We conducted a comprehensive analysis of the associations between postoperative complications and postoperative length of stay in a broad surgical population. To our knowledge, this is the first study to publish precise estimates of risk-adjusted increases in postoperative LOS associated with postoperative complications across a broad range of complications and operations. While this has been suspected by surgeons, confirmatory data and precise estimates are lacking in the literature. All 18 postoperative complications collected by the ACS-NSQIP were significantly associated with increased postoperative LOS in unadjusted analysis and remained significant after adjusting for preoperative risk characteristics and concomitant complications. The increase in postoperative LOS varied by different complications, and prolonged ventilation use, wound disruption, and acute renal failure were associated with the greatest increases in LOS. These findings have implications in resource allocation in the 13.4% of patients who have postoperative complications. If the postoperative complication rates could be reduced by only a few percentage points, this could save many additional inpatient hospitals days.

In addition to identification of postoperative complications that most affect postoperative LOS, this study provides evidence for preoperative characteristics and comorbidities that may modify LOS. While some risk factors for prolonged postoperative LOS are non-modifiable, others may be targeted for patient optimization prior to surgery. A potential target of optimization is underweight patients, who had a 2-day extended stay in the hospital postoperatively compared to normal weight patients. While many patients may be underweight due to systemic comorbidities such as cancer or severe illness, surgeons may counsel patients that improving their weight in a healthy manner may reduce the time they spend in the hospital after an operation by reduction of complications and improved recovery.

Our data are consistent with other smaller studies that have found associations between postoperative complications and length of stay in various subsets of surgeries and using different databases. Out of 18 studies in the literature that we reviewed [4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21], only two did not find a strong association between postoperative complications and increased LOS [4, 16]. In 2017, Mrdutt et al. analyzed 42,365 ACS-NSQIP patients undergoing elective laparoscopic colectomy and found that each postoperative complication increased the LOS [20]. In 2008, Boakye et al. examined the National Inpatient Sample and found that complications after laminectomy doubled postoperative LOS [6]. Additionally, in 2015 Damrauer et al. found similar results while investigating the effect of postoperative complications on postoperative LOS in 6,307 patients undergoing lower extremity bypass surgery using the California State Inpatient Database [7]. That study also found an independent association between postoperative LOS and patient readmission, indicating that finding avenues to reduce postoperative LOS could also have an effect on patient readmission, although further study is needed to determine if that is the case in a general surgical population. These findings suggest that the results of our study in a broad surgical population are in concordance with much of the research already done in individual operations.

There were two studies in our review that did not demonstrate a significant association between postoperative complications and postoperative LOS. In 2014, Krell et al. found that much of the variations in postoperative LOS were not attributable to either preoperative characteristics or postoperative complications in 22,664 patients undergoing inpatient colorectal resections [16]. Adogwa et al. reached the same conclusion in 23,102 patients after lumbar decompression and fusion procedure [4]. However, these studies used a different postoperative LOS outcome, whether patients had a LOS ≥ the 75th percentile of LOS. A binary cutoff could mask some of the effects of complications. This, plus a smaller sample size and more distinct patient populations, could explain the differences in the findings between these two studies and our study.

Our study is novel in that it has four important characteristics not currently found in the studies on postoperative complications and postoperative LOS in the literature: (1) analysis of a broad surgical population; (2) use of very large sample sizes; (3) inclusion of a wide array of postoperative complications; and (4) use of a straightforward analysis that provides average changes in LOS to be expected for each individual postoperative complication after adjusting for preoperative patient factors and other complications. Our data demonstrate that there is a strong association between postoperative complications and postoperative LOS, even when accounting for preoperative characteristics and concurrent complications. The results of this study provide concrete data for guiding decision making and resource allocation, providing hospitals with data to improve quality at their institutions and target complications that can most severely impact patient postoperative LOS. In addition, it provides information that a surgeon can share with patients and their families to better understand the course of their hospital stay.

Strengths of this study include: (1) the use of a large, audited, comprehensive database capturing a representative sample of the national surgery volume; (2) consideration of postoperative complications across many surgical specialties; and (3) inclusion of data over 14 years. However, there are several important limitations of our study to consider. First, we analyzed a broad surgical population, and the associations may vary in more specific surgical subspecialties and operations. In addition, predictor variables and complications analyzed were limited to those measured in the ACS-NSQIP. In this analysis, we included patients who died in the 30-day window post-operatively, which may potentially artificially depress their postoperative LOS. There has also been concern that the “inpatient” variable in ACS-NSQIP may not be standardized between institutions, potentially complicating an inpatient-specific analysis [25]. Finally, while we were able to use a large dataset from the ACS-NSQIP, there may be bias in which institutions participate, favoring large academic centers, and therefore these data may not be applicable to all hospitals.

Our study demonstrates how postoperative complications significantly affect hospital resources, particularly by prolonging the length of stay. These findings stress the importance of careful perioperative management and prompt intervention when complications arise. The varying impact of different complications on length of stay highlights the need for tailored strategies to address specific postoperative challenges. For example, implementing enhanced protocols to prevent and manage prolonged ventilator use, wound disruptions, and acute renal failure could substantially reduce hospital stays and enhance patient outcomes. Furthermore, this study offers valuable data for developing predictive models for risk assessment, which can help healthcare providers allocate resources more efficiently and take proactive measures for high-risk patients [26,27,28,29,30]. Future research should focus on identifying the most effective interventions to lessen the impact of these complications and explore the potential benefits of personalized postoperative care plans.