SAN DIEGO – Readmission rates and the underlying reasons for them vary between medical specialties following surgery in children, a study of national data suggests.
“Hospital readmission is a very hot topic, particularly in light of the Affordable Care Act,” Afif N. Kulaylat, MD, said at the American College of Surgeons/National Surgical Quality Improvement Program National Conference. “Beyond economic costs there are very tangible costs to patients that we must consider. Readmissions have been associated with significant morbidity for patients. They often herald or implicate a postoperative complication. There are also indirect costs to patients and families such as time off from school or work.”
Dr. Kulaylat, of the division of pediatric surgery at Penn State Children’s Hospital, Hershey, Penn., presented findings from a retrospective analysis of NSQIP Pediatric (NSQIP-P), an ongoing collaboration between the ACS and the American Pediatric Surgical Association to improve the care of young patients. The researchers evaluated NSQIP-P data from 2013 and 2014 and focused on unplanned readmission within 30 days, including reasons for readmission based on NSQIP-P readmission categories and ICD-9 readmission codes as categorized by the AHRQ Clinical Classification Software. Multivariate logistic regression was used to evaluate factors associated with unplanned readmission.
Dr. Kulaylat reported results from a cohort of 129,849 patients cared for by 64 NSQIP-P participating hospitals. Among these, the all-cause readmission rate was 4.7%. After excluding patients with planned readmissions, the unplanned readmission rate was found to be 3.9%. From this cohort, 28% required reoperation within 30 days, and the median time from discharge to unplanned readmission was 8 days, with an interquartile range between 3 and 14 days.
Among the procedures captured in NSQIP-P, neurosurgery accounted for the highest readmission rate (10.8%), followed by general/thoracic surgery (5.2%), urology (2.6%), ENT (2%), orthopedic (1.9%), and plastic and reconstructive surgery (1.3%). The most common reason for readmission was surgical site infection at 23%, followed by GI complications such as ileus, obstruction, and constipation (17%); pulmonary-related complications (9%); device-related complications including shunt malfunction (8%); neurologic (7%); pain (6%); other medical diseases (6%); sepsis (5%); electrolytes/dehydration (5%); and urinary tract infection (UTI, 3%). It is estimated that at least two-thirds of unplanned readmissions (63%) were directly related to surgery. “These reasons for readmission and their frequency closely parallel what is seen in adults, with the exception of bleeding complications, which were rare in children compared to adults,” Dr. Kulaylat said.
The top five CPT codes associated with readmissions were laparoscopic appendectomy, laparoscopic gastrostomy tube placement, and three additional codes related to placement and replacement/revision of ventricular shunts/catheters.
Reasons for readmission varied among specialties. For example, among general and thoracic surgery, surgical-site infections (SSI) and GI-related issues dominated, while in neurosurgery SSI and device issues dominated. In urology, UTIs were the most frequent, while ENT had a greater proportion of pulmonary complications. Certain patient variables were also associated with an increased risk of hospital readmission, including comorbidities related to GI, CNS, renal, and immunosuppression and nutrition (P less than .001 for all). The strongest association was the occurrence of a postoperative complication, namely a post-discharge complication.
“The granularity of NSQIP-P can continue to be refined to help predict who is likely to get readmitted or if specific follow-up strategies might identify those headed to readmission,” remarked Robert E. Cilley, MD, a senior author and surgeon-in-chief at Penn State Children’s Hospital. Dr. Kulaylat acknowledged certain limitations of the study, including its retrospective design, the potential for data entry/data interpretation error, and that the researchers were unable to adjust for clustering at the hospital level. Directions of future research include a plan to study readmissions and predictive factors at the procedural level, establish risk-adjusted specialty/procedural-specific benchmarks for readmission rates, and refine the accuracy and reliability of the readmission data. “With these NSQIP-P data there is substantial opportunity for quality improvement as we strive to improve the care of children everywhere,” Dr. Kulaylat said. He reported having no relevant disclosures.