PGMP for Alberta Surgical Patients

Overview

Patients with postoperative hyperglycemia have higher complication rates and worse outcomes than patients with in-target glucose, including longer length-of-stay, more admissions to intensive care, increased readmission rates, and higher risk of infection. There are 35-40% of surgical patients at risk of postoperative hyperglycemia, including 20-30% of patients with diabetes, 10% with undiagnosed diabetes, and another 5-10% without diabetes who will have hyperglycemia. Data demonstrate significant quality and practice gaps that contribute to postoperative hyperglycemia in Alberta; 43% of patients with diabetes had no postoperative glucose measurements and two-third had postoperative hyperglycemia. The multi-disciplinary Perioperative Glycemic Management Pathway (PGMP) has been demonstrated to reduce postoperative hyperglycemia and improve outcomes for surgical patients in pilot-testing. This project will confirm the effects of the PGMP on (1) postoperative glycemic management; (2) length-of-stay; (3) readmission rates and (4) cost savings; compared to usual care in surgical patients admitted for \>24-hours postoperatively. We will implement the PGMP in 12 services in 6 hospitals in Alberta using implementation science and an evidence-informed knowledge translation approach. Impact will be analyzed using a stepped-wedge study design and will include formal evaluation of cost-effectiveness and implementation, clinical, and patient-reported outcomes. The primary efficacy outcome is change in length-of-stay pre/post implementation. Secondary clinical outcomes include the 30-day readmission rate, surgical site infection rate, 30-day mortality, and one-year mortality. We will implement the PGMP for 23,200 patients, and reduce postoperative hyperglycemia and associated complications for 7,665 patients, avoiding the expenditure of \>$24.6 million for a return on investment of 1,810%.

Study design: This stepped-wedge study(69) will evaluate cost effectiveness and implementation, clinical, and patient-reported outcomes. Research question: What are the effects of the PGMP on: (1) postoperative glycemic management; (2) length-of-stay; (3) readmission rates and (4) cost; compared to usual care in surgical patients admitted for ≥24-hours postoperatively? Population: Surgical patients at risk of hyperglycemia, admitted for ≥24-hours postoperatively. Pre-intervention sites serve as control groups. Outcomes: Change in length-of-stay is the primary effectiveness outcome, stratified for patients with diabetes, with and without hyperglycemia and patients without diabetes, with and without hyperglycemia. Secondary clinical outcomes: odds of 30-day readmission, 30-day mortality, and one-year mortality. Data Analysis: The exposure is the PGMP. Patients will be stratified by diagnosis of diabetes (pragmatically defined using Diabetes Canada diagnostic criteria(51)) and postoperative hyperglycemia (categorical). We will use quantile regression to estimate differences in the median length-of-stay and logistic regression to estimate 30-day odds of readmission and mortality, and odds of 1-year mortality, adjusted for age, sex, surgical type, surgical urgency, hospital unit, and comorbidities. Surgical services will serve as their pre-implementation control, and sites that have not undergone implementation will serve as controls for post-implementation sites.(69) Prespecified covariates are age, HbA1c, and surgery type. Cost-effectiveness Analysis: We will conduct a cost-utility analysis comparing current standard of care (pre-implementation)) with the PGMP (post-implementation) over a 90-day post-operative period. Costs will be based on the 2023 Alberta costs of care. Health benefits will be described as Quality-Adjusted Life Years (QALYs) using EQ-5D-3L collected from pre-and post-intervention sites. The QALYs will represent the impact of PGMP on morbidity and mortality. The resultant incremental cost-effectiveness ratio (ICER) (cost of the intervention as well as net system health service utilization costs as numerator and QALYs as the denominator) will establish the incremental cost-utility ratio of the PGMP compared to the prior standard of care. We will evaluate the cost effectiveness from a healthcare and societal perspective using the robust resources of the DON SCN for societal costs of surgical complications from diabetes. Sample Size: Based on the effect size seen in pilot-testing (reduction in length-of-stay, 0.6-1.1 days, variance 2.9 days) and using an estimated intraclass correlation of 0.02 (range 0.01-0.05), 3 steps (Appendix 5) and a period of 3 months, we will require 2 clusters of 75 patients each to demonstrate a significant reduction in length-of-stay with 90% power (alpha 0.05).(70) This accounts for only 35% of patients being eligible for the PGMP (Appendix 6).