The Association Between Number of Hospital Advanced Practice Providers and Surgical Morbidity

Studies assessing the impact of advanced practice providers (APPs), including nurse practitioners and physician assistants, have demonstrated a similar quality of care for patients admitted to the hospital for medical diagnoses. However, no studies have examined the relationship between APP integration and morbidity after cancer surgery. This study assesses the relationship between APP staffing intensity and patient outcomes following major abdominal cancer surgery.

We used 100% national Medicare data (2010-2019) to assess the link between APP staffing intensity and surgical outcomes for patients undergoing major abdominal cancer surgery, including cystectomy, colectomy, hepatectomy, esophagectomy, gastrectomy, and pancreatectomy. The primary exposure was the ratio of APPs per 100 hospital beds, and patients were empirically divided into tertiles. Outcomes included readmission rates and length of stay, adjusted for patient and hospital level factors. As a secondary outcome, we measured 30-day perioperative mortality.

We analyzed 326,547 colectomy patients, 50,400 cystectomy patients, 14,112 esophagectomy patients, 27,152 gastrectomy patients, 15,225 hepatectomy patients, and 46,287 pancreatectomy patients. Surgery at centers with the most advanced practice providers per beds (i.e., the highest tertile) was associated with shorter length of stays for most surgery types analyzed. Unadjusted 30-day readmissions tended to be lower in patients undergoing more complex procedures, such as esophagectomy (21.5% vs. 24.3%; p = 0.006) but not for the less complex operations studied, such as colectomy (13.6% vs. 13.5%; p = 0.22). However, clinical differences in outcomes were lost on analyses controlling for patient and hospital factors (IRR length of stay: 0.98-1.01; p = 0.002-087) (OR readmissions: 0.86-1.01; p = 0.003-0.80).

The relationship between hospital APP staffing intensity and surgical outcomes was varied and heterogenous. However, differences in outcomes were primarily explained by hospital factors. More work is needed to determine process measures associated with the deployment of inpatient APPs.