Approach to patient reported outcome measures selection and implementation in a chronic norovirus clinical efficacy trial for patients after solid organ transplant

Norovirus is the second most common cause of diarrhea among solid organ transplant recipients. There are currently no approved therapies for Norovirus, which can substantially impact quality of life, particularly in an immunocompromised patient population. In order to establish clinical efficacy and support any medication claims related to a patient's symptoms or function, the Food and Drug Administration requires that a trial's primary endpoints be derived from patient reported outcome measures‐ an outcome that comes directly from the patient, without interpretation of the patient's response by a clinician or anyone else. In this paper, we describe our study team's approach to the definition, selection, measurement and evaluation of patient reported outcome measures as part of establishing clinical efficacy of Nitazoxanide for acute and chronic Norovirus in solid organ transplant recipients. We specifically describe our approach to measuring the primary efficacy endpoint‐ days to cessation of vomiting and diarrhea after randomization through 160 days, measured through daily symptom diaries as well as the impact of treatment on exploratory efficacy endpoints, particularly change in the impact of norovirus on psychological function and quality of life.

immunoglobulin, breast milk and nitazoxanide used clinically in limited case series. 1 We conducted "a phase 2 multi-center, prospective, randomized, double-blind study to assess the clinical and antiviral efficacy and safety of nitazoxanide for the treatment of norovirus in hematopoietic stem cell and solid organ transplant recipients (the NNITS study)" to determine the clinical efficacy of nitazoxanide for the management of acute and chronic norovirus in solid organ transplant recipients with secondary aims to assess its virologic efficacy and safety in the same population.
To establish clinical efficacy and support any medication claims related to a patient's symptoms or function, the Food and Drug Administration requires that a trial's primary endpoints be derived from patient reported outcome (PRO) measures-"any report of the status of a patient's health condition that comes directly from the patient, without interpretation of the patient's response by a clinician or anyone else. 3,4 " PROs used in clinical trials measure the effect of a medical intervention the patient received (or didn't receive) on one or more constructs (symptoms or quality of life) that can only be known to the patient. PROs can either be considered in absolute terms (e.g., symptom severity at a single point in time), or in reference to an earlier assessment (change in symptom severity over time).
In this paper, we describe our study team's approach to the selec- In addition, we considered the independence of the measures from each other to avoid asking the same question multiple times, the reliability and validity and other psychometric properties of the measures, patient acceptability of completing the measure-feasibility of completing, types of questions, frequency of administration, interpretability of the scores (what does a significant change in nausea look like), and since this was a clinical trial, how sensitive the measure might be to change over the course of 1 year. It is also best practice to "triangulate" or measure in multiple ways the top 1 or two constructs you are most interested in, in our case norovirus symptoms (diarrhea, vomiting, fever, loss of appetite and nausea). Finally, we considered measures, when possible, that could eventually be used in pediatric as well as adult populations and when possible, had a Spanish translation (See Figure 1). F I G U R E 1 In step 1 of our instrument selection process, we interviewed 10 adult patients who had been admitted to the hospital with norovirus, and who were at least 6 months past a solid organ transplant. In step 2, The NNITS investigative team agreed upon the key PRO constructs for final consideration for the clinical trial.

Approach to Measurement of selected constructs
Step 3 of PRO selection took the recommendations from the patient and clinician stakeholders and began to focus on picking the most sensitive measure based on our research questions, the clinical trial's study design and feasibility for patients to complete and study staff to collect.
of anti-motility agents and medication adherence were also of interest and the study team decided that these would need to be monitored, via self-report on a daily basis to accurately capture the primary endpoint.
We adapted an existing, well-validated daily symptom diary used by study author LK in a behavioral clinical trial of irritable bowel syndrome (NCT00738920).
We created 2 versions of the diary (in English and Spanish) to reduce participant burden and ensure complete collection of data through the long duration of the study: • During active treatment (days 1-28), we used a daily diary to track medication adherence, symptom frequency and severity. The wellvalidated, visual Bristol Stool Form Scale 5,6 was incorporated into the diary to ensure that patient definition of stool consistency was accurately captured. The diary was tested on 5 patients, who underwent cognitive interviews to ensure each item was understood and clear prior to executing in the larger trial. We were particularly interested in ensuring participants clearly understood the direction not to "backfill" the diaries, but rather skip the day and move on to the next day in order to ensure symptom perception was not influenced by recall bias (See Figure 2). • Next, a shorter daily diary was created to capture occurrence of core Norovirus symptoms from Day 29-180-here the same items were captured, but as a dichotomous, yes/no question. This approach reduced participant burden but still allowed us to address the primary outcome of relapse/failure.

Change in mood, anxiety, and function over time
In addition to changes in Norovirus symptoms, we expected that if the medication was working and Norovirus symptoms were improving, there would be subsequent changes in the patient's mood, anxiety, sleep, fatigue, and daily functioning. We also implemented a selfreport questionnaire that focused on gastrointestinal symptoms more broadly-abdominal pain, incontinence and bowel urgency to track alongside the daily symptom diary.

The Patient Reported Outcomes Measurement Information System
(PROMIS), supported by the NIH Common fund was used whenever possible. We saw this as an advantage because many of the measures in the PRO Bank also had Spanish translation, established normative values, had short forms and pediatric versions. The PROMIS Global Health measure had also been used in one study of clostridioides difficile. 7 but otherwise infectious disease studies have had limited use of PROs to date. We chose the following, administered at baseline, day 28, and at each follow up (days 60, 120, 180): • PROMIS global health (8 items) • PROMIS fatigue (7 items) • PROMIS anxiety (6 items) • PROMIS depression (8 items) • PROMIS sleep disturbance (8 items) • PROMIS gastrointestinal (6 of 8 symptom scales)

Change in quality of life over time
Finally, we considered the broader construct of quality of life, defined both generally and as a disease-specific construct.
• The EuroQol-5D, is a 5-item, preference-based QOL measure that will also allow us to QOL limitations as well as gauge the economic impact of treatment in terms of quality-adjusted life year. It has previously been used in inflammatory bowel disease and irritable bowel syndrome clinical trials. 8 • The IBS-QOL scale (IBS-QOL) 9 is a 34 item disease-specific QOL measure that was validated in irritable bowel syndrome and reflects similar areas of concern in this population (e.g., embarrassment, eating behavior, bowel habits, symptom anxiety). 10

Training of clinical team on PRO collection
Once measures were chosen and timelines (Supplemental Material Figure S1) were agreed upon, the clinical research coordinators underwent training around how to optimize PRO collection in the study.
Strategies were developed to encourage patients to become a "citizen scientist," ensuring that each participant understood the critical nature of self-monitoring and the importance of accuracy-for example, we encouraged patients to skip a day rather than backfill their diary, we acknowledged that not all questions/items applied to everyone, and that it was important to hear from patients who did not have these (psychological) problems as well.
We prioritized 3 PRO measures for collection at each assessment, including if patients were to dropout or terminate early-PROMIS GI symptoms, IBS-QOL and EUROQOL.

Approach to Analysis of the PROs
For the diary related endpoints (primary and secondary), data were analyzed as the median time (days) to clinical improvement at Day 28 and through Day160. Kaplan-Meier curves for each treatment arm were also calculated. Diaries were considered complete if >80% of data were recorded (5 days a week or more). Change in all secondary patient-reported outcome measures was assessed using ANCOVA, adjusting for baseline mean values, for each patient-reported outcome measure by group. Study outcome data will be presented elsewhere.

DISCUSSION
In this paper, we describe our approach to patient reported outcome measure selection in a study of patients struggling with acute and  11 and inflammatory bowel disease (IBD) 12 (see ICHOM guidelines for PRO measure use in IBD). 13 There is also precedent for use of PROs in other acute and chronic infectious diseases including COVID-19. 14 and HIV 15 .
We have conveyed the complexity of decision making around PRO selection, demonstrating the importance of starting and ending with the patients, sharing patient experience with science and clinical stakeholders in order to determine the best balance between participant burden and the need to show change/improvement over time.
One of the highlights of the study was our work adapting a validated, daily symptom diary that has been used recurrently in clinical trials of medications treating bowel disorders, particularly diarrhea.
There were advantages and disadvantages to this approach-the primary advantage we had when choosing to adapt versus create was time. Developing a patient facing measurement tool from scratch, while ideal, can take years to validate, especially, as in our case, among a rare patient population.
In conclusion, regardless of the clinical trial outcome, the time and effort we put into selecting our patient reported outcome measures will generate important information on the psychosocial burden and impact of Norovirus on recovery from transplant-and bring attention to the ongoing medical needs of immunocompromised patients for future studies. Further, it also provides a proof of concept for an approach that could be used in studying newer agents in immunocompromised and immunocompetent patients with acute and chronic Norovirus infections.

AUTHOR CONTRIBUTIONS
LK conceptualized, wrote, and revised the paper. MGI obtained funding for the study, conducted the study, and wrote and revised the paper. AD revised the paper. All authors approved the final version of the paper.