Patient-Reported Outcomes as End Points in Nephrology Trials

There is increasing recognition in the nephrology community of the need to shift research, quality metrics, and health policy priorities to patient-reported outcomes (PROs) that align with person-centered care. PROs are more meaningful to patients than biochemical outcomes, reflect their quality of life, and are identified as important outcomes by patients, care partners, clinicians, researchers, and policy makers.¹ There is a call for action by patient advocacy groups, the Kidney Health Initiative, US Food and Drug Administration, and Kidney Disease Outcomes Quality Initiative and Kidney Disease Improving Global Outcomes to prioritize these outcomes for research.^2,3 Keeping the Standard Protocol Items: Recommendations for Interventional Trials PRO Extension recommendations in mind, we discuss some best practices and considerations for trials with PROs as end points in patients with kidney disease (Figure 1).⁴

Figure 1.

Best practices and considerations for trials with PROs as end points in patients with kidney disease. COMET, Core Outcome Measures in Effectiveness Trials; PRO, patient-reported outcome; SONG, Standardised Outcomes in Nephrology.

Patient- and Collaborator-Informed Study Design

Answering the right question for the right patient in the right setting requires using the right trial design. Clinical trials in nephrology should ensure that the intended trial design is aligned with research questions and outcomes that are relevant and meaningful. Thus, involving patients and key partners early in the design process and including a multidisciplinary approach is critical. For example, the ongoing Heart Outcomes Prevention Evaluation (HOPE) Consortium Trial to Reduce Pain and Opioid Use in Hemodialysis incorporated patient perspectives and lived experiences and input from investigative team representing nephrology, psychology, biostatistics, pain, and addiction medicine to inform protocol development, outcome selection, recruitment and retention strategies, results interpretation, and dissemination plans.⁵

Trial Design

As in all clinical trials, the eligibility criteria should aim to maximize broad representation of the patient population to enhance generalizability. This may be challenging in trials evaluating PROs. Patients with heavy symptom burden may be less likely to enroll or engage with the trial intervention, contributing to selection bias. In the A Trial of Sertraline vs. Cognitive Behavioral Therapy for End-stage Renal Disease Patients with Depression (ASCEND) trial for depression in patients on hemodialysis, one third of the patients with depression at screening declined treatment and did not consent.⁶ Among those who enrolled, one third of patients did not accept any treatment for depression even after an engagement interview that aimed to motivate treatment acceptance. The ASCEND trial highlighted the need to develop and test strategies that increase trial participation, intervention acceptance, and engagement.

Design consideration to ensure real-world implementation can facilitate transportability of interventions into routine clinical care. In our Technology-Assisted Stepped Collaborative Care trial for improving fatigue, pain, and/or depression among patients on hemodialysis, we used real-world treatment choices (pharmacotherapy and/or cognitive behavioral therapy), shared decision making, and telemedicine to deliver cognitive behavioral therapy and to enable future clinical adoption, scalability, and resource efficiency.⁷ Another design option is hybrid trials, which can simultaneously evaluate intervention effectiveness and implementation, thus expediting the translation of science and transformation of clinical care paradigms. Alternatively, mixed-method study designs can contextualize quantitative PRO data by collecting qualitative data that may offer insights into barriers and facilitators of intervention from the perspective of patients and key partners. Finally, novel registry-embedded trials using electronic devices for PRO data capture, such as the ongoing Symptom Monitoring with Feedback Trial, can provide pragmatic evaluation of clinical and cost-effectiveness of implementing PRO measurement (PROM) in routine clinical practice.⁸

Control group selection is another important design consideration. In some situations, comparison with an attention control, as opposed to usual care, may mitigate Hawthorne effect or observer bias in which participants may report an improvement in PROs just because of awareness that it is being studied. An attention control can simulate participant interactions similar to the intervention and may enhance interpretability of results. However, attention control groups have caveats. In some trials, an attention control may also cause harm and be an unnecessary expense. An attention control may also be an intervention if the attention received is not usual care, thus biasing results toward null effect and complicating the interpretation of a negative trial.

Finally, protocols to ensure not only participant retention and engagement but also completeness of PRO data collection are vital to minimize bias, especially in this population with high symptom and comorbidity burden, and hospitalizations. Thus, rigorous protocols and methodology for PRO data collection, research personnel training, quality monitoring, transparent reporting, and a priori statistical plans to handle data missingness and intercurrent events (which occur after trial initiation and affect data completeness or interpretation) are essential.

Outcome Selection

The Standardised Outcomes in Nephrology and Core Outcome Measures in Effectiveness Trials initiatives have used multistep, iterative processes to identify evidence-based and consensus-based core outcomes and develop measures informed by patients and key partners.^1,9 Equally important is the identification and alignment of research hypothesis and expected intervention effects with the right PRO domain. For example, in the HOPE Trial, input from patients and key partners informed the primary outcome to be pain interference rather than pain intensity because of its high priority, impact on life, and better generalizability across populations.⁵

PROM Tools

Selecting the ideal PROMs with robust psychometric properties in patients with kidney disease can be difficult because of the plethora and heterogeneity (i.e., PRO-specific versus disease-specific) of available instruments. A systematic review on fatigue measurements in patients on hemodialysis found 43 different tools varying in content dimensions, length, validity, and reliability in hemodialysis patients.¹⁰ To address the need for a short, standardized, and meaningful fatigue measure, the Standardised Outcomes in Nephrology group developed and validated a fatigue measure specifically for patients on hemodialysis.¹¹ Such kidney disease–specific PROMs may capture unique aspects of PROs in this population but may also make it difficult to compare results with other populations. The choice of PROM should also be driven by its responsiveness to treatment and ability to detect clinically meaningful change (i.e., extent of change that makes a real difference in patient's quality of life). A clinically significant change is more relevant than statistically significant change and should be used to determine sample size and interpretation of results. Additional considerations for optimal PROM tools are ease of comprehension, availability of validated non-English translations, floor and ceiling effect, recall bias, and time burden for completion.

Data Collection

Thoughtful planning of mode of survey administration is equally important. Self-administered questionnaires (on paper or digitally) may be challenging to complete for the elderly, minority groups, and those with high comorbidity and may exclude patients with low literacy, visual problems, and diminished dexterity, creating selection bias. Alternatively, computer-assisted telephone interviewing can provide an efficient way of administering these using blinded and trained interviewers and attain high rates of data collection, while limiting selection bias and study burden. Adaptive questionnaires that use contingency branching can personalize the assessment and survey experience to the respondent and provide rich data in a time-efficient manner.

The timing, frequency, and location of PROMs should be considered to reduce patient burden, uphold privacy, enable evaluation of short-term and long-term effects of interventions, and minimize potential confounding from the effects of the hemodialysis. A unique occurrence in patients on hemodialysis is the daily and diurnal variation in symptoms and their temporal relationship to hemodialysis. Conventional, recall-based questionnaires may fail to capture this, resulting in misclassification bias. Ecological momentary assessment is a tool that measures symptoms prospectively and repeatedly in real time, in the patient's natural environment, and may provide high-fidelity ascertainment of symptom severity and variability, particularly those potentially affected by hemodialysis like postdialysis fatigue. However, Ecological Momentary Assessment is resource-intensive and burdensome to patients; thus, its pros and cons should be balanced.

In some situations, proxy PROs reported by care partners based on their observation of patient experience may be valuable to reduce data missingness. Proxy PROs may substitute or complement patient self-assessment in trials where participants may not be able to self-assess PROs on their own—for example, young children, people with cognitive impairment, severe acute illness, or sudden clinical decline. However, careful selection of proxy and use of psychometrically sound questionnaires and methods for eliciting and interpreting proxy responses are needed to ensure concordance in patient-proxy assessments.

In conclusion, PROs as end points for nephrology clinical trials can accelerate change in quality of care and drive health policy changes to improve lives of millions of people living with kidney disease. Careful and innovative research design considerations that are driven by patient voice are needed to fuel this new era of nephrology research.

Disclosures

Disclosure forms, as provided by each author, are available with the online version of the article at http://links.lww.com/JSN/E677.

Funding

None.

Author Contributions

Writing – original draft: Manisha Jhamb, Linda-Marie Ustaris Lavenburg.

Writing – review & editing: Manisha Jhamb, Linda-Marie Ustaris Lavenburg.