Abstract
Aims: The objective of this study was to quantify the treatment costs and revision surgery rates in chronic rhinosinusitis (CRS) patients, with and without nasal polyposis (CRSwNP and CRSsNP), who require treatment with endoscopic sinus surgery (ESS). The additive contributions of nasal polyposis (NP) and revision surgery to 1-year costs were a primary focus.
Materials and methods: Adults (aged 18–64 years) undergoing ESS for CRS in 2012–2015 were identified within the Blue Health Intelligence database and used to estimate revision rates. Patients with ±1 year of enrollment around the index ESS were used to estimate 1-year healthcare expenditures. Revision ESS rates were evaluated via Kaplan-Meier and Cox regression models. Disease-related healthcare and pharmacy expenditures were modeled with generalized linear regression to assess the impact of baseline patient characteristics.
Results: A total of 86,052 patients underwent ESS for CRS (43.5 ± 12.4 years; 49.3% male), and a sub-set of 23,542 patients were available for 1-year healthcare expenditure analysis (44.0 ± 12.1 years; 50.0% male). Revision ESS rates within 1 year were 3.5% in the CRSwNP cohort and 1.6% in the CRSsNP cohort. NP, deviated septum, gender, and region were statistically significant predictors of revision surgery. Mean 1-year treatment expenditures, including the index ESS, were $8,824 for CRSsNP and $11,166 for CRSwNP patients without revision ESS. CRSwNP doubled the risk of revision surgery in the first year after ESS compared with CRSsNP and cost 24% more in the absence of a second procedure. Revision ESS within the first year increased mean 1-year expenditures by $11,150 and $13,139 for CRSsNP and CRSwNP, respectively.
Limitations: The primary limitation was the limited length of follow-up available for estimating revision ESS rates.
Conclusions: In a large commercially insured US population, disease-related expenditures for patients having ESS for CRS are substantial, as are the additive impacts of NP and revision surgery.
Introduction
An estimated 4.9% of the population in the US is afflicted with chronic rhinosinusitis (CRS)Citation1, resulting in an estimated $6.9–$9.9 billion in direct CRS-related healthcare costsCitation2. Within the CRS population, between 19–55%Citation3–8 have nasal polyps (CRSwNP), a condition which has been found to be associated with higher treatment costs and utilizationCitation9. The increased burden in the CRSwNP patients, compared to CRS patients without nasal polyps (CRSsNP), has been attributed to higher rates of related comorbidities, less improvement after surgical treatment, increased symptom severity, and higher objective disease severityCitation10,Citation11. In addition, there is evidence that CRSwNP patients treated with endoscopic sinus surgery (ESS) have higher rates of revision surgery and long-term recurrence of diseaseCitation11–13.
The objective of this study was to quantify the disease-specific cost burden and healthcare utilization in chronic rhinosinusitis (CRS) patients who require treatment with endoscopic sinus surgery (ESS). In particular, we wanted to differentiate the burden for patients with and without nasal polyposis (NP), and with and without revision ESS, in the year following ESS. Since ethmoidectomy is generally advocated in the setting of NP, we chose this as the most appropriate ESS procedure to studyCitation14. We also sought to understand the impact of NP, as well as other patient characteristics, on the rates of revision surgery.
Methods
This was a retrospective study of real world evidence, utilizing a large claims database representative of the insured non-Medicare population in the US. Primary outcomes of interest for this analysis were revision ESS rates and disease-related healthcare expenditures, including prescription medications, within a year after an index ESS surgery.
Inclusion criteria and populations
The population of interest was adults aged 18–64 with an ESS procedure that included ethmoidectomy for CRS, for which the first known ESS visit occurred between 2012 and 2015. In order to synchronize provider and facility claims related to the same admission, but with different dates, we included all claims that overlapped the surgery date, but did not exceed 7 days (6 nights). The population that met these criteria and had at least 1 day of medical enrollment beyond the index ESS visit was called the survival population, and was used for survival models of revision rates. The sub-set of these patients with a full year of both medical and pharmacy insurance enrollment comprised the cost population, and was used for all 1-year cost summaries and models. The CRSwNP cohort included all patients in either population with a diagnosis code of NP at any time between 6 months prior to their index ESS visit through 1 year after the index visit, whereas the CRSsNP cohort did not have an NP diagnosis in this time period.
Data source
The Blue Health Intelligence (BHI) database contains diagnoses, procedures, details on drug prescriptions filled, and all associated expenditures for patients enrolled in Blue Cross Blue Shield health insurance plans. It includes over 10 years of claims experience from over 180 million unique members located within every 5-digit zip code in the US. Eligible patient visits in the BHI database included those between January 1, 2012 and December 31, 2015. The database includes fully integrated, de-identified, individual-level healthcare claims data, including complete payment records. Since the BHI data contains detail for inpatient hospitalization, outpatient visits in any setting of care, and pharmacy claims, including any claims paid under a coordination-of-benefit arrangement, it provides a complete picture of the healthcare resource utilization and expenditures for enrolled patients.
Definitions
Patients having ESS for CRS were defined as those having a paid claim with a Current Procedural Terminology (CPT) code of 31254 or 31255 for partial (anterior) or total (anterior and posterior) ethmoidectomy and an International Classification of Diseases, Ninth Revision (ICD-9) code of 473.xx as the primary diagnosis associated with the visit or as the diagnosis associated with the ESS procedure codeCitation15. The CRSwNP cohort was defined as the sub-set of patients also having an ICD-9 code of 471.xx at any time between 6 months prior to their index ESS through 1 year after discharge, while the complement constituted the CRSsNP cohort.
Disease-related medical expenditures were defined as all healthcare utilization with an associated primary diagnosis of CRS, acute sinusitis (ICD-9 461.xx), or NP. In addition, disease-related prescription drug utilization was identified via therapeutic class, including corticosteroids, antibiotics, nasal agents, and a number of classes associated with related comorbidities or acute illnesses. Many of these types of drugs are typically purchased over the counter, and, thus, would not be included in the calculation of drug expenditures.
Survival models for revision surgery
A Kaplan Meier survival model was utilized to estimate revision surgery rates over time in the ESS population by CRSwNP/CRSsNP cohort. Cox multivariable survival models were then used to explore the impact of additional baseline patient characteristics as predictors of revision surgery. Survival methodology allowed for the capture of revision surgery rates over time, with patients dropping out of the model as their follow-up ended. Patients were censored in these models at the time of their first revision ESS, their last day of medical insurance enrollment, or the end of the study period. All models of revision rates used the survival population, as described above.
Regression model for 1-year disease-related treatment expenditures
Linear regression, with a gamma distribution and log link, was used to model total disease-related 1-year medical and pharmacy expenditures as a function of NP cohort, revision surgery status within 1-year of index ESS, and any additional statistically significant patient characteristics. The log-gamma model takes into account the right-skew of healthcare expenditure distributions without the need for data transformation. The dependent variable was total 1-year medical and pharmacy expenditures, with revision surgery and nasal polyps as primary predictors of interest. Stepwise selection was used to identify any additional explanatory variables that were statistically significant from among patient demographics and comorbidities, region, year, and key clinical outcomes. The cost population, having at least 1 year of medical and pharmacy enrollment pre- and post-index, was used for expenditure models.
All statistical analyses in this study were performed using SAS software, Version 9.2 (SAS Institute, Inc., Cary, NC). All data used to perform the statistical analysis were de-identified and accessed in compliance with the Health Insurance Portability and Accountability Act.
Results
A total of 88,850 patients aged 18–64 had a paid claim of ESS for CRS without any documented prior ESS during their enrollment with BHI. A small number of patients were excluded for having overlapping visits extending beyond 7 days at the index visit (2.0%) or for not having medical coverage extending at least 1 day beyond the index visit (1.1%), leaving 86,052 patients in the survival population for modeling revision ESS rates. The sub-set comprising the cost population, with a full year of both medical and pharmacy insurance coverage after their index ESS, included 23,542 patients that were used for all 1-year cost reporting and modeling.
Baseline characteristics and procedures
The patient population comprised an equal mix of males and females, with age evenly distributed across the inclusion range of 18–64 (mean 44 years), and index ESS procedures also distributed evenly across the study timeframe of 2012–2015. Mean medical plan enrollment after index ESS was 18.1 months in the survival population, allowing a sufficient sample size to follow revision rates through 3 years post-index. Mean total medical coverage in the same population, including pre-surgery enrollment, was 4.4 ± 2.1 years ().
Table 1. Patient characteristics.
Approximately 32% of the patients had a NP diagnosis recorded at the index ESS visit (), while the CRSwNP cohort encompassed 41% of the population due to including patients with an NP diagnosis recorded at any time within an 18-month period.
All patients had CRS recorded at the time of the index ESS visit, as this was a requirement for inclusion in the study. Deviated septum (60%) and turbinate hypertrophy (58%) were the most common secondary diagnoses that were not required, followed by NP and disease of the nasal cavity not otherwise classified (32% each) (). Note that, in a claims database, conditions that are less relevant to the surgery may be under-reported at the surgery visit and, thus, are not necessarily reflective of their prevalence over longer timeframes.
Partial and total ethmoidectomy procedures were recorded in 29.8% and 76.4% of the index ESS visits, with a small percentage of patients having both codes in their records. In addition, multiple concomitant surgical procedures were common, with the most prevalent non-sinus procedure being septoplasty (55.7%) (). The majority (91.3%) of patients with a deviated septum diagnosis at index had a concurrent septoplasty procedure code, as did 11.8% of the patients without a record of the diagnosis.
Table 2. Prevalent concomitant procedures at index ESS (survival population, n = 86,052).
Revision surgery
The Kaplan-Meier estimates of 1-year revision ESS rates were 2.4% overall and were approximately doubled in the CRSwNP cohort compared with the CRSsNP cohort (3.5% vs 1.6%). Cumulative revision rate estimates continued to rise linearly in both cohorts through 3 years beyond the index ESS. The higher rate of revision in the CRSwNP vs CRSsNP cohort was highly significant (p < .0001). These rates are based on the time to a patient’s first revision following the index ESS, and as such do not take into account patients who may have had multiple revision surgeries ().
Figure 1. Cumulative revision surgery rate estimates (survival population, n = 86,052).
The Cox regression model results were similar with respect to highlighting the significance of NP in estimating revision rates (p < .0001, HR = 2.1). In addition, they showed that patients with a deviated septum were less likely to need revision surgery (p < .0001, HR = 0.6), while females had an increased likelihood (p < .0001, HR = 1.2). Geographical differences were also significant, as measured by US census region (p < .0001) (). Asthma was a significant predictor of revision surgery in preliminary statistical models, with a hazard ratio of ∼1.2, but was not included in the final model, because it did not meet the proportional hazards assumption required for Cox regression.
Table 3. Cox model results for revision surgery (survival population, n = 86,052).
Expenditures
Mean 1-year disease-related medical and pharmacy expenditures were $9785 across the entire cost population. Individual patient expenditures were most highly impacted by whether or not the patient had a revision surgery within the year and by whether they were in the CRSwNP or CRSsNP cohort. These two factors were inter-related due to the higher rate of revision surgery in the CRSwNP cohort, but NP was also an independent predictor of increased expenditures, as seen by a $2088 differential in CRSwNP vs CRSsNP patients without revision surgery ().
Table 4. One-year healthcare expenditures (cost population, n = 23,542).
Due to the large sample size, regression models showed a number of significant predictors of 1-year disease-related expenditures. The factors with the greatest impact on expenditures were revisions within the year, US census region, and CRSwNP vs CRSsNP cohort (p-values <.0001). The most significant comorbid conditions were asthma, cystic fibrosis, and deviated septum (p-values <.0001). Additional predictors, all significant at a level of 0.01 or below, and listed in order of statistical significance, were type of insurance, index ESS admission year, hypertension, age, disease of the nasal cavity, hyperlipidemia, allergic rhinitis, diabetes, tobacco use, turbinate hypertrophy, and allergy to analgesic.
Discussion
This study of patients having ESS for CRS in a large insurance database estimated revision rates and elucidated the contributions of baseline patient characteristics to the need for revision as well as to 1-year treatment costs. Estimated revision rates at 1 year, based on a large population of 86,052 patients, were more than double in CRSwNP vs CRSsNP patients (3.5% and 1.6%). A similar study based on a large retrospective database of California patients having a wider range of endoscopic sinus surgery procedures between 2005–2011 reported a raw 1-year revision rate of 2.86% and a 5-year estimate of 8.6% overall, with 9.9% vs 7.9% in CRSwNP and CRSsNP patients, respectivelyCitation13. Another study of Canadian patients with maxillary anstrostomy plus ethmoidectomy procedures showed a 5-year revision rate of 10.6%Citation16. If the cumulative revision rates seen in our study continued to rise linearly beyond our 3-year estimates, the 5-year rate for our study population would be ∼9.3%, suggesting that the differences in these study designs and populations had only a small effect on overall revision rates. We did observe a larger difference in revision rates between the CRSwNP and CRSsNP cohorts than the California study, but the magnitude of this difference very closely mirrored the odds ratio of 2.07 reported for NP as a predictor of revision surgery in the Canadian study.
Additional reports of revision rates in CRS populations with ESS, or in CRSwNP sub-sets, are very similar magnitude to those found in this analysisCitation12,Citation17. Moreover, in a study of 118 patients with severe NP (stage 3 or 4 with a Lund-McKay score ≥16) and median follow-up of 40 months, 47% were advised to undergo revision surgery and 27% had the revision surgery, illustrating a further increase in revision rate with disease severity, the clinical details of which we were unable to discern with claims dataCitation18. Studies that include a mix of simple polypectomy and ESS procedures also report higher revision ratesCitation19,Citation20.
In addition to the impact of NP on revision rates, other factors were found to be influential. While we observed lower revision rates for patients with a diagnosis of deviated septum, Rudmik et al.Citation16 reported a similar result based on concurrent septoplasty, which was highly correlated with the diagnosis in our population. Another study reported a similar correlation between the presence of septal deviation accompanied by septoplasty and improvement in patient reported outcomes after ESS (p = .004)Citation21. We also found that females had an increased likelihood (p < .0001, HR = 1.2) of revision, which is consistent with the results of a logistic regression model (OR =1.20, p < .001) reported by Stein et al.Citation13. Finally, several studies have reported significant differences in revision rates based on geography, with higher differences seen between individual US states and between regions within Alberta, Canada than we saw in this study, based on the larger and more diverse US Census RegionsCitation22,Citation23.
Limitations and strengths
The primary limitation of this study is that the data spans only 4 years, and, thus, the patients’ CRS treatment histories are unknown outside of the 2012–2015 timeframe. The data does provide a very comprehensive picture of a patient’s total healthcare utilization, including prescription medications, while they are enrolled in the contributing insurance plan. However, we cannot rule out ESS surgery prior to a patient’s enrollment period, nor can we estimate revision surgery rates beyond 3 years.
There is a risk in any study of the results being confounded by unmeasured variables, and claims data in particular is limited with respect to clinical detail such as disease severity scores, patient-reported outcomes, lab results, non-prescription (over-the-counter) drug use, and lifestyle factors. Also, because this data includes only commercially insured patients and their dependents, it is not representative of Medicare and Medicaid populations.
A key strength of this study is the very large sample size, which allows for small standard errors of the reported population means. In addition, the data represents “real-world” healthcare utilization, providing a realistic picture of the US insured population without the strict population limitations typically imposed by many clinical trials.
Conclusions
In a large commercially insured US population, the disease-related expenditures for patients having ESS for CRS are substantial, as are the additive impacts of NP and revision surgery. CRSwNP patients are approximately twice as likely as those without NP to require revision ESS in any time period within 3 years of the index ESS. CRSwNP patients incur expenditures $2,342 higher than CRSsNP patients in the first year after index ESS. Although the higher cost associated with NP is partly due to higher revision rates, 1-year expenditures are also $2,088 higher for CRSwNP patients vs CRSsNP patients without revision surgery.
Transparency
Declaration of funding
This study was funded by Intersect ENT.
Declaration of financial/other relationships
TDH is an employee of CTI Clinical Trial and Consulting Services Inc., which is a consultant to Intersect ENT. ASD is a consultant to Intersect ENT, Olympus, Stryker Endoscopy, and Optinose. RPM reports research grants from Intersect ENT, Medtronic, Pfizer, Sanofi, and Genentech. JME peer reviewers on this manuscript have no relevant financial or other relationships to disclose.
Acknowledgments
None reported.
References
- Bhattacharyya N. Incremental health care utilization and expenditures for chronic rhinosinusitis in the United States. Ann Otol Rhinol Laryngol 2011;120:423–7
- Smith KA, Orlandi RR, Rudmik L. Cost of adult chronic rhinosinusitis: a systematic review. Laryngoscope 2015;125:1547–56
- Benninger MS, Holy CE. The impact of endoscopic sinus surgery on health care use in patients with respiratory comorbidities. Otolaryngol Head Neck Surg 2014;151:508–15
- Bhattacharyya N. Clinical and symptom criteria for the accurate diagnosis of chronic rhinosinusitis. Laryngoscope 2006;116(7Pt2Suppl110):1–22
- Mascarenhas JG, da Fonseca VM, Chen VG, et al. Long-term outcomes of endoscopic sinus surgery for chronic rhinosinusitis with and without nasal polyps. Braz J Otorhinolaryngol 2013;79:306–11
- Pearlman AN, Chandra RK, Chang D, et al. Relationships between severity of chronic rhinosinusitis and nasal polyposis, asthma, and atopy. Am J Rhinol Allergy 2009;23:145–8
- Soler ZM, Wittenberg E, Schlosser RJ, et al. Health state utility values in patients undergoing endoscopic sinus surgery. Laryngoscope 2011;121:2672–8
- Tran KN, Beule AG, Singal D, et al. Frontal ostium restenosis after the endoscopic modified Lothrop procedure. Laryngoscope 2007;117:1457–62
- Benninger MS, Holy CE. Endoscopic sinus surgery provides effective relief as observed by health care use pre- and postoperatively. Otolaryngol Head Neck Surg 2014;150:893–900
- Batra PS, Tong L, Citardi MJ. Analysis of comorbidities and objective parameters in refractory chronic rhinosinusitis. Laryngoscope 2013;123(Suppl7):S1–S11
- Deal RT, Kountakis SE. Significance of nasal polyps in chronic rhinosinusitis: symptoms and surgical outcomes. Laryngoscope 2004;114:1932–5
- Koskinen A, Salo R, Huhtala H, et al. Factors affecting revision rate of chronic rhinosinusitis. Laryngoscope Investig Otolaryngol 2016;1:96–105
- Stein NR, Jafari A, DeConde AS. Revision rates and time to revision following endoscopic sinus surgery: a large database analysis. Laryngoscope. 2018;128:31–36
- Orlandi RR, Kingdom TT, Hwang PH, et al. International Consensus Statement on allergy and rhinology: rhinosinusitis. Int Forum Allergy Rhinol 2016;6(Suppl1):S22–S209
- Macdonald KI, Kilty SJ, van Walraven C. Development and validation of an administrative data algorithm to identify adults who have endoscopic sinus surgery for chronic rhinosinusitis. J Otolaryngol Head Neck Surg 2017;46(1):38
- Rudmik L, Xu Y, Alt JA, et al. Evaluating surgeon-specific performance for endoscopic sinus surgery. JAMA Otolaryngol Head Neck Surg 2017;143:891–8
- Mendelsohn D, Jeremic G, Wright ED, et al. Revision rates after endoscopic sinus surgery: a recurrence analysis. Ann Otolo Rhinol Laryngol 2011;120:162–6
- Wynn R, Har-El G. Recurrence rates after endoscopic sinus surgery for massive sinus polyposis. Laryngoscope 2004;114:811–13
- Hopkins C, Slack R, Lund V, et al. Long-term outcomes from the English national comparative audit of surgery for nasal polyposis and chronic rhinosinusitis. Laryngoscope 2009;119:2459–65
- Philpott C, Hopkins C, Erskine S, et al. The burden of revision sinonasal surgery in the UK-data from the Chronic Rhinosinusitis Epidemiology Study (CRES): a cross-sectional study. BMJ Open 2015;5:e006680
- Smith TL, Mace JC, Rudmik L, et al. Comparing surgeon outcomes in endoscopic sinus surgery for chronic rhinosinusitis. Laryngoscope 2017;127:14–21
- Rudmik L, Holy CE, Smith TL. Geographic variation of endoscopic sinus surgery in the United States. Laryngoscope 2015;125:1772–8
- Rudmik L, Bird C, Dean S, et al. Geographic variation of endoscopic sinus surgery in Canada: an Alberta-based small area variation analysis. Otolaryngol Head Neck Surg 2015;153:865–74