*Seth E. Pross, Andrew Layton, Kuo Tong, and *Lawrence R. Lustig

Otology & Neurotology 00:00Y00 Ó 2013, Otology & Neurotology, Inc. Cost of Placement and Complications Associated With Osseointegrated Bone-Conducting Hearing Prostheses: A Retrospective Analysis of Medicare Billing Data *Seth E. Pross, Andrew Layton, Kuo Tong, and *Lawrence R. Lustig *Department of OtolaryngologyYHead and Neck Surgery, University of California San Francisco; and ÞQuorum Consulting, San Francisco, California, U.S.A. Objective: To report the cost of placement and complications related to bone-conducting hearing prostheses (OBHPs) in a Medicare population. Materials and Methods: We performed a retrospective analysis of nationwide Medicare claims data for operative and nonoperative complications associated with the placement of percutaneous OBHPs between the first quarter of 2007 and the second quarter of 2009 for which there were 6 subsequent quarters of follow-up. We used Medicare Standard Analytical Files (SAF), which contain a 5% random sample of Medicare fee-for-service beneficiaries, excluding those that also were enrolled in a managedcare organization. Results: We identified 118 patients who had OBHPs placed in the requisite period. Their complication billing data were analyzed for the six-quarters after initial placement. Seventy patients (59%) had no billing codes for complications or repeat procedures after receiving the implant, whereas 48 patients (41%) had such codes. The total adjusted mean cost with repeat/revision operations or complications was $7,812 per patient compared with $6,733 for those without these issues, an increase of $1,079 or 16%. Discussion: We estimate that complications associated with the implantation of percutaneous OBHPs led to $417,616 in additional costs in the entire Medicare fee-for-service population during the study period and that the total cost of placement of these devices together with the cost of their complications totaled $6,789,248. In conclusion, the Medicare SAF database suggests that complications associated with OBHP increased the overall cost of placement by 16%. Like all surgical procedures, these complications and their associated costs should be taken into account when considering treatment options for patients who experience hearing loss. Key Words: Bone- Anchored Hearing AidVComplicationsVCostVOsseointegrated hearing device. Otol Neurotol 00:00Y00, 2013. Osseointegrated bone-conducting hearing prostheses (OBHPs) are very effective tools to treat patients who have conductive or mixed hearing loss but who cannot be effectively rehabilitated using standard surgical techniques or conventional hearing aids. Patients with chronically draining ears, canal atresia, congenital defects, surgically absent lateral temporal bone, and single-sided deafness may all fall into this category. In addition, these Address correspondence and reprint requests to Lawrence R. Lustig, M.D., Department of OtolaryngologyYHead and Neck Surgery, San Francisco, CA; E-mail: llustig@ohns.ucsf.edu Conflict of interest and source of funding: A. L. and K. T. are employees of Quorum Consulting, Inc., an independent consulting group. Sonitus Medical (San Mateo, CA, USA) provided funding to Quorum Consulting, Inc., for access to and assistance with the analysis of the data from the SAF database. S. P. and L. R. L. received no financial support from Sonitus or Quorum and retained full editorial control of all aspects of this research article. Sonitus Medical Corp. provided no editorial input into this work. Editorial input from Quorum was limited to data access and interpretation only. Supplemental digital content is available in the text. devices are commonly used in patients with unilateral sensorineural deafness (1,2). The most commonly used OBHP devices (the Bone- Anchored Hearing Aid, or BAHA, Cochlear Corp; Sydney, Australia; and Ponto, Oticon Medical; Copenhagen, Denmark) transmit sound from a processor percutaneously through a titanium post that is in the temporal bone. Multiple previous studies have shown that patients experience improved hearing and quality of life with these devices (2Y4). However, the devices are also associated with potential complications. A majority of the reported complications associated with these OBHP implants are due to the skin-implant interface. The most common complications identified in previous studies include skin reactions, implant infections, soft tissue overgrowth of the abutment, failure to osseointegrate, the need for revision surgery, loss of the implant, and abandonment of the implant (1,2,5Y11). There are newer devices that transmit sound transcutaneously, avoiding problems associated with the percutaneous abutment and 1

2 S. E. PROSS ET AL. their associated complications (12,13). However, because these later devices are relatively new, few long-term data are available on their efficacy and complication rates. The goal of this study is to analyze the overall costs that are associated with complications from percutaneous OBHP placement. To accomplish this, we estimated these costs by analyzing a large, nationwide database of claims data from a Medicare population, which includes a 5% random sample of beneficiaries claims of U.S. citizens aged 65 years or older and those with disability. This database has been used extensively as a comprehensive snapshot of health-care use and expenditures among the elderly in the United States in many fields of medicine ranging from cardiac surgery to internal medicine (14Y16). Our analysis was confined to percutaneous OBHP devices because long-term data only exists for these devices. From this database, we estimate that complications associated with percutaneous OBHP placement results in 16% added costs over an 18-month period after surgery in the Medicare population. MATERIALS AND METHODS Data Source This was a retrospective, longitudinal study using Medicare Standard Analytic Files (SAFs) provided by the Centers for Medicare and Medicaid Services. The SAFs contain a 5% random sample of beneficiaries claims, which includes U.S. citizens aged 65 years or older, those eligible for disability, and/or subjects with end-stage renal disease. Excluded are beneficiaries who are also enrolled in a managed-care organization. The SAFs are constructed from weekly data submissions to the National Claims History 100% Nearline File. The SAF database includes medical claims submitted for laboratory tests, inpatient hospital stays, outpatient care, physician care, skilled nursing facility care, home health care, durable medical equipment use, hospice care, and a denominator file that includes beneficiaries demographic characteristics (age, sex, race, etc.). In this study, quarterly beneficiary claims data from the different claims setting were linked via encrypted beneficiary identification numbers at each beneficiary level and followed longitudinally from initial placement and tracked each patient s records for 6 subsequent quarters. Study Population and Baseline Characteristics We analyzed data from patients who received an OBHP between the first quarter of 2007 and the second quarter of 2009. Study patients were identified by querying the SAF database for appropriate billing codes indicating OBHP implantation (see Supplemental Digital Content for details, http://links.lww.com/mao/a167). The codes used were American Medical Association (AMA) Current Procedural Terminology (CPT) codes 69714, 69715, 69717, and 69718; International Classification of Disease, Ninth Revision, Clinical Modification (ICD-9-CM) procedure code 20.95; and Centers for Medicare and Medicaid Services (CMS) Healthcare Common Procedure Coding System (HCPCS) codes L8690 and L8691. Patients were excluded if they did not exhibit continuous enrollment in Medicare Parts A and B for six-quarters after the first claim. To eliminate costs not associated with OBHP procedures and ensure that the complications identified were the result of the hearing prosthesis implantation, complications were excluded if they had significant otolaryngologic related diagnoses including underlying malignancies affecting the ear including otologic cancers, bone and connective tissue malignancies of the face, melanoma and other skin cancers of the ear, cranial nerve and meningeal cancers, and other benign neoplasms of similar areas including neurofibromatosis (see Appendix for ICD-9 codes). Complications were also excluded if the patient had billing codes for other significant otologic surgeries, including radical excision of external auditory canal, temporal bone resection, labyrinthectomy, vestibular nerve section, facial nerve decompression, internal auditory canal decompression, or temporal bone tumor resections (see Supplemental files for CPT codes, http://links.lww.com/mao/a167). To ensure that only complications associated with the surgical procedure were included, complications were excluded if the submitting provider was a podiatrist. Lastly, 1 inpatient hospital stay was identified with an unusually high billing and was therefore excluded from analysis as an extreme outlier because it was felt the costs were not likely associated with OBHP placement. Demographic data were tabulated including age, sex, and race as well as the total number of complications billing codes per patient. Study Outcomes We compiled a list of events identified as complications associated with OBHPs. These events included diagnoses such as skin infection, skin reaction, and wound dehiscence, and procedures such as debridement, skin excision, wound repair, and treatment of abscess. These were identified using selected ICD9 and CPT codes (see Appendix). We summarized the number of complications, time to complication, and allowed charges per complication. The study also identified repeat or revision implantation operations within the six-quarter follow-up period and included these in the cost analysis. RESULTS Study Cohort and Baseline Characteristics The SAF database contained 2,651,558 Medicare beneficiaries covering the first quarter of 2007 through the second quarter of 2009. Of these, we identified 151 patients who had an OBHP procedure or approximately 0.006% of the Medicare population. We excluded 33 patients who did not have six-quarters of continuous enrollment in Medicare Parts A and B. This resulted in a patient sample of 118 who underwent a total of 126 surgical placements. Among these patients, 70 patients (59%) had no further billing codes related to complications or revision/replacement procedures. The remaining 48 beneficiaries (41%) had billing codes for these events (Fig. 1). Demographic characteristics, including age, sex and race, were used to compare our study sample to the SAF population as a whole. The groups were similar in terms of an approximate equal proportion of men and women, the predominance of white race, and an age range primarily between 65 and 80 years. One difference was that blacks compromised less than 1% of our study sample but 10% of the SAF database. In addition, the number of complication codes per patient was about 2.5 times greater in the study group compared with the SAF population as a whole (Table 1).

COST OF PLACEMENT AND COMPLICATIONSRELATED TO OBHPS 3 FIG. 1. Identification of study patients. Cost of Nonrevision Complications A total of 48 of the 118 patients (41%) had billing codes associated with nonrevision surgical complications. Of these 48 patients, a total of 115 claims were submitted, resulting in $41,712 in costs. The most common complication was skin infection, which was seen in 19 (16%) of the study patients and resulted in $4,643 in costs. Nineteen patients (16%) had claims for skin flaps, skin grafts, or adjacent tissue transfer costing a total of $37,062 (Table 2). Although the number of patients and claims for complications remained relatively stable during the six-quarter follow-up period, the sum of costs for complications was greatest during the initial quarter and the quarter following TABLE 1. Patient baseline characteristics Study patients No. (% of total) SAF population No. (% of total) Total: 118 2,651,558 Sex Female 61 (51.7) 1,463,240 (55.2) Male 57 (48.3) 1,188,318 (44.8) Race Asian 0 49,727 (1.9) Black 1 (0.9) 270,065 (10.2) Hispanic 4 (3.4) 66,038 (2.5) North 1 (0.9) 11,322 (0.4) American Native Other 0 48,589 (1.8) Unknown 0 5,867 (0.2) White 112 (94.9) 2,199,950 (83.0) Age G65 23 (19.5) 659,020 (24.9) 65Y79 77 (65.3) 1,340,164 (50.5) Q80 18 (15.3) 652,374 (24.6) Age range 15Y89 0Y98 Age, mean (median) 68.5 (71) 70.99 (71) Complication codes 148 (1.25 per patient) 1,386,692 (0.52 per patient) the surgical placement. Costs decreased markedly during the remainder of the follow-up period (Fig. 2). Cost of Revision Surgery Complications A total of 8 of 118 patients (7%) had billing codes indicating revision surgical complications. The 8 patients submitted 12 claims, resulting in $39,319 in costs (Table 3). The number of claims and costs did not follow any obvious trends during the six-quarter follow-up period (Fig. 3). Overall Costs of Complications Overall for the 118 patients identified in the SAF database who received OBHPs and who were enrolled during the six-quarter follow-up period, 70 (59%) had no further billing codes for complications or revision surgery, whereas 48 (41%) did have billing codes for complications. Of the 70 patients without complications or repeat operations, the total allowed charges were $471,305 or a mean of $6,733 per patient. Of the 48 patients TABLE 2. Nonrevision complications with number of claims and total of allowed charges in decreasing order of number of patients reporting claims Complication group Patients %of patients Claims Total allowed charges Skin infection 19 16 38 $4,643 Skin reactions 9 8 21 $1,546 Skin grafts 8 7 14 $12,673 Adjacent tissue transfer 7 6 11 $19,347 Debridement of skin 6 5 19 $717 Wound 4 3 4 $407 dehiscence/nonhealing Skin flaps 4 3 8 $5,042 Treatment of abscess 4 3 6 $645 Wound repair 2 2 3 $749 Skin Excision 0 0 0 $0 Treatment of 0 0 0 $0 Wound Dehiscence Total 44 37 115 $41,712

4 S. E. PROSS ET AL. FIG. 2. Nonrevision complications by quarter. with complications or repeat operations, the total allowed charges were $374,987 or a mean of $7,812 per patient. The mean increase per patient in the group with complications or repeat operations was $1,079 or a 16% increase in this population (Table 4). Because the dataset represents a random sampling of 5% of the Medicare population, one can estimate that the 118 OBHPs implanted over 2.5 years in this study represents, 2380 implants in the entire Medicare population over this same period or approximately 944 OBHP implantations per year among the nationwide Medicare population during the study period. Complications were seen in 41% of the patients, with an average increase in cost of $1,079 per patient. Therefore, we estimate that complications associated with OBHPs led to $417,616 in additional allowed charges in the entire Medicare feefor-service population yearly during our study period. Based on $7,192 in average total charges per patient with OBHPs and an estimated 944 cases per year in the Medicare population, we further estimate that the cost of placement together with the cost of complications is $6,789,248 in the entire Medicare population during our study period. DISCUSSION Percutaneous OBHPs are an effective treatment for patients with conductive hearing loss who cannot use conventional hearing aids or who have single-sided deafness. However, as multiple prior studies have documented, placement of these devices is associated with potential complications (1,2,6,8Y11,17Y18). This observational, retrospective study followed patients from the initial implantation of the percutaneous OBHP through the six-quarters (18 mo) after initial placement. The study used a population-based, nationally representative claims database to estimate the economic implications of initial implantation, complications, and repeat procedures associated with these devices in the Medicare population during the 2.5-year period from the first quarter 2007 until the second quarter 2009 with a subsequent 1.5-year follow-up period. Our data source represents a random sampling of 5% of the Medicare population, which can be used to extrapolate to the entire Medicare fee-for-service population during the same period. Based on these numbers, we estimate that 944 OBHP implantations were performed yearly among the nationwide Medicare population during the study period, with an average increase in cost of $1,079 per patient from complications or roughly $417,616 total additional allowed charges. The total cost of placement together with the cost of complications is approximately $6,789,248 in the entire Medicare population during our study period. TABLE 3. Revision procedure surgical complications Repeat/ revision codes Description Patients %of patients Claims Total allowed charges 69714 Implantation of 5 4.2 7 $24,993.19 hearing prosthesis with or without mastoidectomy 69715 69717 Replacement of 3 2.5 5 $14,325.51 hearing prosthesis with or without mastoidectomy 69718 20.95 Implantation of electromagnetic hearing device L8690 Auditory 3 2.5 3 $25,095.33 device, includes all internal and external components L8691 Auditory device, external sound processor, replacement Total a 8 6.8 12 $39,318.70 a Total does not equal the sum of numbers above in column.

COST OF PLACEMENT AND COMPLICATIONSRELATED TO OBHPS 5 FIG. 3. Repeat/revision complications by quarter. There is a wide variability of complication rates in the published medical literature associated with OBHPs. For example, the need for revision surgery has been reported as low as 1.3% by Wazen et al. (9) and as high as 18% by Badran et al. (8). Soft tissue reactions have been reported as low as 4.5% by Wazen et al. (9) and as high as 70% (mostly mild) by Reyes et al. (5). This problem has been highlighted by Tjellstrom and Stalfors (10) who explain that studies often suffer from poor follow-up and incomplete data. In this study, we identified a 6.8% revision surgery rate, which falls within the published range. We also report a 37% rate of complications not requiring revision surgery, which is also within the published range and very similar to the 36% of skin reactions among U-shaped grafts reported by Stalfors and Tjellstrom (11). Although several previous studies have reported on the complication rates associated with these procedures, this study represents the first attempt to estimate the total costs of percutaneous OBHPs including postoperative complications and management. Knowledge of the total costs associated with OBHPs can be useful to both physicians and patients in choosing an appropriate treatment plan. In addition, this information will be helpful in future costeffectiveness studies. Toward this end, similar studies of the total costs associated with other surgically implantable devices for hearing rehabilitation (e.g., transcutaneous OBHPs, cochlear implantation, implantable middle ear hearing aids, and other hearing-related procedures) should also be considered. We acknowledge several limitations in this study. First and foremost, the Medicare SAF data are derived from billing data that were not originally designed for use in clinical research. Any complication that was not billed to Medicare cannot be accounted for, potentially underestimating the costs of OBHP placement. In addition, the patients in the SAF database represent only the population using Medicare fee-for-service, which limits the ability to generalize to the wider U.S. population, particularly with regard to pediatric patients. One potential area where this study may lead to an overestimation of the overall costs associated with complications from OBHPs relates to the billing codes used to capture data, inadvertently including complications unrelated to the implanteddevices.wechoseasetoficd-9 and CPT codes that represent common complications from implantation. However, we acknowledge that these codes may have occasionally been used for unrelated medical problems in our study patients. We attempted to control for this by excluding all claims made by podiatrists, while leaving in claims made by otolaryngologists and other medical and surgical specialties. Also excluded from the analysis were patients with diagnostic codes indicating an underlying disease that may generate independent charges (e.g., otologic cancers, bone and connective tissue malignancies of the face, melanoma, and other skin cancers of the ear) and if the patient had billing codes for other significant otologic surgeries (e.g., lateral temporal bone resection). Last, we identified 1 inpatient hospital stay that had an unusually high billing and excluded it from analysis as an extreme outlier, with charges likely unrelated to the OBHP. As documented in Table 1, the complication billing codes were used 2.5 times more frequent among patients who received device implantation than in the wider SAF population, suggesting that the claims were a result of their implantations rather than of baseline billing patterns. An additional group where complication costs may be overestimated includes those with repeat operations because these surgeries may represent implantation of a second contralateral OBHP and not a failure of the first implant. Although there is unfortunately no way to identify contralateral implantations in our data source, bilateral TABLE 4. Total cost with and without revisions No repeat operation or complications With repeat operation or complications No. of patients 70 48 No. of claims 137 210 Total allowed $471,305 $374,987 charges Mean allowed charges per patient $6,733 $7,812

6 S. E. PROSS ET AL. implantation remains uncommon (G5% of cases) (19), and even in those cases where bilateral implants are placed, this is most commonly performed during the same initial operative procedure; thus, we believe this does not significantly alter our findings. There are also several areas where the methodology used may underestimate the overall cost associated with complications for OBHPs. Although the set of complication codes was chosen based on an understanding of common postoperative complications reported in the literature, they may not be representative of all the complications in our study sample. Other ways our study may have underestimated the overall costs include claims made within the global period and outyof-pocket expenses including co-pays or Medicare Part D claims, all of which are not captured in the database. In addition, although the Medicare SAF database represents complete data from 5% of the Medicare fee-forservice population, there is evidence of incomplete billing in the data. Because of the nature of a retrospective analysis, the costs reported in our analysis do not reflect current dollar values. The costs reported in current dollar amounts would be higher. For example, the average Medicare payment for OBHPs and related services in our study group from 2007 to 2009 was $7,192 per patient and we estimate that given the 2013 Medicare fee schedule rates, the average per-patient payment would be $11,151, an increase of $3,959. There is a newer generation of surgically implanted OBHPs that have recently been released that transmit sound transcutaneously, as opposed to percutaneously (12,13). Unfortunately long-term data do not yet exist for these devices to allow an appropriate comparison with percutaneous devices, although one would predict a reduced incidence of skin-related complications and their associated charges. Whether these transcutaneous devices are associated with other complications not seen, offsetting cost savings from reduced skin complications remains to be seen. CONCLUSION This study presents an analysis of the costs associated with placement and complications related to percutaneous OBHPs. Such complications are common, and treating these postoperative conditions has led to significant spending in the Medicare population. This is essential information for physicians who must ascertain costs as they pursue appropriate treatments for their hearingimpaired patients. REFERENCES 1. House JW, Kutz JW Jr. Bone-anchored hearing aids: incidence and management of postoperative complications. Otol Neurotol 2007;28:213Y7. 2. Gluth MB, Eager KM, Eikelboom RH, Atlas MD. Long-term benefit perception, complications, and device malfunction rate of bone-anchored hearing aid implantation for profound unilateral sensorineural hearing loss. Otol Neurotol 2010;31:1427Y34. 3. McDermott AL, Williams J, Kuo M, Reid A, Proops D. Quality of life in children fitted with a bone-anchored hearing aid. Otol Neurotol 2009;30:344Y9. 4. Mylanus EA, van der Pouw KC, Snik AF, Cremers CW. Intraindividual comparison of the bone-anchored hearing aid and air-conduction hearing aids. Arch Otolaryngol Head Neck Surg 1998;124:271Y6. 5. Reyes RA, Tjellstrom A, Granstrom G. Evaluation of implant losses and skin reactions around extraoral bone-anchored implants: a 0- to 8-year follow-up. Otolaryngol Head Neck Surg 2000;122:272Y6. 6. Shirazi MA, Marzo SJ, Leonetti JP. Perioperative complications with the bone-anchored hearing aid. Otolaryngol Head Neck Surg 2006;134:236Y9. 7. de Wolf MJ, Hol MK, Mylanus EA, Cremers CW. Bone-anchored hearing aid surgery in older adults: implant loss and skin reactions. Ann Otol Rhinol Laryngol 2009;118:525Y31. 8. Badran K, Arya AK, Bunstone D, Mackinnon N. Long-term complications of bone-anchored hearing aids: a 14-year experience. J Laryngol Otol 2009;123:170Y6. 9. Wazen JJ, Young DL, Farrugia MC, et al. Successes and complications of the BAHA system. Otol Neurotol 2008;29:1115Y9. 10. Tjellstrom A, Stalfors J. Bone-anchored hearing device surgery: a 3- to 6-year follow-up with life table and worst-case scenario calculation. Otol Neurotol 2012;33:891Y4. 11. Stalfors J, Tjellstrom A. Skin reactions after BAHA surgery: a comparison between the U-graft technique and the BAHA dermatome. Otol Neurotol 2008;29:1109Y14. 12. Mulla O, Agada F, Reilly PG. Introducing the Sophono Alpha 1 abutment free bone conduction hearing system. Clin Otolaryngol 2012;37:168Y9. 13. Huber AM, Sim JH, Xie YZ, Chatzimichalis M, Ullrich O, Röösli C. The Bonebridge: Preclinical evaluation of a new transcutaneouslyactivated bone anchored hearing device. Hear Res 2013;301:93Y9. 14. Clark MA, Duhay FG, Thompson AK, et al. Clinical and economic outcomes after surgical aortic valve replacement in Medicare patients. Risk Manage Healthc Policy 2012;5:117Y26. 15. Yu H, Rubin J, Dunning S, Li S, Sato R. Clinical and economic burden of community-acquired pneumonia in the Medicare fee-forservice population. J Am Geriatr Soc 2012;60:2137Y43. 16. Mercaldi CJ, Ciarametaro M, Hahn B, et al. Cost efficiency of anticoagulation with warfarin to prevent stroke in Medicare beneficiaries with nonvalvular atrial fibrillation. Stroke 2011;42:112Y8. 17. Reyes RA, Tjellstrom A, Granstrom G. Evaluation of implant losses and skin reactions around extraoral bone-anchored implants: a 0- to 8-year follow-up. Otolaryngology Head Neck Surg 2000; 122:272Y6. 18. de Wolf MJ, Hol MK, Mylanus EA, Cremers CW. Bone-anchored hearing aid surgery in older adults: implant loss and skin reactions. Ann Otol Rhinol Laryngol 2009;118:525Y31. 19. Ricci G, Della Volpe A, Faralli M, et al. Results and complications of the BAHA system (Bone-anchored hearing aid). Eur Arch Otorhinolaryngol 2010;267:1539Y45.