The experiences of patients who leave hospital against medical advice: analysis of survey data.

Abstract

Background: Historically, when patients leave hospital against medical advice (LAMA), the focus has often been on non-compliance or other patient-level factors, rather than on how services may be designed to better support these patients. Efforts to better understand why patients LAMA could strengthen the provision of patient-centered care that is responsive to individual needs and values. This study aimed to explore the experiences of Albertan adults who LAMA by examining patient-centered quality indicators (PC-QIs) derived from survey data. We sought to identify actionable insights that may inform service improvements and reduce the risks associated with LAMA discharges.

Methods: We analyzed seven years of survey data, encompassing hospital discharges from April 2016 to March 2023. A random sample of respondents completed the Canadian Patient Experiences - Inpatient Care (CPES-IC) instrument by telephone within six weeks of hospital discharge. From the data, we assessed ten patient-centred quality indicators (PC-QI) which were previously co-created with patient advisors, researchers, and health system administrators. Survey responses/PC-QIs were reported as percent in "top box", as represented by the most positive answer choice. Differences between patients who LAMA and other medical/surgical discharges were assessed.

Results: A total of 144,480 surveys were successfully linked with inpatient records and included for analysis. This included 1,177 (0.9%) respondents who LAMA. In our sample, those who LAMA were predominantly male, younger, had a lower level of educational attainment, and were living with a greater number of comorbid health conditions. They also had lower self-reported levels of physical and mental health and had a longer average length of stay. The LAMA group had significantly lower top-box percentages on all ten of the PC-QIs which we examined. This difference ranged from 20.7% (communicating test results; 51.6% LAMA group vs. 71.3% others) to 29.2% (patient involvement in decisions about their care and treatment; 39.8% vs. 69.0% respectively).

Conclusion: Patients who LAMA reported lower ratings of patient experience across all PC-QIs studied. Our findings may provide actionable, service-related insights into the reasons why patients LAMA. This is important as those who do so may place themselves at increased risk for future unplanned healthcare events, mortality, and morbidity.