Humes 2017.

Methods	3‐arm, double‐blinded, single‐centre, parallel‐arm RCT, with 6 and 10 weeks duration of treatment and follow‐up
Participants	Location: USA (Indiana) Setting of recruitment and treatment: university research clinic Sample size: Number randomised: 164* (53 intervention (audiology best practice, AB), 55 intervention (consumer decides, CD), 55 control (placebo). Number completed: 154 (53 intervention (audiology best practice, AB), 50 intervention (consumer decides, CD), 51 control (placebo); attrition n = 10 (6.0%), ear or health problems n = 3, fitting problem or non‐use n = 2, unable to complete hearing aid selection n = 4, *withdrawal after randomisation and before the fitting session n = 1) Participant (baseline) characteristics: Age: mean 69.1 years (SD 6.1) Gender: 92 male, 72 female Main diagnosis: hearing loss (PTA averaged across 0.5, 1.0, 2.0 kHz = 28.1 dB HL (SD 8.0); high frequency PTA averaged across 1.0, 2.0, 4.0 kHz = 38.8 dB HL (SD 7.9) Other important effect modifiers: none Inclusion criteria: age 55 to 79 years, English as native language, MMSE score > 25, no prior hearing aid experience, pure‐tone audiometry (air) consistent with age‐related hearing loss within the fitting guidelines of this study, bilaterally symmetrical hearing loss. Exclusion criteria: presence of a medically treatable ear condition, bilateral, flat tympanograms, known fluctuating or progressive HL, presence of cognitive, medical or language‐based conditions that limit ability to complete all test procedures, currently or recently taking platinum‐based cancer drugs or mycin‐family antibiotics, previously diagnosed with either multiple sclerosis or Ménière's disease, failure to seek or waive medical evaluation and clearance following hearing evaluation, unwillingness to be randomly assigned to a treatment group.
Interventions	Intervention group, AB and CD (n = 108): active hearing aids (ReSound Alera Mini), behind‐the‐ear, fully digital. Bilateral fits. Fixed directional microphones, dynamic feedback suppression and noise reduction unclear if enabled. AB: fitted using real‐ear measurements according to the NAL‐NL2 target, with adjustments as necessary. Verified via real ear measurements using Audioscan Verifit system. CD: 3 possible prescriptions based on NAL‐NL2 fit to the 3 most common patterns of hearing loss among older adults in the US. Different programmes applying different constant gains across all frequencies (gain values based on chosen typical prescription). Comparator group (n = 51): placebo hearing aids (ReSound Alera Mini), behind‐the‐ear, fully digital. Bilateral fits. Fixed directional microphones (n = 20), omni‐directional microphones (n = 23), dynamic feedback suppression and noise reduction enabled. Programmed to achieve 0 dB insertion gain to control for any occlusion effect. Verified via real ear measurements using Audioscan Verifit system. Use of additional interventions (common to both treatment arms): none up to 6 weeks post‐baseline, then the CD group was offered AB‐delivered hearing aids for a further 4 to 5 weeks trial.
Outcomes	Primary outcome: Profile of Hearing Aid Benefit Secondary outcomes: Connected Speech Test, Hearing Handicap Inventory for the Elderly, Hearing Aid Satisfaction Survey
Funding sources	National Institute on Deafness and Other Communication Disorders R01 DC011771
Declarations of interest	None reported
Notes	AB and CD were combined into the intervention group as hearing aids not service delivery models were of interest. The 6‐week follow‐up measured the primary and secondary outcome measures. Participants paid for their hearing aids.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Based on random number generation prior to study initiation, blocked by unaided Connected Speech Test performance (low, medium, high).
Allocation concealment (selection bias)	High risk	The clinical trial co‐ordinator (CTC) allocating patients had access to the randomisation lists and allocation was not concealed.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	All markings and materials revealing manufacturer or model of the devices were obscured. The CTC was not blinded to patient allocation and there were several opportunities where there was potential for participants to be unblinded.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Outcome assessor (audiologist 4) was blinded to allocation of intervention group. However, it is unclear whether all the participants remained effectively blinded.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All dropouts and loss to follow‐up described and reasonable.
Selective reporting (reporting bias)	Low risk	Data on all outcomes in the published study protocol (NCT01788423) were reported. The study authors provided the full data set for this review upon request.
Other bias	Low risk	Study appears to be free of other sources of bias.