Psychosocial risk factors for hospital readmission in COPD patients on early discharge services: a cohort study

Abstract

Background: Hospital readmission for acute exacerbation of COPD (AECOPD) occurs in up to 30% of patients, leading to excess morbidity and poor survival. Physiological risk factors predict readmission, but the impact of modifiable psychosocial risk factors remains uncertain. We aimed to evaluate whether psychosocial risk factors independently predict readmission for AECOPD in patients referred to early discharge services (EDS).

Methods: This prospective cohort study included 79 patients with AECOPD cared for by nurse led EDS in the UK, and followed up for 12 months. Data on lung function, medical comorbidities, previous hospital admissions, medications, and sociodemographics were collected at baseline; St George's Respiratory Questionnaire (SGRQ), Hospital Anxiety and Depression Scale (HADS), and social support were measured at baseline, 3 and 12-months. Exploratory multivariate models were fitted to identify psychosocial factors associated with readmission adjusted for known confounders.

Results: 26 patients were readmitted within 90 days and 60 patients were readmitted at least once during follow-up. Depression at baseline predicted readmission adjusted for sociodemographics and forced expiratory volume in 1 second (odds ratio 1.30, 95% CI 1.06 to 1.60, p = 0.013). Perceived social support was not significantly associated with risk of readmission. Home ownership was associated with the total number of readmissions (B = 0.46, 95% CI -0.86 to -0.06, p = 0.024). Compared with those not readmitted, readmitted patients had worse SGRQ and HADS scores at 12 months.

Conclusion: Depressive symptoms and socioeconomic status, but not perceived social support, predict risk of readmission and readmission frequency for AECOPD in patients cared for by nurse-led EDS. Future work on reducing demand for unscheduled hospital admissions could include the design and evaluation of interventions aimed at optimising the psychosocial care of AECOPD patients managed at home.

Figure 1

Mean change (95% CI) for St George's Respiratory Questionnaire (SGRQ) by readmission status. Interval estimates with lower and upper limits were plotted for mean SGRQ scores at baseline, 90 days and 12 months by hospital readmission status. At 12 months confidence limits do not overlap showing that HRQOL was significantly worse among readmitted patients compared with patients who were free of readmissions.

Figure 2

Mean change (95%CI) for Hospital Anxiety and Depression Scale (HADS) total score by readmission status. Interval estimates with lower and upper limits were plotted for mean total HADS scores at baseline, 90 days and 12 months by hospital readmission status. At 12 months confidence limits do not overlap showing that mixed anxiety and depression was significantly worse among readmitted patients compared with patients who were free of readmissions.

Figure 3

Mean change (95% CI) for Hospital Anxiety and Depression Scale (HADS) depression scores by readmission status. Interval estimates with lower and upper limits were plotted for mean depression scores on the HADS at baseline, 90 days and 12 months by hospital readmission status. At 12 months confidence limits do not overlap showing that depression was significantly worse among readmitted patients compared with patients who were free of readmissions.