In July 2003, we reallocated our resident workforce to address mandated duty-hour restrictions. In the subsequent academic year (AY), surgical intensive care unit (SICU) service readmission rates (RR) doubled. We hypothesized that a targeted intervention could reduce SICU service RR in academic year (AY) 2004-05.
Methods: This study was conducted at an urban teaching hospital before (AY02-03, period 1), during (AY03-04, period 2), and after (AY04-05, period 3) implementation of the Accreditation Council for Graduate Medical Education guidelines. Demographics, RR, and reason were culled from Project Impact and a complications database. SICU staff (dedicated intensivist, two or three fellows, and six residents) remained constant. In periods 2 and 3 (versus 1), ward residents cross-covered > or = 3 services every 5 to 6 nights (versus every 3 in period 1) with physician assistant support (versus none in period 1). During period 3, a focused transfer phone call, charted care summary, and discharge checkup defined the intervention. Interperiod comparisons were by chi2 and t test analysis; p < 0.05 (versus period 1) defined significance.
Results: In all, 1,570, 1,705 and 1,681 patients were treated in periods 1, 2, and 3, respectively. There were no demographic or APACHE score differences. RRs were 1.4%, 3.0% and 1.2% in periods 1, 2, and 3, respectively. The percentages of readmissions as a result of ward care were 16.7, 41, and 10%, respectively. The most common readmission indication was respiratory (46% in period 1; 51% in period 2, and 80% in period 3) and was associated with an increased proportion of readmission as a result of patient disease (46% in period 1; 41% in period 2; 80% in period 3). Intervention noncompliance preceded 30% of period 3 readmissions.
Conclusion: A targeted intervention can reduce the rate of SICU readmission caused by care inadequacies stemming from a resident reallocation strategy.