Tu Salud Si Cuenta Worksite: Pilot Intervention Study

Obesity has not only been increasing over the past three decades (Wang & Beydoun, 2007; Wang
& Zhang, 2002), it is predicted to increase in the future until by 2048 all American adults
would be overweight (Wang, Beydoun, Liang, Caballero, Kumanyika, 2008). Moreover, the
majority of Mexican American men and women are predicted to be overweight and obese by 2030
(91.1% and 86.7% respectively).

The proposed intervention research study is building from existing intervention strategies
designed for the Mexican American population entitled Tu Salud Si Cuenta. This intervention
highlights in the media successful role models who have changed their food choices and
physical activity levels. In addition the intervention designs environmental changes to help
community members carry out recommended behavior change strategies. Outcomes of interest for
this study include increased physical activity so as to achieve the CDC/ACSM recommendations
of 150 minutes of moderate to vigorous physical activity each week (Haskell et al., 2007),
eating 5 servings of fruits and vegetables each day, reducing food portion sizes, and
subsequent changes in biomarkers of health.

This proposed worksite pilot study will be testing a new feature to the TSSC intervention by
designing worksite promotion intervention strategies to promote physical activity and
healthful food choices through environmental change strategies at the worksite. The same
full scope of behavioral and clinical outcomes of interest will be examined in this pilot.

The participants in this study will be male and female employees of two worksites, aged
18-65 years. A convenience sample will be used for this study. Employees from State
Department of Public Health Region 11 Harlingen Office or City of Brownsville will be
recruited. Supervisors from both sites have provided letters of support for this project and
have agreed to allow recruitment. Employees will be contacted with informational flyers and
e-mails sent by the supervisors providing the letters of support. Interested adults will be
sent an interest letter and informed consent. Data collection will proceed after the
participant completes the informed consent form and provides documentation of physician's
clearance to participate in physical activity.

The purpose of this study is to test the efficacy of a worksite-based physical activity and
healthful food choice, intervention among two worksites in the Lower Rio Grande Valley who
employ primarily Mexican American personnel.

Persons will complete consent forms during an enrollment/pretest period of 4 weeks. This
period will include an electronic pre-test survey using Survey Monkey online software
(Finley, 2008) completed by each participant. Additionally, clinical measures such as blood
pressure, anthropometrics, fasting total cholesterol, triglycerides and glucose will be
collected for each participant. These exams will be conducted by the trained staff from the
clinical research unit and follow the identical procedures used for members of the Hispanic
Health Research Cohort. Finally, in line with the Centers for Disease Control and
Prevention, each participant will be encouraged to visit their personal physician and
discuss participation in the recommended 150 minutes of physical activity per week.

All individuals are given an exit interview with a registered phlebotomist or registered
nurse to review immediate findings. All samples for laboratory testing and storage will be
coded and identifiers removed. Participants will also provided with a written summary
including their height, weight, waist circumference, blood pressure, fasting glucose,
cholesterol and triglycerides. Values out of their expected range are flagged and referrals
made as necessary. This interview is also used to address any questions or concerns that may
be raised.

Descriptive analyses will be used to compare the characteristics of the participants in the
intervention and control groups that we intend to compare with respect to any outcome of
interests, (e.g., blood pressure). Specifically, we will compare the groups with respect to
demographic and socio-economic factors such as age, sex, ethnicity, and education level.
Since the outcome measurements at baseline and the follow up quarterly visit (at 3 months, 6
months, etc) are expected to be correlated, we will use Generalized Estimating Equations
(GEE) models. GEE takes into account the correlation of an outcome measured repeatedly over
time by treating it as a nuisance parameter (Zeger & Liang, 1986). As a result, the
within-subjects correlation is not modeled explicitly. GEE uses a "working correlation" to
treat the within-subjects correlation, and produces standard error estimates that take into
account the correlation of responses with subjects. In addition, we will use Random effects
models (Wolfinger & O'Connell, 1993) to adjust for the heterogeneity of subjects, which can
be due to unmeasured predispositions, such as genetic factors, that account for the
correlation within subjects across time. Estimated effects of a covariate of interest (e.g.,
males vs. females) are adjusted for these individual differences and model estimates should
be interpreted in terms of the change due to the covariate in an individual at a given level
of random subject effect. The random effects model is most useful when inferences about
individual differences are the primary interest. Random effects models have been developed
for both continuous and binary outcome variables (Wolfinger et al., 1993).

The study will take place over a 1 year time frame. Participants will have data collected
prior to the intervention and four posttests at 3-month intervals. The immediate
intervention group will receive the pilot worksite intervention for the initial 12 week
period. The delayed intervention group will receive the pilot worksite intervention
beginning in month 3 of the study.