This report describes recent research at the University of Minnesota on community intervention strategies for obesity reduction. This research has sought innovative ways to reach persons in the community using low-cost programs that result in large-scale changes in diet and exercise behaviors and that ultimately reduce obesity. The programs described involve worksite intervention, intervention by direct mail, and a multifactorial community approach. The Minnesota experience and related research is discussed in terms of three specific questions: 1) How successful are these programs in reaching persons in need? 2) How effective are they in achieving and sustaining weight loss in program participants? and 3) Do they have a favorable effect on the prevalence of obesity in the general population?

Worksite Intervention: The Healthy Worker Project

One community intervention strategy that has recently attracted much attention is worksite health promotion, an approach whose appeal depends on various factors including ease of access to a large audience, the potential for social support, and the potential for cost recovery through lower health care expenses [4]. The Healthy Worker Project was a randomized trial of such an intervention, focusing on weight loss and smoking cessation [5]. Conducted in the Minneapolis-St. Paul, Minnesota, metropolitan area between 1987 and 1990, it involved 32 worksites with an average of 600 employees each. A random sample of 200 employees at each site was surveyed at baseline. Participants were then randomized by site to receive either no treatment or 2 years of intervention consisting of on-site health education classes and an incentive system organized through payroll deduction. At the end of 2 years, the 200 employees assessed at baseline were resurveyed, as were those in a second cross-sectional sample. Primary outcome analysis used body mass index (BMI) as the dependent variable and company as the unit of analysis.

In the 2-year intervention, 1576 persons enrolled in the weight-loss program. This represented approximately 16% of all employees in the treatment firms and 36% of all overweight employees. Women were more likely than men to participate in the program, and blue-collar employees participated at lower rates than did white-collar or clerical employees, although all groups were well represented (Lando HA. Unpublished data).

Weight loss outcomes for program participants were positive, showing a mean loss of 2.2 kg in 6 months. The outcome at 2 years, however, was less impressive. Although mean change in BMI was positively associated with participation in the program (that is, employees at worksites with high participation lost more weight compared with those at worksites with low participation), no net difference was seen between treatment and control worksites. The same project, however, showed a much greater effect on smoking behavior, yielding a significant reduction in prevalence.

The Healthy Worker Project illustrates several points characterizing most worksite programs. The first is that worksite programs can generate high levels of participation. Stunkard and colleagues [6], for example, evaluated an intervention based on competition between worksites or between teams within worksites. In 15 worksites, 20% of the total workforce (1177 employees) participated [6]. Similarly, Erfurt and colleagues [7] recently reported a 60% to 70% participation rate among overweight employees using an intervention consisting of risk-factor screening and a multimodal follow-up strategy. A second point illustrated by the Healthy Worker Project is that weight losses among participants in worksite programs are positive. They are often less in absolute terms than those observed in clinical settings, but this may be due in part to the fact that persons referred to clinical programs are more obese and may be more highly motivated to lose weight [8]. In terms of cost-effectiveness, worksite programs are superior to clinic-based strategies in the short term [9]. Furthermore, although the average weight loss usually achieved in such programs is of questionable value to someone who is clinically obese, important population-wide health benefits could be expected if mean weight could be reduced by this amount (2 kg).

The long-term efficacy of worksite interventions has recently been examined by three studies, one with and two without sustained intervention throughout the study period. Both studies without sustained intervention concluded that worksite programs have difficulty in maintaining weight loss among participants [6, 10]. The program that included a sustained intervention, however, reported a more favorable long-term outcome (mean loss, 2.2 kg after 3 years) [7]. The Healthy Worker Project is the only worksite study ever to evaluate the effect of this type of intervention on the prevalence of obesity in the worksite population as a whole. Although the program was not unsuccessful in achieving this goal, the observation of a dose-response relation between participation and weight change suggests that if participation were enhanced, a more positive result could be attained. Additional studies in this area are needed.

Direct Mail Intervention: Invest in Your Health

A second method of community-based intervention that has even greater reach than worksite health promotion is intervention by direct mail. Although such interventions lack the social support and personal contact of a worksite or clinic-based weight control program, they have the compensating advantages of reaching an audience that includes persons who do not work outside the home and those who are unwilling or unable to make regular clinic visits. The most extensive home correspondence weight-loss program conducted by our research group was used as a component of the obesity intervention efforts of the Minnesota Heart Health Program. This program, called Invest In Your Health, was implemented in three communities in Minnesota and North Dakota, with a total population of approximately 200 000 persons. It was formally evaluated in a suburban community of about 85 000 persons [11]. The Invest In Your Health Weight Loss Correspondence Program was offered by mail to 31 400 community households in two forms. One was a fee-for-service program with a $5 enrollment fee. The other was as a "free" program that required a $60 incentive deposit, which was refunded after successful weight loss. This 6-month program consisted of a self-help weight loss manual, monthly newsletters, and a return-address postcard attached to each newsletter on which participants were asked to record information about weight, diet, and exercise habits.

A total of 1304 persons signed up for this weight-loss program. The $5 fee-for-service version generated about five times the response rate of the incentive program (5.67% compared with 1.09% of households). Assuming that one third of households contained an overweight person, the more popular program attracted the interest of about 15% of households in need. Program participants were more educated than persons in the general population, suggesting that this type of program might not be effective in low-literacy groups. The program also generated a higher response rate from men (37%) and from older adults (33% were 55 years or older) than is characteristic of clinic-based programs. No demographic differences were noted between persons enrolling in the two types of correspondence programs.

The results of this program were evaluated by self-report questionnaire and telephone survey with validation by direct assessment in a subset of the sample. Mean weight change after 6 months averaged about 2 kg for the $5 fee-for-service program and 4 kg for the $60 incentive program. No data were collected on total population effects.

The results of the Invest In Your Health program suggest that direct-mail weight control programs deserve additional study. This study and others [8, 12-14] have shown that modest weight losses can be achieved using educational formats in which the most expensive part of intervention, professional contact, has been entirely removed. A noteworthy characteristic of this study was its attempt to examine the trade-offs in community intervention between recruitment and outcome. Programs with higher individual efficacy (in this case, the incentive program) may be less attractive to potential participants because of cost, convenience, or other considerations. Thus, selection of community-based strategies for weight control may involve evaluation of the relative benefits of small weight losses distributed among larger numbers of persons or larger weight losses concentrated in a smaller group.

Multicomponent Community Trials: The Minnesota Heart Health Program

A discussion of community-based interventions would be incomplete without some description of the outcome of multicomponent community trials. Several of these trials have been conducted to determine whether multifaceted community interventions aimed at health behaviors can reduce the community-wide incidence of cardiovascular or total mortality. These trials include the North Karelia trial in Finland [15], the Stanford three-community [16] and the Stanford five-community studies [17], the Minnesota Heart Health Program [18], and the Pawtucket Heart Health Program [19]. The Minnesota program was a 10-year research and demonstration trial conducted between 1980 and 1990. It involved three matched pairs of communities: two rural centers, two urban centers, and two suburban communities. One member of each pair was given multicomponent cardiovascular risk-reduction education over a period of 7 years. Community education programs included population-wide screening for cardiovascular risk factors, mass media education, risk-reduction classes through adult education systems in the community, worksite interventions, direct mail interventions, school- and church-based interventions, and grocery store and restaurant interventions. Although weight loss was not a primary target of intervention for this trial, it was strongly encouraged for persons with cardiovascular risk factors (that is, elevated cholesterol and blood pressure); avoidance of weight gain was encouraged for all.

The Minnesota Heart Health Program was successful in delivering educational programs. More than 60% of all adults in educated communities visited an education center where their weight was assessed and where they were given direct counseling and referral to community treatment services. More than 100 000 additional direct educational contacts were documented to be related to topics that might have a positive influence on weight (for example, reducing dietary fat and increasing exercise). Unfortunately, the Minnesota Heart Health Program did not show an overall treatment effect. Modest decreases in community-wide weight were seen for some years in the intervention communities, but they were not large or consistent enough to be statistically significant. The most striking result of this study was a substantial secular increase in body weight [3]. On average, persons in this cohort were 3 kg heavier in 1990 than in 1980. A comparison of the Minnesota Heart Health Program experience with that of other published weight data from large community trials suggests that this finding is typical. These trials have distributed intervention efforts over multiple health behaviors. Although some trials have reported positive results for change in blood pressure, cholesterol, smoking, or combined cardiovascular disease risk, none has yet to show convincingly a positive effect for obesity.

Summary and Recommendations

Community interventions for weight loss represent a new area for research. Current data suggest that approaches such as worksite and direct mail interventions can reach more obese persons than can traditional clinic approaches. They can also produce short-term changes in weight at a lower unit cost.

Although future research may take many directions, two that deserve special emphasis are recruitment and primary prevention. Recruitment is essential to the success of community-based interventions because obesity is so pervasive. Without widespread participation, a community effect cannot be achieved. Primary prevention is important because obesity is the end point of a cumulative process that occurs over an extended period and because much of the health damage of obesity may be done during this period of weight gain. It might be argued that community-based approaches, with effects that are modest in magnitude but broad in reach, may be better suited to the prevention of weight gain than to the treatment of obesity.