A novel risk prediction model for 30-day severe adverse events and readmissions following bariatric surgery based on the MBSAQIP database.

BACKGROUND

Although bariatric surgery is safe, some patients fear serious complications.

OBJECTIVES

This retrospective study used the 2015 Metabolic and Bariatric Surgery Accreditation Quality Improvement Project (MBSAQIP) database to evaluate patient outcomes for gastric bypass (GB) and sleeve gastrectomy and to develop a risk prediction model for serious adverse events (SAEs) and readmission rates 30 days after surgery.

SETTING

MBSAQIP national patient database.

METHODS

We created separate exploratory multivariable logistic regression models for SAEs and readmissions. We then externally validated both models using the 2016 MBSAQIP Participant Use Data File.

RESULTS

Significant predictors of SAEs were preoperative body mass index (adjusted odds ratio [AOR] 1.07, P < .0001); GB surgery (AOR 2.08, P < .0001); cardiovascular disease (AOR 1.43, P < .0001); smoking (AOR 1.12, P = .04); diabetes (AOR 1.15, P = .0001); hypertension (AOR 1.17, P < .0001); limited ambulation (AOR 1.48, P < .0001); sleep apnea (AOR 1.12, P = .001); history of pulmonary embolism (AOR 2.81, P < .0001); and steroid use (AOR 1.40, P = .001). Significant predictors of readmissions were GB surgery (AOR 1.81, P < .0001); female sex (AOR 1.26, P < .0001); diabetes (AOR 1.08, P = .04); hypertension (AOR 1.11, P = .004); preoperative body mass index (AOR 1.05, P < .0001); sleep apnea (AOR 1.11, P = .002); history of pulmonary embolism (AOR 2.35, P < .0001); cardiovascular disease (AOR 1.61, P < .0001); smoking (AOR 1.14, P = .01); and limited ambulation (AOR 1.55, P < .0001). External validation supported these covariates, with similar model discriminative power.

CONCLUSIONS

Our exploratory regression models may be used by clinicians to counsel patients about surgical risks, although future external validation should occur in non-North American populations.