Letter to the Editor: Reporting of Data to Inform the Design of a Definitive Trial Re
This letter highlights the study by Henry et al. (2015), which sought to assess whether combination devices with full amplification features and an added sound generation feature provide greater benefit to patients than hearing aids alone. Finding the answer to this question is important and so this study is a welcome addition to the literature. However, the published randomized controlled trial used a pilot trial design that would not provide a definitive answer to the research question. Indeed, the authors concluded by saying “A larger randomized controlled trial is needed to more definitely address these questions.” pp. 51. A small group of 30 patients with an aidable hearing loss were fit binaurally with a pair of commercially available receiver-in-the-canal combination devices and received tinnitus counseling. Half the patients were prescribed the full features of the device, while half received amplification only. Both groups appeared to benefit at 3 months in terms of a reduction in tinnitus symptoms measured by the tinnitus functional index (TFI) and improvement in hearing handicap, measured by hearing handicap inventory for elderly. Over the years, there have been a number of lessons learned about trial design for tinnitus (e.g., Agency for Healthcare Research and Quality 2013; Landgrebe et al. 2012; Tyler et al. 2006). Yet, a recent systematic review reports evidence that such recommendations still tend to be overlooked, even in those studies published subsequently (Hall et al. 2016). Examples include careful justification and reporting of the choice of primary outcome measure and the study sample size. The body of knowledge about what constitutes good trial design highlights several further incremental analyses that could be implemented on the pilot data collected by Henry et al. and which could help to inform what the design of that larger randomized controlled trial might look like. I therefore invite the authors to comment on these important issues so that this knowledge can assist the tinnitus community in designing that good quality trial.
First, it is worthwhile to take a step back to reflect on what specific tinnitus-related symptoms one might expect a combination device to change. The TFI provides a composite measure of eight different symptom domains (the intrusiveness of tinnitus, the patient’s sense of control, cognitive interference, sleep disturbance, auditory issues, relaxation issues, quality of life, and emotional distress). It is reasonable to make the case that some subscale scores should be more responsive to combination hearing aid intervention than others. Indeed Meikle et al. (2012) do discuss this point generally by saying “some.. of the subscales may prove useful as subsidiary outcome measures in studies of treatment efficacy.” For example, tinnitus masking provided by the sound generator should reduce the intrusiveness of tinnitus, while the device may have little effect on sleep quality because it is not worn at night. Moreover, one could also make the case that both combination devices and amplification-only devices should improve auditory issues equally. If this pattern was emerging from the data that was published in Ear and Hearing, then this information could be used to inform what the primary outcome in a definitive trial might be. In particular, scores from TFI subscales that were insensitive (or less sensitive) to patient benefits from combination devices would be diluting the sensitivity of a global TFI score. The data presented in Table 4 by Henry et al. (2015) indicate very wide variability in treatment-related change from person to person. Notably, the range of change was +15 to −82 in the full-features group and −5 to −83 in the amplification-only group.