Baylor University Medical Center Proceedings April 2014, Volume 27, Number 2 | Seite 22
A cohort analysis of the cardiovascular risk factors in the
employees of a pediatric hospital from 2009 to 2012
Lilly Ramphal, MD, MPH, Jun Zhang, MS, and Sumihiro Suzuki, PhD
A retrospective longitudinal cohort regression analysis was completed in
853 of the 3435 employees of Cook Children’s Hospital who participated
all 4 years (2009 to 2012) in an employer wellness program. The presence of the metabolic syndrome (MS) was used as an outcome measure
for the success of the wellness program. Data were stratified by weight,
gender, and ethnicity. The odds ratios and regression analysis showed
a significant decline in MS over the 4 years of the study (P = 0.008), as
well as a significant association between MS and obesity and overweight
status (P < 0.0001), male gender (P = 0.0018), and all ethnic categories
(P < 0.05) except African American ethnicity and the multiple ethnicity
category. Age was strongly associated with risk for MS. Overall, the study
showed that the wellness program significantly decreased the incidence
of MS (P < 0.05).
he metabolic syndrome (MS) is a serious public health
concern defined by interconnected factors that directly
impact the risk of coronary heart disease. MS has been
defined by the International Diabetes Federation (IDF)
and the National Cholesterol Education Program (3). Both
groups present similar criteria for MS, but for purposes of this
study the IDF definition, which put more emphasis on waist
circumference, was used. Waist circumference, independent
of other parameters of MS, is the most significant predictor of
cardiovascular risk (1–3). This study identified trends in the
prevalence of MS in 853 employees who consistently participated in a hospital wellness program from 2009 to 2012.
T
METHODS
The study was approved by the Cook Children’s Hospital
institutional review board. A retrospective longitudinal cohort
regression analysis was completed on 853 (24.8%) volunteers
who had participated in all 4 years of a hospital wellness program
from 2009 to 2012 and had submitted complete demographic
records. A total of 2582 employees were excluded from the
study because they participated inconsistently during the study
period. Participants in the wellness program were incentivized
with an annual monetary award to reduce their cardiovascular
risk by reducing four or more of the measured criteria in 2
consecutive years. Each participant’s demographic information
was obtained from survey health questionnaires collected by the
96
employer’s wellness coordinator. Measured parameters such as
weight, waist circumference, height, and systolic and diastolic
blood pressures were obtained annually by registered nurses. No
formal instruction on how to reduce these parameters was given
to participants. They were referred to their personal doctors for
direction or sought out resources to improve their health.
All data were analyzed by statisticians using SAS. Two
models were used to analyze the variables affecting MS risk:
a conventional logistic regression and a longitudinal generalized estimated equation (GEE) model. Model 1 for the output
analysis is shown below:
Logit[P(MS = 1)] = β0 + β1 ∗ √t + β3 ∗ WeightStatus + β4
∗ Gender + β5 ∗ Ethnicity
Odds ratio estimates were used to evaluate risk. Both analyses
molded in risk factor variables including square root of time,
gender, weight status (normal, overweight, and obese), ethnicities (American Indian, Asian, Black, Hispanic other, Hispanic
White, Multiple, and White), and age groups (18–30, 30–40,
40–50, 50–65, and 65+). In both regression models, 2009 was
the baseline. Normal weight was used as the baseline and compared with overweight and obese categories. The underweight
were excluded due to the very small numbers in this category.
Caucasian (White) ethnicity was set as the baseline and compared to Asian, African American (Black), Hispanic Other, Hispanic White, American Indian, and Pacific Islander ethnicities.
For age, the 18–30 age group was used as the reference group,
and all other age categories were compared to it. All analyses
were performed using SAS (V9.2).
R ESULTS
Table 1 presents the demographic characteristics of the 853
participants who were eligible for longitudinal study. More
From Cook Children’s Hospital, Fort Worth, Texas (Ramphal); and the Departments
of Environmental Health (Ramphal) and Biostatistics (Zhang, Suzuki), the University
of North Texas School of Public Health. Dr. Ramphal is now with Blue Cross Blue
Shield.
Corresponding author: Lilly Ramphal, MD, MPH, Department of Environmental
Health, University of North Texas School of Public Health, 3500 Camp Bowie
Boulevard, Fort Worth, TX 76107-2699 (e-mail: [email protected]).
Proc (Bayl Univ Med Cent) 2014;27(2):96–99