Harnessing the patient voice in prostate cancer research: Systematic review on the use of patient‐reported outcomes in randomized controlled trials to support clinical decision‐making

Abstract Background Given the growing importance of patient‐reported outcomes (PROs) as part of “big data” in improving patient care, there is a need to provide a state‐of‐the‐art picture of the added value of using PROs in prostate cancer (PCa) randomized controlled trials (RCTs). We aimed to synthetize the most recent high‐quality PRO evidence‐based knowledge from PCa RCTs and to examine whether quality of PRO reporting in PCa research improved over time. Methods We conducted a systematic literature search using PubMed, from April 2012 until February 2019. For benchmarking purposes, we also included RCTs identified in our previously published review of RCTs (2004‐2012). Methodology for study identification and evaluation followed standardized criteria and a predefined data extraction form was used to abstract information. PRO quality of the studies was evaluated using the International Society of Quality of Life Research (ISOQOL) recommended criteria. Results A total of 55 new RCTs were published between April 2012 and February 2019. About 24 (43.6%) RCTs were found to be of high‐quality regarding PRO assessments. Of these, 13 (54.2%) have been reported in the most recent European Association of Urology (EAU) PCa Guidelines. Overall QoL and sexual, urinary, and bowel function were the most commonly reported PROs. FACT‐P, EPIC‐26, and EORTC QLQ‐C30 and/or its module PR25 were most frequently used as measurement tools. An overall improvement in the completeness of PRO reporting was noted over time. Conclusion Many PRO trials are currently not included in the EAU guidelines. Our findings suggest that there has to be a better consensus on the use of PRO data for PCa patients, which will then be reflected in the PCa Guidelines and future data collection. Homogeneity in PROs collection and measurement tools will in turn enable “big data” Consortia to increase the patients’ voice in clinical research.


| INTRODUCTION
Receiving a diagnosis of cancer and undergoing its subsequent treatments does not only have a detrimental impact on physical function, but may also affect psychological and social well-being of the individual. Patient-reported outcomes (PROs) addressing these domains of well-being (including both physical and psychosocial components) are increasingly being incorporated in randomized controlled trials (RCTs) to assess the effectiveness of cancer treatments. A recent Cochrane systematic review on psychosocial well-being and care needs of people with cancer 1 concluded that there is a need not only for more uniformity in outcomes and reporting, but importantly also for combining PROs with objective clinical outcomes. These conclusions are also in line with a recent report of the US Food and Drug Administration and the Critical Path Institute-as they are "committed to collaborate with international drug development stakeholders to identify rigorous methods to incorporate the patient perspective into the development of cancer therapeutics." 2 In the context of prostate cancer (PCa), we have previously evaluated the completeness of PRO reporting of 65 RCTs published between January 2004 and March 2012, which reported on PROs. 3 Significant improvements in PRO quality reporting over time were observed and it was estimated that about 20% of PRO RCTs had provided solid PRO data to allow health policy makers and clinicians to make a critical appraisal. Since 2012, over 800 new studies have been published on PCa and PROs, according to PubMed. Since then, PIONEER, an IMI2 funded pan-European public private partnership led by the European Association of Urology (EAU), is in the process of defining core outcome sets for PCa in the context of the patient's treatment pathway-whereby PROs from RCTs as well as real world evidence will be included. Core outcome sets for PCa will provide homogeneity in clinician reported outcomes and patient-reported outcomes measures (PROMs) (in terms of outcomes and their definitions) to help Guidelines Offices with summarizing study findings and generate evidence-based recommendations for the PCa treatment pathway. PIONEER's goal is to ensure the optimal care for all European men diagnosed with PCa by unlocking the potential of "big data" and "big data analytics." 4,5 Given the increasing importance of the use of PROs as part of "big data" in improving patient care, 6 there is a need to provide evidence-based PRO information that may be used to facilitate clinical decision-making. The main objective of this study was to synthetize most recent high-quality PRO data from PCa RCTs, that is, those studies most likely to robustly inform patient care through for example inclusion in the EAU PCa guidelines. The latter is a useful indication of quality of clinical studies with an impact on the PCa patient experience as these urological clinical guidelines are used worldwide to disseminate recommended PCa treatment pathways. The EAU PCa guidelines are updated on an annual basis using a broad and comprehensive literature search and are based on modified version of the Oxford Center for Evidence-based Medicine Levels of Evidence. Any flaws in the evidence used to support any given recommendation are taken into account and hence reflect on the quality of PRO reported information. Details of the search methodology can be found here. 7 A secondary objective of the current study was to assess whether completeness of PRO reporting in PCa research improved over time.

| Search strategy and identification of studies
We conducted a systematic literature search using PubMed, from April 2012 until February 2019. Methodology for study identification and evaluation followed standardized criteria used in the Patient-Reported Outcomes Measurements Over Time In Oncology (PROMOTION) Registry (http://promo tion.gimema.it). Study abstracts and additional records identified through hand searching of the literature were screened by two independent reviewers, following study selection criteria (see further for additional details). Then, full-text articles selected were assessed for eligibility. This was previously described in the systematic review on PROs in PCa RCTs covering the years 2004-April, 2012. 3 For the purpose of the current review, we used the same search terms as in our previous work 3 : ("quality of life" OR "health related quality of life" OR "health status" OR "health outcomes" OR "patient outcomes" OR "depression" OR "anxiety" OR "emotional" OR "social" OR "psychosocial" OR "psychological" OR "distress" OR "social functioning" OR "social wellbeing" OR "emotional" OR "patient reported symptom" OR "patient reported outcomes" OR pain OR fatigue OR "patient reported outcome" OR "PRO" OR "PROs" OR "HRQL" OR "QOL" OR "HRQOL" OR "symptom distress" OR "symptom burden" OR "symptom assessment" OR "functional status" OR sexual OR functioning) AND prostate. The search strategy was restricted to RCTs. In case of multiple publications from the same RCT, all relevant data possibly published in secondary articles were combined.

| Selection criteria
Only English-language reports of RCTs comparing conventional treatments and involving adult men with PCa were included-irrespective of disease stage. The minimum, overall sample size (combined treatment arms) was set at 50 patients. Screening studies or those involving patients with benign | 4041 VAN HEMELRIJCK Et AL. disease were excluded. We did not consider conference abstracts as these typically report insufficient information on PRO methodology and outcomes. RCTs of interventions that were psychological, behavioral, complementary, or alternative were also excluded.
We included all studies evaluating a PRO either as a primary or secondary endpoint-either as a multidimensional QoL outcome or a single dimensional outcome, such as symptoms. Those studies evaluating only treatment adherence or satisfaction were excluded. Details on search strategy and selection process were documented according to the PRISMA guidelines. 8

| Methods of evaluation of studies
Three reviewers (FS, LM, and KB) extracted information from the identified studies. Each study was evaluated independently by two of these reviewers. All data were entered by the reviewers into a password protected online database (REDCap) 9 by completing a predefined electronic-data extraction form (eDEF). Full details on information contained in the PROMOTION eDEF are reported in the appendix. A double-blind data entry procedure was performed as each reviewer completed the eDEF independently. Discrepancies in evaluations were electronically recorded and when disagreements occurred in the evaluation of any item included in the eDEF, the reviewers revisited the paper to reconcile any differences. If no consensus was achieved, a fourth reviewer (FE) was consulted. For every included RCTs, their inclusion in the EAU PCa guidelines was checked by manual searches of the references and corresponding sections.

| Type of data extraction and data analysis
For the purpose of this review, the following types of information were considered: (a) basic trial characteristics, (b) clinical and PRO characteristics, and (c) elements of PRO reporting based on recommendations from the International Society of Quality of Life Research (ISOQOL) 10 and the CONSORT-PRO extension. 11 Quality of PRO reporting was evaluated with the ISOQOL checklist, which comprises a common set of 17 key issues regardless of PRO being a primary or secondary endpoint. Eleven additional issues were considered when a PRO was a primary endpoint of the study. Each item of the ISOQOL checklist was rated as "yes" if documented in the publication (scored as 1) or "no" if not documented (scored 0). To further refine the investigation of the accuracy of reporting, we divided the ISOQOL item addressing the problem of missing data into two (ie reporting the extent of missing data and reporting statistical approaches for dealing with missing data). We thus rated each RCT with a score ranging from 0 to a maximum of 18 (RCT with PRO as a secondary endpoint) or 29 (PRO as primary endpoint); in both cases, the higher the score the better the quality of the PRO reporting. Identification of high-quality PRO studies was based on previously defined criteria. 3 Specifically, we defined as high-quality PRO studies those which, at the same time, satisfied at least two-thirds of the recommended criteria (12 for RCT with PRO as secondary endpoint and 20 for PRO as primary endpoint) and addressed three mandatory issues: study patients characteristics and baseline PRO scores described, documentation of PRO instrument validity, and missing data reported. In addition, we checked whether those studies considered as high-quality RCTs were included in the most recent EAU PCa Guidelines. 12,13 Main characteristics of eligible studies (eg disease stage, type of PRO endpoint) were reported by proportions and means, according to the type of variable. Differences between studies were assessed using the chi-square test. Based on the ISOQOL checklist score, comparisons between RCTs selected for this review and those included in a previous study 3 were performed to examine whether the completeness of PRO reporting in PCa trials has improved over time. We also compared the overall level of PRO reporting according to the CONSORT-PRO extension, between the RCTs published before and after the publication of this guideline. All tests were two-sided and statistical significance was set at α = 0.05. Analyses were performed by SAS software v. 9.4 (SAS Institute Inc).

| RESULTS
A total of 55 new RCTs were published between April 2012 and February 2019 (Figure 1), of which the majority were international trials (ie more than one country) (60%). An overview of trial characteristics, as compared to the data from January 2004 to March 2012 is shown in Table 1. The duration of PRO assessment has increased, with 63.6% of trials reporting a period of more than 1 year as compared to 43.1% of trials previously (P < .006). Most RCTs reported on PROs as a secondary endpoint (70.9% vs 60% previously; P = .212). However, the prevalence of RCTs with a sample size ≥200 has increased over time, with 69.1% of RCTs including more than 200 men with PCa, as compared to 52.3% previously (P = .062).
To assess the clinical impact of the RCTs, we first identified the proportion of high-quality RCTs with respect to PRO reporting, which was 43.6%, an increase since 2012 when it was 20% (P < .01). Table 2 provides more information about these high-quality RCTs (n = 47 reports-describing study results of 24 different RCTs). The majority of trials (15/24) were conducted in the advanced PCa setting (ie locally advanced PCa, metastatic PCa, metastatic castration-resistant and nonmetastatic castration-resistant PCa). Some RCTs included patients with different PCa stages, so that it was not possible to categorize these RCTs according to localized, locally advanced, or metastatic PCa. Six studies focussed exclusively on men with localized PCa. In these high-quality RCTs, PROs encompassed Health-Related Quality of Life (HRQoL) or symptoms such as erectile, bladder, or bowel function. The most frequently used measures were Functional Assessment of Cancer Therapy-Prostate (FACT-P) (9/24), EORTC QLQ-C30 and/or its module QLQ-PR25 (5/24), and Expanded Prostate Cancer Index Composite (EPIC) (5/24).
To further assess the clinical impact of these RCTs, Table 3 reports on the influence of the reported HRQoL on the final treatment recommendation (as defined by the authors of each published RCT). As most RCTs reported clinical and HRQoL specific outcomes separately, Table 3 aims to combine the results as one final recommendation for each RCT. In the metastatic PCa group, new treatment options (ie second-generation antiandrogens like enzalutamide and abiraterone) and the use of docetaxel resulted in significant improvements in both clinical (overall, progression-free, and metastatic-free survival) and HRQoL outcomes.
To date, 14 of the 47 publications have been included in the EAU PCa Guidelines 12,13 -reflecting 13 different RCTs ( Table 2).
In terms of methodology, there was still a lack of information on mode of administration of the PRO tool and methods  of data collection over time (83.6% vs 76.9%; P = .360). However, there was an improvement in the reporting of the evidence for PRO instrument validity and reliability (80% vs 66.1%; P = .007). More RCTs also identified PROs in the trial protocol and post hoc analyses (67.3% vs 20%; P < .001). In general, there was an improvement in terms of reporting methods and results for PROs in PCa RCTs. A detailed overview of the methodological assessment, as compared to our previous systematic review, 3 is provided in Table 4.
Evaluating the level of PRO reporting according to the CONSORT-PRO extension, we observed, for all of the items, an improvement in the studies published after the publication of the CONSORT-PRO extension, compared with those published before. Major differences were observed in two key items: the statistical methods for dealing with missing data were reported in 36.6% of newer RCTs as compared to 21.5% of older RCTs and PRO-specific limitations were discussed in 46.3% vs 32.9% of RCTs, respectively ( As highlighted by Kluetz et al cancer clinical trials have mainly focused on overall survival and measures of tumor growth or reduction to assess the efficacy of a specific treatment. 14 However, the balance between improving disease symptoms and introducing symptomatic toxicity affects how patients function in their daily lives and hence affects their HRQoL. Kluetz and colleagues suggest to focus on three distinct measures of well-defined concepts: symptomatic adverse events, physical function, and disease-related symptoms. As shown in Table 2, these outcomes are measured in several of the included PCa RCTs, however, only the ACTRN12611000 15 and the SPARTAN 16 trials specifically reported on all three outcomes.
It is important to note that the majority of trials used validated tools, with FACT-P, EPIC-26, and EORTC QLQ-C30

Outcomes
The mode of administration of the PRO tool and the methods of collecting data are described

Participant flow
A flow diagram or a description of the allocation of participants and those lost to follow-up is provided for PROs specifically being the most commonly used. Those studies that were included in the EAU guidelines had all used at least one of these three measurement tools. However, these validated HRQoL questionnaires were less common in trials of men with localized PCa. The latter focused more on specific adverse events such as erectile or bowel function. Nevertheless, the CHHiP trial measured general HRQoL with FACT-P 17 and the TROG 03.04 RADAR trial measured QoL with EORTC QLQ-C30 and its PCa module. 18 There seems to be a need for PRO-specific guidance for the different disease-specific stages of PCa. There are emerging international standards 19 to help generate robust data with more focus on patient engagement and the EAU is already undertaking work to develop core outcome sets for PCa, including both clinician-reported outcomes and PROs. [4][5]20 With improved methodology and practice, and increasing patient engagement, high quality and clinically meaningful generation of PRO data will become the norm for PCa clinical studies 21 and help increase their clinical impact in every disease stage. This will ensure that more high-quality PRO studies will also be incorporated in the recommendations of guidelines offices, which also uses robust evidence assessment methods. A modified Grading of Recommendations Assessment, Development and Evaluation (GRADE) is currently used by the EAU. This approach allows for a transparent assessment of how recommendation statements have been developed, whereby overall quality of the evidence along with magnitude of effect, certainty of the results, balance between desirable and undesirable outcomes, impact of patient values, and preferences on the intervention, and certainty of those patient values and preferences result in guideline recommendations. 7 Inclusion of PRO studies into the guidelines thus provides a reflection on the quality of the evidence and the reporting of PROs.
It is important to note that a patient-centred treatment recommendation should consider both clinical (ie survival) and HRQoL outcomes. Considering these HRQoL aspects relies on understanding patients' values, needs, and experience of the disease and integrate them to formulate an optimal treatment strategy. Quality of life remains subjective and validated PROMs are of critical importance to guide these tailored interventions to improve patients' well-being. The rising importance of PROs is already captured for localized disease by a systematic review, coordinated by the EAU PCa Guidelines panel, evaluating the effect of primary treatment on HRQoL. 22 The same PROMs as reported in our systematic review were captured, emphasizing their importance and systematic use. Such evaluation and recommendations are lacking for advanced and metastatic disease, where HRQoL becomes even more important. Methodology for reporting of PRO and HRQoL for these different disease stages does not need to differ, but there is a clear need for a broader consensus on its use in all disease stages so that it can be integrated with the existing clinical outcomes.
In addition to RCT data, it is of interest to note that there is an increased interest in using real world data to support regulatory decision-making-including the EAU guidelines. Information about how a patient feels and functions, P1b. The PRO should be identified in the abstract as a primary or secondary outcome. This percentage was calculated on the basis of all studies (including those explicitly reporting an exploratory evaluation, for which this item would be rated as not applicable). b Items P6a and P6aa were combined in the original CONSORT-PRO extension; however, for the purposes of this study, to better appraise the proportion of RCTs providing evidence on the validity of the PRO instrument but not further describing how this was administered to patients, this item was split into two items.
c In case of studies using multiple PRO measures, we evaluated this as "yes" if at least one measure was validated.
as captured directly from patients themselves, is however also often missing in real world data (eg observational data or hospital data). For example, PROs were only collected in 14% of recent postauthorisation safety studies. 23 Harnessing the patient voice through the use PROs in "big data" indeed implies the need to align PRO measurements (PROMs) between RCTs and real world data. 24 With respect to the methodological assessment of PRO use in PCa RCTs, data from this study suggests that completeness of PRO reporting has improved over the last years, as documented by the higher proportion of high-quality RCTs published from 2012 onward. Indeed, while only 20% of PCa RCTs published between 2004 and 2012 were considered as high quality, this percentage has more than doubled for the more recent RCTs. This may be partly explained by the publication of the ISOQOL recommended standards 10 and the subsequent CONSORT-PRO criteria 11 in 2013, which may have guided and helped investigators to increase the completeness of PRO reporting. Indeed, journal endorsement and author use of CONSORT-PRO extension has been demonstrated to be associated with improved PRO reporting. 25 This study has limitations. First, despite our comprehensive search strategy, it is possible that some RCTs with a PRO component might have been missed. Another limitation is the exclusion of non-English language published papers. However, it is unlikely that such omission would have significantly altered the conclusion of this review. 26 A strength of the current review is that we used a formal, replicable approach to evaluate PRO reporting of PCa RCTs. Since all studies use different reporting criteria and methods, the information was extracted and assessed by two independent researchers. In case of inconsistencies, a third arbiter helped achieving consensus. Also, by using state of the art and well-established international recommendations for PRO reporting, we were able to identify the proportion of studies that are most likely to robustly inform patient care.

| CONCLUSION
We observed an important improvement in the reporting of the evidence for PROs during the last 7 years, of which only a small proportion of high-quality PRO trials made it into the EAU PCa Guidelines. The most commonly used measurements focused on overall HRQoL and were predominantly used in RCTs of men with advanced PCa, whereas for RCTs in men with localized PCa the focus was more on adverse effects. Given the increasing recognition that the patients' voice in clinical research needs to be heard, there is a need for better guidance as to how to include and measure PRO in "big data" and guidelines-an answer which may be delivered by the EAU-led PIONEER Consortium.