This study evaluated the impact of respiratory depression on length of stay and hospital costs, which unlike the impact of general ORADEs on these outcomes, are not well described in the literature [11, 15–20]. United States patients who had ≥ 1 respiratory depression episode had a significantly longer length of stay and a higher cost of hospitalization, compared to patients without opioid-induced respiratory depression. Patients at high risk for respiratory depression (PRODIGY score) with ≥ 1 confirmed respiratory depression episode also had significantly higher hospital costs. In-depth propensity weighted analysis found that patients with ≥ 1 respiratory depression episode in the United States cohort had a 16% higher healthcare cost compared to patients without respiratory depression and a 17% higher healthcare cost excluding patient outliers. Total healthcare costs, which included the sum of fixed and variable costs incurred by the hospital, were significantly increased by patient length of stay, length of stay complicated by occurrence of respiratory depression, longer length of surgery, and procedure type. Importantly, respiratory depression identified by continuous capnography and pulse oximetry monitoring was critical, since patients with respiratory depression experienced exponentially increased healthcare costs as length of stay increased. In contrast, in the absence of respiratory depression episodes, increased length of stay was associated with increased healthcare cost, but this association was linear.
The hospital general care unit or ward remains the site for an alarmingly high number of acute cardiorespiratory compromise events . About 290,000 in hospital cardiac arrests occur in the United States each year of which 40% have a respiratory insufficiency etiology. These events are usually preceded by a period of 6–8 hours of a gradual change in vital signs, which are not detected with traditional ‘spot-check’ based monitoring as is in place today [29, 30]. A majority of opioid-induced perioperative respiratory complications therefore occur in the under-monitored hospital ward and are associated with serious patient outcomes, including anoxic brain injury and mortality, as well as legal claims with significant financial burdens [13, 31]. Universal adoption of continuous monitoring systems is an attractive option, however the initial resource expenditure, challenges of alarm fatigue and lack of interventions based on alarm data remain at large. Here the PRODIGY trial allowed early identification of and stratification of patients at risk for respiratory impairment and is a first crucial step to improving patient outcomes and reducing healthcare cost. Other risk scores have been developed to identify patients at risk for ORADEs  however PRODIGY is a novel score to identify patients at risk specifically for opioid-induced respiratory depression . Similar to our work, other trials have reported that ORADEs, a majority of which are respiratory, are associated with increased healthcare utilization, longer hospital length of stay, higher 30-day readmission, and increased healthcare costs [14, 15, 19–21, 31]. Studies have also demonstrated the utility of continuous pulse oximetry on the ward, where up to 90% of postoperative hypoxemia episodes go undetected by intermittent spot-check monitoring . In one study, continuous pulse oximetry on the ward reduced rescue events and ICU transfers, and hence decreased healthcare costs . Similarly, after implementing continuous capnography to monitor patients receiving intravenous patient controlled analgesia opioids on the hospital ward, Stites and colleagues reported a 50% reduction in the incidence of opioid-induced respiratory depression rescue using rapid response teams, and a 79% decrease in transfers to higher levels of care, both of which are costly endeavors . The PRODIGY trial confirmed a 46% incidence of opioid-induced respiratory depression using continuous pulse oximetry and capnography, which has been shown to detect respiratory depression better than pulse oximetry alone [4, 5, 8, 34]. Given the high frequency of respiratory depression and our findings that it increases healthcare utilization and cost, reducing the incidence of respiratory depression by utilizing continuous oximetry and capnography and allowing for early detection of respiratory depression episodes has the potential to improve patient outcomes and decrease healthcare utilization.
The additive cost burden of respiratory depression of $3,237 (16%) increase (with outliers) and $3,200 (17%) increase (without outliers) in hospitalization costs per PRODIGY trial analysis, is somewhat less than the $4350-$8225 [14, 15, 17, 20] (27–47%) range, reported in the literature. The more conservative cost burden estimate, as found by PRODIGY trial, may be explained by PRODIGY being a prospective trial that used continuous capnography and oximetry monitoring to identify opioid-induced respiratory depression, and strict adherence to inclusion and exclusion criteria, vs. literature studies [14, 15, 17, 20] which were retrospective in nature, relying on claims analyses, and coded instances of ORADEs, likely missing milder (and less costly) cases of ORADE.
Although other studies have reported differences in cost and healthcare outcomes for ORADEs, the factors contributing to these outcomes are not well described [14, 15, 19–21, 31]. Our analysis identified patient characteristics that significantly impacted length of stay and cost. Use of multiple opioids, longer, high risk, or open surgery, respiratory depression, and medical conditions including chronic heart failure, hypertension, and sepsis, all contributed to increased length of stay. Interestingly, the PRODIGY score accounts for chronic heart failure and opioid naivety when determining patient risk for respiratory depression . Importantly, respiratory depression contributed to both length of stay and cost, highlighting its importance in determining patient outcomes. The findings of this trial may be of particular interest to payers (e.g., CMS), organizations related to quality measurement and reporting (e.g., National Quality Forum), and hospital administration, highlighting the unmet need in the quality of care for post-surgical patients on opioids, and the potential need to institute quality metrics to improve outcomes and reduce costs in this patient population. Finally, our analysis excluded outlier patients who had very high costs or an extended length of stay, and confirmed that inclusion or exclusion of these patients did not alter our main findings. Therefore, the increases in length of stay and cost for patients with respiratory depression are not due to a subset of patients requiring extended care or costly interventions but reflect differences between typical ward patients with and without respiratory depression.
Our work is limited by the fact that we included a portion of United States hospitals from our trial cohort, though PRODIGY also enrolled in Asia and Europe. While this may limit the generalizability of our data, we included a substantial number of United States patients and hospitals of various types and sizes. Our analysis evaluated the actual hospital cost incurred, including both fixed and variable costs, and did not rely on diagnosis related group payment data. However, this analysis was limited to the total hospital cost per patient, preventing identification of specific factors that may have contributed to increased hospital cost for patients with respiratory depression. Determination of opportunity cost and productivity loss associated with increased length of stay was out of the scope of this analysis but would be a valuable addition to future studies. Furthermore, an actual calculation of the ‘break-even’ cost of the institution of monitoring versus the cost of respiratory depression events is out of the scope of this work, but a planned future analysis.