Risk Factor Trajectory and Variability: Predictors of Vascular Disease?
Dawn C Schwenke PhD MS
Phoenix
Phoenix , AZ
Funding Period: October 2003 - December 2005

BACKGROUND/RATIONALE:
The prevalence of diabetes among veterans is high and increasing. Obesity increases risk of type 2 diabetes, suggesting that nationwide trends for increasing prevalence of obesity will further increase the incidence of type 2 diabetes. The high prevalence of type 2 diabetes among veterans and potential for a rapid increase in incidence of type 2 diabetes is of grave concern because type 2 diabetes increases risk of cardiovascular and microvascular diseases. These diseases account for significant morbidity, and cardiovascular diseases are the single leading cause of death among Americans. While factors that increase risk of cardiovascular and microvascular disease have been identified, average levels of these risk factors can only partly explain differences in risk among persons with type 2 diabetes. Understanding reason(s) for differences in cardiovascular and microvascular disease risk among persons with type 2 diabetes is critically important to veterans due to the high prevalence of type 2 diabetes among veterans and the substantial morbidity and mortality caused by vascular diseases.

OBJECTIVE(S):
Our objective is to determine whether consideration of temporal trends for risk factors in addition to average values of these risk factors improves the ability to predict risk of clinical manifestations of cardiovascular and microvascular disease in veterans with type 2 diabetes. As our long-term goal is to develop a cost-effective method for identifying veterans with type 2 diabetes at highest risk for cardiovascular and microvascular disease, we are including in our explanatory models measures shown to be associated with increased cardiovascular and/or microvascular risk that are relatively inexpensive and commonly measured during standard clinical care. Thus, we are considering as risk factors body mass index, as an index of relative obesity; white blood count as an index of inflammation/infection, serum albumin and blood hemoglobin as measures of nutritional status, blood pressure, blood glucose, plasma cholesterol and triglyceride, and LDL and HDL cholesterol.

METHODS:
This project is a longitudinal observational study that will make use of clinical data collected during the course of standard clinical care of veterans with type 2 diabetes at the Carl T. Hayden VA Medical Center. We will use Cox proportional hazards models to model risk of macrovascular and microvascular events as a function of a risk factors for cardiovascular disease. A series of sequential models will be considered so that we can identify the contribution of trajectory and variability in risk factors independent of average values of these risk factors.

FINDINGS/RESULTS:
Since the time of the last report, the computer disks needed for this study have been installed. IRMD has pulled from VISTA files and loaded into the database all of the data needed for this project that could be gotten from VISTA files. As the various VISTA files were loaded into the database, we continued development and testing of methods for identifying errors in the clinical data. We have also worked extensively with the coders to refine our protocols for identifying macrovascular and microvascluar events as well as comorbid conditions. Vital signs (including weight and height, needed to calculate BMI) were among the last data that were loaded into the database. When the last data for vital signs had been loaded (June 2005), we discovered that while all of the other data elements needed for this project were available for the 12 year time period that is needed, data for vital signs was severely deficient, with near complete data only for the last 2.5 years and limited data for the prior 5 years. We had not anticipated this, and had not requested funds for abstraction of vital signs from paper charts. Since that time, we have divided our efforts between exploring ways to accommodate this largely missing and critically important data (present in all of the sequential models we will consider) and writing grant proposals to support abstraction of height and weight from paper charts.

IMPACT:
Health care costs for persons with diabetes or cardiovascular disease are high, and these costs are greater for persons with both diseases. Health care resources are limited. If this project is successful, it may permit identification of veterans at greatest risk of cardiovascular and microvascluar disease so that expensive screening, monitoring, and preventive efforts can be targeted to these persons in greatest need. Focusing these efforts where they are most needed should improve patient health and may reduce costs by preventing cardiovascular events requiring expensive interventions. Because diabetes and cardiovascular disease are the most important cause of mortality in Americans, including veterans, the results should also be useful for civilians.

PUBLICATIONS:
None at this time.

For most recent citations (updated monthly):

DRA: Chronic Diseases
DRE: Etiology
Keywords: Cardiovasc’r disease, Diabetes, Risk factors
MeSH Terms: