The Effects of Service-Delivery Model and Purchase Price on Hearing-Aid Outcomes in Older Adults: A Randomized Double-Blind Placebo-Controlled Clinical Trial

Abstract

Objectives: The objectives of this study were to determine efficacy of hearing aids in older adults using audiology best practices, to evaluate the efficacy of an alternative over-the-counter (OTC) intervention, and to examine the influence of purchase price on outcomes for both service-delivery models.

Design: The design of this study was a single-site, prospective, double-blind placebo-controlled randomized trial with three parallel branches: (a) audiology best practices (AB), (b) consumer decides OTC model (CD), and (c) placebo devices (P). Outcome measures were obtained after a typical 6-week trial period with follow-up 4-week AB-based trial for those initially assigned to CD and P groups.

Setting: Older adults from the general community were recruited via newspaper and community flyers to participate at a university research clinic.

Participants: Participants were adults, ages 55-79 years, with mild-to-moderate hearing loss. There were 188 eligible participants: 163 enrolled as a volunteer sample, and 154 completed the intervention.

Intervention(s): All participants received the same high-end digital mini-behind-the-ear hearing aids fitted bilaterally. AB and P groups received best-practice services from audiologists; differing mainly in use of appropriate (AB) or placebo (P) hearing aid settings. CD participants self-selected their own pre-programmed hearing aids via an OTC model.

Primary and secondary outcome measures: Primary outcome measure was a 66-item self-report, Profile of Hearing Aid Benefit (Cox & Gilmore, 1990). Secondary outcome measure was the Connected Speech Test (Cox, Alexander, & Gilmore, 1987) benefit. Additional measures of hearing-aid benefit, satisfaction, and usage were also obtained.

Results: Per-protocol analyses were performed. AB service-delivery model was found to be efficacious for most of the outcome measures, with moderate or large effect sizes (Cohen's d). CD service-delivery model was efficacious, with similar effect sizes. However, CD group had a significantly (p < .05) lower satisfaction and percentage (CD: 55%; AB: 81%; P: 36%) likely to purchase hearing aids after the trial.

Conclusions: Hearing aids are efficacious in older adults for both AB and CD service-delivery models. CD model of OTC service delivery yielded only slightly poorer outcomes than the AB model. Efficacious OTC models may increase accessibility and affordability of hearing aids for millions of older adults. Purchase price had no effect on outcomes, but a high percentage of those who rejected hearing aids paid the typical price (85%).

Trial registration: Clinicaltrials.gov: NCT01788432; https://clinicaltrials.gov/ct2/show/NCT01788423.

Figure 1.

CONSORT flow diagram for the ABCD randomized clinical trial.

Figure 2.

Comparison of prescribed (NAL-NL2) real-ear targets (filled circles) in dB SPL and measured (unfilled circles) real-ear levels for left (top) and right (bottom) ears for the audiology best practices (AB) group. The stimulus was a 65-dB SPL speech signal (“carrot passage”) for the speech-mapping measures from the Verifit test system. Symbol = mean value; error bar = ±1 standard deviation.

Figure 3.

Means for real-ear insertion gain measured in the left (top) and right (bottom) ears of the placebo (P) group. Data are shown separately for each microphone-setting P subgroup: those with directional microphones (Placebo–Dir) and those with omnidirectional microphones (Placebo–Omni). The “target” gain for the P group is 0 dB. In addition, the mean measured real-ear insertion gain for the audiology best practices (AB) group is shown (X). All real-ear insertion gain values shown are for 65-dB-SPL pure-tone input stimuli. Error bars = ±1 standard deviation.

Figure 4.

The X, Y, and Z audiograms from Ciletti & Flamme (2008), being among the most common audiometric configurations, were used to pre-match the hearing aids for the consumer decides/over-the-counter (CD) participants to NAL-NL2 targets.

Figure 5.

Median air-conduction pure-tone thresholds for left (top) and right (bottom) ears of the audiology best practices (AB), consumer decides/over-the-counter (CD), and placebo device (P) groups.

Figure 6.

Means for the PHABglobal outcome measure for each of the six subgroups evaluated in the 3 × 2 factorial analysis. Unaided (black bars) and aided (red bars) PHAPglobal scores are provided, as well as the difference between these two scores, PHABglobal (green bars). Data are for N = 154, with 53 audiology best practices (AB), 51 consumer decides/over-the-counter (CD), and 50 placebo device (P) participants included. Purchase-price subgroups are “red” (reduced; $600) and “typ” (typical; $3,600). Error bars = +1 standard error; PHABglobal = difference between aided and unaided scores of PHAPglobal (Profile of Hearing Aid Performance, average of the five communication-related subscales: Familiar Talkers, Ease of Communication, Reverberation, Reduced Cues, and Background Noise).

Figure 7.

Means for the PHABavds outcome measure for each of the six subgroups evaluated in the 3 × 2 factorial analysis. Unaided (black bars) and aided (red bars) PHAPavds scores are provided, as well as the difference between these two scores, PHABavds (green bars). Data are for N = 154, with 53 audiology best practices (AB), 51 consumer decides/over-the-counter (CD), and 50 placebo device (P) participants included. Purchase-price subgroups are “red” (reduced; $600) and “typ” (typical; $3,600). Error bar = +1 standard error; PHABavds = difference between aided and unaided scores of PHAPavds (Profile of Hearing Aid Performance, average of Aversiveness of Sound and Distorted Sound subscales).

Figure 8.

Means for the CST benefit outcome measure for each of the six subgroups evaluated in the 3 × 2 factorial analysis. Unaided (black bars) and aided (red bars) CST scores are provided, as well as the difference between these two scores, CST benefit (green bars). Data are for N = 154, with 53 audiology best practices (AB), 51 consumer decides/over-the-counter (CD), and 50 placebo device (P) participants included. Purchase-price subgroups are “red” (reduced; $600) and “typ” (typical; $3,600). All CST scores are shown in rationalized arcsine units (RAU). Error bars = +1 standard error; CST benefit = difference between aided and unaided Connected Speech Test scores.

Figure 9.

Means for the HHIE benefit outcome measure for each of the six subgroups evaluated in the 3 × 2 factorial analysis. Unaided (black bars) and aided (red bars) HHIE scores are provided, as well as the difference between these two scores, HHIE benefit (green bars). Data are for N = 154, with 53 audiology best practices (AB), 51 consumer decides/over-the-counter (CD), and 50 placebo device (P) participants included. Purchase-price subgroups are “red” (reduced; $600) and “typ” (typical; $3,600). All HHIE scores are total scores (emotional and social subscales combined). Error bar = +1 standard error; HHIE benefit = difference between aided and unaided Hearing Handicap Inventory for the Elderly scores.

Figure 10.

Means for the Hearing Aid Satisfaction Survey (HASS) satisfaction (left ordinate) and usage (right ordinate) outcome measures for each of the six subgroups evaluated in the 3 × 2 factorial analysis. Measures of satisfaction with hearing features and function, HASShaf (light gray bars); satisfaction with dispenser-related issues, HASSdisp (black bars); and daily hearing-aid usage from the datalogger (gray bars) are provided. Data are for N = 154, with 53 audiology best practices (AB), 51 consumer decides/over-the-counter (CD), and 50 placebo device (P) participants included. Purchase-price subgroups are “red” (reduced; $600) and “typ” (typical, $3,600). Error bars = +1 standard error.

Figure 11.

Comparison of real-ear output to NAL-NL2 targets after hearing aids for consumer decides/over-the-counter (CD; left) and placebo device (P; right) participants were reprogrammed to match targets following audiology best practices (AB) procedures in Session 3. Mean and ±1 standard deviation are shown in each panel. Top panels show data for the left ear and bottom panels for the right ear. The stimulus was a 65-dB SPL speech signal (“carrot passage”) for the speech-mapping measures from the Verifit test system.

Figure 12.

Mean PHABglobal and PHABavds scores obtained from the consumer decides/over-the-counter (CD) group (left panel; n = 48) and the placebo device (P) group (right panel; n = 46) during Sessions 3 (black bars) and Session 3B (gray bars). The left-pointing arrows show the mean values for the original audiology best practices (AB) group from Session 3 for the lefthand outcome measure in each panel, and the right-pointing arrows show the corresponding Session 3 AB values for the righthand outcome measure in each panel. Error bar = +1 standard error; PHABglob = difference between aided and unaided scores of PHAPglobal (Profile of Hearing Aid Performance, average of the five communication-related subscales: Familiar Talkers, Ease of Communication, Reverberation, Reduced Cues, and Background Noise); PHABavds = difference between aided and unaided scores of PHAPavds (PHAPavds = Profile of Hearing Aid Performance, average of the Aversiveness of Sound and Distorted Sound subscales). *Significant (p < .05) paired-sample t tests in each panel.

Figure 13.

Mean CST benefit, in rationalized arcsine units (RAU), and Hearing Handicap Inventory for the Elderly benefit (HHIE ben) scores obtained from the consumer decides/over-the-counter (CD) group (left panel) and the placebo device (P) group (right panel) during Sessions 3 (black bars) and Session 3B (gray bars). The left-pointing arrows show the mean values for the original audiology best practices (AB) group from Session 3 for the lefthand outcome measure in each panel, and the right-pointing arrows show the corresponding Session 3 AB values for the righthand outcome measure in each panel. Error bar = +1 standard error; CST ben = difference between aided and unaided Connected Speech Test scores. *Significant (p < .05) paired-sample t tests in each panel.

Figure 14.

Mean HASShaf and HASSdisp scores obtained from the consumer decides/over-the-counter (CD) group (left panel) and the placebo device (P) group (right panel) during Sessions 3 (black bars) and Session 3B (gray bars). The left-pointing arrows show the mean values for the original audiology best practices (AB) group from Session 3 for the lefthand outcome measure in each panel and the right-pointing arrows show the corresponding Session 3 AB values for the righthand outcome measure in each panel. Error bar = +1 standard error; HASShaf = Hearing Aid Satisfaction Survey, items concerning hearing aid features; HASSdisp = Hearing Aid Satisfaction Survey, items concerning dispenser-related processes. *Significant (p < .05) paired-sample t tests in each panel.

Figure 15.

Mean usage outcome (left ordinate) and Practical Hearing Aid Skills Test–Revised (PHAST-R) scores (right ordinate) obtained from the consumer decides over-the-counter (CD) group (left panel) and the placebo device (P) group (right panel) during Sessions 3 (black bars) and Session 3B (gray bars). Note that the PHAST-R proportions have a maximum of 1.0 but are multiplied by 10 for appropriate scaling in this figure. The left-pointing arrows show the mean values for the original audiology best practices (AB) group from Session 3 for the lefthand outcome measure in each panel and the right-pointing arrows show the corresponding Session 3 AB values for the righthand outcome measure in each panel. Error bar = +1 standard error. *Significant (p < .05) paired-sample t tests in each panel.

Figure 16.

Percentage of participants, assigned at the start of the trial to placebo device (P) and consumer decides/over-the-counter (CD) group and who participated in the extra 4-week trial with audiology best practices (AB) service delivery, who ultimately kept their hearing aids after Session 3B. This is plotted against these same participants' plans to keep the hearing aids when asked at the conclusion of the initial 6-week trial and prior to revealing their group.

Figure 17.

The X, Y, and Z audiograms from Figure 4 (filled symbols) and the median audiograms of the subgroups of consumer decides/over-the-counter (CD) participants (unfilled symbols) who had chosen hearing aids programmed to match NAL-NL2 targets for the X, Y, and Z audiograms. Only data for the left ears of the CD participants are shown here for clarity.