Association of Chronic Disease With Patient Financial Outcomes Among Commercially Insured Adults

Abstract

Importance: The bidirectional association between health and financial stability is increasingly recognized.

Objective: To describe the association between chronic disease burden and patients' adverse financial outcomes.

Design, setting, and participants: This cross-sectional study analyzed insurance claims data from January 2019 to January 2021 linked to commercial credit data in January 2021 for adults 21 years and older enrolled in a commercial preferred provider organization in Michigan.

Exposures: Thirteen common chronic conditions (cancer, congestive heart failure, chronic kidney disease, dementia, depression and anxiety, diabetes, hypertension, ischemic heart disease, liver disease, chronic obstructive pulmonary disease and asthma, serious mental illness, stroke, and substance use disorders).

Main outcomes and measures: Adjusted probability of having medical debt in collections, nonmedical debt in collections, any delinquent debt, a low credit score, or recent bankruptcy, adjusted for age group and sex. Secondary outcomes included the amount of medical, nonmedical, and total debt among individuals with nonzero debt.

Results: The study population included 2 854 481 adults (38.4% male, 43.3% female, 12.9% unknown sex, and 5.4% missing sex), 61.4% with no chronic conditions, 17.7% with 1 chronic condition, 14.8% with 2 to 3 chronic conditions, 5.4% with 4 to 6 chronic conditions, and 0.7% with 7 to 13 chronic conditions. Among the cohort, 9.6% had medical debt in collections, 8.3% had nonmedical debt in collections, 16.3% had delinquent debt, 19.3% had a low credit score, and 0.6% had recent bankruptcy. Among individuals with 0 vs 7 to 13 chronic conditions, the predicted probabilities of having any medical debt in collections (7.6% vs 32%), any nonmedical debt in collections (7.2% vs 24%), any delinquent debt (14% vs 43%), a low credit score (17% vs 47%) or recent bankruptcy (0.4% vs 1.7%) were all considerably higher for individuals with more chronic conditions and increased with each added chronic condition. Among individuals with medical debt in collections, the estimated amount increased with the number of chronic conditions ($784 for individuals with 0 conditions vs $1252 for individuals with 7-13 conditions) (all P < .001). In secondary analyses, results showed significant variation in the likelihood and amount of medical debt in collections across specific chronic conditions.

Conclusions and relevance: This cross-sectional study of commercially insured adults linked to patient credit report outcomes shows an association between increasing burden of chronic disease and adverse financial outcomes.

Figure 1.. Predicted Probability of Credit Outcomes by Number of Chronic Conditions

Each line reports the predicted probability of each debt outcome by the number of chronic conditions. Results are from 2-part exponential hurdle models (for medical debt in collections, nonmedical debt in collections, and any delinquent debt) and logit regression models (for low credit score and any bankruptcy). Additional covariates in all models include sex and age-band fixed effects. Standard errors are robust. All estimates for chronic condition categories greater than 0 are significantly different from 0 chronic conditions (all P < .001). Error bars indicate the 95% CI for each estimate. Full regression estimates available in eTable 3 in the Supplement.

Figure 2.. Average Debt Among Individuals With Nonzero Debt by the Number of Chronic Conditions

Each line reports the estimated average debt (in dollars) for each debt outcome among individuals with any nonzero debt in that category by the number of chronic conditions. Results are from 2-part exponential hurdle models; additional covariates in all models include sex and age-band fixed effects. Standard errors are robust. All estimates for chronic condition categories greater than 0 are significantly different from 0 chronic conditions (all P ≤ .001). Error bars indicate the 95% CI for each estimate. Full regression estimates available in eTable 4 in the Supplement.

Figure 3.. Estimated Increase in the Probability of Having Medical Debt in Collections by Type of Chronic Condition

Each column reports the marginal increase in the absolute predicted probability (in percentage points) of having medical debt in collections for each chronic condition relative to not having that condition. Results are from 2-part exponential hurdle models (1 for each chronic condition); additional covariates for all models include sex and age-band fixed effects. Standard errors are robust. All point estimates are significantly different from the reference category of not having each chronic condition (P < .001). Error bars indicate the 95% CI for each estimate. CHF indicates congestive heart failure; CKD, chronic kidney disease; COPD, chronic obstructive pulmonary disease.

Figure 4.. Estimated Increase in Dollar Amount of Medical Debt in Collections by Type of Chronic Condition Among Individuals With Nonzero Medical Debt in Collections

Each bar reports the marginal increase in predicted medical debt in collections (in dollars) for each chronic condition relative to not having that condition for individuals with any medical debt in collections. Results are from 2-part exponential hurdle models (1 for each chronic conditions); additional covariates for all models include sex and age-band fixed effects. Standard errors are robust. All point estimates are significantly different from the reference category of not having each chronic condition (P < .001). Error bars indicate the 95% CI for each estimate. CHF indicates congestive heart failure; CKD, chronic kidney disease; COPD, chronic obstructive pulmonary disease.