PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
 
Obesity (Silver Spring). Author manuscript; available in PMC 2010 December 22.
Published in final edited form as:
PMCID: PMC3008358
NIHMSID: NIHMS90371

The Association of Obesity and Cervical Cancer Screening: A Systematic Review and Meta-Analysis

Abstract

Obese women are at an increased risk of death from cervical cancer, but the explanation for this is unknown. Through our systematic review, we sought to determine whether obesity is associated with cervical cancer screening and if this association differs by race. We identified original articles evaluating the relationship between body weight and Papanicolaou testing in the United States through electronic (PubMed, CINAHL, and the Cochrane Library) and manual searching. We excluded studies in special populations or those not written in English. Two reviewers sequentially extracted study data and independently extracted quality using standardized forms. 4,132 citations yielded 11 relevant studies. Ten studies suggested an inverse association between obesity and cervical cancer screening. Compared to women with a normal BMI, the combined odds ratios (95% CI) for Papanicolaou testing were 0.91 (0.80 to 1.03), 0.81 (0.70 to 0.93), 0.75 (0.64 to 0.88), and 0.62 (0.55 to 0.69) for the overweight and class I, class II, and class III obesity categories, respectively. Three of 4 studies that presented the results by race found this held true for white women, but no study found this for black women. In conclusion, obese women are less likely to report being screened for cervical cancer than their lean counterparts, and this does not hold true for black women. Less screening may partly explain the higher cervical cancer mortality seen in obese white women.

INTRODUCTION

The incidence of and mortality from cervical cancer have decreased considerably with the advent of Papanicolaou (Pap) testing for cervical cancer screening (13). Despite the availability of this effective screening, obese women in the United States experience higher mortality from cervical cancer (4). Recent observational studies suggest that obese women may receive screening for cervical cancer less frequently than their counterparts of normal weight (511). Cervical cancer mortality is higher for black women (12), and some studies report that race may modify the possible association between obesity and cervical cancer screening (911).

To determine if these recent observations are consistent, we conducted a systematic review and meta-analysis to study the hypothesis that obese women are less likely to receive screening for cervical cancer with Pap testing than those with normal weight. We also studied the effect of race on the relationship between obesity and cervical cancer screening.

METHODS

Data Sources

We searched the PubMed, CINAHL, and Cochrane Library electronic databases from inception to November 2006 to identify original articles which evaluated the relationship between body weight and cervical cancer screening in the United States using search terms for cervical cancer screening, cervical cancer, and body weight (Appendix Tables 1–3). We also manually searched 11 key medical journals from August 2006 through November 2006 and the references of key and included articles. All reviewers were physician investigators and included a senior obesity researcher, an investigator with systematic review experience, and a post-doctoral epidemiology trainee with relevant clinical experience. Two reviewers conducted title and abstract reviews independently, and conflicts were resolved by consensus.

Study Selection

We included published studies if they reported the prevalence of Pap testing by body weight in adults ≥ 18 years of age. Exclusion criteria included the following: not written in English, conducted outside of the United States since other countries may have different screening guidelines, no original data, did not occur in humans, or screening studied in special populations since there may be different screening expectations for some populations (e.g., participants presenting to a cancer screening clinic, HIV positive patients, those with a history of cervical cancer, and those involved in a study of interventions to improve screening).

Data Extraction

Two reviewers sequentially extracted the data on population characteristics, the exposure, and the outcome using standardized data abstraction forms. The quality of the studies was evaluated independently by two authors using a quality form based on the STROBE Statement, Checklist of Essential Items Version 3 (September 2005) (13). Disagreements were resolved by consensus. One study included body mass index in its analysis but did not explicitly report results for Pap smear prevalence by body weight (14); the author kindly provided these results upon our request.

Data Analysis

First, we summarized the results qualitatively. We created tables to qualitatively describe the included studies and to display the results of the studies that evaluated race as an effect modifier.

We conducted meta-analyses for the sufficiently homogenous studies that: 1) had nationally representative survey data and 2) reported body mass index (BMI) in 5 standard categories according to World Health Organization criteria (15) and the National Institutes for Health (16): normal: 18.5–24.9 kg/m2, overweight: 25–29.9 kg/m2, class I obesity: 30–34.9 kg/m2, class II obesity: 35–39.9 kg/m2, and class III obesity: ≥ 40 kg/m2. We contacted the authors for articles which used nationally representative data but did not report results for Pap testing by BMI in these 5 standard categories; no authors were able to provide the necessary quantitative results. One paper displayed cervical cancer screening by BMI graphically, and upon our request, the author provided numeric results (11).

Using the DerSimonian and Laird method (17), we used random effects models to calculate combined odds ratios and 95% confidence intervals for Pap testing by BMI category (compared with normal BMI). For the study that reported adjusted proportions (10), we calculated odds ratios. One study only reported results stratified by race (white and black) (9), and we incorporated these into the analyses as two separate studies. The metainf command in STATA was used to determine the effect of the exclusion of any one study on the combined estimate.

We tested for heterogeneity among the studies using a chi-square test (18) (with α set to ≤ 0.10 due to the low power of this test with only a small number of studies) and an I2 statistic (19). An I2 value of > 50% signified “substantial heterogeneity” (18). A random effects model was chosen to account for between-study variability.

We tested for publication bias using the tests of Begg and Mazumdar (20) and Egger and colleagues (21). All analyses were completed using STATA (StataCorp. 2005. Stata Statistical Software: Release 9. College Station, TX: StataCorp LP.)

RESULTS

Literature Search Results

Our overall search strategy addressed a broader question regarding the association between obesity and screening for cervical, breast, and colon cancer. Out of 4,132 titles identified in the initial search, eleven articles met the inclusion criteria and addressed cervical cancer screening (Figure 1). The most common reason for exclusion at the article review level was failure to address the key question. Four (811) of the 11 studies were sufficiently homogeneous to include in the meta-analyses (i.e., used nationally representative survey data and provided information for cervical cancer screening by standard categories of BMI). Four studies (811) reported the relationship between obesity and report of Pap testing by race.

Figure 1
Search Strategy

Qualitative Synthesis

The 11 included studies (170,689 participants) are described in Table 1. All studies used BMI to estimate adiposity. Five studies defined the outcome as Pap testing in the last 3 years, three studies as Pap testing in the last two years, and 3 studies as Pap testing in the last year. Eight of the 11 studies (73%) were based on nationally representative surveys, the National Health Interview Survey (NHIS), Behavioral Risk Factor Surveillance System (BRFSS), or Health and Retirement Survey (HRS). None of the studies used overlapping data. Most subjects were Caucasian. One study contained only participants between the ages of 50 and 64 years (9). Overall absolute screening proportions ranged from 71% to 85%.

Table 1
Description of Included Studies

Ten of the 11 (91%) studies reported at least one inverse association between Pap testing and higher BMI category which was statistically significant in 7 of the 11 studies. The study that found no inverse association between BMI and Pap testing differed from the other studies in that the investigators adjusted for receipt of breast cancer screening and used 27.3 kg/m2 as the cut-off for overweight (22).

Effect of Race

Four studies (811) examined the effect of race on the relationship between BMI and report of Pap testing. Overall screening rates were similar for whites (68.6 to 86%) and blacks (67.4 to 88%) when reported (8, 9, 11). Three (911) of the 4 studies found that higher BMI was associated with lower likelihood of Pap testing among white women but not among black women (Table 2). The study that did not find a significant interaction reported a borderline-significant P value of 0.076 for the interaction.

Table 2
Associations of BMI with Papanicolaou Testing by Race

Quantitative Assessment of Cervical Cancer Screening by BMI

Four studies (811) used nationally representative surveys and reported BMI in 5 standard categories. The combined results of these studies (98,389 participants) found that BMI was inversely associated with cervical cancer screening. Compared to the normal BMI group, the combined odds ratios (95% confidence interval) for Pap testing by BMI category were 0.91 (0.80 to 1.03), 0.81 (0.70 to 0.93), 0.75 (0.64 to 0.88), and 0.62 (0.55 to 0.69) for the overweight and class I, class II, and class III obesity categories, respectively (Figure 2.). The tests of homogeneity (P > 0.10) and I2 statistics indicated statistical heterogeneity. There was no statistical heterogeneity found for the class III obesity category.

Figure 2
Combined Odds Ratios for Papanicolaou Testing by BMI Category

Sensitivity Analyses

For the overweight BMI category, the combined odds ratio reached statistical significance upon exclusion of the estimate for black participants from one study (9). Additionally, for the class II and class III obesity categories, the estimate for the combined odds ratio did not reach statistical significance upon exclusion of the estimate for the white participants from the same study (9). In addition, removal of the largest study (N=72,889) (8) resulted in a non-significant combined odds ratio for the class I and II obesity categories. In all cases, the removal of any of the studies did not lead to a qualitatively different point estimate of the combined odds ratio.

Publication Bias

There was no evidence of publication bias statistically or on visualization of funnel plots.

Quality of Included Studies

We did not use a formal grading system, but the quality of the included studies was generally adequate. Ten of the 11 studies (91%) relied on self-reported BMI and cancer screening. Most studies (91%) adjusted for key confounders including age, race/ethnicity, co-morbid conditions, general health status, smoking status, and socioeconomic factors such as education. One study did not adjust for any confounding factors (23). Survey response rates ranged from 55% to 97% in the 7 of 10 survey studies that reported response rates (6, 7, 911, 14, 22). Most studies used nationally representative surveys and did not report the validity of the surveys used. Further quality information is provided in Appendix Table 4.

DISCUSSION

This systematic review demonstrates a strong, graded, inverse relationship between BMI and the likelihood of undergoing screening for cervical cancer. Compared to their lean counterparts, women with class III obesity are approximately 40% less likely to undergo cervical cancer screening. In black women, we did not find any association between obesity and Pap testing based on the 4 studies (811) that provided the results necessary to study this association.

Our results are especially important as the prevalence of the subtype of cervical cancer, cervical adenocarcinoma, appears to be increasing (24). Obesity, through its hormonal actions, may play a role in the pathogenesis of cervical adenocarcinoma (25) and has been found to be associated with adenocarcinoma of the cervix in some observational studies (26, 27).

There are several possible patient- and physician-related barriers to Pap testing for overweight and obese women. First, obese women may delay medical care (28) because of negative body image (29), embarrassment (11, 28), a perceived lack of respect from health care providers, or because they want to avoid unwanted weight loss advice (23). Second, obesity may be a marker for less favorable health behavior of which avoidance of cancer screening may be one facet (8, 10). Third, there may be differences in attitudes toward screening on the part of obese women (10). Fourth, obesity-related co-morbid conditions may hinder performance of purely preventive services (7, 30, 31). Fifth, providers have reported difficulty in providing care for obese women (23) and in particular, there may be technical difficulties in performing Pap testing in obese women (10, 23, 32) due to anatomy and inadequate equipment. Sixth, physicians have been shown to be reluctant to perform pelvic exams on reluctant patients, and obese women are often reluctant to undergo pelvic exams (29). Finally, physicians may suffer from biases against obese women that make them less likely to recommend routine screening (3335).

Despite a higher incidence and mortality from cervical cancer among black women (12), in our systematic review, obesity was not associated with Pap testing in this group. This may be due to public health efforts to increase screening among minorities in the 1990s (5, 22) or to racial differences in obesity-related body image (3638). In particular, it has been reported that white but not black women who are overweight or obese are more likely to feel worthless (39) and that obese black women may be less likely to report embarrassment or discomfort related to screening (11).

Limitations mostly reflect limitations of the published literature. We were only able to include 4 of 11 studies based on the studies being sufficiently homogenous with respect to patient characteristics and obesity outcomes (i.e., nationally representative data and BMI in 5 standard categories). However, ten of the 11 of the studies found an inverse relationship between BMI and report of cervical cancer screening. The one study that did not find a relationship between BMI and Pap testing (22) differed from the other studies in two ways: 1) the study adjusted for report of breast cancer screening which could obscure an association between BMI and cervical cancer screening; and 2) the authors dichotomized BMI using a cut-off of 27.3 kg/m2 to differentiate between normal-weight and overweight/obese women which would lead to the inclusion of thinner women in the overweight group (as compared to using BMI ≥30 kg/m2 as a cut-off) thus diluting the contribution of excess weight.

We found statistical evidence of heterogeneity for the overweight through class II obese categories of BMI. One source may be age since the studies varied substantially in the age of their participants. One study excluded women < 50 years of age (9), whereas the average age for 2 studies was 46 years (8) and 39 years (10). Furthermore, there is some evidence that elderly women undergo cervical cancer screening less than younger women (40). Age may also modify the relationship between obesity and cervical cancer screening. However, we were unable to explore heterogeneity using meta-regression given the limited number of studies in our meta-analyses.

Additional limitations of our review deserve mention. First, most of the included studies were cross-sectional and cannot establish causality. Second, we relied on the use of observational studies which are susceptible to residual and unmeasured confounding. Third, publication bias may be an issue. However, we searched and included studies where the association of BMI and Pap testing was only one of the many factors evaluated, and we contacted authors for data when we thought BMI may have been evaluated. Therefore, we do not expect publication bias to be a major factor in this analysis. Fourth, self-reported height and weight may introduce recall bias which could differ by survey type (telephone versus in-person), age, and BMI (41). However, the correlation between measured and self-reported weight and height is adequate (41), and misclassification of BMI would affect the estimates of the odds ratios but leave the inferences intact. Finally, most of the included studies also relied upon self-report of cervical cancer screening. Although not studied in nationally representative samples, the positive predictive value of 1-year recall for Pap testing has been shown to range from 33–75% and the negative predictive value of 1-year recall for Pap testing from 85–94 % (42). Thus, cervical cancer screening results may be inflated above their actual rates, but this should not affect the observed association between BMI and report of cervical cancer screening.

Our study also has several strengths. This is the first systematic review with meta-analysis exploring the relationship between obesity and cervical cancer screening. We comprehensively searched multiple electronic databases in addition to hand searching. Also, we contacted authors for data and obtained additional results from 4 studies for use in qualitative and quantitative analyses. Finally, the meta-analyses were based on nationally representative surveys and thus are generalizable to the U.S. population.

CONCLUSIONS

Inadequate cervical cancer screening may partly explain the increased cervical cancer mortality observed in obese women and in particular, among obese white women. Since cervical cancer is preventable, clinicians should be aware of this disparity in evaluating their own practices. Future research should address why obese women are less likely to undergo cervical cancer screening in an effort to develop solutions and also further explore the association between obesity and cervical cancer screening among black women.

ACKNOWLEDGEMENTS

Dr. Maruthur was supported by a training grant (5 T32 HL007024-31) from the National Heart, Lung, and Blood Institute, National Institutes of Health (NIH), and Dr. Brancati was supported by a mid-career investigator award (5 K24 DK062222-05) from the National Institute of Diabetes and Digestive and Kidney Diseases, NIH. There was no project-specific funding.

We thank the following individuals for contributions to the study: Eliseo Guallar, MD, DrPH (Johns Hopkins University, Baltimore, MD); Nancy K. Amy, PhD (University of California, Berkeley, CA); Jeanne Ferrante, MD (University of Medicine and Dentistry of New Jersey Robert Wood Johnson Medical School and New Jersey Medical School, Newark, NJ); Christina Wee, MD, MPH (Beth Israel Deaconess Medical Center, Boston, MA); and Marilyn Winkleby, PhD (Stanford School of Medicine, Stanford, CA). No compensation was given to those acknowledged.

Footnotes

SUPPLEMENTARY MATERIAL Appendix Tables 1–3 contain the terms used in searching the electronic databases. Appendix Table 4 describes the quality of the included studies. All files are in Microsoft Word document format. Supplementary Material is available at www.nature.com/obesity.

DISCLOSURES The authors have no conflicts of interest to disclose.

REFERENCES

1. Screening for squamous cervical cancer: Duration of low risk after negative results of cervical cytology and its implication for screening policies. IARC working group on evaluation of cervical cancer screening programmes. Br Med J (Clin Res Ed) 1986 Sep 13;293(6548):659–64. [PMC free article] [PubMed]
2. Sasieni PD, Cuzick J, Lynch-Farmery E. Estimating the efficacy of screening by auditing smear histories of women with and without cervical cancer. The National Coordinating Network for Cervical Screening Working Group. Br J Cancer. 1996 Apr;73(8):1001–5. [PMC free article] [PubMed]
3. Screening for cervical cancer: Recommendations and rationale. Jan, 2003. homepage on the Internet. Available from: http://www.ahrq.gov/clinic/3rduspstf/cervcan/cervcanrr.htm.
4. Calle EE, Rodriguez C, Walker-Thurmond K, Thun MJ. Overweight, obesity, and mortality from cancer in a prospectively studied cohort of U.S. adults. N Engl J Med. 2003 Apr 24;348(17):1625–38. [PubMed]
5. Amonkar MM, Madhavan S. Compliance rates and predictors of cancer screening recommendations among Appalachian women. J Health Care Poor Underserved. 2002 Nov;13(4):443–60. [PubMed]
6. Coughlin SS, Uhler RJ, Hall HI, Briss PA. Nonadherence to breast and cervical cancer screening: What are the linkages to chronic disease risk? Prev Chronic Dis. 2004 Jan;1(1):A04. [PMC free article] [PubMed]
7. Fontaine KR, Faith MS, Allison DB, Cheskin LJ. Body weight and health care among women in the general population. Arch Fam Med. 1998 Jul–Aug;7(4):381–4. [PubMed]
8. Fontaine KR, Heo M, Allison DB. Body weight and cancer screening among women. J Womens Health Gend Based Med. 2001 Jun;10(5):463–70. [PubMed]
9. Ostbye T, Taylor DH, Jr, Yancy WS, Jr, Krause KM. Associations between obesity and receipt of screening mammography, Papanicolaou tests, and influenza vaccination: Results from the Health and Retirement Study (HRS) and the Asset and Health Dynamics among the Oldest Old (AHEAD) study. Am J Public Health. 2005 Sep;95(9):1623–30. [PubMed]
10. Wee CC, McCarthy EP, Davis RB, Phillips RS. Screening for cervical and breast cancer: Is obesity an unrecognized barrier to preventive care? Ann Intern Med. 2000 May 2;132(9):697–704. [PubMed]
11. Wee CC, Phillips RS, McCarthy EP. BMI and cervical cancer screening among white, African-American, and Hispanic women in the United States. Obese Res. 2005 Jul;13(7):1275–80. [PubMed]
12. United States cancer statistics: 1999–2003 incidence and mortality web-based report. U.S. Department of Health and Human Services, Centers for Disease Control and Prevention and National Cancer Institute; Atlanta: homepage on the Internet. Available from: www.cdc.gov/uscs.
13. STROBE statement: Checklist of essential items. version 3. Sep, 2005. [homepage on the Internet]. Available from: http://www.strobe-statement.org/PDF/STROBE-Checklist-Version3.pdf.
14. Winkleby MA, Snider J, Davis B, Jennings MG, Ahn DK. Cancer-related health behaviors and screening practices among Latinos: Findings from a community and agricultural labor camp survey. Ethn Dis. 2003 Summer;13(3):376–86. [PubMed]
15. World Health Organization . Obesity: Preventing and managing the global epidemic. part 1: The problem of overweight and obesity. World Health Organization; Geneva: 2004. Report No.: 894.
16. The evidence report. National Institutes of Health; 1998. NHLBI Obesity Education Intiative Panel on the Identification, Treatment, and Management of Overweight and Obesity in Adults. Clinical guidelines on the identification, treatment, and management of overweight and obesity in adults. Report No.: 98-4083.
17. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials. 1986 Sep;7(3):177–88. [PubMed]
18. Deeks JJ, Higgins JPT, Altman DG, Higgins JPT, Green S, editors. Cochrane handbook for systematic reviews of interventions 4.2.5 [updated May 2005]; Section 8. 2005. Analyzing and presenting results.
19. Higgins JP, Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat Med. 2002 Jun 15;21(11):1539–58. [PubMed]
20. Begg CB, Mazumdar M. Operating characteristics of a rank correlation test for publication bias. Biometrics. 1994 Dec;50(4):1088–101. [PubMed]
21. Egger M, Davey Smith G, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ. 1997 Sep 13;315(7109):629–34. [PMC free article] [PubMed]
22. Simoes EJ, Newschaffer CJ, Hagdrup N, et al. Predictors of compliance with recommended cervical cancer screening schedule: A population-based study. J Community Health. 1999 Apr;24(2):115–30. [PubMed]
23. Amy NK, Aalborg A, Lyons P, Keranen L. Barriers to routine gynecological cancer screening for white and African-American obese women. Int J Obes (Lond) 2006 Jan;30(1):147–55. [PubMed]
24. Bray F, Carstensen B, Moller H, et al. Incidence trends of adenocarcinoma of the cervix in 13 European countries. Cancer Epidemiol Biomarkers Prev. 2005 Sep;14(9):2191–9. [PubMed]
25. Ursin G, Pike MC, Preston-Martin S, d'Ablaing G, 3rd, Peters RK. Sexual, reproductive, and other risk factors for adenocarcinoma of the cervix: Results from a population-based case-control study (California, United States) Cancer Causes Control. 1996 May;7(3):391–401. [PubMed]
26. Parazzini F, La Vecchia C. Epidemiology of adenocarcinoma of the cervix. Gynecol Oncol. 1990 Oct;39(1):40–6. [PubMed]
27. Lacey JV, Jr, Swanson CA, Brinton LA, et al. Obesity as a potential risk factor for adenocarcinomas and squamous cell carcinomas of the uterine cervix. Cancer. 2003 Aug 15;98(4):814–21. [PubMed]
28. Olson CL, Schumaker HD, Yawn BP. Overweight women delay medical care. Arch Fam Med. 1994 Oct;3(10):888–92. [PubMed]
29. Adams CH, Smith NJ, Wilbur DC, Grady KE. The relationship of obesity to the frequency of pelvic examination: Do physician and patient attitudes make a difference? Women & health. 1993;20:45–57. [PubMed]
30. Ostbye T, Yarnall KS, Krause KM, Pollak KI, Gradison M, Michener JL. Is there time for management of patients with chronic diseases in primary care? Ann Fam Med. 2005 May–Jun;3(3):209–14. [PubMed]
31. Yarnall KS, Pollak KI, Ostbye T, Krause KM, Michener JL. Primary care: Is there enough time for prevention? Am J Public Health. 2003 Apr;93(4):635–41. [PubMed]
32. Ferrante JM, Chen PH, Jacobs A. Breast and cervical cancer screening in obese minority women. J Womens Health (Larchmt) 2006 Jun;15(5):531–41. [PubMed]
33. Hebl MR, Xu J. Weighing the care: Physicians' reactions to the size of a patient. Int J Obes Relat Metab Disord. 2001 Aug;25(8):1246–52. [PubMed]
34. Price JH, Desmond SM, Krol RA, Snyder FF, O'Connell JK. Family practice physicians' beliefs, attitudes, and practices regarding obesity. Am J Prev Med. 1987 Nov–Dec;3(6):339–45. [PubMed]
35. Teachman BA, Brownell KD. Implicit anti-fat bias among health professionals: Is anyone immune? Int J Obes Relat Metab Disord. 2001 Oct;25(10):1525–31. [PubMed]
36. Stevens J, Kumanyika SK, Keil JE. Attitudes toward body size and dieting: Differences between elderly black and white women. Am J Public Health. 1994 Aug;84(8):1322–5. [PubMed]
37. Anderson LA, Eyler AA, Galuska DA, Brown DR, Brownson RC. Relationship of satisfaction with body size and trying to lose weight in a national survey of overweight and obese women aged 40 and older, united states. Prev Med. 2002 Oct;35(4):390–6. [PubMed]
38. Flynn KJ, Fitzgibbon M. Body images and obesity risk among black females: A review of the literature. Ann Behav Med. 1998 Winter;20(1):13–24. [PubMed]
39. Wee CC, McCarthy EP, Davis RB, Phillips RS. Obesity and breast cancer screening. J Gen Intern Med. 2004 Apr;19(4):324–31. [PMC free article] [PubMed]
40. Sirovich BE, Welch HG. The frequency of Pap smear screening in the United States. Journal of General Internal Medicine. 2004;19(3):243–50. 03/61. [PMC free article] [PubMed]
41. Rowland ML. Self-reported weight and height. Am J Clin Nutr. 1990;52:1125–33. [PubMed]
42. McGovern PG, Lurie N, Margolis KL, Slater JS. Accuracy of self-report of mammography and pap smear in a low-income urban population. American Journal of Preventive Medicine. 1998 Apr;14(3):201–8. [PubMed]