect of the PHE on the delivery/receipt of the cholesterol screening among community dwelling Medicare recipients who received reimbursement for the PHE compared to Medicare recipients receiving usual care. While this study was limited in its generalizability to other ambulatory populations, it was specifically designed to evaluate the effect of the PHE on delivery/receipt of clinical preventive services and therefore directly addressed Key Question 2. The four cross-sectional studies were noted to have several limitations, including inability to completely control for potential confounding in several of the studies, lack of detail in studies' descriptions of the PHE and study populations, and the potential for recall bias in studies based on participant interviews/questionnaire responses (Tables 2, 6 and 7).

Randomized controlled trials. The RCT studied 455 Medicare recipients for 24 months beginning in 1995.⁵⁵ This study was a Medicare demonstration project in which the PHE was delivered in the context of a “preventive care package” in which patients received clinical preventive services including annual history and physical on at least an annual basis, delivered by both a nurse and a physician. The comparison group received clinical preventive services as customarily offered in their physicians practices.⁵⁵ Delivery/receipt of cholesterol screening was improved by randomization to the preventive care package in this study (Cohen's d (95% (CI)): 0.02 (0.00,0.04). This study was potentially limited in its lack of generalizability to non-Medicare populations (Tables 2 and 7, Evidence Table 4a).

Observational studies. Observational studies evaluating the delivery/receipt of cholesterol screening included four cross-sectional studies and two observational studies with pre-post designs performed from 1993–2003. Study populations included evaluated a variety of study subjects including female residents in Ontario, Canada who completed the National Population Health Survey reporting their use of annual examinations with answers linked to their use of services in a national health insurance plan,⁶⁷ a cross-sectional audit of outpatient billing claims for adults seen at least once by a primary care provider classified by visit type (visits for preventive care vs. acute care),⁶⁶ Mexican-American participants in a telephone and door-to-door survey designed to assess access to and use of ambulatory health care,⁶⁹ patients randomly selected from 44 ambulatory outpatient clinics who completed a survey to ascertain their receipt of preventive services in the context of “checkup physical examinations” versus other types of visits,³¹ patients in an ambulatory family practice residency clinic in which physicians participated in a quality improvement program to enhance the delivery of the “health maintenance examination” and clinical preventive services,⁷¹ and family practice residents and faculty physicians using a practice-based teaching model to increase resident compliance with USPSTF guidelines.⁷² In four cross-sectional studies, receipt of the PHE was positively associated with the delivery/receipt of cholesterol screening while both pre-post studies reported neutral results (Table 8, Evidence Tables 4b–c). Several limitations were noted among these observational studies including inability to completely control for potential confounding in several of the studies, lack of detail in studies' descriptions of the PHE or study populations, and the potential for recall bias in studies based on participant interviews/questionnaire responses (Table 2).

Colon Cancer Screening

Summary of findings. Six studies (including two randomized controlled trials and four observational studies) assessed the association of receipt of the PHE with delivery/receipt of colon cancer screening. Both the delivery of fecal occult blood testing (6 studies) and sigmoidoscopy (4 studies) were studied. The delivery of all types of colon cancer screening among studies was treated as a single outcome. The best available evidence to assess this outcome was comprised of two randomized controlled trials, performed from 1988 to 1995, and it was deemed to be of “high” quality based on standard criteria. These studies reported large positive effects of the PHE on the delivery/receipt of fecal occult blood testing. While these studies were specifically designed to assess this outcome, one was noted to have poor description of the study population and the PHE itself. Both studies were limited by their focus on Medicare populations and patients receiving care in the Veterans Affairs setting.

Findings

Strength and limitations of the evidence. Six studies assessed the effect of the PHE on delivery/receipt of colon cancer screening. The best available evidence assessing this outcome, emanating from two RCTs, received an overall grade of “high.” In grading the evidence, one study was felt to have serious limitations in quality. However, these studies did not have important inconsistency in terms of the direction of the results, had sufficient data to ascertain results, and both studies demonstrated a strong association between the intervention and the outcome (Table 6). These two studies evaluated the effect of the PHE on the receipt of colon cancer screening among Medicare enrollees and patients of a Veterans Affairs medical center compared to similar patients receiving usual care. While these studies were limited in their generalizability to other ambulatory populations, they were specifically designed to evaluate the effect of the PHE on receipt of clinical preventive services and therefore directly addressed Key Question 2. These studies only evaluated the delivery of fecal occult blood testing in the setting of the PHE (Tables 2, 6 and 7).

Randomized controlled trials. Two randomized controlled trials performed from 1988 to 1995 studied Medicare recipients and patients attending a VA medical center (over 1000 patient combined total).^{44, 55} Follow up of patients ranged from 24 months to five years. In one study, a Medicare demonstration project, the PHE was delivered in the context of a “preventive care package” in which patients received clinical preventive services including annual history and physical on at least an annual basis, delivered by both a nurse and a physician. The comparison group received clinical preventive services as customarily offered in their physicians practices.⁵⁵ In the second study, patients attending a VA medical center were randomized to receive the PHE in the context of a “health promotion clinic” versus usual care.⁴⁴ In both studies, receipt of the PHE improved delivery/receipt of fecal occult blood testing with large positive effect sizes (Cohen's d (95% CI): 1.19(1.17,1.21) and 1.07 (1.05,1.08), respectively).^{44, 55} (Table 7, Evidence Table 5a). Limitations of these studies included their potential limited generalizability to non-Medicare or VA populations (Table 2).

Observational studies. Observational studies evaluating the delivery/receipt of colon cancer screening included one retrospective cohort study, two cross-sectional studies, and one observational study with pre-post design performed from 1997–2003. Study populations included patients seen in community practices who interacted (or did not interact) with a touch-sensitive computer system placed in primary care practices to promote the delivery of preventive services,⁶⁵ a cross-sectional audit of outpatient billing claims for adults seen at least once by a primary care provider classified by visit type (visits for preventive care vs. acute care),⁶⁶ patients from randomly selected community practices agreeing to complete a questionnaire and medical record review to assess their receipt of a “periodic health examination” and their receipt of recommended clinical preventive services,⁷⁰ and patients in an ambulatory family practice residency clinic in which physicians participated in a quality improvement program to enhance the delivery of the “health maintenance examination” and clinical preventive services.⁷¹ Both cross sectional studies reported receipt of the PHE was positively associated with delivery/receipt of both sigmoidoscopy and fecal occult blood testing, as did the pre-post study. The retrospective cohort study reported mixed results (Table 8, Evidence Tables 5b–d). Several limitations were noted among these observational studies including poor description of the study populations, inability to completely control for potential confounding in several of the studies, studies not specifically designed to answer Key Question 2 (Table 2).

Mammography

Summary of findings. Twelve studies (including one RCT and eleven observational studies) assessed the association of receipt of the PHE with delivery/receipt of mammography. The best available evidence to assess this outcome was comprised of one RCT and one retrospective cohort study. These studies were performed in 1988 and 1998, were deemed to be of “low” grade based on standard criteria. The PHE had a small positive effect on the receipt of mammography in the RCT study, while it had mixed effects in the observational study. While the RCT was limited in its generalizability to non-Medicare populations, it was specifically designed to assess the effect of the PHE on this outcome. In contrast, the retrospective cohort study was not specifically designed to assess this outcome, did not employ a detailed description of the PHE, and was potentially limited by inadequate adjustment for residual confounding.

Findings

Strength and limitations of the evidence. The best available evidence assessing this outcome, emanating from one RCT and one RCT analyzed as a retrospective cohort study, received an overall grade of “low.” In grading the evidence, these studies were felt to have some serious limitations in quality and important inconsistency in the direction of results. However the data were not deemed to be sparse, and the studies did not appear to have high probability of reporting bias (Table 6). The RCT evaluated the effect of the PHE on the receipt of mammography among community dwelling Medicare recipients who received reimbursement for the PHE compared to Medicare recipients receiving usual care. While this study was limited in its generalizability to other ambulatory populations, it was specifically designed to evaluate the effect of the PHE on receipt of clinical preventive services and therefore directly addressed Key Question 2. The other RCT studied the effectiveness of a computerized touch screen system employed in primary care practices to improve rates of preventive screening. While the study compared patients seen in primary care practices randomized to employ the touch screen system versus patients seen in practices not employing the touch screen system, they performed a retrospective chart review to assess whether patients in the intervention and control groups had received a “health maintenance examination” during the past year. Thus, for the purposes of our analyses, this study was analyzed as a retrospective cohort study (of persons exposed versus not exposed to the health maintenance examination) without regard to the study's randomized intervention. While this was a study of adult patients of all ages being seen in representative primary care practices, it was not designed to directly address Key Question 2 (Tables 2, 6 and 7).

Randomized controlled trials. The RCT studied 455 Medicare recipients for 24 months beginning in 1988.⁵⁵ This study was a Medicare demonstration project in which the PHE was delivered in the context of a “preventive care package” in which patients received clinical preventive services including annual history and physical on at least an annual basis, delivered by both a nurse and a physician. The comparison group received clinical preventive services as customarily offered in their physicians practices.⁵⁵ In this study, receipt of the PHE improved delivery/receipt of mammography with a small positive effect size (Cohen's d (95% CI): 0.14(0.12,0.16) (Tables 2 and 7, Evidence Table 6 a).⁵⁵ Inferences from this study are potentially limited to non-Medicare populations.

Observational studies. Observational studies evaluating the delivery/receipt of mammography included one RCT analyzed as a retrospective cohort study, eight cross-sectional studies, and two observational study with pre-post design performed from 1988–1998. Study populations included patients seen in community practices who interacted (or did not interact) with a touch-sensitive computer system placed in primary care practices to promote the delivery of preventive services,⁶⁵ respondents to a California telephone survey who were contacted to assess their access to preventive services and satisfaction with preventive services,⁶⁸ data from the National Ambulatory Medical Care and National Hospital Ambulatory Medical Care Surveys in which physicians completed forms describing reasons for ambulatory visits (including general medical visits or gynecological) and the receipt of preventive services,⁶⁴ female residents in Ontario, Canada who completed the National Population Health Survey reporting their use of annual examinations with answers linked to their use of services in a national health insurance plan,⁶⁷ a cross-sectional audit of outpatient billing claims for adults seen at least once by a primary care provider classified by visit type (visits for preventive care vs. acute care),⁶⁶ Mexican-American participants in a telephone and door-to-door survey designed to assess access to and use of ambulatory health care,⁶⁹ patients randomly selected from 44 ambulatory outpatient clinics who completed a survey to ascertain their receipt of preventive services in the context of “checkup physical examinations” versus other types of visits,³¹ 93 physicians in an ambulatory practice network surveyed to recall the content of non-acute care visits with women age 40–75 years seen in their practices,⁷⁵ patients from randomly selected community practices agreeing to complete a questionnaire and medical record review to assess their receipt of a “periodic health examination” and their receipt of recommended clinical preventive services,⁷⁰ patients in an ambulatory family practice residency clinic in which physicians participated in a quality improvement program to enhance the delivery of the “health maintenance examination” and clinical preventive services,⁷¹ and family practice residents and faculty physicians using a practice-based teaching model to increase resident compliance with USPSTF guidelines.⁷² The RCT analyzed as a retrospective cohort study, performed in 1998, reported mixed associations of the PHE with receipt of mammography. Seven cross-sectional studies reported a positive association of receipt the PHE with receipt of mammography, while one reported a negative association. Both pre-post studies reported no statistically significant effect of the PHE on improving mammography rate (Table 8, Evidence Tables 6b–d). Several limitations were noted among these observational studies including poor description of the study populations, inability to completely control for potential confounding in several of the studies, studies not specifically designed to answer Key Question 2, and the potential for recall bias in studies based on participant interviews/questionnaire responses (Table 2).