Lifestyle Modification to Improve Blood Pressure Control in Individuals With Diabetes: Is physician advice effective?

Data from the 1998 NHIS (21) were analyzed. The NHIS is a national household survey of nonmilitary and noninstitutionalized people in the U.S., sponsored by the National Center for Health Statistics of the Centers for Disease Control and Prevention (22). For the sample adult core (23), one adult per family was randomly selected to respond to a Computer-Assisted Personal Interview questionnaire. The sample was selected by a complex sampling design involving stratification, clustering, and multistage sampling with a nonzero probability of selection for each person.

A new module that incorporated questions about health behavior of adults—the adult prevention module (23)—was added to the 1998 survey. This study combined data from the sample adult core and the adult prevention module. Final weights were constructed to reflect the unequal probability of selection and adjust for nonresponse and poststratification. Details about the methodology of the 1998 NHIS are available online (21–23).

The sample adult core of the 1998 NHIS collected data on CVD risk factors such as hypertension, high cholesterol, obesity, physical activity levels, and smoking, besides data on diabetes. However, the sample adult prevention module collected data on physician advice and patient behavior related only to hypertension and hypercholesterolemia. Consequently, the remaining analysis in this article includes only data on self-reported behavior of individuals with and without diabetes in response to physician advice to modify hypertension-related lifestyles.

The proportion of patients who were advised by a physician to modify their lifestyle or prescribed medications to improve blood pressure control was based on the number of patients who answered “yes” to the following questions:

1. Has a doctor or other health professional ever advised you to go on a diet or change your eating habits to help lower your blood pressure?

2. Because of your high blood pressure, has a doctor or other health professional ever advised you to exercise?

3. Was any medicine ever prescribed by a doctor for your high blood pressure?

The proportion of patients who were adherent to a physician’s advice to modify their lifestyle or were taking prescribed medications to improve blood pressure control was based on the number of patients who answered “yes” to the following questions:

1. If you were told you had high blood pressure and you were advised to go on a diet or change your eating habits to help lower your blood pressure, are you now following this advice?

2. If you were told you had high blood pressure and you were advised to exercise to help lower your blood pressure, are you now following this advice?

3. If you were told you had high blood pressure and any medication was prescribed by a doctor for your high blood pressure, are you now taking any medicine prescribed by a doctor for your high blood pressure?

Data were collected on age, sex, race/ethnicity, education, and household income as the percentage of federal poverty level. Other demographic data included marital status, census region, employment status, and perceived health status. Additional data were collected on the number of physician visits in 1998 and the presence of comorbid conditions such as hypertension, high cholesterol, and coronary artery disease. Three age categories were created: 18–44, 45–64, and ≥65 years. Four racial/ethnic categories were used: white, black, Hispanic, and other. Three levels of education were created: less than high school graduate, high school graduate, and more than high school graduate.

Household income was reported as the ratio of family income to poverty threshold as follows: >400%, 200–399%, 125–199%, and <125%. Marital status was dichotomized as married and not married. Similarly, employment status was grouped as employed and unemployed. Health status was based on the individual’s current opinion of their health in comparison to 12 months previously (better, same, or worse than 12 months ago). For this analysis, perceived health status was categorized as better/same and worse.

Contact with a physician was based on the individual’s report of how many visits he or she made to a primary care physician (general practice physician, family medicine physician, or internal medicine physician) within the previous 12 months. Three categories were created to identify the number of physician visits within the previous 12 months: one, two or three, and four or more visits.

STATA statistical software (24), which accounts for the multistage sampling, clustering, and stratification design of national surveys such as the NHIS (25), was used for statistical analyses and to generate population estimates. Three separate analyses were performed. First, sociodemographic characteristics and proportion advised by a physician to modify lifestyle in individuals with and without diabetes who also had hypertension and at least one physician visit in the previous 12 months were compared using χ² statistics.

Second, multiple logistic regression was used to determine the likelihood of physician advice and patient adherence by diabetes status controlling for age, sex, race/ethnicity, education, household income, marital status, census region, employment, health status, and number of physician visits. The third set of analyses were restricted to individuals with diabetes and hypertension who had at least one physician visit and were advised by a physician to modify their lifestyle. In this group of patients, multiple logistic regression was used to determine whether the likelihood of patient adherence differed by age, sex, and race/ethnicity, controlling for other previously listed covariates.

In 1998, 32,440 individuals aged ≥18 years completed the NHIS interview, giving an overall response of 74%. Of this number, 1,906 people had diabetes, not including women with gestational diabetes. Table 1 compares the characteristics of people with diabetes (n = 1,609) with those of people without diabetes (n = 19,672) among people who had one or more physician visit in 1998. Individuals with diabetes were older, more likely to be black, less educated, poorer, and more likely to be from the southern U.S. Similarly, individuals with diabetes were more likely to report being unemployed, report worse health status, have multiple physician visits, and have comorbid hypertension, hypercholesterolemia, and established coronary artery disease. Both groups did not differ significantly by sex or marital status.

Table 2 compares the prevalence of physician advice and patient’s reported adherence among individuals with hypertension and diabetes (n = 989) and individuals with hypertension but without diabetes (n = 5,030). Physicians were more likely to advise individuals with diabetes to lose weight and increase physical activity for blood pressure control. In addition, physicians were more likely to prescribe medications to control blood pressure in people with diabetes. On patient adherence, people with diabetes were more likely to report losing weight and taking medications for blood pressure control in response to a physician’s advice. On the contrary, individuals with diabetes did not differ significantly from individuals without diabetes in the proportion who reported engaging in physical activity in response to a physician’s advice.

Table 3 shows the likelihood of physician advice and patient adherence by diabetes status controlling for covariates. Among people with hypertension who had one or more physician visit and were advised by a physician to modify their lifestyle (n = 5,655), people with diabetes were twice as likely as people without diabetes to receive advice from a physician or be prescribed antihypertensive medications. Among a subgroup of patients who reported adherence to physician advice (n = 2,880), people with diabetes were two times more likely to report taking prescribed medications. However, the likelihood of adherence to a physician’s advice to increase physical activity or lose weight did not differ significantly by diabetes status.

Table 4 shows the results of analyses restricted to individuals with diabetes and hypertension who had at least one physician visit and received advice from a physician to modify their lifestyle or take medications (n = 769). In this subset of people with diabetes, patient adherence to physician recommendations did not differ significantly by sex or race/ethnicity. However, there were significant differences in patient adherence by age. Individuals 18–44 years old were six times less likely to adhere to weight loss recommendations and three times less likely to report taking medications to control high blood pressure compared with people ≥65 years of age.

This study has three major findings. First, the advice of a physician seems effective at modifying hypertension-related lifestyles in people with diabetes. Second, people with diabetes appear more likely to adhere to a physician’s advice to take medications compared with people without diabetes. Third, controlling for covariates, adherence to a physician’s advice seems to differ significantly by age but not by sex or race/ethnicity.

Patients with diabetes appeared to modify behavior in response to the advice of physicians. This finding contradicts prevailing physician assumptions that most patients are unwilling to change negative health habits and that counseling is ineffective in modifying CVD risk behavior (14–18). Prior studies have shown that physician perception about the effectiveness of counseling affects how often patients get counseled (26), the duration of counseling (27), and the use of effective counseling techniques (28). Therefore, the results of this study, which suggest that physicians are effective at getting individuals with diabetes to modify CVD risk behavior, should encourage physicians to offer preventive care counseling to more patients with diabetes.

Adherence to a physician’s advice to lose weight and increase physical activity by individuals with diabetes did not differ by sex or race/ethnicity. This finding contradicts the findings of the 1996 Surgeon General’s report on physical activity (19) and another Centers for Disease Control and Prevention study (20), which reported that women and individuals from minority ethnic groups were less likely to engage in physical activity. In addition, unlike the Surgeon General’s report (19), which suggested that older individuals were less likely to engage in physical activity, this study found that the proportion of individuals engaging in physical activity did not differ by age. Therefore, previous assumptions about which patients are less likely to engage in physical activity for CVD risk reduction may not be applicable to patients with diabetes.

Lifestyle counseling did not seem effective at promoting CVD risk modification in younger patients with diabetes and hypertension. Controlling for factors likely to affect behavior, individuals with diabetes in the 18- to 44-year-old age-group were six times less likely to report losing weight and three times less likely to take antihypertensive medications in response to a physician’s advise. On one hand, it may be that primary care physicians are not providing effective theory-based brief counseling interventions, as recommended by the U.S. Preventive Services Task Force for a variety of reasons (29). Alternatively, it may be that physician counseling interventions in primary care for this group of patients need to be redesigned to cater to their unique needs. Regardless of the cause, there is a need to assist younger patients to adopt and maintain healthy lifestyles to reduce the complications and deaths that result from diabetes and hypertension.

Another important finding was that patients with diabetes in this study had similar demographic characteristics, access to health care, and comorbidity characteristics as patients from other national studies (5,30–32). In addition, the prevalence of CVD risk factors and the pattern of physician advice and patient adherence to advice were similar to those from previous studies (5,13), which suggests that results of this study can be generalized to the adult population with diabetes in the U.S.

Analogous to previously reported findings (13), physicians’ advice and patient adherence were significantly influenced by the presence of established CVD or CVD risk factors. People perceived to be at low risk seemed less likely to be advised by a physician and were less likely to adhere to advice if given. Such patterns of behavior in physicians and patients need to be discouraged because they may minimize the effectiveness of preventive care counseling in low-risk patients with diabetes, who may benefit from CVD lifestyle modification as much as high-risk patients.

Finally, despite substantial evidence on the benefit of exercise training, the results of this study suggest that patients with diabetes may be less likely to engage in exercise training than other lifestyle interventions such as weight loss. For example, in contrast to increased adherence by individuals with diabetes to a physician’s advice to lose weight or take medications, adherence to a physician’s advice to exercise more did not differ by diabetes status. In addition, among individuals with diabetes, the lower adherence to exercise recommendations in response to a physician’s advice did not differ by age, unlike the case with other recommendations. Therefore, brief counseling strategies that are effective at increasing physical activity levels across age, sex, and racial/ethnic groups need to be implemented in primary care, particularly for individuals with diabetes.

The results of this study should be interpreted in light of the following limitations. First, it is important to recognize the ambiguity of the term “physician advice.” In most national surveys, physician advice is used interchangeably with physician counseling despite the fact that the U.S. Preventive Services Task Force has clear guidelines for what constitutes counseling (33). Therefore, reports from national surveys such as the Behavorial Risk Factor Surveillance Study (BRFSS) and NHIS, which use similar sets of questions to determine national estimates of physician advice on preventive services and patient adherence, need to be interpreted cautiously.

Second, the individual’s understanding of advice or counseling needs to be clarified, particularly among individuals with diabetes who may receive advice on diet, weight loss, and physical activity from a variety of sources. Whether these factors affect self-reported estimates of counseling and adherence to counseling is unclear; however, these inconsistencies need to be considered as questions for future national surveys are developed.

A third limitation is the way adherence to physician advice was measured. The relatively weak measures of patient behavior may overestimate the true prevalence of behavior or adherence. However, because of use of similar questions by other nationally representative surveys (13,20), the questions on patient behavior remain useful for generation of hypotheses that can be tested by more precise measures. Additionally, whereas the prevalence estimates may be inflated, the relative differences between individuals with and without diabetes and among individuals with diabetes are likely to be reliable.

A fourth limitation is that the NHIS did not collect data on the frequency, duration, intensity, or sustainability of exercise in survey participants. However, longitudinal studies have shown that behavior modification due to physician counseling is sustained for up to 1 year of follow-up (34,35). Additionally, a recent meta-analysis of 54 randomized trials found that aerobic exercise was associated with blood pressure reduction regardless of the frequency or intensity of exercise (9). Finally, there is the potential for recall bias. However, this is unlikely to be significant in this study because previous studies have documented the validity of self-reported information on diabetes, CVD risk factors, and health promotion habits (36,37).

In conclusion, despite these potential limitations, the above results remain important. This study suggests that among adults with diabetes and hypertension, physician advice to lose weight and take antihypertensive medications is effective at modifying hypertension-related lifestyles regardless of sex or race/ethnicity.

L.E.E. is supported by grant 1K08HS11418 from the Agency for Health Care Research and Quality, Rockville, Maryland.