NRP/513 Week 5: Research Critique, sample paper

A Negative Primary Result and a Promising Subgroup: A Structured Critique of the ACHIEVE Hearing Intervention Trial for a Primary Care Audience

[Student Name]

University of Phoenix

NRP/513: Clinical Applications of Theory and Research

Week 5 Assignment

[Instructor Name]

[Date]

The clinical framing is a composite written for a model paper.

In Week 4, I examined three designs on hearing loss and cognition. This week I appraise one study in depth: the ACHIEVE randomized trial (Lin et al., 2023). I chose it because it is the strongest test so far of whether treating hearing loss changes cognitive decline, and because its mixed results are easy to misread in either direction. I use the CONSORT 2010 statement (Schulz et al., 2010) to judge reporting and the GRADE approach (Guyatt et al., 2008) to judge certainty.

Summary of the Study

Across four U.S. field sites, the investigators enrolled 977 people in their seventies and early eighties whose mild to moderate hearing loss had never been treated and whose cognition was still largely intact, and split them by chance between two arms. One arm received audiologist care: fitted hearing aids, counseling about hearing and other listening devices. The other met with a health educator for sessions on healthy aging. Enrollment drew on two pools, 238 volunteers already followed in the Atherosclerosis Risk in Communities study, an older group with more risk factors and lower cognitive scores at entry, and 739 people recruited fresh from the community. The main endpoint was change in a composite cognitive score over three years. Pooling both pools, the two arms declined at almost identical rates. A sensitivity analysis written into the protocol in advance pointed to slower decline with hearing care among the higher-risk volunteers and no difference among the fresh recruits (Lin et al., 2023).

The Question

The question is clear and important: does a hearing intervention reduce cognitive decline in older adults with hearing loss? It follows directly from observational evidence of an association between hearing loss and dementia. The question is appropriately answered with a randomized trial.

Design and Randomization

A randomized controlled trial is the strongest design for testing whether an intervention causes a difference. CONSORT asks whether the report describes how the random sequence was generated and how allocation was concealed (Schulz et al., 2010); the report addresses both. Random assignment should balance known and unknown confounders between groups, and the published baseline characteristics appear similar across arms.

Blinding

Masking participants was impossible, since anyone fitted with hearing aids knows it. That is unavoidable for this kind of intervention. Blinding of outcome assessors is therefore important, since those administering cognitive tests could be influenced by knowing a participant's group, and the reader should confirm how the trial handled this. The use of an active control, health education with similar contact time, is a strength: it controls for the social contact and attention that could themselves affect cognition.

An attention-matched control means that any difference found would be due to hearing care, not to the extra visits that came with it.

Sample

The sample of 977 was large for a behavioral intervention trial, drawn from four U.S. sites. Its limits affect applicability. Participants were mostly White, 88%, which limits generalizability to the diverse older adults in many primary care practices. They were volunteers willing to wear hearing aids and attend sessions, likely more motivated than typical patients. Exclusion of those with substantial cognitive impairment was appropriate for a prevention question but means results do not apply to people already impaired.

The Intervention

The intervention was thorough, delivered by audiologists with counseling and follow-up. This strengthens the test of the question but limits applicability to settings without such resources. A primary care practice recommending over-the-counter devices without audiologist support would be delivering a different, less intensive intervention.

Outcomes

The primary outcome, three-year change in a global cognitive score, is appropriate and measured with standardized tests. A dementia diagnosis usually takes longer than three years to emerge, so the trial was never designed to show an effect on dementia itself. Secondary outcomes, including domain-specific cognition and measures of communication and quality of life, offer a fuller picture, and readers should look at them rather than only the headline cognitive result.

Analysis

The primary analysis followed the prespecified plan and found no significant difference. The sensitivity analysis comparing the two cohorts was prespecified, which lends it more weight than a post hoc subgroup analysis. Still, subgroup findings carry higher risk of chance results than primary analyses, and the higher-risk cohort was small, 238 participants. The finding generates a hypothesis worth testing but is not proof of benefit.

Certainty Using GRADE

Guyatt et al. (2008) rate certainty by asking five questions of the evidence: how likely it is biased, whether studies agree, how directly they address the question, how precise the estimates are and whether unpublished results might be missing. For the question of whether hearing intervention slows cognitive decline in older adults generally, a single large trial with low risk of bias and a null primary result provides moderate-certainty evidence of no substantial effect over three years. For the question of benefit in higher-risk older adults, the evidence rests on a subgroup of one trial and would be rated low certainty because of imprecision and the need for replication.

Participant Flow and Adherence

CONSORT asks reports to show how many participants were screened, randomized, followed and analyzed in each arm (Schulz et al., 2010). ACHIEVE screened 3,004 people to randomize 977, a ratio that shows how selective enrollment was. Readers should also look at how many participants in the hearing arm actually wore their devices regularly, since a null result could partly reflect low use rather than lack of effect. Adherence is a question the report addresses and that matters directly to my clinical problem, which is continued use.

Funding and Conflicts

The trial was funded by the National Institutes of Health, and hearing aid manufacturers did not fund it, which reduces concern about sponsor influence on a result that could have favored their products.

Applicability to My Clinic

My clinic's older patients are more diverse than the trial sample, often have multiple chronic conditions and would rarely receive audiologist-delivered care at the intensity studied. Some resemble the higher-risk cohort. The trial cannot tell me whether hearing aids will protect their cognition, but it does not undermine the value of hearing care for communication and quality of life.

What a Clinician Can Conclude

A responsible conclusion is that hearing care should be offered for its established benefits in communication and daily life, that patients should not be told hearing aids prevent dementia and that older adults at higher risk of cognitive decline might benefit cognitively, a possibility that further trials should test.

Questions I Would Ask the Authors

I would ask how many hours a day intervention participants wore their devices and whether cognitive change differed by level of use.

Conclusion

ACHIEVE is a well-designed, well-reported randomized trial with an attention-matched control, appropriate outcomes and a prespecified analysis plan. Its primary result was null, and its promising finding in higher-risk older adults rests on a smaller, less certain subgroup. For primary care, it supports offering hearing care for communication while speaking honestly about what is and is not known about cognition.

References

Guyatt, G. H., Oxman, A. D., Vist, G. E., Kunz, R., Falck-Ytter, Y., Alonso-Coello, P., & Schünemann, H. J. (2008). GRADE: An emerging consensus on rating quality of evidence and strength of recommendations. BMJ, 336(7650), 924-926. https://doi.org/10.1136/bmj.39489.470347.AD

Lin, F. R., Pike, J. R., Albert, M. S., Arnold, M., Burgard, S., Chisolm, T., Couper, D., Deal, J. A., Goman, A. M., Glynn, N. W., Gmelin, T., Gravens-Mueller, L., Hayden, K. M., Huang, A. R., Knopman, D., Mitchell, C. M., Mosley, T., Pankow, J. S., Reed, N. S., . . . Coresh, J. (2023). Hearing intervention versus health education control to reduce cognitive decline in older adults with hearing loss in the USA (ACHIEVE): A multicentre, randomised controlled trial. The Lancet, 402(10404), 786-797. https://doi.org/10.1016/S0140-6736(23)01406-X

Schulz, K. F., Altman, D. G., & Moher, D. (2010). CONSORT 2010 statement: Updated guidelines for reporting parallel group randomised trials. BMJ, 340, Article c332. https://doi.org/10.1136/bmj.c332