Influence of income on in-hospital mortality in older adults during the first wave of the COVID-19 pandemic: results from the EDEN-33 study.

Abstract

Objectives: To determine whether income was associated with unexpected in-hospital mortality in older patients treated in Spanish public health system hospital emergency departments.

Material and methods: Fifty-one public health system hospital emergency departments in Spain voluntarily participated in the study. Together the hospitals covered 25% of the population aged 65 years or older included in all patient registers during a week in the pre-pandemic period (April 1-7, 2019) and a week during the COVID-19 pandemic (March 30 to April 5, 2020). We estimated a patient's gross income as the amount published for the postal code of the patient's address. We then calculated the standardized gross income (SGI) by dividing the patient's estimated income by the mean for the corresponding territory (Spanish autonomous community). The existence and strength of an association between the SGI and in-hospital mortality was evaluated by means of restricted cubic spline (RCS) curves adjusted for 10 patient characteristics at baseline. Odds ratios (ORs) for each income level were expressed in relation to a reference SGI of 1 (the mean income for the corresponding autonomous community). We compared the COVID-19 and pre-pandemic periods by means of first-order interactions.

Results: Of the 35 280 patients attended in the 2 periods, gross income could be ascertained for 21 180 (60%), 15437 in the pre-pandemic period and 5746 during the COVID-19 period. SGIs were slightly higher for patients included before the pandemic (1.006 vs 0.994; P = .012). In-hospital mortality was 5.6% overall and higher during the pandemic (2.8% pre-pandemic vs 13.1% during COVID-19; P .001). The adjusted RCS curves showed that associations between income and mortality differed between the 2 periods (interaction P = .004). Whereas there were no significant income-influenced differences in mortality before the pandemic, mortality increased during the pandemic in the lowest-income population (SGI 0.5 OR, 1.82; 95% CI, 1.32-3.37) and in higher-income populations (SGI 1.5 OR, 1.32; 95% CI, 1.04-1.68, and SGI 2 OR, 1.92; 95% CI, 1.14-3.23). We found no significant differences between patients with COVID-19 and those with other diagnoses (interaction P = .667).

Conclusion: The gross income of patients attended in Spanish public health system hospital emergency departments, estimated according to a patient's address and postal code, was associated with in-hospital mortality, which was higher for patients with the lowest and 2 higher income levels. The reasons for these associations might be different for each income level and should be investigated in the future.