Addictive Behaviors, Vol. 16, pp. 151-160. Printed in the USA. All rights reserved.

1991 Copyright

0306-4603/91 $3.00 + .OO 0 1991 Pergamon Press plc

DEMOGRAPHIC, KNOWLEDGE, PHYSIOLOGICAL, AND BEHAVIORAL VARIABLES AS PREDICTORS OF COMPLIANCE WITH DIETARY TREATMENT GOALS IN HYPERTENSION THOMAS

L. SCHMID and ROBERT W. JEFFERY University

of Minnesota

LYNN ONSTAD George Washington

University

SHEILA A. CORRIGAN University

of Mississippi

Abstract - The Hypertension Prevention Trial (HPT) was a multicenter, randomized trial testing calorie control, sodium restriction, and potassium increases in the prevention of hypertension in 841 men and women. Thirty four variables that were potentially related to changes in urine sodium, urine potassium, and weight were examined individually and together across 3 years of treatment and maintenance sessions. Univariate and multivariate analysis did not reveal a consistent pattern of variables associated with successful attainment of treatment goals or failure to do so. Baseline levels of urine sodium, urine potassium, and weight were the variables most consistently associated with compliance. Household composition, such as number of people living in the house or marital status and control over selection and preparation of foods, was associated with compliance in sodium reduction treatments. Number of complaints about the diets were positively associated with noncompliance. Attendance at treatment sessions was generally associated with compliance especially for weight loss. Practical implications of these results are discussed.

High blood pressure is a pervasive health problem, affecting over 24 million Americans (National Center for Health Statistics, 1985), and the number of people under treatment has risen significantly over the last decade (Lenfant & Roccella, 1984). High blood pressure has long been associated with serious cardiovascular complications such as stroke, congestive heart failure, and kidney disease (Kannel, Castellis, McNamara, McKee, & Feinleib, 1972; Kannel, Wolf, Verter, & McNamara, 1970; Moyer, Heider, Pevey, & Ford, 1978) and is considered one of the three major risk factors for premature mortality (Kannel & Gordon, 1974; Pooling Project Research Group, 1978). Pharmacologic treatment has proven effective for those with severe as well as “mild” hypertension (cf. Strasser & Ganten, 1987). The use of antihypertensive drugs is, however, not without problems, including known and unknown side effects and significant cost to the consumer. With the move to treat many more people, including those with blood pressure levels previously considered normal (80-90 mm Hg), interest in the development of nonpharmacologic methods for treatment and prevention of hypertension has increased. Dietary management is seen by many to offer the best potential for nonpharmacologic management and prevention (Tillotson, Winston, & Hall, 1984). Dietary control of sodium, potassium, and weight has been suggested as a “reasonable” initial approach to hypertension management (Joint National Committee on Detection, 1986; Kaplan, 1985).

The Hypertension Prevention Trial was supported by research grants (ROl HL 25192, ROl HL 25 194, ROl HL 25201, ROI HL 25202, ROI HL 26585, and ROl HL 26841) from the National Heart, Lung and Blood Institute, Bethesda, Maryland. Requests for reprints should be sent to Thomas L. Schmid, currently with the Community Health Promotion Branch, Mail Stop K 46, Centers for Disease Control, 1600 Clifton Road, Atlanta, GA 30333. 151

152

THOMAS

Table 1. Number of participants

L. SCHMID et al.

randomized conditions Treatment

Men Women

to each of the experimental

group

NA

NAK

NACAL

CAL

CTL

124 72

127 68

84 45

94 31

120 76

Notes. NA, Sodium Reduction Group; NAK, Sodium Reduction and Potassium Enhancement Group; NACAL, Sodium Reduction and Weight Reduction Group; CAL, Weight Reduction Group; CTL, Control Group.

Although prevention enjoys advantages over treatment, most current prevention protocols share many of the problems associated with medical procedures. For example, it is difficult to determine who will be attracted to these programs, who will participate fully, and who will benefit from participation. Such difficulties are commonly discussed under the rubric of compliance and often include differential outcomes in relation to patient-related characteristics, as well as behaviors and responses made during treatment. In dietary treatments, compliance by the participant often requires more complex behaviors than in medical management (German, 1988). Maintaining a sodium-restricted or potassium-enchanced diet, for example, entails successfully making food selections in a marketplace in which 7,000 new items are added each year. Factors related to dietary compliance may thus be different from those predicting medical compliance, and learning more about these factors may be useful in guiding the development of effective nonpharmacologic therapy. The purpose of this study was to examine a variety of behavioral and demographic variables individually and in combination, to learn more about what factors may help discriminate between participants who will meet dietary treatment goals from those who do not, in the context of hypertension prevention. It was hoped that factors might be identified that might help in determining who would be the best candidates for a nonpharmacologic approach to hypertension and what barriers might exist for successful compliance. These variables come from two sources: information at baseline and information obtained during the intervention process. Baseline variables included physiologic variables such as baseline weight and baseline sodium excretion, demographic attributes such as marital status, knowledge about treatment goals, and relevant health behaviors. Although they are routinely recorded, research has not generally shown a strong relationship between demographic variables and outcomes. For instance, Glanz (1980) found no consistent relationship with demographic variables and dietary compliance. Evers, Bass, Donner, and McWhitney (1987) found that men and those with lower levels of education did better at salt avoidance. Glasgow, McCaul, and Schafer (1986) found no association with age, but women reported more barriers for adherence to diabetic regimens. Webb et al. (1984) found that dietary complaince could not be predicted from “demographic characteristics, duration of disease, knowledge or beliefs.” The last group of variables are based on behaviors that occurred during intervention and are referred to as process variables. Process variables such as attendance have been associated with treatment success. For instance, Jeffery et al. (1984) found that attendance at weight loss classes is an important variable in successful dieting. Timely and meaningful feedback has also been strongly associated with success in a variety of treatments including sodium restriction (Kaplan et al, 1982) and adherence to hypertension medicine regimens (Kirscht, Kirscht, & Rosenstock, 1981).

Dietary treatment goals in hypertension

1.53

METHODS

Data are from the Hypertensions Prevention Trial (HPT), a randomized clinical trial that tested four different dietary management procedures. As shown in Table 1, 841 free living adults were randomized into one of five groups, including a no-treatment control and four dietary interventions. Dietary interventions included the following: sodium restriction (NA), sodium restriction and potassium enhancement combined (NAK), weight reduction (CAL), and a combined weight reduction and sodium restriction group (NACAL). Subjects were stratified by weight. Men greater than 110% and women greater than 105% ideal weight according to the 1959 Metropolitan tables were assigned to the high-weight strata (Metropolitan Life Insurance Co., 1959). Only high-weight strata subjects were randomized to the treatments that included weight reduction. Both high- and normal-weight strata subjects were randomized to the other treatments. Participants were men and women ages 2549 with diastolic blood pressure between 78 and 90 mm Hg. They were predominately male (65.3%), white (82.2%), well educated (97% high-school graduates, 53.7% college graduates), and young (mean age 38.6 years). The majority (66%) had a family history of hypertension, and 16.6% were cigarette smokers. Treatment groups were equivalent in terms of age, education, race, weight, living with a partner, and family history of hypertension. Interventionists were trained by the Nutritional Education Resource Center, which was responsible for developing the common treatment protocol. Training of interventionists included two 2-day sessions, including orientation to the HPT goals and guided practice through the HPT materials. Intervention sessions encouraged participation by subjects and their families and included topics such as behavior change methods, food selection, ordering in restaurants, assertiveness, unlearning food cues, and self-monitoring. Groups met for at least an hour each week for the first 10 weeks (intensive intervention--E) and every other month thereafter (Maintenance). Measurement

and analysis

Participants in the HPT collected overnight urine samples and were weighed at semiannual follow-up examinations (FU l-FU 6) at 6-month intervals throughout the 3-year program. Overnight urine collections were performed on randomly assigned days of the week. This was because urine sodium and potassium levels are very sensitive to daily food consumption, and consumption patterns vary considerably, especially between weekdays and weekends. Treatment outcomes are reported in terms of weight (lbs) and 8- and 24-hr urine concentrations of sodium (r&q) and potassium (mEq). See Meinert, Borhani, and Langford (1989) for full details of the experimental design, procedures, and baseline results of HPT. For each outcome and independent variable, a model was fit containing four terms: baseline value of the outcome, the independent variable, treatment group, and the interaction between the independent variable and treatment group. No evidence of interaction was observed, and, thus, unified analyses over treatment groups were done rather than conducting analyses separately with each treatment group. Analysis, therefore, is based on combined average follow-up measures of 24-hr urine sodium excretions in the NA, NAK, and NACAL treatments, average urine potassium excretion in the NAK group and combined average weights in the NACAL and CAL groups. A review of the literature reveals that definitions of Compliance are controversial and unresolved (cf. Epstein, 1984; German, 1988). For the purposes of this article, we use progress toward or attainment of dietary treatment goals as indices of compliance. As a first level of exploratory analysis, univariate and multiple linear regressions were conducted comparing 34 baseline and process variables with urine sodium excretion, urine

154

THOMAS

Table 2. Demographic

and lifestyle predictor variables of mean follow-up urine sodium excretion, potassium excretion, and weight from univariate regressions

Variable 1. 2. 3. 4. 5. 6.

7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20 21. 22.

L. SCHMID et al.

Age (years) Sex (I = male, 2 = female) Married (1 = yes, 2 = no) Race (1 = white, 2 = nonwhite) No. in household Income $25.999-unknown Job category (1 = prof, 2 = nonprofl Education (years) BP taken in last year (1 = yes, 2 = no) Ever smoked (1 = yes, 2 = no) Ever drank alcohol (1 = yes, 2 = no) Family history of hypertension (1 = yes, 2 = no) Hospital in last 5 years (1 = yes. 2 = no) Baseline weight (lb) No. meals away from home Meal preparer (1 = self alone, 2 = other) Vitamins (1 = yes, 2 = no) Use salt in cooking (I = yes, 2 = no) Use salt at table (1 = yes. 2 = no) Use salt substitute (I = yes. 2 = no) Regular exercise (1 = yes. 2 = no) Last time bp was taken 24 months 7-12->24 months 13-24->24 months

Urine sodium (mEq/E hr) coeff P -0.01 -2.20 1.40 -0.09 0.65

0.944 0.109 0.501 0.960 0.161 0.822

Urine potassium (mEq/E hr) coeff P

coeff

P

-0.04 -3.72 -0.41 - 1.66 -0.36

-0.22 1.56 3.25 4.78 0.34

0.043 0.428 0.086 0.014 0.506 0.598

0.501 10.001 0.686 0.103 0.206 0.103

3.68 4.29

2.75 2.50

urine

Weight (lb)

-0.50 -2.18

2.10 -0.87

0.171 0.014

- 1.50 0.37

0.092 0.084

-2.22 0.05

0.193 0.904

-2.53 -0.12

0.132 0.929

0.08 -0.74

0.931 0.317

-0.09 -0.71

0.963 0.647

0.633

-0.41

0.604

-0.40

0.799

0.134

- I .71

0.028

-0.98

0.530

1.14 0.13 0.05

0.501