Today smartphones are an integral part of most people’s daily lives. Pairing hearing instruments to Bluetooth for streaming digital content, taking handsfree phone calls etc. in theory felt like it would extend the time a user would wear their hearing aids.

So we asked the question

Could it be true that people with Bluetooth wireless hearing devices wear them longer?

Then we analyzed data from a few thousand appointments and the findings were impressive.

A colleague recently suggested that people who have hearing aids with Bluetooth wireless connectivity wear them more than people who don’t. She asserted that hands-free telephone use and streaming for music, television and podcasts encouraged longer wearing times.

Presumably, this is because people will wear a device for more hours per day if it provides a greater range of useful functionality in more listening environments. This is consistent with earlier suggestions (Kochkin, 2010) of Multiple Environment Listening Utility (MELU). MELU was defined as, “the percent of listening situations in which the patient was satisfied or very satisfied.” If you think about it, having access to handsfree phone calls for any Android or Apple device, and streaming all kinds of digital content opens two very common and satisfying listening environments. Therefore, it seemed like her idea had some validity.

Working in the product development facility of a major hearing instrument manufacturer provides direct access to the engineers who build those products and who 

are required by regulation to undertake post-market clinical follow-up. That follow-up means access to basic information about our hearing instruments out in the field. Such access may include, the hearing instrument model, average hours of use per day, the technology level, wireless protocols or battery type. Of course, there are limits. No information is available to identify the wearer or the fitter of the products. The available data is strictly anonymous.

Using “hours of use” per day as a dependent variable it may be possible to assess which of factors of the hearing instrument were consistent with longer or shorter wearing times. Our trial was a carefully undertaken retrospective analysis of data from several thousand hearing instruments out in the real world. This record of findings will hopefully pique the curiosity of others to continue this line of inquiry to determine if some or all of these results are determinative of wearing time. 

Wearing time should be considered a strong indicator of success among hearing instrument wearers. We understand that zero-wearing time is, by definition, an abysmal failure. Hearing aids in a drawer or charger are of no use to anyone. Of sheer necessity, it is clear the more severe the hearing loss, the longer use time is likely to be. Overall, for a large group of randomly selected fittings, variables that impact wearing time or that correlate with wearing time should become evident even when the extreme cases of no wearing time or near-total wearing time are eliminated. Based on the results of retrospective analysis one cannot say that characteristics which correlate with wearing time are predictive of wearing time. Though wearing time can’t be known before fitting a hearing aid, hardware features such as wireless technology or battery type are known upfront. If a given wireless or battery technology correlates with longer wearing time, you may wish to consider including them when choosing hearing aids. 

For example, if you see a high correlation between Bluetooth-enabled hearing aids or rechargeable hearing aids and longer wearing times, wouldn’t you be more inclined to recommend hearing aids with those features?

The initial analysis of this data was published in a Hansaton white paper, “What type of hearing aids are worn longer per day?” (Rule, 2021). That white paper covered the largest questions about individual variables affecting wearing time. This deeper dive analysis will cover some of the interactions between relevant variables under test.

Data sets of anonymous fitting records were pulled from three large mature hearing instrument markets: USA, Germany & France. We used only the data from Receiver in the Canal (RIC) instruments to minimize potential hardware variables beyond those under study. The information in the data sets included only the following parameters: Hearing instrument model, battery type, wireless type, technology level and hearing threshold loss.

Hearing instrument models

The two Discover RICs are identical except the Jump R is rechargeable with a Li-Ion battery. The two Fit devices are also equivalent except that the Discover product uses Bluetooth wireless instead of the older FM wireless option (HiBAN*). It is not certain if these results would be typical for similarly equipped products from other manufacturers, but that may be the case. Therefore, from this point on the three product models will be referred to as BT/Lion, BT/ZAir and HiBAN/ZAir. 

The technology levels that were compared:

The technology levels in each row are equivalent across platforms in terms of hearing instrument performance. The technology levels will be referred to as levels 3, 5, 7 and 9 from now on.

The data retrieved from the post-market clinical follow up does not include hearing threshold levels per se as it would potentially provide confidential purchaser information. Instead, the data was filtered by general categories of low and high-frequency hearing loss using the same threshold ranges as standard audiograms. The range of hearing losses under investigation was limited to those well within the fitting ranges of the hearing instrument models in question. This was intended to minimize confounders such as under-fitted devices which may be worn selectively or those that would be prone to chronic feedback. The final categories were as follows. The term “Mild” hearing loss category to be used going forward refers to all records defined as having combinations of normal or mild hearing loss in the low frequencies and mild to moderate hearing loss in the high frequencies. The term “Moderate” hearing loss category was defined as all records with combinations of hearing losses including mild or moderate in the low frequencies and moderate to moderately severe hearing loss in the high frequencies. Thus, all the records in the dataset could be lumped into either the “Mild” or “Moderate” hearing loss category.

Wearing time was obtained by downloading the datalogging of the devices along with the start and end dates. The hearing instruments do not keep track of

time by hour or day. Instead, they only sum the number of seconds that they are turned on. The start date and the end date are then known from the computer running the fitting software. All of these factors can lead to anomalous data points at the very low or very high end of the hours per day results. To minimize the impact of the above noted potential outliers, all data sets used in the analysis included only records that fell into the 10th to the 90th percentiles of each data set. For example, this eliminated records of devices with very limited wearing times (<1 hour/day) or excessively long (> 23 hours/day) wearing times. The actual middle 80 percentile ranges varied slightly by the data set. But they typically ranged from 1.5 to 3 hours/day at the bottom end to 13 to 14 hours/day at the top end. In other words, they were fairly representative of average hearing aid wearers.

As mentioned in the Hansaton white paper (Rule, 2021), the first analysis of the data revealed the main effects on wearing time for battery type and technology level as well as an apparent interaction between the two. This paper reports the effects that were found when the interactions of multiple variables were assessed for wearing time.

The first statistical test was a Three-Way ANOVA (n=5436) which included the two levels of wireless transmission x four tech levels and two levels of hearing loss (HTL).

Table 1. Analysis of Variance 

The overall ANOVA was highly significant. There were also significant main effects of both wireless technology (p = 0.0225) and device technology level (p < 0.0001). There was a significant interaction between the two variables of wireless technology and technology level (p = 0.0201). However, there were no significant main

effects or interaction effects related to HTL. Therefore, HTL will not be analyzed further.

Further insights into the main effects and interactions between the wireless and tech level variables can be seen in the Figures below. We can begin with the simplest case of main effects and then work up to the Interaction effect. Beginning with Figure 1.

Figure 1.

Table 2. 

The top and bottom graphs of Figure 1 are showing the same data in two different ways. The dark blue bars represent the mean wearing time per day by Technology Level (TL) for the Bluetooth products and the lighter blue

bars show the same information for the HiBAN wireless products. The statistics above tell us there are two significant differences in this data set. 

1. Among people who wear the Moxi Fit devices (312 Z/A) the instruments using HiBAN were worn longer in hours per day. With the exception of TL7, the mean difference is close to half an hour per day.
2. The people who wore higher TL products, regardless of the wireless type, wore them significantly longer than those who wore the lower TL products. 

As seen in Table 2, the mean wearing time difference between Bluetooth type TL9 and Bluetooth TL3 was 1.24 hours per day. Similarly among HiBAN devices, the TL9 products were worn an average of 1.02 hours longer per day than TL3. This second point is easier to see on the bottom panel of Figure 1. The line graph on the bottom is another way of looking at the data. The slope of each line shows the mean difference and direction of wearing time between Bluetooth and HiBAN devices. Absolute TL differences are very clear by comparing each colored line to the other three. The difference between the yellow and blue lines represents the difference in mean wearing time from TL9 down to TL3.

This indicates that among hearing aid wearers who only used Z/A batteries, the people with HiBAN wireless devices wore their instruments significantly longer than those who had Bluetooth enabled devices. Furthermore, people who had higher tech level hearing instruments wore them significantly longer than those with lower tech level instruments. The mean wearing time difference among people who wore instruments with Z/A batteries across wireless technologies can be seen in Table 2. The difference scores are in the bottom row of Table 2. The mean difference in wearing time is very close to half an hour per day for HiBAN instruments at all tech levels except TL7. In the case of TL7, there was no meaningful difference in wearing time. The right hand column of Table 2 shows the difference in mean wearing time from TL3 (shortest) to TL9 (longest) to be about an hour for HiBAN wearers and an hour and a quarter for Bluetooth users.

Having looked at wearing time by wireless communication and tech level for Z/A batteries it would be desirable to make similar comparisons across devices that had only Li-Ion batteries. However, we have never matched Li-Ion batteries with HiBAN wireless technology. As such, that comparison isn’t possible. Therefore, the next step is to compare hearing aid models with different battery types while holding the wireless technology constant, using Bluetooth only. The wearing time results for people who owned Bluetooth instruments but differed by using either Li-Ion or Z/A batteries are shown in Figure 2.

Figure 2.

Table 3.

The ANOVA for battery type x Tech Level was highly significant (p < 0.0001) for (n = 1948) hearing aid wearers with BlueTooth devices. The main effects of battery type (p < 0.0001) and Tech Level (p = 0.0020) were both significant. There was no significant interaction effect. The clearest difference visible in Figure 2 and Table 3 is the increased wearing time with the rechargeable Li-Ion devices compared to the Z/A instruments. The dark blue bars (Figure 3, top) show the mean wearing time per day by TL for the Li-Ion battery products and the lighter blue bars show the same information for the 312 Z/A wireless products. The stats tell us there are two significant differences in this data set.

1. Among people who wear the Discover devices (Bluetooth) the Li-Ion instruments were worn significantly more hours per day. The Li-Ion TL3 products were worn 2 hours longer/day and the TL9 Li-Ion instruments were worn an average over an hour longer/day.
2. The people who wore higher TL products regardless of the battery type wore them significantly longer than those who wore the lower TL products.

You can see this again on the bottom graphic a bit more clearly. The mean difference between Li-Ion type TL9 and Li-Ion TL3 was about 0.3 hours per day. Similarly, among 312 z/a devices, the TL9 products were worn an average of about 1.25 hours longer per day than TL3.

The line graph on the bottom is another way of looking at the data. The slope of each line shows the mean difference and direction of wearing time between Li-Ion and 312 Z/A devices. Absolute TL differences are very clear by comparing each colored line to the other three. The difference between the yellow and blue lines represents the difference in mean wearing time from TL9 down to TL3.

Putting together all three combinations of battery type by wireless type across the four technology levels. Everything is plotted together in Figure 3.

Figure 3.

In general, it is clear that combining Bluetooth wireless with a rechargeable Li-Ion battery yielded the highest mean wearing time by a wide margin at every tech level. Combining Bluetooth wireless with a Zinc-Air battery yielded the lowest mean wearing time at every tech level except for TL7. Finally, tech level seems to have very little clinically relevant impact on wearing time for the Bluetooth Li-Ion devices. But it has a substantial impact on the devices with a Zinc-Air batteries.

The condition with the greatest mean wearing time was the TL9, Bluetooth instrument with a rechargeable Li-Ion battery (9.76 hours/day). The condition with the lowest mean wearing time was the TL3, Bluetooth instrument with a replaceable Zinc-Air battery (7.41 hours/day). That is a mean wearing time difference of 2.35 hours/ day. You could say it’s 2 hours and 21 minutes more per day that the average client will wear one device over another. This begs the question why? Due to the design limitations of this study, it’s not possible to answer that question without further research, which I encourage. But if we conjecture a bit about human nature there is one plausible explanation worth considering; hearing aid cost.

The option with the lowest up-front cost is the TL3 device with Zinc-Air batteries. The options with the lowest cost of ownership after purchase are the rechargeable Li-Ion devices. The option with the lowest up-front cost but the highest cost of ownership after purchase is the TL3, Bluetooth, and Zinc-Air instrument. It has roughly the same upfront cost as the TL3 Zinc-Air HiBAN device. But because Bluetooth transmission has a much higher current drain than HiBAN, Bluetooth devices will go through disposable batteries more quickly. Thus, the more you wear it, the more expensive it is to own. On the other hand, once you pay the upfront cost for a rechargeable device, there is no penalty for hours of Bluetooth transmission. It is not a given that this is the reason or the only reason for the results of this study. But the explanation does fit the data. People who are cost conscious, purchase TL3 devices because they are the least expensive. However, being cost conscious should not end at the point of purchase. If someone can’t afford a high-end hearing aid feels the pain every time they swap out batteries, they may wear the hearing

instruments less often to reduce battery purchases. But if there is no cost for hours of use as with rechargeable devices, people with TL3 devices can wear their hearing instruments as long as those with TL9 devices. This type of pattern is evident in Figure 3.

In answer to the question at the opening of this paper, “Do people wearing Bluetooth wireless devices wear them longer?”, yes and no. It seems that question is an oversimplification of the factors involved in wearing time. At first glance in Figure 3, where all conditions are shown together, the Bluetooth devices appear to be worn longer than other wireless devices. But it is not that simple. The interactions between tech level, type of battery, and wireless transmission are more telling. In the end, the longest wearing times appear to correlate much more highly with rechargeable devices and higher technology levels. But looking at Figure 3 where all permutations that were tested can be seen, rechargeable Li-Ion batteries seem to be the factor most closely associated with longer wearing time followed by higher technology levels.