A latent class modeling approach to evaluate behavioral risk factors and health-related quality of life

The Behavioral Risk Factor Surveillance System (BRFSS) monitors multiple health indicators related to 4 domains: risky behaviors, health conditions, health care access, and use of preventive services. When evaluating the effect of these indicators on health-related quality of life (HRQOL), conventional analytical methods focus only on individual risks and thus are not ideally suited for analyzing complex relationships among many health indicators. The objectives of this study were to 1) summarize and group multiple related health indicators within a health domain by using latent class modeling and 2) analyze how 24 health indicators in 4 health domains were associated with 2 HRQOL outcomes to identify Rhode Island adult populations at highest risk for poor HRQOL.

The 2008 Rhode Island BRFSS, a population-based, random-digit-dialed telephone survey, collected responses from 4,786 adults aged 18 years or older. We used latent class modeling to assign 24 health indicators to high-, intermediate-, and low-risk groups within 4 domains. The effects of all risks on HRQOL were then assessed with logistic regression modeling.

The latent class model with 3 classes fitted the 4 domains best. Respondents with more health conditions and limited health care access were more likely to have frequent physical distress. Those with more health conditions, risky behaviors, and limited health care access were more likely to have frequent mental distress. Use of preventive health services did not affect risk for frequent physical or mental distress.

The latent class modeling approach can be applied to identifying high-risk subpopulations in Rhode Island for which interventions may have the most substantial effect on HRQOL.