Abstract
Background:
To date, the cost of surgical care is largely measured by charges or payments, both of which are inadequate. Actual cost data from the hospital’s perspective are required to accurately quantify the financial return on investment of engaging in quality improvement. Our objective was to define the cost of individual, 30-day post-operative complications using robust cost data from a diverse group of hospitals.
Methods:
Using clinical data derived from the American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP), this retrospective study assessed postoperative complications for patients who underwent surgery at one of four diverse hospitals in 2016. Actual direct and indirect 30-day costs were obtained and the adjusted cost per complication was determined.
Results:
From the 6,387 patients identified, the three complications associated with the highest independent adjusted cost per event were prolonged ventilation ($48,168, 95% CI $21,861, $74,476), unplanned intubation ($26,718, 95% CI $15,374, $38,062), and return to the OR ($20,258, 95% CI $13,537, $26,978). The three complications associated with the lowest independent adjusted cost per event were UTI (−$372, 95% CI −$1,336, $592), superficial SSI ($2,473, 95% CI −$256, $5,201) and VTE ($7,909, 95% CI −$17,903, $33,721). After colectomy, the adjusted independent cost of anastomotic leak was $10,195 (95% CI$ 5,941, $14,449) while the cost of postoperative ileus was $10,205 (95% CI $6,259, $14,149).
Conclusions:
By using cost data from four diverse hospitals, the actual hospital costs of complications were estimated. These data can be used by hospitals to estimate the financial benefit of reducing surgical complications.
INTRODUCTION
Surgical services account for >40% of inpatient health care spending in the U.S1 and projected to account for over 7% of the U.S. GDP by 2025.2 One of the primary drivers of surgical costs are complications which dramatically increase the intensity of healthcare utilization, including increases in testing (e.g., labs, imaging), treatments (e.g., invasive interventions, reoperation), and clinician services (e.g., nursing care, consultants).3,4 For example, an anastomotic leak following colon surgery requires additional labs, cross sectional imaging, and almost always, an invasive intervention (percutaneous drainage or return to the operating room [OR]) which increases length of stay and increase the total cost of that episode of care.3,4
Hospitals work to improve surgical quality and reduce costs in several ways, including engaging in local quality improvement initiatives, clinical quality registries, and collaboratives (i.e., groups of hospitals working together to improve care). However, hospitals are unable to accurately quantify the financial benefit of their efforts as current cost estimates of individual complications are inadequate for several reasons. First, financial estimates are frequently based on one of two approaches, both of which fail to reflect the cost of a complication. One of these is to assign the payer payment as a proxy for cost, (e.g., Medicare’s Diagnosis Related Group [DRG]); another is to use hospital charges, which vary significantly from hospital to hospital and do not reflect the actual costs that hospitals experience. Second, if actual cost data are available, they are limited to estimates using single institutional cost information and are not focused on individual complications.
Estimating the actual hospital costs of complications is important for hospitals to understand the return on investment of undertaking quality improvement work and participating in registries and collaboratives to reduce surgical complication rates, all of which require substantial hospital investment and resources.5–7 However, robust estimation of the actual hospital cost of individual complications have generally not been attempted. Therefore, our objective was to estimate the actual hospital cost of individual 30-day postoperative complications by merging detailed cost and clinical data from a diverse group of hospital types (community, comprehensive community, academic medical center).
METHODS
Data Source and Study Population
Between January 1, 2016 and December 31, 2016, patients who underwent any general surgery, colorectal, otolaryngology, gynecology, neurosurgery, orthopedic, urology or vascular procedure at one of four hospitals the American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP) were included in this study. The four hospitals included a quaternary care, 894-bed academic medical center, a 392-bed comprehensive community hospital, a 198-bed community hospital, and a 159-bed community hospital. Each hospital is located in a separate geographic location serving different patient populations.
The details of the ACS NSQIP, including the sampling strategy, data abstraction procedures, variables collected, outcomes and structure have been detailed elsewhere.8–17 In brief, hospitals collect standardized and audited clinical data on patient demographics, preoperative variables, and postoperative complications for a predefined sample of patients. Each hospital has clinical data abstractor(s) who use standardized definitions to collect and report data to ACS NSQIP. Onsite data audits are regularly performed. Patient follow up is 30-days from the index operation, irrespective of whether the patient is an inpatient, has been discharged to another facility, or has been readmitted to another hospital. Patients are followed up by surgical clinical reviewers at each participating hospital who examine the medical record, query involved clinicians, and directly contact patients when needed to ascertain the required ACS NSQIP data elements.
Cost data
Actual, 30-day internal cost data were obtained from the finance department at each hospital for each patient reported by the hospitals to ACS NSQIP. Costs were summed for each individual patient based on the index hospital stay and all subsequent inpatient or outpatient encounters that may have occurred within 30-days from the date of surgery. Total cost was defined as the sum of the fixed direct (e.g., nurse salaries), variable direct (e.g., drugs, medical supplies), fixed indirect (e.g., information technology, medical records), and variable indirect (e.g., housekeeping, food services) costs.
Outcomes
ACS NSQIP collects data on postoperative complications whether or not the event occurred at the same or a different facility within 30-days from the day of the index surgery.18,19 Thirty-day complications were assessed using standardized definitions and included prolonged ventilation, unplanned intubation, cardiac events, sepsis, superficial Surgical Site Infection (SSI), deep/organ space SSI, renal failure pneumonia, venous thromboembolism (VTE) (deep vein thrombosis and pulmonary embolism), urinary tract infection (UTI), return to the operating room (OR), and readmission. Cardiac events included both myocardial infarction and cardiac arrest. In addition, anastomotic leak and postoperative ileus events were evaluated in a separate cohort of patients undergoing colorectal surgery as part of the ACS NSQIP Procedure Targeted program, given that colorectal procedures are frequently the focus of quality improvement initiatives.20
Statistical Analysis
Unadjusted costs for each complication were calculated by assessing the difference between the total median cost among patients with a single complication event (i.e., no other complications except the one being examined) and among patients without any complication events. This was performed separately for each individual complication. For example, to calculate the cost of prolonged ventilation, the total median cost among patients who experienced a prolonged ventilation event (in the absence of any other complication) was subtracted from the total median cost among patients who experience no complication events.
Adjusted costs of individual complications were calculated using median regression with robust standard errors at the facility level. A single model was constructed that included preoperative factors (ASA class, diabetes, body mass index [BMI], dyspnea, congestive heart failure [CHF], chronic obstructive pulmonary disease [COPD], and sepsis), surgical details (procedure type, emergency surgery status), and postoperative factors (inpatient or outpatient admission status). The exposure variables of interest were 30-day complications. All surgical complications except readmission and return to the OR were included in the same model. Readmission and return to the OR are intermediate outcomes and were evaluated in separate, individual models. Cost of individual complications generated from the median regression analysis reflects the added total cost of a particular complication (net of all other covariates in the model, including other complications if they were in the model) on the adjusted median cost. For example, unplanned intubation raised costs at the median by $26,718 (95% CI 15,374, 38,062) after controlling for other factors including other complications in the model.
Since anastomotic leak and postoperative ileus complication data were only available after colectomy, a separate model in this cohort of patients was constructed to estimate costs. Deep/organ space SSI was excluded from this model as it clinically could possibly represent the same complication event as anastomotic leak.
Sensitivity analyses were performed using an alternative modeling strategy with gamma regression to assess the robustness of our results. Gamma regression was chosen as it accounts for the non-normal and right-skewed nature of cost data.21,22 Gamma regression with log link is a commonly used method for health care cost analysis.23,24 Given that the coefficients from the log-gamma model do not have a straightforward interpretation, our analytic approach requested marginal effects25 in terms of dollars. For each complication, it estimated predicted mean cost from the model if all cases in the data had the complication (maintaining all other covariates as they are) and also estimated the predicted mean cost from the model if all cases did not have the complication (again, with other variables taken at their actual values). The difference of the marginal effects is the cost of the complication. Estimates are not additive with this approach. Additional analyses were also performed excluding emergency cases, only including general and colorectal procedures and including interaction terms of clinically relevant complications (prolonged ventilation and unplanned intubation, sepsis and deep/organ space SSI, sepsis and pneumonia).
All tests were two-sided and the significance level was set at P<0.05. All analyses were performed using STATA/MP 14.1 (College Station, TX) and SAS version 9.4 (SAS Institute, Cary, NC). The Northwestern University Institutional Review Board deemed the study exempt from human subjects review
RESULTS
From four hospitals, 6,387 patients were identified, the majority of whom underwent an orthopedic (38.8%), general surgery (15.3%) or colorectal (13.9%) procedure. There were 607 patients who underwent colon surgery and were monitored as part of the ACS NSQIP Procedure Targeted program and were included in the separate analysis of anastomotic leak and postoperative ileus events. Overall, the median age was 61 years (IQR=21). The majority of patients were ASA class I or II (61.4%) and were inpatients (61.8%). Most patients underwent a non-emergent operation (93.2%). Additional patient and procedure details are presented in Table 1.
Table 1.
Characteristics of patients who underwent surgery at one of four hospitals in 2016 (n=6,387).
| Patient Characteristics | |
|---|---|
| Age, median (IQR) | 61 (21) |
| N (%) | |
| Gender (%) | |
| Female | 3745 (58.63) |
| Male | 2642 (41.37) |
| Race (%) | |
| White | 5323 (83.34) |
| Black | 400 (6.26) |
| Asian | 140 (2.19) |
| American Indian or Alaska Native | 10 (0.16) |
| Other/Unknown | 514 (8.05) |
| Diabetes (%) | |
| Insulin | 282 (4.42) |
| Oral | 569 (8.91) |
| No | 5536 (86.68) |
| BMI (kg/m 2 ) (%) | |
| Underweight (<18.5) | 79 (1.24) |
| Normal (18.5–24.9) | 1385 (21.68) |
| Overweight (25.0–29.9) | 1997 (31.27) |
| Class 1 Obese (30.0–34.9) | 1514 (23.70) |
| Class 2 Obese (35.0–39.9) | 781 (12.23) |
| Class 3 Obese (≥ 40.0) | 631 (9.88) |
| Dyspnea (%) | |
| At rest | 3 (0.05) |
| Moderate exertion | 96 (1.50) |
| No | 6288 (98.45) |
| Congestive heart failure (%) | |
| Yes | 50 (0.78) |
| No | 6337 (99.22) |
| History of COPD (%) | |
| Yes | 206 (3.23) |
| No | 6181 (96.77) |
| ASA classification (%) | |
| I | 519 (8.13) |
| II | 3402 (53.26) |
| III | 2292 (35.89) |
| IV–V | 174 (2.72) |
| Functional Status (%) | |
| Independent | 6364 (99.64) |
| Dependent | 23 (0.36) |
| Preoperative sepsis | |
| Sepsis | 68 (1.06) |
| SIRS | 121 (1.89) |
| None | 6198 (97.04) |
| Pre-operative Renal failure | |
| Yes | 7 (0.11) |
| No | 6380 (99.89) |
| Surgery Setting (%) | |
| Inpatient | 3949 (61.83) |
| Outpatient | 2438 (38.17) |
| Emergency Surgery | |
| Yes | 432 (6.76) |
| No | 5955 (93.24) |
| Procedure Type | |
| Orthopedics | 2481 (38.84) |
| Gen surgery | 975 (15.27) |
| Colorectal | 887 (13.89) |
| Breast | 380 (5.95) |
| Gynecology | 361 (5.65) |
| Neurology/Spine | 338 (5.29) |
| Urology | 251 (3.93) |
| Vascular | 239 (3.74) |
| Hepatopancreatobiliary | 236 (3.7) |
| Ear/Nose/Throat | 160 (2.51) |
| Foregut | 79 (1.24) |
SIRS, systemic inflammatory response syndrome
Unadjusted complication rates are presented in Table 2. In the overall cohort, unadjusted complication rates ranged from 0.4% (renal failure) to 4.9% (readmission). Among the 607 patients who underwent colon surgery, the incidence of anastomotic leak was 4.5% and postoperative ileus was 18.6%.
Table 2.
Unadjusted thirty-day complication rates after surgery (n=6387).
| Complication | Number of Events | Complication Rate (%) |
|---|---|---|
| Prolonged ventilation | 38 | 0.6 |
| Unplanned intubation | 39 | 0.61 |
| Cardiac | 27 | 0.42 |
| Renal failure | 26 | 0.41 |
| Pneumonia | 55 | 0.86 |
| Sepsis | 59 | 0.94 |
| VTE | 100 | 1.57 |
| Deep/Organ Space SSI | 80 | 1.25 |
| UTI | 78 | 1.22 |
| Superficial SSI | 66 | 1.03 |
| Readmission | 310 | 4.85 |
| Return to operating room | 135 | 2.11 |
| Anastomotic leak* | 27 | 4.45 |
| Postoperative Ileus* | 113 | 18.62 |
Colectomy only, N=607
VTE, venous thromboembolism; SSI, surgical site infection; UTI, urinary tract infection
The unadjusted cost of individual complications are presented in Table 3. The three complications associated with the highest median cost were prolonged ventilation ($46,237), unplanned intubation ($42,487), and cardiac event ($24,017). The three complications associated with the lowest median cost were UTI ($3,847), superficial SSI ($6,477), and readmission ($8,524). The two intermediate complications assessed were associated with a median cost of $8,524 for readmission and $15,166 for return to the OR. After colon surgery, the associated median cost of anastomotic leak was $18,903 and postoperative ileus was $15,797.
Table 3.
Unadjusted and adjusted thirty-day total cost of individual complications after surgery
| Single Complication | Multiple Compilations | Median Regression | |||||
|---|---|---|---|---|---|---|---|
| Complication | Number of Events∞ N, (%) | Unadjusted Cost* ($) | Number of Events N, (%) | Unadjusted Cost* ($) | Adjusted Cost | 95% Confidence Interval | P value |
| Prolonged ventilation | 2 (0.03) | 46,237 | 36(0.56) | 102,608 | 48,168 | (21861, 74476) | <0.001 |
| Unplanned intubation | 7 (0.11) | 42,487 | 32(0.5) | 89,388 | 26,718 | (15374, 38062) | <0.001 |
| Cardiac event | 13 (0.2) | 24,017 | 14(0.22) | 62,584 | 15,109 | (9601, 20618) | <0.001 |
| Renal failure | 12 (0.19) | 17,729 | 14(0.22) | 116,381 | 18,528 | (17076, 19981) | <0.001 |
| Pneumonia | 24 (0.38) | 16,905 | 31(0.49) | 77,036 | 9,401 | (5878, 12925) | <0.001 |
| Sepsis | 8 (0.13) | 18,158 | 51(0.8) | 35,392 | 12,440 | (5905, 18974) | <0.001 |
| DVT/VTE | 72 (1.13) | 9,720 | 28(0.44) | 89,926 | 7,909 | (−17903, 33721) | 0.548 |
| Deep+Organ Space SSI | 41 (0.64) | 17,990 | 39(0.61) | 37,681 | 12,135 | (6321, 17949) | <0.001 |
| UTI | 62 (0.97) | 3,847 | 16(0.25) | 30,034 | (372) | (−1336, 592) | 0.449 |
| Superficial SSI | 48 (0.75) | 6,477 | 18(0.28) | 23,716 | 2,473 | (−256, 5201) | 0.076 |
| Readmission† | 134 (2.10%) | 8,524 | 176(2.76) | 17,494 | 8,020 | (4597, 11444) | <0.001 |
| Return to OR† | 38 (0.59%) | 15,166 | 97(1.52) | 30,305 | 20,258 | (13537, 26978) | <0.001 |
| Anastomotic leakǂ | 9 (1.48) | 18,903 | 18(2.97) | 46,048 | 10,195 | (5941, 14449) | <0.001 |
| Ileusǂ | 48 (7.91) | 15,797 | 65(10.71) | 42,545 | 10,205 | (6259, 14149) | <0.001 |
Number of events is lower here because the complication had to occur in isolation without any other complications.
Unadjusted costs determined by subtracting median cost of the single complication in isolation or if it occurred with other complications from the median cost if no complication occurred
Estimated from separate models that did not include other complications
Colectomy only N=607
VTE, venous thromboembolism; SSI, surgical site infection; UTI, urinary tract infection
We next estimated the independent cost of individual complications after adjustment for preoperative, intraoperative, and postoperative factors including the occurrence of other complications (Table 3). The three complications associated with the highest adjusted cost were prolonged ventilation ($48,168, 95% CI $21,861, $74,476), unplanned intubation ($26,718, 95% CI $15,374, $38,062), and renal failure ($18,528, 95% CI $17,076, $19,981). The three complications associated with the lowest adjusted cost were UTI (-$372, 95% CI -$1,336, $592), superficial SSI ($2,473, 95% CI -$256, $5,201), and VTE ($7,909, 95% CI -$17,903, $33,721). For the two intermediate outcomes assessed, the adjusted cost was $8,020 (95% CI $4,597, $11,444) for readmission and $20,258 (95% CI $13,537, $26,978) for return to the OR. In the colon surgery only model, the adjusted cost of anastomotic leak was $10,195 (95% CI $5,941, $14,449), while the postoperative ileus cost was $10,205 (95% CI $6,259, $14,149).
Sensitivity analyses were performed to test the robustness of our main findings to differences in our analytic approach. First, an alternative modeling approach using gamma regression revealed qualitatively similar results for some but not all complications. However, the estimates were unstable in the gamma regression models. Since, the total cost distribution demonstrated a minimal rightward skew, we therefore choose to use median regression. In separate sensitivity analyses, after excluding emergency cases, there were no qualitative differences in risk adjusted costs (Table 4). Finally, risk adjusted costs when focused just on general and colorectal procedures were also similar, with one notable exception (Table 4). Specifically, return to the OR was 20,258 (95% CI 13,537, 26,978) in the overall model and $33,818 (95% CI 29,694, 37,941) for general surgery and colorectal procedures. Models which included clinically relevant interaction terms were either not significant or resulted in qualitatively similar findings, and therefore are not reported.
Table 4.
Sensitivity analyses of the risk adjusted total thirty-day costs after excluding emergency cases or including only general and colorectal procedures.
| Median Regression without emergency procedures | Median Regression with general surgery and colorectal procedures only | |||||
|---|---|---|---|---|---|---|
| Complication | Adjusted Cost ($) | 95% Confidence Interval | P value | Adjusted Cost ($) | 95% Confidence Interval | P value |
| Prolonged ventilation | 48,263 | 38073, 58452 | <0.001 | 45,312 | 41353, 49272 | <0.001 |
| Unplanned intubation | 26,732 | 15154, 38310 | <0.001 | 20,550 | 20107, 20994 | <0.001 |
| Cardiac event | 15,056 | 13183, 16930 | <0.001 | 14,222 | 6468, 21977 | <0.001 |
| Renal failure | 18,555 | 17184, 19926 | <0.001 | 23,154 | 15241, 31068 | <0.001 |
| Pneumonia | 9,374 | 5901, 12847 | <0.001 | 3,536 | −7270, 14342 | 0.521 |
| Sepsis | 8,320 | 5361, 11279 | <0.001 | 13,033 | 9905, 16160 | <0.001 |
| VTE | 7,926 | −13649, 29502 | 0.471 | 13,214 | 6860, 19567 | <0.001 |
| Deep/Organ Space SSI | 12,148 | 8600, 15696 | <0.001 | 14,315 | −11272, 39902 | 0.273 |
| UTI | (346) | −715, 22 | 0.066 | (43) | −1703, 1618 | 0.96 |
| Superficial SSI | 2,462 | −172, 5095 | 0.067 | 2,477 | 1653, 3301 | <0.001 |
| Readmission* | 8,024 | 4369, 11679 | <0.001 | 7,225 | 2935, 11514 | 0.001 |
| Return to OR* | 20,258 | 13570, 26946 | <0.001 | 33,818 | 29694, 37941 | <0.001 |
| Anastomotic leak† | 10,389 | 4491, 16288 | 0.001 | 10,195 | 5942, 14449 | <0.001 |
| Postoperative Ileus† | 10,187 | 7540, 12835 | <0.001 | 10,205 | 6260, 14149 | <0.001 |
Estimated from separate models that did not include other complications
Colectomy only N=607
VTE, venous thromboembolism; SSI, surgical site infection; UTI, urinary tract infection
DISCUSSION
Using data from different hospital types, we sought to estimate the actual cost of individual 30-day complications after a wide range of surgical procedures. The complications with the highest risk-adjusted cost per complication were associated with organ dysfunction: prolonged ventilation, unplanned intubation, renal failure, and cardiac event. Other complications, which occurred with considerably higher frequency, were associated with moderate cost per event and included readmission, return to the OR, anastomotic leak, and postoperative ileus. To our knowledge, this is the most comprehensive, generalizable assessment of the actual hospital costs of individual postoperative complications to date and uniquely enables hospitals to estimate the return on investment of their quality improvement efforts.
A number of prior studies have attempted to evaluate the cost of complication. In a study that used ACS NSQIP data merged with institutional direct cost data, Dimick and colleagues estimated cost of complications in categories (e.g., infectious, cardiovascular, respiratory).26 This study reported the highest complication cost was associated with respiratory events (adjusted cost, $52,466). Similar to our study, their analysis adjusted for patient, procedure, and postoperative factors including the occurrence of other complications. However, individual complications were not assessed, precluding estimating cost savings of targeted efforts which focus on the structural and process of care related to an individual postoperative complication event. In a more recent study by Healy et al., complication costs were estimated based on total direct and indirect costs of care at a single academic medical center.27 They report complications increased the cost of care, on average, by nearly $20,000. Seven individual complications were assessed. In addition, their focus was on relative changes in the hospital profit margin rather than absolute estimated costs of these complications. While instructive, this study was from a single academic medical center limiting a broader generalizability.
In our study, we comprehensively evaluated the unadjusted and adjusted cost of 14 complications using high-quality clinical complication data and total internal cost data from four diverse hospitals. When considered as individual events, we found that several complications were associated with a substantial cost burden. These high-cost complications represented organ dysfunction events, particularly respiratory events, as demonstrated by Dimick et al.26 For example, the most costly complication was prolonged ventilation which was associated with a median adjusted cost of $48,168. This is nearly 2-fold higher when compared to the next most costly complication (unplanned intubation, $26,718), and 5-fold higher than others (e.g., pneumonia, $9,401).
When assessing complication cost in the context of their overall financial burden, it is important to also consider their frequency and potential preventability. Low cost but frequent, modifiable events may represent a more important target for hospital quality improvement efforts. For example, one of the most common complications after colon surgery is postoperative ileus. Although this event was associated with a lower adjusted cost ($10,205), it occurred in nearly 1 in 5 colon resection patients in our study. Prior work has demonstrated that enhanced recovery pathways can greatly reduce colorectal complications such as ileus.28 Therefore for a hospital commonly performing colon surgery, ileus may represent a more impactful cost saving target while other centers may have high rates of postoperative pneumonia and tailored process improvement interventions focused on pulmonary events could be an appropriate focus.
Additional work has reported costs of complications estimated based on hospital charges or insurance payments. Charges have little association with actual costs of care.29 Insurance reimbursement is widely variable depending on payer and hospital-specific negotiations, limiting its utility in cost analyses.29 Medicare payments are additionally limited in that they are based on DRGs, blunting the relationship between individual complications and costs.30 It is also important note that cost data is not publicly available. Increased transparency and reporting of healthcare costs at individual hospitals would allow patients, providers, and payers the ability to better understand and select high value hospitals.
There are several limitations to consider. First, there is unclear generalizability to other hospitals as our data may be unique to region and specific hospital cost structure. We attempted to address this limitation by including a diverse group of hospital types from different regions including small community, comprehensive community and a single large academic medical center. Second, methodological challenges exist when assigning cost to individual events as complications are often not mutually exclusive. We attempted to account for this by adjusting for all complications in the same model and estimating their individual contribution to the cost of care. We also addressed complications which may represent the same postoperative event by performing separate analyses, (e.g., reoperation, readmission). Furthermore, the complications included in the same median regression model are additive. Third, this study only addressed hospital-based costs, not costs that may have occurred outside that specific hospital. Costs of care outside the index hospital is relatively infrequent and unlikely to bias our results.31,32 Nevertheless, including any external costs would only increase the magnitude of our findings and therefore, our results can be considered conservative estimates. Fourth, this study does not evaluate hospital level variation in complication costs. This may bias the finding in the direction of the higher volume facilities which include the 894-bed academic medical center and the 392-bed comprehensive community hospital. Finally, there are substantial secondary costs associated with regulatory constraints, for example value-based purchasing evaluations, of certain complications (e.g., catheter associated urinary tract infections). These efforts are not captured by our data.
Conclusion
In one of the first studies to detail actual cost of individual complications, we found organ failure or dysfunction events were uncommon but associated with the greatest cost per event. More frequent, potentially preventable events, such as ileus and SSI, were associated with lower cost per event. This study defines a replicable methodology which can be efficiently implemented in other hospitals for useful local estimates of the cost of complications. These data can be used to estimate the financial benefit of engaging in local quality improvement initiatives and participation in registries and quality improvement collaboratives.
Funding:
The authors report no conflicts of interest, financial or otherwise, related to this work. This work is supported by the Agency for Healthcare Research and Quality (R01HS024516) and Blue Cross Blue Shield. RPM is supported by the Agency for Healthcare Research and Quality (K12HS026385) and an Institutional Research Grant from the American Cancer Society (IRG-18-163-24).
REFERENCES
- 1.Health Care Cost Institute. 2017. Health Care Cost and Utilization Report (February 2019). Available at: https://www.healthcostinstitute.org/images/easyblog_articles/276/HCCI-2017-Health-Care-Cost-and-Utilization-Report-02.12.19.pdf. Last accessed 2/23/2020.
- 2.Munoz E, Munoz W 3rd, Wise L National and surgical health care expenditures, 2005–2025. Ann Surg. 2010;251(2):195–200. [DOI] [PubMed] [Google Scholar]
- 3.Eappen S, Lane BH, Rosenberg B, et al. Relationship between occurrence of surgical complications and hospital finances. JAMA. 2013;309(15):1599–1606. [DOI] [PubMed] [Google Scholar]
- 4.Vonlanthen R, Slankamenac K, Breitenstein S, et al. The impact of complications on costs of major surgical procedures: a cost analysis of 1200 patients. Ann Surg. 2011;254(6):907–913. [DOI] [PubMed] [Google Scholar]
- 5.Jean RA, Bongiovanni T, Soulos PR, et al. Hospital Variation in spending for Lung Cancer Resection in Medicare Beneficiaries. Ann Thorac Surg. 2019. [DOI] [PubMed] [Google Scholar]
- 6.Bennett KM, Kent KC, Schumacher J, et al. Targeting the most important complications in vascular surgery. J Vasc Surg. 2017;65(3):793–803. [DOI] [PubMed] [Google Scholar]
- 7.Pradarelli JC, Healy MA, Osborne NH, et al. Variation in Medicare Expenditures for Treating Perioperative Complications: The Cost of Rescue. JAMA Surg. 2016;151(12):e163340. [DOI] [PubMed] [Google Scholar]
- 8.NSQIP A User Guide for the 2016 ACS NSQIP Participant Use Data File. 2017; https://www.facs.org/~/media/files/quality%20programs/nsqip/nsqip_puf_userguide_2016.ashx. Accessed March 5, 2019.
- 9.Bilimoria KY, Cohen ME, Merkow RP, et al. Comparison of outlier identification methods in hospital surgical quality improvement programs. J Gastrointest Surg. 2010;14(10):1600–1607. [DOI] [PubMed] [Google Scholar]
- 10.Cohen ME, Dimick JB, Bilimoria KY, et al. Risk adjustment in the American College of Surgeons National Surgical Quality Improvement Program: a comparison of logistic versus hierarchical modeling. J Am Coll Surg. 2009;209(6):687–693. [DOI] [PubMed] [Google Scholar]
- 11.Cohen ME, Ko CY, Bilimoria KY, et al. Optimizing ACS NSQIP modeling for evaluation of surgical quality and risk: patient risk adjustment, procedure mix adjustment, shrinkage adjustment, and surgical focus. J Am Coll Surg. 2013;217(2):336–346 e331. [DOI] [PubMed] [Google Scholar]
- 12.Dimick JB, Ghaferi AA, Osborne NH, et al. Reliability adjustment for reporting hospital outcomes with surgery. Ann Surg. 2012;255(4):703–707. [DOI] [PubMed] [Google Scholar]
- 13.Dimick JB, Osborne NH, Hall BL, et al. Risk adjustment for comparing hospital quality with surgery: how many variables are needed? J Am Coll Surg. 2010;210(4):503–508. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Hall BL, Hamilton BH, Richards K, et al. Does surgical quality improve in the American College of Surgeons National Surgical Quality Improvement Program: an evaluation of all participating hospitals. Ann Surg. 2009;250(3):363–376. [DOI] [PubMed] [Google Scholar]
- 15.Ingraham AM, Richards KE, Hall BL, et al. Quality improvement in surgery: the American College of Surgeons National Surgical Quality Improvement Program approach. Adv Surg. 2010;44:251–267. [DOI] [PubMed] [Google Scholar]
- 16.Khuri SF. The NSQIP: a new frontier in surgery. Surgery. 2005;138(5):837–843. [DOI] [PubMed] [Google Scholar]
- 17.Sellers MM, Merkow RP, Halverson A, et al. Validation of new readmission data in the American College of Surgeons National Surgical Quality Improvement Program. J Am Coll Surg. 2013;216(3):420–427. [DOI] [PubMed] [Google Scholar]
- 18.Merkow RP, Bilimoria KY, McCarter MD, et al. Post-discharge venous thromboembolism after cancer surgery: extending the case for extended prophylaxis. Ann Surg. 2011;254(1):131–137. [DOI] [PubMed] [Google Scholar]
- 19.Merkow RP, Ju MH, Chung JW, et al. Underlying reasons associated with hospital readmission following surgery in the United States. JAMA. 2015;313(5):483–495. [DOI] [PubMed] [Google Scholar]
- 20.NSQIP A User Guide for the 2016 ACS NSQIP Procedure Targeted Participant Use Data File. 2017; https://www.facs.org/~/media/files/quality%20programs/nsqip/pt_nsqip_puf_userguide_2016.ashx. Accessed March 5, 2019.
- 21.Austin PC, Ghali WA, Tu JV. A comparison of several regression models for analysing cost of CABG surgery. Stat Med. 2003;22(17):2799–2815. [DOI] [PubMed] [Google Scholar]
- 22.Malehi AS, Pourmotahari F, Angali KA. Statistical models for the analysis of skewed healthcare cost data: a simulation study. Health Econ Rev. 2015;5:11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Blough DK, Madden CW, Hornbrook MC. Modeling risk using generalized linear models. J Health Econ. 1999;18(2):153–171. [DOI] [PubMed] [Google Scholar]
- 24.Manning WG, Basu A, Mullahy J. Generalized modeling approaches to risk adjustment of skewed outcomes data. J Health Econ. 2005;24(3):465–488. [DOI] [PubMed] [Google Scholar]
- 25.Basu A, Rathouz PJ. Estimating marginal and incremental effects on health outcomes using flexible link and variance function models. Biostatistics. 2005;6(1):93–109. [DOI] [PubMed] [Google Scholar]
- 26.Dimick JB, Chen SL, Taheri PA, et al. Hospital costs associated with surgical complications: a report from the private-sector National Surgical Quality Improvement Program. J Am Coll Surg. 2004;199(4):531–537. [DOI] [PubMed] [Google Scholar]
- 27.Healy MA, Mullard AJ, Campbell DA Jr., et al. Hospital and Payer Costs Associated With Surgical Complications. JAMA Surg. 2016;151(9):823–830. [DOI] [PubMed] [Google Scholar]
- 28.Grass F, Slieker J, Jurt J, et al. Postoperative ileus in an enhanced recovery pathway-a retrospective cohort study. Int J Colorectal Dis. 2017;32(5):675–681. [DOI] [PubMed] [Google Scholar]
- 29.Arora V, Moriates C, Shah N. The Challenge of Understanding Health Care Costs and Charges. AMA J Ethics. 2015;17(11):1046–1052. [DOI] [PubMed] [Google Scholar]
- 30.Services CfMaM. MS-DRG Classifications and Software. 2018; https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/AcuteInpatientPPS/MS-DRG-Classifications-and-Software.html. Accessed February 7, 2019.
- 31.Brooke BS, Goodney PP, Kraiss LW, et al. Readmission destination and risk of mortality after major surgery: an observational cohort study. Lancet. 2015;386(9996):884–895. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Tsai TC, Orav EJ, Jha AK. Care fragmentation in the postdischarge period: surgical readmissions, distance of travel, and postoperative mortality. JAMA Surg. 2015;150(1):59–64. [DOI] [PubMed] [Google Scholar]