Determinants associated with completion of postdischarge follow-up survey among multimorbid patients: a secondary analysis of the non-randomised clinical In-HospiTOOL trial

Importance Postdischarge surveys are critical in collecting patient-reported experience measures (PREMs) and patient-reported outcome measures (PROMs), but response rates vary and are often low. Objective To assess determinants that are associated with survey completion by phone in a complex medical care setting. Design Secondary analysis of a prospective controlled interrupted time series analysis. Setting As part of the non-randomised controlled In-HospiTOOL trial, a survey was conducted to gather data on PREMs and PROMs in multimorbid patients from seven hospitals in Switzerland. Participants 31 103 medical acute care hospitalisations among seven intervention hospitals who were eligible for the survey. Interventions Over a 6-month pre-intervention phase (August 2017 through January 2018) and a subsequent 12-month intervention phase (February 2018 through January 2019), patients were contacted by phone 30 days after hospital admission. Main outcomes and measures The primary outcome was responsiveness to the survey. We assessed group differences between responders and non-responders, and associations of patient characteristics with survey completion were estimated using generalised estimating equations. Results Of 31 103 eligible patients, 25 557 (82.2%) completed the survey 30 days after hospital admission. Responders were marginally older than non-responders (median (IQR) age, 73 (60–82) years vs 72 (57–82); standardised mean difference (SMD), −0.08), were more likely to be Swiss (81.9% vs 74.4%; SMD, −0.18), to have private healthcare insurance (22.9% vs 17.9%; SMD, 0.12), to be living at home before admission (85.7% vs 78.6%; SMD, 0.18) and to be less frail (67.4% vs 59.1%; SMD, 0.18). A longer length of stay (OR 0.98; 95% CI 0.97 to 0.99), discharge to a non-home institution (OR 0.50; 95% CI 0.46 to 0.54) and rehospitalisation within 30 days (OR 0.78; 95% CI 0.68 to 0.89) is associated with a decreased responsiveness. Conclusions The study shows that achieving a high survey response rate among vulnerable acute care patients is feasible, which in turn allows for the effective identification of key determinants and enhances the collection of information on patients’ experiences and outcomes. Trial registration number ISRCTN83274049 .