Cochlear Implant Utilization in the United States: How Are We Doing?

Hi, everyone. Thank you so much for joining this session. My name is Kevin Zahn. I am one of the cochlear implant surgeons at Northwestern University. I’m also the director of the cochlear implant program there. I’m excited to share with you this talk which is going to focus on utilization in the United States and how we are currently doing. These are my disclosures. I do some consulting work for these three companies. So the two major objectives that I want to get into here, the first one is really just to review the current utilization literature and why the underutilization that I’ll show here is such a major public health issue. and as well, the latter half of the talk will also talk about how audiology use and cochlear implant use relate to one another across the country, and some of the research that we’ve done here. So, first, getting into some of the background information and what’s going on, we know that the US population is getting older. And so it’s estimated that the population over 65 is going to double by 2050. And since hearing loss and age are intimately related, and we definitely see more severe forms of hearing loss the older we get we’d expect a very large population of hearing loss individuals as the population continues to age. And currently, as it stands, there are over 2 million Americans that have severe to profound hearing loss thresholds. and this has gained a lot of traction in the past few years that hearing loss is the number one modifiable risk factor for dementia. There’s some exciting work going on right now looking at large clinical trials looking at whether or not hearing aids and intervening and hearing intervention will slow down the progression of developing dementia. So before I get into the details of some of the utilization studies that have been published out there, I wanted to start with some of the issues with trying to do research in cochlear implant utilization. the challenges are that there’s not a lot of research tools out there that can capture all of the hearing loss patients in the United States, as well as all of the cochlear implant surgeries that are also performed. And so it makes it challenging when some of the studies you’ll see out there are single institution or multi institution reviews. they’ll say something like, oh, in the last couple of years we’ve done this many evaluations, and these were the number of patients that underwent cochlear implantation. But the problem with that is that obviously that’s not capturing everybody in that area. That’s simply just relating to the number of patients that show up to that institution for that particular study. And so when we look at databases that could potentially capture all this claims data and administrative databases do offer a glimpse into some of this stuff because they can amass an enormous amount of patients as well as claims. And then you can file through that to look at how many cochlear implants are done, for example within those databases. And so there’s two sort of major utilization studies, and they also actually just use market data. And so this 5.6% utilization rate was looking at 2010 market data published by Donna Sorkin who’s the executive director of the American Cochlear Implant Alliance. and this looked at what is the possible denominator that the denominator in this number is number of estimated individuals with severe to profound hearing loss thresholds. And going from that you would estimate that those are probably patients that would qualify for a cochlear implant. And then in the numerator, you can put in the number of cochlear implants that are performed according to this proprietary market data research called idata. A similar study was then repeated in 2022, looking at more recent data. And they went up to 2015 in their data. And the white box that’s circled or highlighted in red shows the number of untreated individuals who met traditional audio metric criteria based off of estimates. And so 1.16 million was that, for example, in 2015 was the estimated denominator or the pool of candidates based off of traditional Medicare criteria and or traditional bilateral cochlear implant criteria. And so they found a rate of 12.7% in their data. And that number was slightly increasing on a year to year basis. However, I would say that given that these studies are looking only at traditional candidates with bilateral sensory neural hearing loss, I think one of the big issues is that candidacy criteria have expanded quite dramatically in the past few years. For that reason, these studies likely are vast underestimates for all of the actual candidates that now we can potentially treat. And so, as I said, indications have expanded quite rapidly and utilization remains quite low, unfortunately. and so what are some of those expansions? Right. So this is just a quick slide, sort of summarizing some. I think some of the biggest changes in the past ten years or so. roughly ten years ago the first FDA hybrid candidacy expansion was created. So we really changed the audiometric thresholds that we look at now for candidacy. And looking at this figure on the right from Johns Hopkins we’re now implanting individuals with audiometric thresholds that look like the red line, for example, that patients that have normal or near normal hearing in the low frequencies and precipitously severe to profound hearing loss thresholds in the higher frequencies where a lot of that speech information is contained and why patients are struggling. And we know that these patients do well with the cochlear implants, really. these audiometric criteria are how we view all cochlear implant candidates when we now think about candidacy from a purely audiometric perspective in 2019, a huge increase in potential candidates with the single sided deafness and asymmetric hearing loss population groups these now are potential candidates for cochlear implantation. And so you think about somebody, for example, who lost their hearing on one side due to sudden sensorineural hearing loss, for example. those patients are now potential candidates, especially if they have bad tinnitus, to be able to hear with two ears again and potentially drastically improve their quality of life. In 2022, Medicare candidacy criteria expanded quite dramatically. I won’t get into the details of that, but it’s now a lot easier to qualify and this is certainly trying to catch up with current clinical care as far as how we are qualified patients and determining candidacy in these patients. And then in January 2024 this really isn’t a major candidacy criteria increase, but this is the release of the minimum speech test battery 3 (MSTB-3), which is a really nice guiding document on how we how we think about candidacy and how we are standardizing the candidacy criteria, you know, candidacy determination process across the country, because there’s a significant amount of variability between clinics currently on how we determine candidacy. And so really that sort of gets into how the confusion of who gets a cochlear implant. Right. And this is very, very tough and challenging you know because it’s hard, it’s challenging enough, even amongst some cochlear implant clinics and providers, to keep up with the times. But it’s very tough for the general public as well as non-otolaryngology providers, as well as even otolaryngology providers and audiology providers and hearing dispensaries and anybody in the healthcare field to really understand when is it appropriate to refer somebody for a cochlear implant evaluation and get them plugged in sooner rather than later. This likely has significant contributions to why there is decreased penetration of cochlear implants and underutilization presentation. So going back to this original slide here that I looked at earlier, I think the real denominator is likely significantly larger now that we’ve included single sided deafness, asymmetric hearing losses hybrid criteria and Medicare candidacy expansions, as we’ve discussed previously. So what are some other reasons for underutilization? I’ve sort of hinted at provider familiarity as a big one. It’s very tough to keep up with all of these changing indications and expanding indications and so that naturally leads to under referral. And perhaps patients that were, would have been cochlear implant candidate ten years ago are now finally getting referred because of some reason here and there. So poor patient familiarity is also a big issue. Right. I don’t think my personal opinion is that cochlear implants are not very well penetrated into our national consciousness. there has been some increased media coverage and representation if you will, in the past few years. Certainly there’s been amongst the younger generations, there’s been some reality dating tv shows with a couple of candidates that have cochlear implants and have really done a good job of increasing the visibility to understanding patients with hearing loss and cochlear implants and how life changing that can be for these patients. But either way, there’s still a considerable amount of hearing loss stigma at present. and so we looked at 5.6% earlier. So certainly if you imagine only 5.6% of patients that need glasses that have them, that actually get one, get a pair, then certainly our day to day life and how safe it is to drive around is going to be very different. And so how we think about hearing and how we think about vision and how we think about other medical issues. There’s certainly a significant stigma with getting hearing loss treated, wearing a hearing aid and so on and so forth. We do know that socioeconomic disparities play a big role in first getting access to cochlear implant care and hearing care in general, but also with just finding out about health literacy and finding out about cochlear implantation and knowing when to be referred as well as obviously logistical issues and a whole sort of other socioeconomic reasons for why cochlear implants are not as heavily used in lower socioeconomic groups. inadequate insurance coverage is certainly an issue although I would argue that most at least for traditional candidates private payers and Medicare do a pretty decent job with paying for these devices now. but certainly we experience this a lot with single sided deafness and asymmetric hearing loss patients where we’ll run into some issues with private payer coverage and so that just depends on the payer at hand. Lack of surgeon availability and CI audiologists and hearing healthcare providers familiar with cochlear implantation is a big issue as we’ll see. And certainly cochlear implant care can be quite involved in the first year or so and leading up to the surgery itself. And so patients that live 150 miles away from a cochlear implant center are going to be less likely to want to seek care because this is a considerable logistical burden for them, financial burden for them to undergo all of this care. Certainly there’s been more attention focused at focused on telehealth and remote programming, those kind of things to reach these patients a lot better and many, many other reasons not listed here that also exist. So this is the study that I wanted to highlight. this is something that we did at Washington University in St. Louis. I wanted to highlight my other co authors here, doctor Angela Mazul, Dr. Dorina Kallogjeri and Doctor Craig Buchman. and so one of the first questions that we wanted to look at was, does utilization vary across the country? Now the first two studies that I looked at I shared with you guys earlier that was the entire population as a whole, right? So there’s no geographic information associated with that aside from just the entire population as a whole. and so we, you know, we would expect that utilization does vary across the country, but that’s never been shown before on a geographic basis. And so we’re excited to share some of that data and look at this question into a bit more detail. Question number two is, does audiology utilization affect CI utilization? And so audiology utilization has also never been studied in this way and certainly never been paired up with cochlear implant utilization. To look at how these variables affect each other. And so we’ll get into some of those results here shortly. And so to try to address this issue as best as possible we looked at, we used the PearlDiver Mariner claims database, which is a private research firm that creates this database. And they have all payers included in it and they compile claims data from all payers Medicare, Medicaid cash, uninsured payers, uninsured patients, private payers, obviously, government insurances, those kind of things. and they track a ton of claims across the country. And they have 93.3 million unique adults in their database for which we had access to. And so they had in this date with these 93.3 million unique adults, that was 1.06 billion claims that we queried for bilateral sensorineural hearing loss. And so specific emphasis here on sensorineural and as well as bilateral as well. Right. So prior studies typically look at all forms of hearing loss, including conductive mix, those kind of things, or at least they don’t specify. but we try to be a bit more specific here with our diagnostic inclusion criteria and diagnostic audiology. we sort of created a variable with these three cpt claims. So to encompass diagnostic audiometry with or without speech audiometry. And cochlear implantation is nice just because it’s just one claim, so it makes it nice and clean. 69930 for cochlear implant use. So in PearlDiver what was really nice about it is that it has three digit zip code data for where the patient lives. So this is specific to the patient location, not where they received care. And so you’ll see when I show some of the maps that those maps pertain specifically to where that patient lives. And all of the values and kind of the numbers that I present are relative to the population of that particular zip code. and so what we were able to do with three digit zip code data was that then I was able to merge it with census data based off of five digit zip codes. And in that way, this gave us some sense of socioeconomic measures, such as the number of people living in that zip code under the 100% poverty level, for example, or the number of high school graduation what’s the rate of high school graduation in that zip code? And what is the rate of uninsured patients in that zip code? So these are three variables of socioeconomic measures that we looked at and tried to correlate with other things. And then we defined utilization itself as what was the cumulative incidence at five years of a particular event. And so for a cochlear implant, for example, then it would be the starting event was right after someone was diagnosed and labeled with a sensorineural hearing loss claim. Then we’d follow them through time. And what was the incidence of cochlear implant for that particular zip code at five years? And obviously someone who got one would be censored out of the database. And so the inverse of that would give us the cumulative incidence. And so the cumulative incidence is nice because it gives us a proportion of that particular population with this particular variable of interest. So cochlear implant use or diagnostic audiometry use. And then mapping was done in ArcGIS Pro and statistics were run in SPSS v28 and R Studio. So the results we identified 6.92 million adults with a sensorineural hearing loss claim. And of those 6.92 million, 73.3% had a diagnostic audiology claim. So that was nice to see that the vast majority of these were verified with a diagnostic audiogram of some sort. and so that makes the kind of the diagnosis a bit more believable. and of that of the 6.92 million of patients had a cochlear implant claim. When we pared down all the three zip codes available for analysis, there were quite a few that had very, very few patients or were completely unoccupied. And so we filtered those out just for patients zip codes with greater than 1000 sensorineural hearing loss claims. And so we ended up with 804 three digits of zip codes for analysis, for geographic analysis. And when we tried to get a sense of what was the capture rate. So how good is PearlDiver at capturing patients? we compared that to census estimates. So we looked at a number of patients existed in PearlDiver for a particular zip code that we used, and then the number of patients that exist in the census for that particular age range as well. And we found on average that pearl diver did a pretty good job with 41% capture rate and that also did vary across the country. So the first thing that I’d like to highlight is that there is significant variation in cochlear implant utilization across the country. And that’s not a surprise here. But this is what that looks like across the country here. And so this is again, these are all the three digit zip codes that were available for analysis. Some of these are going to be blank just because there’s not enough data. So we see that some of these mountain states here but again, these values are, this is shown in quartile. So the darkest blue is a top quartile that most cochlear implants utilize. And then the almost white here is the bottom quartile or very, very little cochlear implant use. And you see just from gestalt that you’ll see higher cochlear implant utilization in the midwestern states, across the United States relatively lower across the eastern United States and the western United States. There are some areas sporadically placed throughout here that also match up with known high volume cochlear implant centers. for example, Oklahoma you know, St. Louis, Iowa you know, there’s various centers in Ohio that are, you know, known large cochlear implant centers relative to their population. And you’ll see in some other areas, for example, the southeastern United States, that there’s not a single three digit zip code in these states, for example, that that sit in the top quartile of cochlear implant use. and so that’s just kind of interesting to see this mapped out across the United States. we talked about how, you know, we talked about this first point here. We also tried to correlate you know some of the socioeconomic variables that we looked at. And we tend, we tend to find that the top utilization quartiles existed in zip codes that had overall higher patient counts, so higher population density, so closer to cities, obviously and lower poverty rates compared to bottom quartile zip codes. When we look at, so this figure pits audiology use as well against cochlear implant use. And these are again, relative values, relative ratios between the two. And once again, so in the blue, in the pure blue here that those are areas of relatively higher cochlear implant use. And in the pure sort of pink here, those are areas of relatively higher audiology usage. And the dark blue are areas of both which are obviously, in my opinion, the ideal situation here. but you’ll see again, once along the eastern United States, that there’s a lot of areas with more pink and higher relative audiology use than cochlear implant use. And you see a lot more cochlear implant use in the midwestern states of the United States. we also looked at what the rate of diagnosis was. We looked at the rate of sensori neural hearing loss diagnosis in our zip code. and not to anyone surprise, that correlated very strongly with audiology utilization. That makes sense. The more audiograms you perform in a particular area, the higher rate of sensori neural hearing loss diagnosis that you can then diagnose. that made a lot of sense with a correlation coefficient of 0.62. but then looking at, you would expect higher rates of sensorineural hearing loss diagnosis, perhaps higher rates of cochlear implant use. And we actually found the opposite. We found a weak negative correlation in areas with high rates of sensorineural hair loss diagnosis with cochlear implant use. At the very least, we can say that audiology utilization does not necessarily correlate at all with cochlear implant utilization. and there might be areas of even negative relationship there that the more you use audiograms and diagnose patients with hearing loss, the less relative cochlear implant use that exists. That really is just to say that there’s a lot of potential under referral going on here in areas potentially under access as well, of cochlear implant surgeons and CI audiologists as well. So based on these data, we estimated that our utilization rate that we estimated was around 6% for traditional CI candidacy. And again, this is based off of census survey estimates and other studies like Goman and Lin, who used census data and survey data to estimate that approximately 5.5% of all hearing loss patients have severely profound hearing loss thresholds. And so these numbers and percentages were important to create the 6% calculation for utilization. I just wanted to briefly point out here these are seven zip codes that in particular, have really high top quartile of audiology use, but bottom quartile of cochlear implant use. And, you know, here in Illinois, Springfield, Illinois, is closer to where. Where I live. and so just, you know, these areas, I think, warrants, you know, potentially a further investigation as to far as, that’s why there’s so much more relative audiology use than there is cochlear implant use. limitations. So these are some of the big ones, obviously, with claims data, the internal validity of the data is always in question. but you’re never able to. With claims data, there is no severity of hearing loss that exists. so we have sensory hearing loss, we don’t know that’s mild or moderate severe. So we don’t know the distribution of that. So we rely on estimates based off of census data and survey data. you know, a three digit zip code is a very, very. Can be a very large area of land. And so one of the challenges with three digit zip code is that you can have an area of very high poverty and areas of very low poverty as well. And they’re all mixed together. And so that really prohibited us from more granular socioeconomic analysis. And really that’s why I didn’t show a lot of socioeconomic data, because a lot of it just didn’t correlate with anything. and that’s probably not because there’s not a relationship that exists. It’s probably just because we have the tools in which to adequately measure these things. because of the large geographic nature of three zip codes. And then, you know with a private, proprietary research database, we don’t know just how adequately covered certain payers are in certain areas versus others. And so we have to rely on an unknown degree of payer contributions to a database that may shift over time. Some of the main takeaways here that CI utilization varies greatly throughout the country but overall remains low across the country. there is a negative, weak negative correlation between hearing loss diagnosis and cochlear implant utilization, which suggests a significant component of under referral. And certainly future efforts should definitely focus on improving awareness amongst providers and patients of all of the. The indications for cochlear implants and the wonderful benefits that they can provide to patients with hearing loss. So thank you so much for your attention today.