Inclusion and exclusion criteria for preventive services during office visits. Codes refer to reason for visit classification for ambulatory care (RVC code),282930 which is used in the National Ambulatory Medical Care Survey and National Hospital Ambulatory Medical Care Survey surveys, except where noted that an International Classification of Diseases, Ninth Revision (ICD-9) diagnosis code is used. EKG, electrocardiogram; NOS, not otherwise specified; Pap, Papanicolaou; and PSA, prostate-specific antigen.
Development of the multivariate regression model.
Rates of preventive health examinations (PHEs) and preventive gynecological examinations (PGEs) by age group. The vertical lines represent 95% confidence intervals around percentage estimates.
Proportions of preventive care services ordered or delivered at preventive health examinations (PHEs) and preventive gynecological examinations (PGEs). The vertical lines represent 95% confidence intervals around percentage estimates. Pap indicates Papanicolaou smear; PSA, prostate-specific antigen.
Ateev Mehrotra, Alan M. Zaslavsky, John Z. Ayanian. Preventive Health Examinations and Preventive Gynecological Examinations in the United States. Arch Intern Med. 2007;167(17):1876–1883. doi:10.1001/archinte.167.17.1876
Preventive health examinations (PHEs) are controversial, and limited data are available on their use and content.
We conducted a retrospective analysis of 8413 ambulatory visits from January 1, 2002, to December 31, 2004, for PHEs and preventive gynecological examinations (PGEs) by adults in the National Ambulatory Medical Care Survey and National Hospital Ambulatory Medical Care Survey. Population estimates were obtained from the Current Population Survey. We estimated rates of PHEs and PGEs by patients' demographic characteristics, the frequency of 8 preventive services provided at these visits, and total costs of PHEs and PGEs at Medicare reimbursement rates.
An estimated 44.4 million adults per year (20.9%; 95% confidence interval [CI], 18.2%-23.6%) received a PHE, and 19.4 million women per year (17.7% of adult women; 95% CI, 14.9%-20.4%) received a PGE, together accounting for 8.0% of all ambulatory visits. The PHE rates varied by region (Northeast vs West: relative risk, 1.58; 95% CI, 1.17-2.14) and insurance type (those without vs those with private insurance or Medicare: relative risk, 0.51; 95% CI, 0.40-0.65). Preventive services occurred at 52.9% (95% CI, 48.8%-57.0%) of PHEs and 83.5% (95% CI, 80.7%-86.3%) of PGEs, but only 19.9% (95% CI, 18.4%-21.5%) of 8 preventive services occurred at a PHE or PGE. The annual costs of these visits were approximately $7.8 billion.
PHEs and PGEs are among the most common reasons adults see a physician. These visits frequently include preventive services, but most preventive services are provided at other visits. These findings provide a foundation for continuing national deliberations about the use and content of PHEs and PGEs.
The value of many preventive health services is well established, but the role of preventive health examinations (PHEs) (also called periodic health evaluations) for health promotion and screening of disease risk factors and subclinical illness remains controversial.1,2 Surveys have found that two-thirds of patients and physicians believe it is important for adults to receive a PHE each year,3,4 and routine PHEs are believed to strengthen physician-patient relationships.5,6 Approximately 80% of commercial insurance plans pay for annual PHEs.7 In 2005, Medicare introduced the “Welcome to Medicare” examination, a PHE for beneficiaries during their first year of enrollment.8 However, annual PHEs or preventive gynecological examinations (PGEs) (also called annual pelvic or gynecological examinations) are not recommended by major North American clinical organizations.9,10 These organizations have instead recommended a more individualized package of preventive services for patients.10,11 Some observers have raised concerns that preventive visits are an inefficient use of physicians' time, potentially interfering with timely access for sick patients.12,13 Others have emphasized the value of providing preventive care during nonpreventive visits.14
Despite this controversy, limited previous research has assessed which patients receive PHEs and PGEs and the care provided at these visits. Women receive a significant fraction of mammograms and Papanicolaou smears at preventive visits,15 and receipt of a PHE is associated with increased use of evidence-based preventive care and cancer screening.16,17 However, discretionary or unnecessary laboratory testing is frequently ordered at these appointments.18,19
Using nationally representative data on ambulatory visits, we addressed the following questions: How many adults receive a PHE or PGE each year and what factors predict their receipt? Which types of physicians are seen for PHEs or PGEs? What is the content and cost of these examinations?
Data on ambulatory visits from January 1, 2002, to December 31, 2004, were obtained from the National Ambulatory Medical Care Survey (NAMCS) and the outpatient component of the National Hospital Ambulatory Medical Care Survey (NHAMCS).20 Together, these surveys are designed to be nationally representative of visits,21,22 have been validated against other data sources,23 and have been used to assess national trends in ambulatory visits24 and PHEs and PGEs.19,25,26 Because this study used de-identified data, it was deemed exempt from review by the Harvard Medical School Human Studies Committee.
The NAMCS collects data on health care services provided by office-based physicians. The NHAMCS assesses services offered at hospital outpatient departments (excluding military, Veterans Affairs, and federal hospitals), including hospital-owned clinics that are geographically separate. From 2002 to 2004, the participation rate of physicians sampled for the NAMCS was 64.7% to 70.4% and the participation rate in the NHAMCS by sampled hospitals was 69.3% to 74.8%.27- 33
Standard 1-page encounter forms were completed by the physician (17%), office staff (30%), an outside coder using the medical record (32%), or some combination of these respondents (15%) for 30 random patient visits during an assigned reporting week. Item nonresponse rates were 5% or less in both surveys. Our analysis focused on patient demographic characteristics, up to 3 coded reasons for a visit,28- 30 up to 3 coded diagnoses,34 and laboratory tests, radiographs, and counseling services provided or ordered at a visit.
Data on the age, sex, race or ethnicity, region, and insurance status of US adults were obtained from the Current Population Survey (CPS) (March 2002-2004 supplements).35 Although the NAMCS/NHAMCS and the CPS have categories for private insurance, Medicare, or Medicaid, they differ in their other insurance classifications.36,37 To make the surveys more comparable, we created an “other” classification, defined as not having private insurance, Medicare, or Medicaid, which includes the uninsured, workers' compensation, military health, and other classifications.
This study focused on ambulatory visits to physicians by adults (aged ≥18 years). Adapting previous definitions,25,26 we identified a visit as a PHE if (1) the patient's major reason for the visit identified was “general medical examination” or (2) a diagnosis code for the visit was a general medical examination (code V70.0 or V70.9).34 We identified a visit as a PGE if (1) the patient's major reason for the visit was “gynecological examination” or (2) a diagnosis code for the visit was an annual pelvic examination (code V72.3).34 The PGE definition excludes visits for prenatal care, birth control, or repeat or abnormal Papanicolaou smear results. If a visit met the criteria for both PHE and PGE, we classified it as a PGE. The NAMCS and NHAMCS sample visits rather than patients, so we cannot determine whether a specific patient had more than 1 PHE or PGE in a given year.
We excluded PHEs or PGEs with dermatologists, general surgeons, psychiatrists, orthopedic surgeons, otolaryngologists, or ophthalmologists unless the physician self-identified as the patient's primary care physician, which was rare. Visits with only a nurse or medical assistant were also excluded, but visits at which a nurse practitioner or physician assistant was seen were included. Excluded visits accounted for less than 1% of each type of preventive visit.
Of the 79 312 sampled visits for the NAMCS and the 101 861 for the NHAMCS over 3 years, 5387 met our definition of a PHE and 3026 of a PGE. Using the sampling weights provided for each of these 8413 visits, we calculated national estimates of ambulatory visits to physicians' offices and hospital-based clinics for these visits in a single year.
Some previous evaluations of PHEs have focused on their use for apparently healthy patients.10,11 To better understand the health status of patients who received PHEs and PGEs, we identified visits in which patients had any of 7 chronic conditions listed as 1 of their 3 diagnosis codes, including asthma (codes 493.0-493.9), cerebrovascular disease (codes 430.0-437.9), diabetes mellitus (codes 250.0-250.9), hypertension (codes 401.0-405.9), ischemic heart disease (codes 410.0-414.9), chronic obstructive pulmonary disease (codes 492.0-492.9 and 496.0-496.9), and depression (codes 296.0-296.9, 309.0-309.9, and 311.0-311.9).
We examined 8 preventive or counseling services that can be ordered or performed at PHEs and PGEs and are listed on encounter forms used by the NAMCS and NHAMCS, focusing on populations most likely to benefit as defined by the US Preventive Services Task Force38: (1) mammograms among women 40 years or older, (2) prostate-specific antigen testing among men aged 50 to 70 years (recognizing the controversy over its use39), (3) Papanicolaou testing among women aged 18 to 65 years, (4) cholesterol testing among women 45 years and older and men 35 years and older, (5) smoking cessation counseling among current smokers, (6) weight loss counseling among obese patients, (7) exercise counseling, and (8) nutrition counseling. Exercise counseling and nutrition counseling were assessed among patients with obesity, hyperlipidemia, hypertension, or diabetes mellitus. We also examined 4 routine tests (complete blood cell counts, serum electrolytes, urinalysis, and electrocardiogram) listed on the encounter form that are not routinely recommended as screening tools and, thus, may be overused in preventive visits.18,19 If 1 of the reasons for a visit suggested a clinical indication for a test (eg, electrocardiogram for chest pain), we excluded these tests from the analysis. Criteria for exclusion are listed in eFigure 1.
To estimate costs associated with PHEs and PGEs, we used the 2006 Medicare physician and laboratory payment schedules for an “initial preventive examination” (Healthcare Common Procedure Coding System code G0344)40 and analyses of related tests based on the mean nonfacility reimbursement rate in the United States.
Rates of PHEs and PGEs were determined by first concatenating the 3 years of NAMCS/NHAMCS data and estimating the number of such examinations per year. We then divided this yearly estimate by the total population or relevant demographic subgroups from the CPS to estimate rates of visits per person per year. Because CPS sample weights are adjusted to match independent population controls, our standard errors reflect variability from the NAMCS and NHAMCS surveys but not from the CPS.
To assess patient factors that could influence the receipt of a PHE or PGE, we fit a separate multivariate logistic regression model for each of the 2 types of preventive visits. We concatenated the CPS data and the NAMCS/NHAMCS data so that people who received a preventive visit were represented twice in the combined database (once from the NAMCS/NHAMCS and once from the CPS), while those who had no PHE or PGE were represented once (from the CPS). In the concatenated data set, the odds (conditional on demographic characteristics) of having been from the NAMCS/NHAMCS component of the data set is an estimate of the probability of having received a PHE or PGE. Thus, odds ratios in the logistic regression can be directly interpreted as relative risks (probability ratios). Further explanation on the regression model is listed in eFigure 2. The CPS does not contain information on chronic illnesses, so these variables were not included in the multivariate models.
To assess which specialties provided PHEs and PGEs, we examined the specialty recorded in the NAMCS data; the NHAMCS does not record physician specialty. We calculated the proportion of PHEs and PGEs that included selected tests and counseling services. Among all visits that included preventive services, we also estimated the proportions that were associated with PHEs or PGEs. Using these estimates and the mean Medicare reimbursement rates, we estimated overall costs associated with these visits.
All analyses were performed using survey procedures in SAS statistical software, version 9.2 (SAS Institute Inc, Cary, North Carolina) to account for the complex design of the NAMCS and NHAMCS. We report 2-tailed P values or 95% confidence intervals (CIs) for all statistical tests.
Of the estimated 792.1 million ambulatory visits per year by adults in the United States between 2002 and 2004, we estimate that 63.5 million (8.0%; 95% CI, 7.4%-8.6%) were for either a PHE or a PGE and, of these, 6.6% occurred in hospital outpatient departments. An estimated 44.4 million adults per year (20.9% of the population; 95% CI, 18.2%-23.6%) received a PHE and an estimated 19.4 million women per year (17.7% of adult women; 95% CI, 14.9%-20.4%) received a PGE. The average PHE lasted 23.1 (95% CI, 21.8-24.4) minutes, and the average PGE lasted 20.5 (95% CI, 19.3-21.8) minutes, vs 18.2 (95% CI, 17.8-18.6) minutes for other types of visits.
Three physician specialties performed more than four-fifths of all PHEs and PGEs. Of all PHEs, general and family practitioners performed 37.1% (95% CI, 31.9%-42.4%), general internists performed 35.9% (95% CI, 30.1%-41.8%), and obstetricians/gynecologists performed 8.2% (95% CI, 5.5%-10.9%). Of all PGEs, general and family practitioners performed 13.1% (95% CI, 10.2%-16.0%), general internists performed 4.9% (95% CI, 2.2%-7.7%), and obstetricians/gynecologists performed 81.5% (95% CI, 77.8%-85.2%). Visits by adults for either a PHE or a PGE accounted for 9.5% (95% CI, 9.1%-13.0%) of visits to general and family practitioners, 11.0% (95% CI, 9.1%-13.0%) of visits to general internists, and 25.7% (95% CI, 22.8%-28.7%) of visits to obstetricians/gynecologists.
Unadjusted rates of PHEs were lowest in young adults (5.4% of men aged 18-24 years vs 45.0% of men ≥75 years), but PGEs were most common in young women (Figure 1). The rate of PHEs varied across racial/ethnic groups, insurance type, and region of the country (Table 1). In multivariate models controlling for other demographic factors, significant predictors of receipt of a PHE were older age; or having health insurance; and living in the Northeast (vs West) (Table 2). Predictors of receipt of a PGE were similar.
Of those who received a PHE or PGE, 74.6% (95% CI, 71.8%-77.5%) and 57.4% (95% CI, 52.8%-61.9%), respectively, had been seen in the clinic in the previous 12 months for another visit, and 35.5% (95% CI, 31.8%-39.1%) and 5.6% (95% CI, 4.1%-7.2%), respectively, were reported to have 1 or more of 7 chronic illnesses.
Receipt of at least 1 of 8 preventive services for eligible patients occurred in 52.9% (95% CI, 48.8%-57.0%) of PHEs and 83.5% (95% CI, 80.7%-86.3%) of PGEs. However, only 19.9% (95% CI, 18.4%-21.5%) of these preventive services occurred in either a PHE or a PGE. For example, mammograms ordered at PHEs and PGEs accounted for 22.9% (95% CI, 17.2%-28.6%) and 44.7% (95% CI, 38.6%-50.8%) of all mammograms, respectively (Figure 2). In contrast, of all visits with weight reduction counseling, only 8.8% (95% CI, 5.1%-12.5%) were PHEs and 1.1% (95% CI, 0.2%-1.9%) were PGEs.
Rates of ordering 1 of 4 routine tests (electrocardiogram, urinalysis, complete blood cell count, or electrolyte level) ranged from 11.0% of PHEs and 0.1% of PGEs for an electrocardiogram to 24.9% of PHEs and 25.1% of PGEs for a urinalysis (Table 3). The total costs of PHEs and PGEs and associated preventive services were approximately $5.2 and $2.6 billion per year, respectively, with the average PHE and PGE with related laboratory and radiology services costing $116 and $136, respectively (Table 4).
From 2002 to 2004, approximately 44 million US adults per year received a PHE and 19 million US adult women received a PGE. Visits for either preventive examination accounted for approximately 1 in 12 adult ambulatory visits. The PHE and PGE are, thus, 2 of the most common reasons adults see a physician and together exceed the number of annual visits for either acute respiratory infections (30 million) or hypertension (48 million).41 The estimated $7.8 billion spent annually on PHEs and PGEs was nearly equal to the $8.1 billion spent on all breast cancer care in 2004.42 Despite the controversy over their value, these findings highlight the significant role that PHEs and PGEs play in the US health care system.
Not surprisingly, we found lower rates of PHEs among young adults and those who did not have Medicare, private insurance, or Medicaid. The lower rate of PHEs in the West vs the Northeast may reflect differing clinical practice styles. A previous study4 found that physicians in Boston were more likely than physicians in San Diego, California, to report performing annual PHEs. Statements by the American Medical Association11 have focused on preventive visits for healthy persons and, in our sample, about two-thirds of adults who received a PHE and more than 90% of women who received a PGE did not have a common chronic illness recorded.
No consensus exists on the appropriate content of preventive visits.2 As in previous studies,2,16,25,43,44 we found that PHEs and PGEs are an important venue for providing some evidence-based preventive services. In particular, most mammograms were ordered and Papanicolaou smears were obtained during these visits, particularly PGEs, although annual Papanicolaou smears may not be necessary for many women.45 In contrast, we found that most counseling services occurred outside either PHEs or PGEs. Interestingly, we found that weight reduction counseling occurs more frequently than tobacco counseling at applicable visits, even though the latter may have a much greater health impact.46 Many of the PHEs in our study included routine tests, such as complete blood cell counts or urinalyses, that do not clearly improve patient outcomes and, thus, may be unnecessary.47
Furthermore, three-quarters of patients who received a PHE had been seen for other reasons in the previous year, and only 1 in 5 of the 8 preventive and counseling services we studied were provided at PHEs and PGEs. These results confirm previous findings that preventive care frequently occurs at visits for immediate care or chronic illness14,48,49 and support the concept advocated by some that provision of preventive care outside PHEs or PGEs should be emphasized.9,14 While “adding a minute for prevention” is feasible for counseling services and possibly the ordering of preventive tests,9,14 it is more difficult to add a pelvic examination and associated Papanicolaou smear to a nonpreventive visit. This may explain why more than 60% of screening Papanicolaou smears were obtained at preventive visits.
Our findings can inform future recommendations about the frequency of visits for PHEs and PGEs and the feasibility of providing a PHE annually to all adults in the United States, as many patients and physicians would prefer.3,4 If every adult were to receive a PHE annually, we estimate the US health care system would need to provide up to 145 million more visits every year. According to the Council on Graduate Medical Education,50 the average full-time primary care physician (excluding obstetricians/gynecologists) has 1600 patients in his or her panel. Primary care physicians, on average, devote 31 hours per week to ambulatory care and work 47 weeks per year.51 Therefore, if PHEs and PGEs last 23.1 and 20.5 minutes, respectively, on average, then together these types of visits would account for 41% of all direct patient time for primary care physicians, even under the unlikely assumption that a physician could deliver all recommended preventive services in a single 23-minute visit.52 Together, these findings underline the need to maintain other means of delivering preventive care beyond annual PHEs or PGEs.
Our study has several potential limitations. We may have underestimated the number of PHEs and PGEs because the national surveys we analyzed do not include the 57.5 million visits (approximately 6.7% of all US ambulatory visits) to clinics affiliated with the Department of Veterans Affairs.53 We do not know how many women received both a PHE and PGE in a given year, and our cross-sectional data sets do not allow us to determine how many patients receive physical examinations in consecutive years. While many physicians believe that a preventive visit is an important mechanism to detect subclinical illness and to improve physician-patient relations,5,6 we do not know how frequently PHEs and PGEs fulfill these roles. It is unknown how many patients visited their physician for a preventive visit for reasons beyond preventive care. In an analysis of 95 PHE visits, 91% of patients raised additional nonpreventive care concerns.43 The race and ethnicity data should be interpreted with caution, because the NAMCS and NHAMCS rely on the reporters' perceptions of patients' race, while race is self-reported in the CPS. Respondents to the CPS underreport enrollment in Medicaid, and this may have biased our findings on how insurance status predicted receipt of a physical examination.54 Our findings also depend on the accurate completion by physicians or their staff of the encounter forms used by the NAMCS and NHAMCS. A comparison of direct observation with NAMCS data found that test ordering was generally more accurately recorded than counseling services.23 Last, our cost calculations are only approximations because they do not account for the differing payment rates among private payers and Medicaid.
Approximately 64 million US adults per year visited a physician from 2002 through 2004 for a PHE or a PGE. Our findings provide a foundation for continuing national deliberations about whether adults should see physicians for these examinations and about their appropriate content.
Correspondence: Ateev Mehrotra, MD, MPH, Division of General Internal Medicine, University of Pittsburgh, School of Medicine, 230 McKee Pl, Ste 600, Pittsburgh, PA 15213 (firstname.lastname@example.org).
Accepted for Publication: May 15, 2007.
Author Contributions:Study concept and design: Mehrotra, Zaslavsky, and Ayanian. Acquisition of data: Mehrotra. Analysis and interpretation of data: Mehrotra, Zaslavsky, and Ayanian. Drafting of the manuscript: Mehrotra. Critical revision of the manuscript for important intellectual content: Zaslavsky and Ayanian. Statistical analysis: Zaslavsky. Study supervision: Ayanian.
Financial Disclosure: None reported.
Funding/Support: Dr Mehrotra’s salary was supported by National Research Service Award 5 T32 HP11001-15 from the Health Resources and Services Administration.
Role of the Sponsor: The Health Resources and Services Administration had no role in the design and conduct of the study; in the collection, analysis, and interpretation of the data; or in the preparation, review, or approval of the manuscript.
Previous Presentation: This study was presented at the annual meeting of the Society for General Internal Medicine; April 29, 2006; Los Angeles, California.
Additional Information: The eFigures are available at http://www.archinternmed.com.