Obesity Disparities in Preventive Care: Findings From the National Ambulatory Medical Care Survey, 2005–2007

Authors

  • Tina Hernandez-Boussard,

    1. SCORE (Stanford Center for Outcomes Research and Evaluation), Department of Surgery, Stanford School of Medicine, Stanford, California, USA
    Search for more papers by this author
  • Shushmita M. Ahmed,

    1. SCORE (Stanford Center for Outcomes Research and Evaluation), Department of Surgery, Stanford School of Medicine, Stanford, California, USA
    Search for more papers by this author
  • John M. Morton

    Corresponding author
    1. SCORE (Stanford Center for Outcomes Research and Evaluation), Department of Surgery, Stanford School of Medicine, Stanford, California, USA
    Search for more papers by this author

(morton@stanford.edu)

Abstract

Obesity and its consequences are a major health concern. There are conflicting reports regarding utilization of preventive health-care services among obese patients. Our objective was to determine whether obese patients receive the same preventive care as normal weight patients. Weighted patient clinic visit data from the National Ambulatory Medical Care Survey (NAMCS) were analyzed for all adult patient visits with height/weight data (N = 866,415,856) from 2005 to 2007. Preventive care practice patterns were compared among different weight groups of normal, obese, and morbidly obese. Obese patients received the least number of preventive exams with a clear gradient present by weight. Obese patients were significantly less likely to receive cancer screening including breast examination (normal weight, reference, obese, odds ratio (OR), 0.8), mammogram (obese OR, 0.7), pap smear (obese OR, 0.7), pelvic exam (obese OR, 0.8), and rectal exam (obese OR, 0.7). The obese population also received less tobacco (obese OR, 0.7) and injury prevention education (obese OR, 0.7), yet significantly more diet, exercise, and weight reduction education. Significant differences in clinic practice patterns relative to normal weight patients were also evident with more physician referral (obese OR, 1.2) and less likely to see physician at the index clinic visit (obese OR, 0.8) and less likely to receive psychotherapy referral (obese OR, 0.6). Significant gaps in preventive care exist for the obese including cancer screening, tobacco cessation and injury prevention counseling, and psychological referral. Although obese patients received more weight-related education, this emphasis may have the consequence of de-emphasizing other needed preventive health measures.

The prevalence of overweight and obese adults in developed countries has been steadily increasing and in the United States 32% of men and 35% of women are obese (1,2,3). Several studies have shown the association between BMI and increased rates of cardiovascular disease, diabetes, cancer, and depression (4,5,6,7,8). Given the strong association between obesity and comorbidity, there has been emphasis on providing preventive health care to the obese.

However, there is conflicting evidence regarding preventive health utilization rates of disease in overweight and obese patients compared to normal-weight counterparts (9,10,11,12,13,14,15). Some studies found significant disparities in receipt of preventive care among obese compared to nonobese patients. Recently, Chang et al. concluded that obesity disparities of care do not exist in selected Veterans Affairs (VA) clinics and Medicare populations for eight, selected quality of care measures (15) and Yancy et al. reported that preventative services were higher in VA obese compare to normal-weight patients (16). However, this study did not account for the full range of recommended preventive services as set forth by the US Preventive Services Task Force.

Currently, no national, all-payor data for a comprehensive examination of preventive services by weight category exist. Given these recent findings, our study objective was to further amplify national utilization patterns of cancer screening, health education, and clinic practice patterns in different weight categories.

Methods and Procedures

National Ambulatory Medical Care Survey (NAMCS) is conducted annually by the National Center for Health Statistics and uses a large, nationally representative sample of patient visits to nonfederal, office-based physician practices and medical care services in the United States (17). The survey includes information regarding the patient's chief reason for visit, up to three diagnoses and procedures, details regarding the clinic visit, patient demographic characteristics, and future planned treatment. The National Center for Health Statistics weighs each patient visit to enable extrapolation for national estimates for all data elements contained in the survey. Patient height and weight measurements obtained during the clinical visit were used to calculate BMI (calculated as weight in kilograms divided by the square of height in meters) and assign patients to one of four categories: normal-weight (BMI 18.6–24.9), overweight (BMI 25–29.9), obese (BMI 30–39.9), and morbidly obese (BMI 40 or greater).

Preventive measures were selected from NAMCS and categorized as diagnostic/screening procedures, counseling/education, and visit disposition. These are services captured in the survey as dichotomous variables, which are not included in the coded procedures section. Only measures recorded for each year of study period (2005–2007) were included. Patient eligibility for each measure was based on recommended ages for clinical preventive services set by the US Preventive Services Task Force to obtain (18). Individual measures, eligibility criteria, and percent of normal-weight patients receiving services are shown in Table 1.

Table 1.  Prevalence of diagnostic services, health education, and visit disposition for normal-weight patients
inline image

Data were analyzed using the statistical program SAS (version 9.2; SAS Institute, Cary, NC). Pregnant females and patients less than 18 years of age were excluded from the analyses. Only patients with both weight and height measurements taken during their clinic visit were included. All analyses used sample weights to account for differential probabilities of selection into the sample, nonresponse, and noncoverage. Association between access to preventive service measures and weight status was examined using multivariate logistic regression models with receipt of service as the dependent variable and weight category as an independent variable. Separate models were developed for each service, based on gender and age-appropriateness of the measure. All models were adjusted for confounders: race/ethnicity, age, gender, comorbidities, region, and insurance status. Two-sided statistical tests were used and P < 0.05 was considered statistically significant and multiple comparisons were performed on the data. Odds ratios (ORs) are reported for overweight, obese, and morbidly obese patients compared to normal weight patients.

Results

The likelihood of receiving diagnostic services, health education, and visit disposition for normal-weight patients are listed in Table 1. Sample sizes and selected demographic characteristics are shown in Table 2 which includes a total of 866,415,856 patient visits (weighted). Thirty-three percent of the population was overweight, 30% were obese, and 8% were morbidly obese. While these proportions are higher than nationally reported population estimates, this difference is likely due to the increased need for clinical services in the obese population.

Table 2.  Sample size and selected patient demographics: US, 2005–2007
inline image

Compared to normal weight patients, there were significant differences in age across all weight groups; however these differences were not clinically important. There was an increasing percentage of female across the categories: overweight (52.8% P < 0.01), obese (56.4% P < 0.01), and morbidly obese (72.3% P = 0.08) and a decreasing percentage of non-Hispanic white (49.7%, P = 0.16; 46.8%, P < 0.01; 48.3%, P = 0.22, respectively). There were no differences in percent of private insurance or northeast region across weight categories.

Multivariate weighted logistic regression models demonstrate that in all three preventive health-care categories, weight-related disparities exist. Table 1 shows the proportion of normal-weight individuals receiving the health service and information regarding their visit disposition. Figure 1 displays the odds of receiving recommended care by weight-category based on the multivariate logistic regression models adjusted for covariates, with receipt of care in normal weight patients used as the referent value. For the majority of the services, a decreased gradient of receipt of care was seen as weight categories increased compared to the normal weight patients.

Figure 1.

Differences in diagnostic screening tests offered across the three weight categories. Triangles represent overweight patients, diamonds represent obese patients, and squares represent morbidly obese patients. *P < 0.05; P < 0.01. CI, confidence interval; PSA, prostate-specific antigen.

For cancer screening services, skin exams were less likely received by overweight (adjusted OR, 0.86; 95% confidence interval (CI), 0.76–0.98) and morbidly obese (OR, 0.71; CI, 0.56–0.90). Obese and morbidly obese were significantly less like to receive a breast exam (OR, 0.80; CI, 0.68–0.95 and OR, 0.50; CI, 0.38–0.64), mammography (OR, 0.71; CI 0.50–1.00 and OR, 0.54; CI, 0.36–0.80), pap test (OR, 0.71; CI, 0.58–0.87 and OR, 0.62; CI, 0.46–0.82), pelvic exam (OR, 0.77; CI, 0.64–0.91 and OR, 0.51; CI, 0.40–0.66), and rectal exam (OR, 0.68; CI, 0.53–0.87 and OR, 0.59; CI, 0.39–0.91) compared to normal weight patients, respectively. Obese and morbidly obese received more glucose tests (OR, 1.27; CI, 1.07–1.50 and OR, 1.30; CI, 1.01–1.68) and glycosylated hemoglobin tests (OR, 2.11; CI, 1.62–2.75 and OR, 3.31; CI 2.28–4.81) compared to normal weight patients (Figure 1). There were no differences in receipt of depression screening, prostate-specific antigen, and cholesterol tests (data not shown).

Disparities existed in health education services for all weight categories (Figure 2). Overweight patients were less likely to receive tobacco education (OR, 0.70; CI, 0.56–0.86) and obese and morbidly obese patients were less likely to receive injury prevention information (OR, 0.65; CI, 0.43–1.00 and OR, 0.32; CI, 0.16–0.67), respectively. Diet and nutrition education was significantly higher in overweight (OR, 1.21; CI, 1.05–1.40), obese (OR, 1.75; CI, 1.51–2.03), and morbidly obese (OR, 3.16; CI, 2.59–3.85) as was weight reduction education in overweight (OR, 4.69; CI, 3.12–7.07), obese (OR, 15.37; CI, 10.13–23.32), and morbidly obese (OR, 30.20; CI, 18.67–48.84). Exercise education was higher in obese (OR, 1.59; CI, 1.35–1.87) and morbidly obese (OR, 2.13; CI, 1.72–2.65). Figure 2 summarizes the receipt of health education services by weight category, significantly different compared to normal weight patients for overweight (P < 0.01) and morbidly obese (P < 0.01).

Figure 2.

Differences in health education offered across the three weight categories. Triangles represent overweight patients, diamonds represent obese patients, and squares represent morbidly obese patients. *P < 0.05; P < 0.01. CI, confidence interval.

Characteristics regarding clinic practice patterns revealed many important differences by weight category (Figure 3). Compared to normal weight patients, obese, and morbidly obese patients were more likely to be referred to another physician (OR, 1.24; CI, 1.05–1.45) and OR, 1.37; CI, 1.09–1.72), see a nurse practitioner (OR, 1.34; CI, 1.01–1.76 and OR, 1.66; CI, 1.09–2.54) and have a scheduled return appointment (OR, 1.17; CI, 1.05–1.29 and OR, 1.61; CI, 1.38–1.88), respectively. In addition, obese and morbidly obese patients were significantly less likely to receive psychotherapy referral (OR, 0.56; CI, 0.35–0.90 and OR, 0.43; CI, 0.22–0.87), other mental health counseling (OR, 0.57; CI, 0.39–0.84 and OR, 0.43; CI, 0.22–0.87), or see a physician at the index visit (OR, 0.79; CI, 0.62–1.00 and OR, 0.47; CI, 0.34–0.64) compared to normal weight patients. Morbidly obese patients were also less likely to have planned telephone follow-up (OR, 0.56; CI, 0.34–0.91) compared to normal weight patients.

Figure 3.

Differences in office visit disposition across the three weight categories. Triangles represent overweight patients, diamonds represent obese patients, and squares represent morbidly obese patients. *P < 0.05; P < 0.01. CI, confidence interval.

Table 3 summarizes the distribution of preventive care across the different weight categories. Overall, normal-weight individuals received a significant greater number of diagnoses or screening services compared to overweight, obese and morbid obese patients (5.25, 5.10, P < 0.0001; 5.19, P = 0.0596; 5.08, P < 0.0001, respectively). An inverse relationship was seen regarding health education services. Normal weight patients received a significant lower number of health education services (0.58) compared to overweight (0.62, P < 0.0001), obese (0.82, P < 0.0001), and morbid obese (1.04, P < 0.0001). There were no differences of the time patients spent in their clinical visit by weight category.

Table 3.  Mean number of preventive services and time spent in clinical visit by weight category
inline image

Discussion

This study provides important US national data regarding comprehensive diagnostic screening, health education, and visit disposition by weight categories in the United States. Previous reports based on selected populations have shown obesity-related discrepancies regarding preventive services (9,10). The lower use of preventive services may be attributed to both patient and clinician attitudes alike (11,19,20,21,22). Unlike previous studies, this study focuses on a US nationally representative population that is not predominately elderly (Medicare) or male (VA) with clinically derived BMI data and a more expansive menu of age/gender recommended preventive health measures.

These current NAMCS data highlight significant gaps in receipt of the services by the obese and morbidly obese for the majority of diagnostic screening services and tobacco and injury education. However, obese patients did receive more weight-related services, such as testing for diabetes and cholesterol and exercise, diet and weight reduction education. Minutes spent in the visit did not vary by weight (data not shown) indicating that the patient's weight is seen as a priority and other important preventive care measures are crowded out by the emphasis on weight alone.

This survey highlights the disparities in clinical visits by patient weight. The strongest disparities seen are in female cancer screening tests and rectal exams. This gap in services is alarming given that there have been studies showing strong associations between cancer incidence and increased BMI (5,8). The higher mortality rates in obese patients could be in part due to later detection due to lower screening prevalence.

Patients with higher than normal weight are less likely to see a physician and more likely to be referred to another physician. These data confirm previous studies that clinicians prefer not to manage obesity. Given the health consequences associated with obesity, it is surprising that the morbidly obese were less likely to have scheduled follow-up appointments compared to their normal weight counterparts.

Beyond cancer screening and health education, obesity disparities existed for clinic practice patterns namely that obese patients were more likely to have referral to another physician and less likely to see a physician at the index clinic visit, have referral for psychotherapy, and to have scheduled follow-up appointments compared to their normal weight counterparts.

Our study has several limitations. First, because of NAMCS's sampling design and weights, the data presented in this report offer accurate estimates of clinical visits. However, because the data available for analysis does not include more recent years, there could have been recent changes in the accessibility of preventive measures for overweight and obese patients. As data are de-identified, it is not possible to follow patients longitudinally, and the data reported are taken from one encounter and we do not know the number of visits a patient makes in a year. However, given the large sample number, the results are representative of clinical services. In addition, NAMCS is based on physician's self reports, and these reports may be affected by the physician's preferences and desires to be seen as an unbiased provider. Lastly, our analyses depend on both height and weight patient information. If the height measurement was missing for a patient, they were excluded from the analyses which can introduce selection bias.

This study indicates a quality gap for obese patients. Although primary prevention of obesity should be a public health initiative, tertiary prevention of the progression of obesity and its manifestations should also be emphasized particularly for tobacco cessation education and mental health treatment. For the obese patient, better coordination of care and dedicated focus to emphasizing prevention and treatment of all relevant disease is needed.

Disclosure

The authors declared no conflict of interest.

Ancillary