Predicting and preventing readmissions in kidney transplant recipients

Abstract

A lack of research exploring post-transplant process optimization to reduce readmissions and increasing readmission rates at our center from 2009 to 2013 led to this study, aimed at assessing the effect of patient and process factors on 30-d readmission rates after kidney transplantation. This was a retrospective case-control study in adult kidney transplant recipients. Univariate and multivariate analyses were utilized to assess patient and process determinants of 30-d readmissions. 384 patients were included; 30-d readmissions were significantly associated with graft loss and death (p = 0.001). Diabetes (p = 0.049), pharmacist identification of poor understanding or adherence, and prolonged time on hemodialysis prior to transplant were associated with an increased risk of 30-d readmissions. After controlling for risk factors, readmission rates were only independently predicted by pharmacist identification of patient lack of understanding or adherence regarding post-transplant medications and dialysis exposure for more than three yr (OR 2.3, 95% CI 1.10-4.71, p = 0.026 and OR 2.1, 95% CI 1.22, 3.70, respectively), both of which were significantly modified by history of diabetes. Thirty-d readmissions are attributable to both patient and process-level factors. These data suggest that a lack of post-transplant medication knowledge in high-risk patients drives early hospital readmission.

“Early” and “Late” Hospital readmissions in the first year after kidney transplant at a single center

“…These findings are comparable with previous reports where the leading cause of readmissions was graft dysfunction and infection . However, in other studies surgical complications comprised the leading indication (15%‐25%) for readmissions after KT compared to readmission for infection (10%‐22%), metabolic (including electrolyte) abnormalities (3.5%‐22%) or graft dysfunction (14%‐21%) . Thus, there appears to be variation between centers not only in the readmission rate but also in the frequency of specific indications for readmission.…”

“…However, demographics of the population in these studies are often limited to recipients with Medicare as the primary insurer and lack granular data on causes of graft loss, recipient death, and magnitude and center variation of missing variables . Published studies also lack clarity regarding the consistency of caregivers and clinical protocols, and the amount of transplant dedicated institutional resources, which may influence the frequency and trends of readmissions . Such differences likely account for the substantial center‐to‐center variation in readmission rate .…”

“…Published studies also lack clarity regarding the consistency of caregivers and clinical protocols, and the amount of transplant dedicated institutional resources, which may influence the frequency and trends of readmissions. 3,4,8,9,14,15 Such differences likely account for the substantial center-to-center variation in readmission rate. 3 Furthermore, no US studies have analyzed readmission patterns over time.…”

“…17 The ex- (14%-21%). 8,9,15,18 Thus, there appears to be variation between centers not only in the readmission rate but also in the frequency of specific indications for readmission. While scrutiny is placed on the 30-day readmission period, our results suggest that 90 days may be more appropriate for the KT population, especially in light of our data demonstrating similar indications for readmissions during the early and late periods.…”

“…Published studies reporting readmission rates after KT are either based on data from large national databases linked to Medicare claims with center-to-center variation of the many parameters influencing readmissions including non-uniform clinical protocols and resources, or involve smaller cohorts of transplant recipients from single-institution studies. 3,4,6,[8][9][10] Tx LOS is not included in this calculation. The time frame is 1 y from initial readmission, while including initial readmission.…”