Trial Outcomes in Glomerular Diseases

In this issue of CJASN, the Standardized Outcomes in Nephrology (SONG) initiative reports results from a series of workshops with the goal of identifying a set of common clinically meaningful outcomes across all glomerular diseases (¹). For the nonclinical trialist, an “outcome” is what is being compared across treatment arms to assess efficacy and safety (e.g., “proteinuria remission status at week 8” or “change in estimated glomerular filtration rate by month 12”). Carter et al. suggest progress across trials is stymied by different trials reporting and analyzing different outcomes, and that often patients and caregivers have minimal input into which endpoints are most important and should be studied.

On the one hand, glomerular disease represents a heterogenous group of conditions and perhaps there is value in using disease-specific outcomes. For example, it would not seem wrong to only use outcomes relevant to patients with FSGS in a trial limited to such patients (say, if the method of action was thought to only potentially benefit patients with FSGS). On the other hand, as the authors argue, there are many similarities in the patient experience across different glomerular diseases. Even though they do not make this point directly, harmonizing outcomes across glomerular diseases could also provide an opportunity to streamline and scale up glomerular disease trials. If we had a potential therapy with a method of action that could be used to treat more than one type of glomerular disease, there would be a benefit in performing “pooled” trials where patients with multiple diseases (such as basket, umbrella, and platform trials) are enrolled (²). Such a trial would likely require a common set of outcomes relevant to all included glomerular diseases. This can improve glomerular disease trials overall by making them more feasible by bringing similar populations together, making it easier to reach recruitment enrollment goals, and improving and streamlining the institutional knowledge needed for successful trials.

The SONG initiative conducted a series of in-person focus groups and workshops with patients, care partners, and health professionals from 19 countries around the world. Quantitative analyses were based on the Delphi method: a multistage/iterative approach where participants rank order in which outcomes are most important to them, and then outcomes are removed in order of lowest importance before the process repeats until the top handful of outcomes emerge. The final stage identified a set of “core outcomes” that were “critically important to all stakeholder groups,” and the authors recommend that these be “reported in all [glomerular disease] trials.” Again, the goal was to identify outcomes most relevant to all glomerular diseases. For example, as one might expect, a relapse/remitting pattern was found to be less relevant in IgA nephropathy than in Minimal Change Disease (and thus not the top-tiered outcome to be reported in all glomerular disease trials), and “need for dialysis or transplant” was not as relevant for Minimal Change Disease as it was for IgA. The most significant “core outcomes” included kidney function, disease activity, death, life participation, and cardiovascular disease.

These outcomes seemed relatively well expected. Patients with glomerular disease, caregivers, and providers agree that better kidney function and living a longer and more fulfilling life are good, and active disease and cardiovascular disease are bad. Still, there are a few interesting themes worth emphasizing.

Patients consistently described “life participation” (e.g., “the ability to participate in meaningful life activities”) as more important than individual symptoms such as edema, fatigue, and anxiety. It would be interesting to see additional and more nuanced examples from the interview transcripts and meeting notes of what patients mean by this, but it appears to be, broadly speaking, being able to return to a regular routine, enjoy hobbies, and attend important events—although not necessarily be completely symptom free. The SONG team identified the need for a standardized and validated instrument, and I agree; I am unaware of a validated instrument that measures this concept precisely. For readers unfamiliar with the term, a “validated instrument” usually refers to a standardized questionnaire where respondents are asked a set of questions about their quality of life, and they respond often on a “Likert scale,” from “strongly disagree” to “strongly agree,” etc. For example, the SF-36 asks, “Does your health now limit [moderate activity such as moving a table, pushing a vacuum cleaner, bowling, or playing golf?]” and respondents reply “yes, limited a lot,” “yes, limited a little,” or “no, not limited at all” (³). Developing new instruments is cumbersome. This requires (1) focus groups to identify the right content and wording of questions, (2) a test sample of respondents to test how well the items correlate together, (3) psychometric analyses to develop a scoring algorithm, and (4) a validation test to correlate the finalized test scores against other existing instruments/measures that we would expect the scores to be correlated with.

The authors highlight an ongoing effort to develop a “life participation” instrument from measures used in the transplant population (⁴), and hopefully the SONG initiative can leverage patients in these meetings to help determine if these instruments are relevant to their concept of “life participation.” I worry existing surveys look to be more closely tailored to different aspects of functional independence, and this might be distinct from “life participation”; it will be interesting to see what they will find. We have only observed modest differences in global and physical patient-reported outcome measures when comparing disease activity or comparing those with or without edema (⁵,⁶). Although the importance of patient-reported outcomes are well recognized (⁷), perhaps we need more precise measures that more directly ask patients if they have missed out on an activity they wanted to do that day by, for example, using end of day assessments via text or through an app.

Although I somewhat disagree with the need for a life participation instrument specific to glomerular disease, this seems like a construct important for many chronic diseases. It also seems like meetings such as these would be an excellent opportunity to ask patients if the existing instruments are capturing what they mean by “life participation.”

Similar to the “life participation” issue, patients described an interesting nuance for the outcome of “disease activity.” They eschewed the usual “relapse/remission” dichotomy for disease activity, and they interpreted this on a much broader continuum, creating similar measurement issues to those for life participation. However, generally speaking, continuous endpoints have greater statistical power than categorical ones. Therefore, it is easier to detect a difference in means with limited sample size than a difference in frequencies/percentages. However, it is also important to know that continuous differences are clinically important. It is convenient that the core outcome set mostly identifies continuous outcomes.

To be clear though, the authors identify that this level of detail, determining which specific measure to use, is the next step in the initiative. The goals at this stage were to identify the key “domains.”

It is worth emphasizing that the authors do not appear to be suggesting that these are the endpoints that should be explicitly included as primary endpoints for drug approval. Nor are they making distinctions between earlier phase II or later phase III trials—just that these endpoints should be collected and reported in trials.

Still, they suggest perhaps “core outcomes in trials to be mandated by funders and regulators” but also stress the importance of a core outcome being “cheap” and “easy” and that it should “impose little measurement burden on the patient.” This can be a difficult balance. Readers are likely well aware that there are fewer clinical studies and clinical trials within nephrology relative to other disciplines (⁸). Clinic and research staff tend to have less clinical trial–related experience than in other disciplines. It is possible that this kind of “scaling up” to include standardized reporting across glomerular diseases may lead to long-term improvements in the quality and number of trials in nephrology broadly and glomerular diseases specifically. However, in the short term, mandatory reporting of certain outcomes could also create extra administrative burden, and it may be difficult to assess which measures have the lowest response burden when the patient interviews and meetings, by design, will only include the most motivated patients and researchers (e.g., an inherent and almost unavoidable selection bias reflecting the most motivated folks participating in the workshops). I would recommend including research study staff in future meetings like these as they may have unique insights into the data quality and collection of these instruments. They may be able to balance these recommendations with what is feasible and what we can do well.

Disclosures

J. Troost is supported in part by the National Center for Advancing Translational Sciences (NCATS) for the Michigan Institute for Clinical and Health Research grant UL1TR002240.

Funding

None.