The prevalence of overweight and obesity in Texas is among the highest in the United States, where overweight and obesity rates have increased steadily in the past two decades (5-7). Sixty-four percent of adult Texans are overweight or obese, and more than 39% of fourth-graders, 37% of eighth-graders, and 29% of eleventh-graders are overweight or at risk for overweight (5). Obesity is more prevalent among women, African Americans, Hispanic Americans, and people with low education levels and income. Many racial and ethnic groups continue to face disproportionate risks for obesity-related chronic diseases and associated decreased life expectancy (8,9).

Recent speculation attributes the increasing obesity prevalence among low-income populations to socioeconomic status, personal choices, established cultural and family patterns, and environmental factors (10-13). Low-income communities are particularly vulnerable because nutritious foods can be expensive or difficult to find, whereas less healthy, energy-dense foods are readily accessible and affordable (11). Compared with higher income neighborhoods, poor neighborhoods may lack well-maintained sidewalks and streets, safe outdoor spaces, or exercise facilities and therefore be less conducive to physical activity (14).

Parents are primary sources of information for children who are learning about healthy eating practices, so family environments are important intervention settings for establishing lifelong dietary habits (14-16). Studies investigating familial aggregation of obesity show that family eating environments link parental adiposity and dietary intake with children's adiposity and dietary intake (17). 

A request for proposals to address overweight and obesity from United Way of Metropolitan Tarrant County prompted university faculty and community organizations to collaborate on designing and implementing Healthy Weigh. As phase 1 was ending, development of a participant leadership group led to more balanced partnerships among participants, community organizations, and investigators, and Healthy Weigh evolved into a community-based participatory research effort. Texas Christian University Human Subjects Institutional Review Board approved the Healthy Weigh program. The previous sidebar summarizes the Healthy Weigh interventions; participants helped design and evaluate phase 2.

Cornerstone Community Center (CCC), a faith-based community organization, provided the physical facility for Healthy Weigh, assisted with weekly reminder calls to participants, and was the primary source of participant referrals and community volunteers. CCC serves a community at high risk for overweight and obesity. Factors for overweight and obesity included the median household income ($20,000), federal poverty level (44.6% below the poverty level), and people older than 25 years lacking high school diplomas (42%) (18). The population in CCC's catchment area is 35% African American and 26% Hispanic (18). 

Phase 1 was offered twice in 2003 (summer and fall) and consisted of 12 weekly sessions, prescreening, and postscreening. Participants chose to enroll in the English- or Spanish-language version of phase 1. Community focus groups provided input on program design, evaluation, and recruitment. Recruitment flyers were distributed door-to-door in target neighborhoods and through schools and CCC.

Of 282 people screened for phase 1, most participants were female (72%) and Hispanic (82%) or African American (12%). Many of the participants were families; 46% of participants were younger than 19 years,  and 9% were older than 60 years. Many participants had limited proficiency in English or Spanish. Based on body mass index, 84% of adults were overweight or obese, and 50% of children and adolescents were overweight or at risk for overweight.

The research design included evaluation of program process objectives and participant outcome objectives based on Healthy People 2010 (1), Recommendations to Increase Physical Activity in Communities (19), and Dietary Guidelines for Americans (20). Following are the outcome objectives (21) and program process objectives for Healthy Weigh Phase 1:

Outcome objectives

Process objectives

Program evaluation was determined by interview, direct measurement, and self-report (Table 1).

The university research team (two registered nurses and one registered dietitian) directed program implementation, which was carried out by registered nurses, a registered dietitian, paid staff, community volunteers, and supervised students from nursing, nutritional sciences, kinesiology, social work, and medicine programs. More than 160 undergraduate and graduate students from two universities served as educators, exercise leaders, meal preparation and service coordinators, table-talk leaders, child care workers, health screeners, and research assistants. Sixty percent of the students were volunteers, and most of the remaining students participated to meet course requirements. Many program participants also served as volunteers by helping with set up, clean up, and interpretation.

Seven framework utility standards address information needs of evaluation users, who were defined as the funding agency and investigators in phase 1. Ways in which utility standards were met included the following:

There were several ways in which the evaluation could have been improved to meet utility standards (Table 2). Because other community stakeholders (CCC, program participants, and staff) were not identified as evaluation users and did not help design the evaluation plan, their information needs were unknown and not addressed. Monitoring the balance of needs is essential to developing effective community-based participatory research partnerships (22). Establishing evaluation criteria with partners could have promoted ongoing monitoring of relationships, frequent acknowledgment of strengths and growth, and rapid response to actual and potential problems.

Three framework feasibility standards were developed to ensure effective and practical use of program resources. Most phase 1 evaluation procedures were practical, adapted for the community setting, and politically viable, and they effectively used time and resources. Use of familiar equipment (e.g., scales, stadiometer, measuring tape, sphygmomanometer) and noninvasive methods for data collection fostered participant acceptance of measurements and made it easier for students to help conduct screenings. Concerted efforts to prevent phase 1 activities from interfering with CCC's daily operations and early negotiation for using equipment and space helped gain and sustain CCC's cooperation in evaluation procedures. Proactive communication with CCC's staff and university, funding agency, and community–partner organizations also contributed to political viability. Ongoing budget management based on financial requirements of the program, university, and funding partner helped ensure that resources were used effectively to achieve program processes and outcomes.

Structure–process–outcome analysis highlighted ways the evaluation process could have better met the feasibility standards related to use of resources and political viability (Tables 2–4). Investigators frequently filled staff roles because of a shortage of staff members, lack of availability of staff members, or staff members' lack of skills. Resolving staff structure issues would have permitted more consistent staff assignments and potentially increased skills and job satisfaction. Greater attention to staff structure would have contributed to more efficient and effective use of resources for participant screening and overall program administration.

Two process variables also affected feasibility. Knowing more about why participants attended some sessions and not others could have promoted more cost-effective use of human and material resources needed for program physical activity and nutrition classes, family meals, and food-bag preparation. Systematically collecting data about participants who dropped out of the program might have helped improve retention and program design. Early clarification and monitoring of community partners' roles would have provided a framework for increased collaboration and communication among partners, making the evaluation process more politically viable.

Eight framework propriety standards address legal and ethical characteristics and how the evaluation process affects all the participants involved. Although the lack of formal agreements among all key parties involved in the evaluation was a weakness, overall the evaluation process met propriety standards. Concern for participant welfare was evident in design of evaluation tools and procedures, the willingness to adapt the research design to meet participant needs, the commitment to building relationships with participants and CCC staff, and adherence to financial accounting standards.

Informed consent and other screening tools were appropriate for participants with limited English proficiency. Documents were available in English and Spanish and were prepared at the fourth- to fifth-grade reading level in English, with reader-friendly formatting and adequate white space (23). Because one consent form per family was used, parents of large families and people with limited literacy did not have to handle multiple documents. Family members could choose to read the informed consent or have it read to them. Child care for preschool-aged children allowed families to concentrate on completing the informed consent.  

Participant privacy was maintained at levels consistent with community expectations. Staff members explained screening results and gave families copies of results in appropriate languages. Referrals for follow-up health care were based on prescreening findings. Efforts were made to obtain health care for participants who lacked regular sources of care.

When formative evaluation showed that participants were unfamiliar with setting goals, we modified the screening protocol to include teaching participants about setting appropriate goals for healthy eating, physical activity, or weight management. Fifty summer participants asked to enroll in the fall program, so we revised the research design. Relationships with participants and their progress toward improved physical activity and healthy eating patterns were more important to overall program goals than strict adherence to the original research design.

All financial reporting requirements of the university and United Way were met. Maintaining and properly cleaning the CCC facility, equipment, and supplies during and after each Healthy Weigh meeting also addressed fiscal responsibility standards.

Twelve framework accuracy standards assess correctness of evaluation findings. The evaluation process for phase 1 grant objectives (see previous lists of outcome and process objectives) met these standards.

Structure–process–outcome analysis revealed how the phase 1 evaluation process could have measured the findings more completely and accurately (Tables 2–4). Collection of qualitative data was not as consistent and systematic as planned because of human resource problems. For example, although many table-talk conversations were recorded in log books, different staff members facilitated and documented the talks each week. Some weeks, too few staff members were available to document all talks, or available staff members lacked skills to document Spanish-language table talks. These factors contributed to uneven and inconsistent qualitative data collection. Most questions posed for improving the phase 1 evaluation process could have been answered by consistent, systematic collection and analysis of field observations, personal narratives, and table-talk conversations. Systematic identification, documentation, and analysis of best practices would have accomplished the following: