Abstract
Purpose :
Visual impairment (VI) and visual disorders often co-occur with other disabling chronic conditions. This research identifies chronic disease patterns and their relationship to VI in a nationally representative sample.
Methods :
Using data from 387,780 participants aged 18 years and older of the National Health Interview Survey (NHIS) 2002-2014, we employed latent class analysis (LCA) technique to place participants into subgroups with different combinations of thirteen self-reported chronic illnesses. LCA is data driven and the groups are derived from participants who are homogenous in their disease patterns. Nine LCA models ranging from two to ten classes were evaluated. Chronic illness class membership was then used as a predictor of self-reported VI status using logistic regression and controlling for covariates.
Results :
LCA identified 5 different groups with 70% of the participants belonging to the “healthy group”. The other four groups represented various degrees and patterns of multi-morbidity. The “hypertensive group” (19.6% of the total sample) had high prevalence of hypertension (63%), arthritis (48%), and diabetes (20%). The “respiratory conditions group” (4.4%) had high prevalence of emphysema (48%) and asthma (46%). The “heart disease group” (3.6%) had high prevalence of coronary heart disease (70%), stroke (15%) and other heart conditions (85%). The “severely impaired group” (1.8%) had much higher prevalence of hypertension (89%), other heart conditions (68%), arthritis (83%), emphysema (50%), diabetes (48%) and hearing impairment (54%).
After controlling for survey design, age, gender, race, Hispanic origin, education, marital status, insurance, income, BMI, smoking and alcohol drinking status, compared to the healthy group, participants in all four groups had elevated risk of reporting VI: heart disease group (OR 3.19, 95% CI [2.92, 3.48]); hypertensive group (3.28 [3.10, 3.48]); respiratory condition group (4.08 [3.76,4.43]); and severely impaired group (10.2 [9.2, 11.28]).
Conclusions :
LCA is effective in identifying pattern of chronic conditions in the population. Individuals in all four multi-morbidity groups were at significantly elevated risk of VI. Assessing the multi-morbidity patterns and its relationship with VI is helpful to reduce healthcare utilization and improve quality of life in these high-risk populations.
This abstract was presented at the 2019 ARVO Annual Meeting, held in Vancouver, Canada, April 28 - May 2, 2019.