Do Adults Change Their Lifestyle Behaviors after a Chronic Disease Diagnosis?

Background and Purpose: The health belief model emphasizes the importance of cues in motivating behavior change. A specific type of cue, labeled as “teachable moment” (TM), describes a naturally occurring health or life event that could motivate individuals to spontaneously engage in risk-reducing health behaviors. While chronic disease is a major threat to quality of life, disease diagnosis could serve as a critical TM for patients to initiate behavior change. However, relevant empirical evidence is limited. This study aimed to examine whether people are more likely to initiate health behavior change including smoking cessation, reduction in drinking, and becoming physically active after a diagnosis of diabetes, heart disease, chronic lung disease, and cancer, respectively.

Methods: Data came from 1996 to 2010 waves of the Health and Retirement Study (HRS). Study sample consisted of 11,439 participants aged 50 through 80 who were free of chronic diseases at baseline. Chronic disease diagnoses were self-reports of physical diagnoses. Current smoking was ascertained from answers of “yes” to “do you currently smoker?” Drinking level was classified into two categories (moderate vs. excessive) based on the 2010 Dietary Guidelines for Americans. Physically active was defined as engaging in vigorous physical activity three times or more per week. Matched case control difference-in-differences (DD) estimator was performed to investigate the impact of chronic disease diagnosis on the likelihood of behavior change. First, individuals with the focal chronic disease diagnosis (the “treatment group”) were matched to those free of chronic diseases throughout the study (the “comparison group”) using nearest-neighbor matching with Mahalanobis distance. The independent effect of a chronic disease can then be obtained by netting out the secular trends (i.e., behavior changes observed among individuals free of the chronic disease) from changes observed among those with the chronic disease. The model estimated can be written as follows, log(Y_jgt) = ρDiagnos_jt + λTime_t + β(Diagnos_jt * Time_t) + δZ_j + ε_jgt  where β is the DD estimator. To facilitate the interpretation of interaction terms in logistic regression, relevant post-estimation procedure in Stata (margins command) was used.

Results: Individuals were less likely to smoke after a disease diagnosis, with the largest effect reported for chronic lung disease (OR=.61, or -.067 in probability, suggesting 67 per 1000 persons quit smoking), cancer (OR=.50, or -.064 in probability), diabetes (OR=.61, or -.047 in probability), and heart disease (OR=.72, or -.031 in probability). Similarly, individuals were less likely to report excessive drinking, with the largest effect reported for chronic lung disease (OR=.65, or -.035 in probability), followed by cancer (OR=.72, or -.030 in probability), diabetes (OR=.72, or -.026 in probability), and heart disease (OR=.74, or -.025 in probability). However, 97 per 1000 persons with cancer, and 79 per 1000 persons with chronic lung disease decreased vigorous physical activity after diagnosis.

Conclusions and Implications: Chronic disease diagnosis may be an important TM for improvement in multiple health behaviors simultaneously. Health promotion programs in midlife and older adulthood could improve effectiveness if targeted at those with a newly diagnosed chronic disease.