SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS AND PROCEDURES
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGMENTS
  8. DISCLOSURE
  9. REFERENCES

We examined 5-year trends in BMI among obese primary care patients to determine whether obesity-related education such as nutrition counseling or a weight management program was associated with declines in BMI. Veterans with BMI ≥30 kg/m2 and ≥1 primary care visits in fiscal year 2002 were identified from the Veterans Health Administration's (VHA) national databases. Outpatient visits from fiscal year 2002–2006 for nutrition counseling, exercise, or weight management were grouped into five categories varying in intensity and duration: (i) intense-and-sustained, (ii) intense-only, (iii) irregular, (iv) limited, and (v) no counseling. Generalized estimating equation assessed associations between obesity-related counseling and BMI trend (annual rate of BMI change fiscal year 2002–2006) among cohort members with complete race/ethnic data (N = 179,881). Multinomial logistic regression compared intensity and duration of counseling among patients whose net BMI increased or decreased by ≥10% vs. remained stable. Compared with patients receiving “intense-and-sustained” counseling, the BMI trend of those receiving “intense-only” or “irregular” counseling was not significantly different, but patients receiving “no counseling” or “limited counseling” had significantly higher rates of decreasing BMI (−0.12 and −0.08 BMI per year; P < 0.01, respectively). This was especially true for veterans in their 50–60s, compared with the oldest veterans who were most likely to lose weight. In contrast, younger veterans (18–35 years) were least likely to lose weight; their BMI tended to increase regardless of counseling intensity and duration. Enhanced efforts are needed to detect and combat increasing weight trajectories among veterans who are already obese, especially among those aged 18–35 who are at greatest risk.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS AND PROCEDURES
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGMENTS
  8. DISCLOSURE
  9. REFERENCES

Obesity is a modifiable risk factor for a variety of chronic illnesses, including several leading causes of death (1). In response to dramatic increases in obesity rates, clinical practice guidelines have urged health-care providers to prevent and treat obesity more aggressively (1,2,3,4,5,6). While several articles have been published about utilization trends of bariatric surgery and antiobesity medications over the past decade (7,8,9), little is known about the utilization and impact of health behavior education and counseling delivered during routine clinical practice. An emerging evidence base suggests that most primary care patients do not receive a diagnosis of obesity when warranted (10,11,12,13,14). In addition, research suggests that brief physician counseling for obesity occurs infrequently, and when it does occur, it is not very effective (11,15).

We recently reported on obesity diagnosis and care practices within a large integrated health-care system, the Veterans Health Administration (VHA). Similar to other studies conducted in the private sector, we found that only about half of obese patients from primary care clinics received a formal diagnosis of obesity (14). In addition, only about one third received obesity-related education or counseling over a 5-year period (14). This follow-up study extends our earlier work to (i) examine trends in BMI among obese patients from primary care clinics over a 5-year period and (ii) determine whether obesity-related education such as nutrition counseling, exercise therapy, or a comprehensive weight management program was associated with declines in BMI. As limited longitudinal research suggests that weight trends may vary by age, we examined BMI trends among five age groups.

Because systematic reviews have indicated that most obese patients require, at a minimum, “moderate intensity” counseling (i.e., more than 1 encounter per month for the first 3 months) as well as longer term maintenance support to successfully lose weight and maintain weight loss (6), we hypothesized that obesity-related education or counseling that met criteria for “moderate”-level intensity and sustained maintenance would be more likely to be associated with declines in BMI over a 5-year period, controlling for sociodemographic and clinical characteristics.

METHODS AND PROCEDURES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS AND PROCEDURES
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGMENTS
  8. DISCLOSURE
  9. REFERENCES

Data sources and population

We utilized sociodemographic, diagnostic, anthropometric, mortality, utilization, and pharmacy data from fiscal years (FY) 2002–2006 obtained from national VHA administrative and clinical databases including the National Patient Care Database, the Vital Status File, the Corporate Data Warehouse, and the Pharmacy Benefits Management Database.

The cohort for this analysis was drawn from an earlier study examining obesity care practices in the VHA (14). Patients were included in the original study if they had one or more primary care visits in 6 of the VHA's 21 regionally integrated care networks and were obese in FY2002 based upon BMI (≥30 kg/m2) calculated from weights and heights recorded in the VHA's electronic medical record during routine clinical encounters (14). A formal diagnosis of obesity was not required because some VHA facilities allowed patients to self-refer for such services during the years covered by the study. In addition, the literature suggests that a substantial number of obese patients do not receive a formal diagnosis of obesity when warranted (11,12).

Out of 933,084 adult primary care patients who had at least one weight recorded in FY2002 and at least one height recorded in FYs2002–2006 allowing calculation of BMI, 330,802 (35.4%) met study criteria for obesity in FY2002 (14). Of these, 264,667 survived through FY2006 and received active Veterans Affairs health care, defined as 1 or more outpatient visits in at least 3 of the 4 follow-up years (FYs2003–2006). For the present analyses examining the impact of obesity-related counseling or education on weight trends, we further excluded 36,698 patients who had a diagnosis of cancer other than nonmelanoma skin cancer and those who received an antiobesity medication or underwent bariatric surgery at any time during the 5-year study period (n = 1,400). We also excluded 3,323 who had an insufficient number of quarterly BMI values or whose BMIs were censored, yielding a cohort of 223,246.

Sociodemographic and clinical characteristics

Sociodemographic data included age, sex, race/ethnicity, and marital status. Age is not normally distributed among veterans served by the VHA, given the timing of past US military conflicts and age restrictions placed on draftees and volunteers. Age was therefore categorized into five age groups based on its distribution in our cohort: 18–35 years, 36–49 years, 50–60 years, 61–72 years, and 73 years or older. Data obtained from VHA administrative databases are generally accurate and complete. Race/ethnicity is the exception; rates of missing data have been high since 2003 due to a change in its manner of collection (16,17). We also collected VHA priority status, which consists of nine categories related to disability and income (18). We grouped patients as follows: priority status 1: patients with 50% or more service-connected disability and no copayment requirement, priority status 2–6: patients with copayment requirements for medications, and priority status 7–9: patients with copayment requirements both for medical care and medications.

We also included distance from patients' homes to their primary source of VHA care as an indicator of geographic barrier to obesity care. Distance was calculated as the straight line distance between the patient's residence and most frequented facility ZIP code centroids by using the latitude and longitude coordinates for each (19,20).

We assigned patients to obesity class I (30–34 kg/m2), II (35–39 kg/m2), or III (≥40 kg/m2) according to their baseline BMI (see details below). We used outpatient and inpatient ICD-9-CM codes (21) recorded in FY2002 to assess the presence or absence of 10 comorbidities that might enhance the probability of receiving treatment for obesity-related care and seven psychiatric conditions that might pose barriers to obesity care (22,23,24). To control for overall chronic illness burden, we counted the number of medication classes used by each patient (25). Using ICD-9-CM codes, we also created an indicator variable denoting whether patients had received an ICD-9-CM diagnosis of obesity during FY2002. Individuals so diagnosed may have been more likely to have received a physician referral for obesity-related counseling, as opposed to individuals who were not diagnosed, and who may have received obesity-related counseling through self-referral.

Obesity education and counseling

We identified obesity-related counseling and education by two methods: (i) clinic stop codes, a categorization scheme used by the VHA to classify the type of clinic in which each outpatient visit occurred and (ii) Level I Current Procedural Terminology (CPT-4) codes and Healthcare Common Procedure Coding System (HCPCS) Level II codes, a uniform coding system maintained by the American Medical Association and Centers for Medicare and Medicaid Services (26) and used by the VHA to identify outpatient medical services and procedures. We included any codes related to visits at obesity-related clinics and/or visits for which obesity-related services were provided (e.g., education or counseling about weight control, nutrition counseling, or physical fitness/exercise). Because attendance at a single session of a comprehensive weight management class might be coded with both a clinic stop code and an HCPCS Level II code, only one obesity-related service or visit was counted per day.

Based on the number and pattern of outpatient visits for individual or group education or instruction in nutrition, exercise, or weight management received in FYs2002–2006, we grouped them into the following five categories varying in intensity and duration: (i) intense-and-sustained—patients who attended at least 6 sessions within a 3-month period and for whom the time between the first and last visits spanned at least 18 months, (ii) intense-only—patients who attended at least 6 sessions within a 3-month period, (iii) irregular—the lag between the patients' first and last sessions was at least 18 months, but visits never met criteria for intensity, (iv) limited—patients who attended one or more sessions that never met criteria for intensity or sustainability, and (vi) no counseling—patients who never attended any sessions for obesity-related counseling or education.

BMI trend and net change outcomes

To examine the association between obesity-related counseling and BMI trend, we used BMI derived from heights and weights obtained during routine clinical encounters. These data are stored in facility information systems and uploaded into the Corporate Data Warehouse; validation work indicates that some of these height and weight values probably reflect data entry errors (27). Therefore, we used an iterative process to eliminate or control for height and weight outliers while avoiding suspect BMI values.

In specifying the original cohort of obese primary care patients, we removed biologically “implausible” values (i.e., weights ≤70 lbs or ≥700 lbs and heights ≤48 inches or ≥84 inches) (14). We then divided each of the five study years into quarters and determined the median value for weights recorded during each quarter for every patient, yielding up to 20 quarterly median weights. Because heights are recorded less frequently than weights, we identified the modal value among all heights recorded for each patient in FYs2002–2006. When multiple modes were present, we averaged the modes if the difference between them was 3 or fewer inches and excluded cases with more extreme differences.

Our conservative estimate of baseline BMI was derived from BMIs calculated using the median of the quarterly median weights for FY2002 and modal height in FYs2002–2006. We then calculated quarterly BMI values for each of the remaining quarters for which a median weight was available, using the same modal height. We also calculated net BMI change by subtracting the last available quarterly BMI value from the first quarterly BMI value. Because clinical trials suggest that a loss of 5–10% of body weight, which corresponds to a 5–10% reduction in BMI, lowers risk for heart disease and other chronic illnesses (1), we categorized patients into three groups: (i) those whose initial BMI decreased by 10%, (ii) those whose initial BMI increased by 10% or more, and (iii) a “stable” group whose initial BMI did not change by 10% or more in either direction. Given the observational nature of the study, we elected to use the more conservative outcome of 10%, but also assessed 5% change for sensitivity analyses.

Analytic plan

We characterized cohort members in terms of their obesity care and BMI change by calculating descriptive statistics. Chi-square and t-tests assessed bivariate associations among the sociodemographic and clinical characteristics and the five obesity-related counseling intensity and duration categories. To address our primary objective of assessing whether repeated measures of BMI over a 5-year period (FYs2002–2006) in an obese population varies by the intensity or sustainability of obesity-related counseling, we used the generalized estimating equation (GEE) technique to assess the relationship of obesity-related counseling with the quarterly rate of change in BMI. GEE is an efficient estimation approach for analyzing correlated longitudinal data, while being robust to the misspecification of the correlation assumption (28).

In our GEE analyses, dependent variables included the available quarterly BMIs during FYs2002–2006. For patients who did not receive any obesity-related counseling, we included all of their quarterly BMIs in the GEE analysis. For those patients who received one or more obesity-related counseling visits, we censored any quarterly BMI value that was available prior to their first counseling visit. Missing or censored quarterly BMI values were not imputed. Instead, our GEE analyses were based on the observed BMI values, which yielded unbiased estimates under the missing-at-random assumption.

More specifically, in our GEE specification, the means of baseline BMI values and the quarterly rate of change in BMI were modeled as a linear function of the five obesity counseling intensity and duration categories and the sociodemographic and clinical covariates. The relationship between obesity-related counseling and the rate of change in BMI was assessed by the coefficients associated with a time—by–obesity-counseling interaction. To account for within-subject correlation among the repeated measures of BMI as well as to ensure that the results were robust to the correlation structure, several correlation structures were explored. Due to a moderate amount of missing race/ethnicity data (19.4%), we compared results including vs. excluding race/ethnicity variables from the analysis. The results were similar whether or not individuals with missing race/ethnicity data were included. Because of difficulties in interpreting the results for those with missing predictors under the missing-at-random assumption, GEE analyses were completed only for those individuals who had complete predictors (N = 179,881).

To further examine whether the association between obesity-related counseling categories and BMI trend varied by age group, we conducted GEE analyses including interaction terms for age—by–obesity-related counseling categories and time—by–age—by–obesity-related counseling categories. We also conducted stratified GEE analyses for each of the five age groups. To gain further insight regarding patients' BMI trends over the 5-year period, we also conducted post hoc analyses using multinomial logistic regression to compare patient characteristics and intensity and duration of obesity-related counseling among those whose BMI increased or decreased by 10% or more vs. those whose BMI was stable. We also assessed 5% changes in sensitivity analyses. P values ≤0.05 were considered statistically significant. All analyses were conducted using SAS version 9.2 (20).

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS AND PROCEDURES
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGMENTS
  8. DISCLOSURE
  9. REFERENCES

Baseline description of cohort

In our cohort of 223,246 primary care patients determined to be obese based on BMI in FY2002, most patients were male (94.1%) and approximately two-thirds were married (Table 1). Their mean (SD) age in FY2002 was 60 (12.2) years. Patients' average baseline BMI was 34.6 (4.4) kg/m2; 147,736 (66.2%) had class I, 52,242 (23.4%) had class II, and 23,268 (10.4%) had class III obesity. The patients had an average of 2.4 (1.4) medical comorbidities and 0.4 (0.8) major psychiatric conditions, and they were prescribed medications from an average of 3.7 (1.9) medication classes. Of the 17 comorbid conditions, hypertension (83.3%), hyperlipidemia (78.1%), diabetes (45.1%), osteoarthritis (43.9%), low back pain (40.8%), ischemic heart disease (39.3%), gastrointestinal reflux disease (32.9%), and depression (27.4%) were most common.

Table 1.  Sociodemographic and clinical characteristics of primary care patients with BMI > 30 kg/m2 (N = 223,246) by intensity and duration of obesity-related counseling received FY2002–2006
inline image

Bivariate analysis of obesity-related counseling

Frequency and duration of obesity counseling was significantly related to each of the sociodemographic and clinical variables evaluated (P < 0.001; Table 1). Within categories, all relationships were significantly different at P < 0.001 relative to “no counseling,” except that patients who received “intense-and-sustained counseling” did not differ significantly in Hispanic or other race/ethnicity from those that received “no counseling.” Otherwise, patients who received “intense-and-sustained” obesity-related counseling were more likely to be female and African American and to have class III obesity at baseline, but less likely to be in the two oldest age groups, to be married, to be in priority status groups 7–9, to live more than 30 miles away from the nearest VHA facility, to have class I obesity at baseline, and to have an ICD-9-CM diagnosis of obesity in FY2002, compared with patients who received “no counseling” (Table 1). Patients who received “intense-and-sustained” obesity-related counseling were also diagnosed with a significantly greater number of psychiatric comorbidities and prescribed a significantly greater number of medication classes in FY2002.

Multivariable GEE analyses

Covariates associated with baseline BMI and BMI change. Consistent with theory, the GEE results were robust to the correlation assumption. Holding all other covariates constant, baseline BMI was inversely associated with increasing age (Table 2). In addition, baseline BMI was greater among African American patients, among those who had more psychiatric diagnoses, and among those who were required to make copayments for medications only (i.e., priority status 2–6).

Table 2.  Patient characteristics and obesity-related counseling associated with baseline BMI and annual rate of BMI change over time from FY2002–FY2006 among primary care patients with BMI ≥ 30 kg/m2 (N = 179,881)
inline image

Holding all other covariates constant, age was inversely associated with the magnitude of the annual rate of BMI change (i.e., change in BMI units per year; Table 2). That is, the covariate-adjusted rate of BMI change per year decreased with increasing age for each of the four youngest age groups, relative to those 73 years and older: 0.50 BMI units (18–35 years), 0.37 BMI units (36–49 years), 0.30 BMI units (50–60 years), and 0.17 BMI units (61–72 years). In addition, the covariate-adjusted annual rate of BMI change was significantly greater (i.e., in terms of either gaining more or losing less weight, relative to the comparison group) among those who were male vs. female (0.03 BMI units per year; P < 0.01), were married vs. unmarried (0.02 BMI units per year; P = 0.0001), lived more than 30 miles from their most frequently used VHA facility vs. lived fewer than 30 miles away (0.02 BMI units per year; P < 0.0001), and had an obesity diagnosis vs. had no obesity diagnosis in FY2002 (0.14 BMI units per year; P < 0.0001). Holding all other covariates constant, the annual rate of BMI change was significantly less (i.e., in terms of gaining less weight or losing more weight, relative to the comparison group) among those who were Hispanic relative to White (−0.02 BMI units per year; P < 0.01), had a lower baseline BMI in FY2002 (−0.02 BMI units per year, P < 0.0001), and had fewer psychiatric comorbidities (−0.04 BMI units per year; P < 0.0001).

Obesity counseling intensity and duration categories associated with BMI trend. The covariate-adjusted BMI trend of patients who received “intense-and-sustained” obesity-related counseling did not significantly differ from those who had “intense-only” (−0.04 BMI units per year; P = 0.43) or “irregular” obesity counseling (−0.04 BMI units per year; P = 0.22; Table 2). In contrast, patients who received “no counseling” or only “limited” obesity-related counseling had significantly lower baseline BMI values (i.e., their baseline BMIs were significantly lower by 0.21 BMI unit; P = 0.01 for both groups) and a significantly lower rate of increasing BMI (i.e., their BMIs increased at a significantly lower rate: 0.12 BMI unit less per year for the “no counseling” group and 0.08 BMI unit less per year for the “limited” group; P < 0.01 for both groups), compared with patients who received “intense-and-sustained counseling.”

When obesity-counseling—by–age group interaction terms were added to the GEE model, the association between BMI trend and “intense-and-sustained” obesity-related counseling differed significantly between the oldest age group (73+ years old) and those who were in their 50s or 60s. Whether compared with “limited counseling” or “no counseling,” the difference in the adjusted rate of annual BMI change for those who received “intense-and-sustained counseling” was significantly greater in the oldest age group compared with those 50–60 years old (−0.25 BMI units vs. −0.23 BMI units per year, respectively) or those 61–72 years old (−0.23 BMI units vs. −0.22 BMI units per year, respectively).

Multivariable GEE analysis stratified by age groups

GEE analyses conducted separately for each age group indicated that category of obesity counseling frequency or duration was not associated with BMI trend for the youngest or the oldest age group. Although counseling category was significantly associated with BMI trend in the three intermediate age groups, it was in the unexpected direction, with BMI increasing over the 5-year period most among those who received the most intensive counseling. Visual examination of age-stratified plots (Figure 1) indicates generally increasing covariate-adjusted BMIs for the three youngest age groups, a relatively stable trend for the 61–72 age group and decreasing trend for those 73 and older.

image

Figure 1. Unadjusted and covariate-adjusted trends in BMI over time from FY2002–2006 by obesity-related counseling categories stratified by five age groups among primary care patients with BMI ≥30 kg/m2 (N = 223,246). Asterisk denoting heavy black line indicates overall covariate-adjusted averages, colored lines indicate unadjusted averages.

Download figure to PowerPoint

Net BMI change and association with obesity-related counseling and age

With regard to 10% net BMI change, most patients (82.8%) fell into the “stable” category. Only 10.4% of the patients experienced decreases of 10% or more of their initial BMI, while 6.8% of these patients had their initial BMI increase by 10% or more over the 5-year period. In multinomial logistic regression analysis for cohort members with complete race/ethnic data (N = 179,881), the association between BMI decreasing by 10% or more, relative to remaining stable, and obesity-related counseling intensity and duration was weak or nonsignificant (Table 3). In contrast, the odds of experiencing a 10% or greater increase in BMI, relative to remaining stable, increased as the intensity or duration of obesity-related counseling increased, with the highest relative odds (odds ratio (OR) = 1.73; 95% confidence interval (CI):1.52, 1.96) noted for those who received “intense-and-sustained counseling” compared with those who received “no counseling.”

Table 3.  Multinomial logistic regression model predicting ≥10% increase or decrease in BMI from FY2002–2006 among obese primary care patients served by the VHA (N = 179,881)
inline image

The odds of experiencing a 10% or greater decrease in BMI, relative to remaining stable, were significantly less likely for each of the four youngest groups, compared with those 73 years or older, controlling for all other variables (P < 0.0001; Table 3). In addition, the relative odds of BMI increasing by 10% or more were inversely related to age (18–35 years: OR = 7.77; 95% CI: 6.93, 8.70; 36–49 years: OR = 4.67; 95% CI: 4.26, 5.13; 50–60 years: OR = 3.55; 95% CI: 3.25, 3.87; and 61–72 years: OR = 1.98; 95% CI: 1.81, 2.17; P < 0.0001).

Using a smaller net BMI change of 5% or more, over half of the patients (55.2%) fell into the “stable” category, while 25.5% experienced decreases of 5% or more of their initial BMI and 19.3% had increases of 5% or more over the 5-year period. The results of the multinomial logistic regressions predicting increases of 5% or more were virtually identical to those for 10% or more. In comparison, the models predicting 5% decreases in net BMI change differed slightly from those predicting 10% or greater decreases. While the association between BMI decreasing by 10% or more and obesity-related counseling was significant for the “limited” and “irregular” counseling categories only, the association between BMI decreasing by 5% or more, relative to remaining stable, was nonsignificant for all four categories of counseling intensity and duration.

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS AND PROCEDURES
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGMENTS
  8. DISCLOSURE
  9. REFERENCES

Obese patients who received obesity-related counseling that was more intense and sustained had significantly greater increases in 5-year BMI compared with those who received no counseling or obesity counseling that was limited. This finding was especially pronounced among veterans who were primarily in their 50s and 60s, compared with the oldest veterans who were most likely to lose weight. In contrast, younger veterans (18–35 years) were least likely to lose weight; their BMI tended to increase regardless of counseling intensity and duration.

Holding all other sociodemographic and clinical characteristics constant, baseline BMI values were inversely related to age. Moreover, 5-year BMI trends exhibited a curvilinear pattern in relation to age. Although the BMI values of the three youngest age groups increased over the 5-year period, the BMIs of those who were 61–72 years of age were relatively stable, while those who were 73 years or older experienced declines in BMI values. This pattern occurred even though we excluded patients who died or were diagnosed with cancers other than nonmelanoma skin cancer during the study period, although it is possible that some elders may have experienced involuntary weight loss due to conditions other than cancer or frailty syndrome (29).

Only a few studies have assessed longitudinal changes in BMI; limited data suggest that BMI tends to be relatively stable over the short term (30,31). Recently published population-based data from 8,548 Canadians indicate that over 5 years, most maintained their baseline BMI classification; when change did occur, BMI category was more likely to increase than decrease (30). On average, men below the age of 45 and women below the age of 55 gained ∼0.45 kg/year, but weight leveled off with increasing age and actually decreased over time in the oldest age groups (30). An even larger prospective population-based study with an 11-year follow-up of over 45,000 Norwegians found that the prevalence of overweight and obesity increased, especially among the youngest age groups, while those who were 70 years or older more typically experienced weight loss (32). These curvilinear patterns are consistent with our findings and other research indicating that aging is associated with a decrease in skeletal muscle mass and an increase in body fat (33). Lean body mass declines progressively during adult life and the rate of decline tends to speed up in later years (34).

Examining net BMI change indicated, however, that some patients did lose clinically significant amounts of weight. Although most patients' weights were “stable,” one-tenth of patients were able to lose 10% or more of their initial body weight, an amount that has been found to be associated with preventing or controlling chronic illnesses such as type 2 diabetes mellitus and cardiovascular disease (35,36,37). Patients in the four youngest age groups, however, were significantly less likely to lose 10% or more of their initial body weight compared with those 73 years of age or older. Our finding that obese patients who were 18–35 years old had a 7.77 greater odds than the oldest veterans of their BMI increasing by 10% or more over the 5-year period is especially alarming. Although we could not ascertain how recently these younger veterans may have separated from military service, our findings suggest that veterans may be at greatest risk for weight gain during the initial years of reintegrating into civilian life. Compared with their older counterparts, recently discharged veterans may be more prone to adverse lifestyle changes as they learn to readapt to a more obesogenic environment. Many of these veterans may need extra support to help them navigate the early postdeployment period and learn how to moderate caloric intake in response to changes in level of physical activity.

We also found that obesity-related counseling was not strongly associated with declines in BMI. In contrast, the percentage of patients who gained 10% or more of their body weight increased as obesity-related counseling increased in frequency and duration. Our findings suggest that patients who gained weight during the 5-year study period tended to receive obesity-related counseling more frequently and/or over a longer period of time than patients who lost weight or whose weight was stable. Although this might be seen as discouraging, it is possible that the less successful patients may have gained even more weight if they had not received continued counseling.

Although our study suggests that obesity-related counseling was not generally associated with weight loss among our cohort of obese primary care patients, it is important to note that during most of the study observation period, obesity care varied considerably within the VHA. Fewer than half of the VHA medical centers who responded to administrative surveys in 2001–2002 indicated that they provided weight management programs for their patients (L.C. Kahwati, personal communication). The term “program” was widely interpreted and ranged from simple dietary referrals to more intensive, multidisciplinary programs. Therefore, it is probable that the content of obesity counseling varied as well, even if patients did receive the full number of counseling sessions recommended for moderate intensity counseling. Indeed, most obesity-related counseling visits identified for this study involved individual or group visits for nutrition or dietetic services. Only a small number of visits (N = 6,831) were assigned the CPT-4 code for “weight management class” (S9449), which may or may not have been comprehensive or evidence based.

It is also possible that patients with early success in losing weight elected to continue weight-loss or weight-maintenance efforts on their own, while less successful patients tended to continue formal treatment within the health-care system. Compared with women, studies suggest that men are less likely to engage in formal weight-loss programs and more likely to prefer to undertake weight-loss efforts on their own (38,39), and our study population was in large part male. Furthermore, patients who received “intense-and-sustained counseling” had a greater number of psychiatric comorbidities and were prescribed medications from a greater number of medication classes. Although only 4% of patients with psychiatric comorbidities were exposed to potentially weight-promoting medications, the possible contribution of such medications to observed BMI gains deserves further exploration. Another possibility is that patients who failed to lose weight, but continued to attend obesity-related counseling or education sessions, derived other benefits such as opportunities to receive social support from other veterans and staff.

Beginning in FY2006, the VHA began to implement a comprehensive, evidence-based, national stepped-care program for identifying and managing overweight and obesity in primary care, the MOVE! program (40). Although an early 2007 administrative evaluation of the MOVE! program also found that duration of treatment was not associated with weight loss and that number of visits was inversely related to success in losing weight (L.C. Kahwati, personal communication), it is hoped that as the program expands and stabilizes, substantially more overweight and obese veterans will learn about it and benefit from its evidence-based practices.

We restricted our focus to health behavior education and counseling for obesity, primarily because fewer than 1% of obese veterans received weight-loss medications or bariatric surgery during the observation period (14). This low level of utilization may have been partially due to systemic barriers, including variations in local formulary prescribing practices or policies and limited number of approved VHA bariatric centers during the study period (L.C. Kahwati, personal communication). Studies also suggest that men are less likely to use weight-loss medications and undergo bariatric surgery (41,42). The adverse effect profile for both forms of treatment, however, may make them a less desirable choice for many veterans, who tend to be older and sicker than the trial participants in which those treatments were tested and approved (L.C. Kahwati, personal communication). Regardless, weight-loss medication and bariatric surgery are recommended as options for obese patients with increased risk who have failed to lose weight following adequate trials of lifestyle interventions (1). Studies suggest that weight-loss medications have an additive effect when combined with counseling (43,44) and that lifestyle interventions may help maintain or enhance weight outcomes in bariatric surgery (45), underscoring the need for future surveillance of their utilization and impact in usual care.

This retrospective observational cohort study has several limitations, including its reliance on heights and weights recorded during routine clinical practice, which may contain some data entry errors or other measurement errors (27). Although we used several strategies to eliminate, or at least minimize, the impact of extreme outliers, we acknowledge that errors may remain, but there is no evidence to suggest that they are not randomly distributed. Our inability to capture potentially important confounders from the administrative data, such as socioeconomic status, education level attainment, and employment status, is another limitation that must also be acknowledged.

Because clinicians' notes are not captured in the VHA's national administrative data, we were not able to identify any instances of brief physician counseling that may have occurred during routine visits or of patients' reports of using commercial weight-loss programs or self-directed weight-loss attempts. As noted earlier, systematic reviews suggest that brief physician counseling is not sufficient in and of itself to achieve clinically meaningful weight loss (6,15), although it may serve to motivate patients to try to lose weight. In addition, the out-of-pocket expenses for most commercial programs typically run as high as $1,200 for 3 months, costs that may be unrealistic for veterans using the VHA, who tend to be less well off than veterans who receive care outside the VHA or than nonveterans (46). At least one survey has indicated that few VHA users engage in such programs (L.C. Kahwati, personal communication). Furthermore, patients who are not provided evidence-based guidance may be more likely to use weight-loss strategies that are ineffective or even harmful (e.g., skipping meals, fasting, or taking unregulated dietary supplements). Another major study limitation is our inability to assess patients' levels of motivation and social support, which clearly play important roles in determining patients' adherence and response to obesity care.

Although our cohort of predominantly male adult outpatients is not generalizable to the population at large, our findings provide interesting insights about men, who tend to be underrepresented in obesity research. While not representative of the entire VHA, our large cohort was identified from a population of 1.5 million primary care patients served by six geographically-diverse care networks, representing about 20% of the VHA's 5.4 million patients. It is important to emphasize that this observational study did not randomly assign patients to counseling intensity and duration and that there were significant differences at baseline (such as, for example, patients who received “intense-and-sustained counseling” were less likely to have class I obesity).

This observational study, however, provides valuable information regarding the routine management of obesity among primary care patients in a large integrated health-care system committed to responding to a persistent and growing health problem, the types of patients who do and do not receive such care, and their outcomes. These pre-implementation data from FY2002–2006 may be used to gauge the future impact of the MOVE! program. Similar evaluations should be repeated now that MOVE! has been in place for 5 years, and future studies should try to incorporate better assessments of patient characteristics and motivation, especially among individuals who continue to gain weight in spite of treatment. Our results suggest that enhanced efforts may be needed to detect and combat increasing weight trajectories among veterans who are already obese, especially those who are 18–35 years of age who appear to be at greatest risk. Considering the potential for reduced cardiovascular and metabolic risk, these are likely to be resources worth developing and deploying.

ACKNOWLEDGMENTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS AND PROCEDURES
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGMENTS
  8. DISCLOSURE
  9. REFERENCES

The research reported here was supported by the Department of Veteran Affairs, Veterans Health Administration, Health Services Research and Development Service (project no. IIR 05–121). The authors thank Leila C. Kahwati, MD, MPH, Deputy Chief Consultant for Preventive Medicine, Office of Patient Care Services, Veterans Health Administration, for her review and suggestions. All of the authors except Dr Wang are employees of the Department of Veterans Affairs. The views expressed in this article are those of the author(s) and do not necessarily represent the views of the Department of Veterans Affairs.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS AND PROCEDURES
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGMENTS
  8. DISCLOSURE
  9. REFERENCES
  • 1
    NHLBI Obesity Education Initiative. Clinical guidelines on the identification, evaluation, and treatment of overweight and obesity in adults: the evidence report. Washington, DC, National Institutes of Health, 1998 June.
  • 2
    National Task Force on the Prevention and Treatment of Obesity. Overweight, obesity, and health risk. Arch Intern Med 2000;160:898904.
  • 3
    Nawaz H, Katz DL. American College of Preventive Medicine Practice Policy statement. Weight management counseling of overweight adults. Am J Prev Med 2001;21:7378.
  • 4
    Lyznicki JM, Young DC, Riggs JA, Davis RM; Council on Scientific Affairs, American Medical Association. Obesity: assessment and management in primary care. Am Fam Physician 2001;63:21852196.
  • 5
    National Task Force on the Prevention and Treatment of Obesity. Medical care for obese patients: advice for health care professionals. Am Fam Physician 2002;65:8188.
  • 6
    McTigue KM, Harris R, Hemphill B et al. Screening and interventions for obesity in adults: summary of the evidence for the U.S. Preventive Services Task Force. Ann Intern Med 2003;139:933949.
  • 7
    Mehrotra C, Serdula M, Naimi TS et al. Population-based study of trends, costs, and complications of weight loss surgeries from 1990 to 2002. Obes Res 2005;13:20292034.
  • 8
    Davis MM, Slish K, Chao C, Cabana MD. National trends in bariatric surgery, 1996–2002. Arch Surg 2006;141:714; discussion 75.
  • 9
    Stafford RS, Radley DC. National trends in antiobesity medication use. Arch Intern Med 2003;163:10461050.
  • 10
    Stafford RS, Farhat JH, Misra B, Schoenfeld DA. National patterns of physician activities related to obesity management. Arch Fam Med 2000;9:631638.
  • 11
    McAlpine DD, Wilson AR. Trends in obesity-related counseling in primary care: 1995–2004. Med Care 2007;45:322329.
  • 12
    Ma J, Xiao L, Stafford RS. Underdiagnosis of obesity in adults in US outpatient settings. Arch Intern Med 2009;169:313314.
  • 13
    Ma J, Xiao L, Stafford RS. Adult obesity and office-based quality of care in the United States. Obesity 2009; e-pub ahead of print 5 February 2009.
  • 14
    Noël PH, Copeland LA, Pugh MJ et al. Obesity diagnosis and care practices in the Veterans Health Administration. J Gen Intern Med 2010;25:510516.
  • 15
    Tsai AG, Wadden TA. Treatment of obesity in primary care practice in the United States: a systematic review. J Gen Intern Med 2009;24:10731079.
  • 16
    Sohn MW, Zhang H, Arnold N et al. Transition to the new race/ethnicity data collection standards in the Department of Veterans Affairs. Popul Health Metr 2006;4:7.
  • 17
    Long JA, Bamba MI, Ling B, Shea JA. Missing race/ethnicity data in Veterans Health Administration based disparities research: a systematic review. J Health Care Poor Underserved 2006;17:128140.
  • 18
    Gao J, Wang Y, Engelhardt J. Logistic analysis of veterans' eligibility-status changes. Health Serv Manage Res 2005;18:175185.
  • 19
    U.S. Census Bureau. (2000). Population and housing profile, U. S. census STF1 <http:factfinder.census.govservletDTGeoSearchByListServletds_nameDEC_2000_SF1_U&_langen&_ts272882185413>. Accessed 6 October 2009.
  • 20
    SAS Institute Inc., SAS 9.2, SAS Institute Inc.: Cary, NC, 2008.
  • 21
    World Health Organization. International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM). American Medical Association Press: Chicago, 2006.
  • 22
    Kerr EA, Hofer TP, Hayward RA et al. Quality by any other name?: a comparison of three profiling systems for assessing health care quality. Health Serv Res 2007;42:20702087.
  • 23
    Cornell J, Pugh JA, Williams JW Jr., Kazis L, Lee A, Parchman M, Zeber J, Pederson T, Montgomery KA, Noël PH. Multimorbidity clusters: clustering binary data from a large administrative medical database. Applied Multivariate Research 2008;12:163182.
  • 24
    Zeber JE, Copeland LA, Amuan M, Cramer JA, Pugh MJ. The role of comorbid psychiatric conditions in health status in epilepsy. Epilepsy Behav 2007;10:539546.
  • 25
    Schneeweiss S, Seeger JD, Maclure M et al. Performance of comorbidity scores to control for confounding in epidemiologic studies using claims data. Am J Epidemiol 2001;154:854864.
  • 26
    Buck CJ. Saunders 2008 ICD-9-CM, Volumes 1, 2, and 3, HCPCS Level II and CPT Professional Edition Package. Saunders: St. Louis, MO, 2007.
  • 27
    Noël PH, Copeland LA, Perrin RA et al. VHA Corporate Data Warehouse height and weight data: opportunities and challenges for health services research. J Rehabil Res Dev 2010;47:739750.
  • 28
    Zeger SL, Liang KY, Albert PS. Models for longitudinal data: a generalized estimating equation approach. Biometrics 1988;44:10491060.
  • 29
    Huffman GB. Evaluating and treating unintentional weight loss in the elderly. Am Fam Physician 2002;65:640650.
  • 30
    Hopman WM, Leroux C, Berger C et al.; CaMos Research Group. Changes in body mass index in Canadians over a five-year period: results of a prospective, population-based study. BMC Public Health 2007;7:150.
  • 31
    Heo M, Faith MS, Pietrobelli A. Resistance to change of adulthood body mass index. Int J Obes Relat Metab Disord 2002;26:14041405.
  • 32
    Drøyvold WB, Nilsen TI, Krüger O et al. Change in height, weight and body mass index: Longitudinal data from the HUNT Study in Norway. Int J Obes (Lond) 2006;30:935939.
  • 33
    Beaufrère B, Morio B. Fat and protein redistribution with aging: metabolic considerations. Eur J Clin Nutr 2000;54 Suppl 3:S48S53.
  • 34
    Forbes GB, Reina JC. Adult lean body mass declines with age: some longitudinal observations. Metab Clin Exp 1970;19:653663.
  • 35
    Tuomilehto J, Lindström J, Eriksson JG et al.; Finnish Diabetes Prevention Study Group. Prevention of type 2 diabetes mellitus by changes in lifestyle among subjects with impaired glucose tolerance. N Engl J Med 2001;344:13431350.
  • 36
    Knowler WC, Barrett-Connor E, Fowler SE et al.; Diabetes Prevention Program Research Group. Reduction in the incidence of type 2 diabetes with lifestyle intervention or metformin. N Engl J Med 2002;346:393403.
  • 37
    Tzotzas T, Evangelou P, Kiortsis DN. Obesity, weight loss and conditional cardiovascular risk factors. Obes Rev 2011;12:e282e289.
  • 38
    Neumark-Sztainer D, Sherwood NE, French SA, Jeffery RW. Weight control behaviors among adult men and women: cause for concern Obes Res 1999;7:179188.
  • 39
    Tsai AG, Wadden TA, Pillitteri JL et al. Disparities by ethnicity and socioeconomic status in the use of weight loss treatments. J Natl Med Assoc 2009;101:6270.
  • 40
    Kinsinger LK, Jones KR, Kahwati LK, Harvey R, Burdick M, Zele V, Yevich SJ. Design and dissemination of the MOVE! Weight Management Program for Veterans. Prev Chron Dis 2009:6:16.
  • 41
    Cawley J, Rizzo JA. One pill makes you smaller: the demand for anti-obesity drugs. Adv Health Econ Health Serv Res 2007;17:149183.
  • 42
    Davis MM, Slish K, Chao C, Cabana MD. National trends in bariatric surgery, 1996–2002. Arch Surg 2006;141:7174.
  • 43
    Wadden TA, Berkowitz RI, Womble LG et al. Randomized trial of lifestyle modification and pharmacotherapy for obesity. N Engl J Med 2005;353:21112120.
  • 44
    Digenio AG, Mancuso JP, Gerber RA, Dvorak RV. Comparison of methods for delivering a lifestyle modification program for obese patients: a randomized trial. Ann Intern Med 2009;150:255262.
  • 45
    Papalazarou A, Yannakoulia M, Kavouras SA et al. Lifestyle intervention favorably affects weight loss and maintenance following obesity surgery. Obesity (Silver Spring) 2010;18:13481353.
  • 46
    Morgan RO, Teal CR, Reddy SG, Ford ME, Ashton CM. Measurement in Veterans Affairs Health Services Research: veterans as a special population. Health Serv Res 2005;40:15731583.