Optimizing Complication Self-Reporting Methodologies Improves Standard of Care and Quality

Abstract

Introduction

We utilized a single adverse event electronic self-reporting platform whose use was distinguished by three levels of faculty supervision, each at a separate hospital within our health system.

Methods

The 5-y study population included 83,885 surgical cases, 10,822 complications, 691 deaths and 3779 cases with complications collected from three independent hospitals within our health system. Each hospital reviewed cases with distinctly different levels of rigor: Site #1(Intense): a comprehensive 1-1½ h weekly meeting with resident case reporting/analysis, and in-person supervised attending review; Site #2 (Mild): resident case reporting/analysis, and occasional supervised review; Site #3 (Minimal): resident case reporting/analysis and no supervised review. Complication and mortality rates, standard of care, case mix index, length of stay and contribution margins were evaluated. Complications were treated as polynomial ordered logistic regression and modeled as logarithm of rate of complications per operation as outcome.

Results

Complications, deaths, and # cases with complications were significantly underreported at Site #2 and Site #3. Complication and mortality rates decreased 54% and 59% at Site #1, 8% and 36% at Site #2 and increased at Site #3. The rate (%) of “cases with complications” reported was greatest for Site #1 at 5.7%. There was a 35% overall improvement in the standard of care and a reduction in length of stay by 1.83 d at Site #1 with no differences in case mix index. An improvement in resident critical thinking was observed along with a reduction in judgment and communication errors.

Conclusions

Self-reporting of complications and mortalities is a valid data collection and quality improvement method when it includes a standardized electronic platform and rigorous in-person, attending review. Concurrent faculty scrutiny improves quality and is a mandatory component of the review process.