Hospital charges associated with “never events”: comparison of anterior cervical discectomy and fusion, posterior lumbar interbody fusion, and lumbar laminectomy to total joint arthroplasty

patients without such complications across 5 common spinal and orthopedic procedures. Cost projections by health care providers will need to incorporate expected costs of added care for patients experiencing such complications, assuming that the cost burden of such events continues to shift from payers to providers.

O n October 1, 2007, the Centers for Medicare and Medicaid Services (CMS) required all hospitals to report both primary and secondary diagnoses when submitting discharge claims. By July 31, 2008, CMS designated 8 specific hospital-acquired adverse events as "never events," which should thus never occur following spinal surgery (https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/HospitalAcqCond/index.html). Three months later, reimbursement restrictions were added for catheter-associated urinary tract infections, certain surgical site infections (SSIs), and manifestations of poor glycemic control. 2 The following year, CMS expanded the ruling to include deep vein thrombosis (DVT) and pulmonary embolism (PE) following total joint arthroplasty ( Fig.  1). 6,[9][10][11]15 Although CMS has yet to expand the definition of never event to DVT/PE following spinal surgery, new rulings such as the addition of "radiological never events" in 2011 are evidence of the ongoing expansion of the program.
Due to the never event ruling, CMS as well as some private insurers insist that the costs of managing these complications be borne by hospitals and health care providers, rather than billings to health care payers for additional care required in their management. 9 As payers continue to shift financial responsibility for these events to hospital providers through bundled payments and populationbased contracting, detailed data are necessary to evaluate the incidence and cost of such adverse events.
Thus far, data regarding the expected costs of never events following surgical procedures have been limited. 6,8,10,15 In the present study, we used a billing database to compare the relative charges for patients experiencing SSI, DVT, and PE following anterior cervical discectomy and fusion (ACDF), posterior lumbar interbody fusion (PLIF), lumbar laminectomy (LL), total knee replacement (TKR), and total hip replacement (THR).

Patient Selection
The California State Inpatient Database (CA-SID) from 2008 to 2009 was used for analysis. Determining admissions to hospitals in the state of California using the discharge data from the CA-SID between the years 2008 and 2009 was performed. This database is a subset of the Healthcare Cost and Utilization Project (HCUP) from the Agency for Healthcare Research and Quality. 14 HCUP records statewide inpatient data, ambulatory surgery, and services data, as well as emergency department data. While California reports all 3 through its CA-SID, only inpatient data were used in the analysis.
CA-SID is an all-payer administrative database including 100% of the state's inpatient discharge records from nonfederal hospitals. Each CA-SID record includes: 1) basic demographic information, including age, sex, race, and primary payer; 2) up to 30 International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) diagnosis codes; 3) up to 21 ICD-9-CM procedure codes; and 4) outcome information, such as discharge disposition, length of stay, and inpatient charges. This database does not include outpatient charges such as diagnostic studies performed as an outpatient or outpatient physical therapy, nor does it include a breakout of direct versus indirect costs or other detailed analysis of cost assessment.
Patients with more than 1 of the examined primary procedure codes were excluded from the final study population with the exception of those with codes for both PLIF and LL, who were subsequently included under the PLIF analysis. Additionally, patients lacking total charge data were excluded from final analyses.

Cost Analysis
Charges for treatment and subsequent inpatient hospitalization from primary surgery through 3 months postoperatively were calculated in patients with uncomplicated procedures and compared with those complicated with DVT, PE, or SSI. Charges were adjusted for inflation to 2015 US dollars utilizing the Consumer Price Index inflation calculator provided by the Bureau of Labor Statistics. 11

Results
A total of 157,431 patients were identified as having 1 index procedure. Of these, 136,371 patients had valid total charges, representing the final study population. There were 15,729 patients with ACDF, 19,706 patients with PLIF, 15,217 patients with LL, 56,467 patients with TKR, and 29,252 patients with THR (Table 1).
In this study population, 130,670 patients did not suffer any of the examined complications up to 3 months postoperatively. In contrast, 1510 patients experienced DVT as the only complication (1.1%), 350 patients had only PE (0.25%), and 591 patients had only SSI (0.43%).
The incidence of the examined adverse events was lowest for ACDF (0.6% DVT, 0.1% PE, and 0.03% SSI;

Discussion
Hospital charges associated with a single complication following spinal surgery are substantial, with an increase in charges ranging from 1.7 to 4.3 times greater than those of patients without a complication. Although CMS has already determined that SSI is a never event following spinal and joint arthroplasty procedures, they have not yet expanded DVT/PE following spinal surgery into the never event category. The results of this study provide data regarding the projected economic repercussions if CMS or private payers determine that DVT/PE fit into their definition of a never event following ACDF, LL, and PLIF.
This investigation revealed that SSI/DVT/PE occurred in 1.8% of all procedures analyzed and resulted in a substantial increase in the cost of care of up to 4.3 times the charges of an uncomplicated case. Additionally, the charges associated with SSI/DVT/PE following spine surgery appear to be even higher than those associated with TKR and THR. To the best of our knowledge, this is the first report analyzing the cost of these specific adverse events following spinal and orthopedic surgery procedures.
The high costs associated with adverse events places a burden on the health care system, and the debate about who should bear the burden of paying for such costs is controversial. As payers continue to shift responsibility for these events to hospitals and physicians through bundled payments and population-based contracting, additional financial stress on hospitals and providers is occurring. Although it is not clear at this time whether this trend will continue, data projecting cost effects are potentially helpful to all entities engaged in health care provision.
A previous investigation similarly attempted to examine the effects of the never event ruling on neurological surgery procedures. Teufack et al. 15 performed a prospective investigation revealing a future reduction in physician and facility reimbursement at a tertiary care center with up to an 8.9% decrease in reimbursement for physicians due to never events. The concern is a decreased ability to care for complex patients due to economic constraints, further restricting care for those most in need.
An investigation of never events in otolaryngology surgery patients undergoing radical neck dissection found a prevalence of never events ranging from 0.2% to 5.0%. They also revealed that the occurrence of never events was associated with an increase in hospital length of stay of 5.6 days and additional charges of $49,153 compared with hospitalizations without a never event. 8 A study of patients following radical cystectomy revealed a never-event rate of 2.42%. Vascular-catheter infections (1.25%) were the most common. Black race and comorbidities increased the likelihood of a never event. The presence of any never event increased the average length of stay by 15 days, total costs by $37,000, and in-hospital mortality by 5.8%. 6 Another consideration is the potential disparity between the economic and clinical implications of complications. 1,5 While this study and others have examined the detrimental economic effects of complications, 6,8,15 numerous studies have also examined the clinical effects of these complications. 4,7,12 In many cases, the economic effects of a complication may be substantial, while the clinical effects on final outcome may be limited, such as in the case of DVT following ACDF. Furthermore, the complications examined in this investigation will continue to inevitably occur in some patients and should not be referred to as never events. Withholding reimbursement following the occurrence of a complication must be carefully considered by both payers and caregivers to ensure that financial risks are balanced fairly, especially for institutions taking care of higher risk populations.
This study has several limitations. The retrospective study design may be viewed as a limitation; however, the size and robust nature of the database lend to the strength of the data. Additionally, the CA-SID may not be gener-alizable to all spinal procedures in the US, as there are large geographic variations in spine surgical procedures and in the use of fusion surgery. 3,13,14,16 In addition, given the design of this database, it is not possible to include costs associated with outpatient management of complications, such as outpatient antimicrobial therapy, international normalized ratio levels, and physical therapy. This may result in underestimation of the total costs associated with managing the studied complications.
Furthermore, an explanation of why DVT has highercost spine procedures compared with arthroplasty procedures is not available with this analysis, although increased vigilance by spine surgeons for bleeding complications such as epidural hematomas that may require prolonged hospitalization or placement of inferior vena cava filters provides a possible explanation. Furthermore, why SSI was comparatively costly for patients with LL or why DVT costs more than PE in patients undergoing spine surgery remains unanswered with the current analysis. One possible explanation for SSI costs would be wound vac therapy, which may be used more commonly for lumbar spine wounds compared with cervical or extremity wounds. More specific analysis of the variability between charges associated with adverse events across procedures would require a more detailed clinical database. HCUP, which provided data for this investigation, collects overall hospital charges as reported by the hospitals themselves to the collecting agency without a breakdown of diagnosis-related group-and current procedural terminology-based fees or direct versus indirect costs.
In addition to these limitations, nearly 15% of cases were excluded from the analysis due to lack of total charge information. This is at least partly attributable to patient admissions to Kaiser Permanente hospitals where the charges are not required to be submitted for database collection. We also did not assess other never events, such as falls, medication errors, and wrong site surgery.