Early Hospital Readmission: The Canary in the Coal Mine?
Naylor et al examined the incidence of EHR in 5437 kidney transplant recipients from Ontario, Canada, across the years of 2002-2014.2 The overall incidence was 20.8%, which is lower than the reported 31.6% incidence in a large American study.3 In their article, Naylor et al speculates on potential reasons for lower EHR in the Ontario study, including patient and system-level differences. However, are there other reasons not highlighted in their study resulting in lower EHR? Lubetzky et al reported that 28.3% of readmissions in their retrospective cohort may be prevented by an improvement in access to outpatient services.4 Could the care in Ontario already use these strategies? Is the care in the emergency department (ED) different for kidney transplant recipients in Ontario, resulting in lower readmission rates? McElroy et al detail a 37.4% incidence in ED visits by abdominal transplant recipients, of which, 33% occurred within the first month.5 Do EDs in Ontario use streamlined processes and better communication among team members to decrease the necessity of readmission to hospital?
When Naylor et al examined trends in EHR over the years, they found no significant changes. Despite an older recipient and donor population with more comorbidity, particularly diabetes, the incidence of readmission remained constant at 21%.2 Other factors, including recipient age, donor type, and length of stay, changed over the years, but the incidence of EHR within these subgroups did not. The most common cause of readmission in their cohort was failure/rejection of the graft and nonspecified complications of the procedure. Failure/rejection decreased over time as procedural complications became more common. These data may be biased as it does not consider secondary diagnoses, and occasionally, there is more than 1 problem responsible for a patient’s readmission. Furthermore, coding subjectivity can impact what is assigned as primary versus secondary admission diagnoses.
Naylor et al found that most readmissions occurred early (<8 days) after discharge from hospital. Longer length of stay and age older than 60 years was associated with a higher risk of EHR.2 McAdams-DeMarco et al reported that frailty was associated with EHR.6 These signals may function as clinical clues to identify at-risk patients. Early hospital readmission varied among the 6 Ontario transplant centers, ranging from 16% to 27%. This is not unique, and such variability within transplant programs has been reported previously.1 This variability does allow the opportunity to identify quality clues within and across different sites. Does collaboration, sharing, or even centralization improve or impede the outcome of patients receiving a kidney transplant?
What benchmark are we striving for in EHR? Kidney transplant recipients have lower EHR compared with liver and simultaneous kidney-pancreas transplants, but these are still far greater than other surgical procedures.