Interpretation of clinical trial results is limited without a defined CMT for within-patient change or between-group differences of study endpoints. Pooled data (n =100) from COMET trial (NCT02782741) comparing avalglucosidase alfa (AVA) vs. alglucosidase alfa (ALG) in LOPD was used to estimate within-patient and between-group CMTs for FVC (%predicted) and 6MWT.

In the CM 9ER trial, nivolumab plus cabozantinib (N+C) was associated with both increased survival and improved HRQoL in 1L aRCC patients (pts) at 23.5 months follow-up when compared to sunitinib (S). We investigated the association between HRQoL and overall survival (OS) and progression-free survival (PFS).

TTD in HRQoL is a common endpoint in oncology trials. A previous targeted literature review (TLR; Skaltsa et al. VIH 2018) of TTD definitions in NSCLC trials revealed that time to first deterioration without confirmation was the most frequently used definition with only few studies requiring confirmation, and none definitive deterioration. Our aim is to identify current practices and assess potential updates. An update of the previously conducted TLR of TTD definitions in NSCLC trials was performed. Searched databases included Ovid Embase and Medline. Searches were restricted from 2017 to March 2022 and publications included in previous TLR excluded. ClinicalTrials.Gov Registry was consulted for potentially further information on the TTD definition in the identified trials. The searches identified 89 publications on NSCLC with 72 (covering 45 trials) included in the TLR. First deterioration was reported in 44% (20/45) of trials, confirmed deterioration in 36% (16/45), and definitive deterioration in 11% (5/45). No TTD definition was reported in 18% (8/45) of trials. The number of thresholds was reported for 38 trials, with only one reporting >1 thresholds. Nine trials (20%; 9/45) included death as an event or as confirmation, while no trial considered disease progression as an event. No trial reported sensitivity or supplementary analyses for the TTD in terms of censoring rules, or model assumptions, or reported TTD results aligned with the estimand framework or clearly specified the handling of intercurrent events. Time to first deterioration remains the most frequent TTD definition in NSCLC trials but confirmed and definitive deterioration have started gaining ground. Although few studies do report more than one definition, most trials mention a single type of TTD definition and threshold. No sensitivity or supplementary analyses were reported. Analyses reported in clinical trial-related articles are not aligned with recommendations for addressing intercurrent events and missing data.

The landscape for first-line (1L) aRCC is rapidly evolving, with P+A and N+C recommended as standard of care irrespective of risk group by the European Association of Urology and the European Society for Medical Oncology. P+A and N+C have similar modes of action and demonstrated a significant efficacy benefit versus sunitinib (S), although no head-to-head data exist. As aRCC significantly impacts HRQoL, understanding HRQoL benefits of these 2 treatments is of interest to inform clinical decision making. An anchored MAIC was conducted using patient-level data from the CheckMate 9ER trial (NCT03141177; N+C vs S) and aggregate published data from the KEYNOTE-426 trial (NCT02853331; P+A vs S). Outcomes included Functional Assessment of Cancer Therapy Kidney Cancer Symptom Index – Disease Related Symptoms (FKSI-DRS) and EQ-5D visual analog scale (EQ VAS) due to limited published HRQoL data from KEYNOTE-426. Hazard ratios for time to first and confirmed deteriorations (TTFD and TTCD, respectively) and baseline to week 30 least squares mean differences in these outcomes were re-estimated for CheckMate 9ER using a weighted population and indirectly compared with those in KEYNOTE-426 via a Bayesian framework. A total of 651 CheckMate 9ER patients (pts) were matched to 861 KEYNOTE-426 pts using age, region, risk group, sites of metastatic disease, and prior nephrectomy. Results from the MAIC favored N+C versus P+A in all outcomes with statistically significant differences for FKSI-DRS and TTFD in EQ VAS score (Table).Table: 668POutcomeMAIC results, N+C vs P+ATTFD, HR (95% CrI) EQ-5D VAS0.73 (0.55–0.96)aTTCD, HR (95% CrI) EQ-5D VAS0.72 (0.52–1.01) FKSI-DRS0.48 (0.33–0.69)aChange from baseline at week 30,LSMD (95% CrI) EQ-5D VAS2.55 (–0.88 to 5.98) FKSI-DRS1.85 (0.96–2.74)baThe 95% CrI does not contain 1. HR < 1 favors N+C over P+A. bThe 95% CrI does not contain 0. LSMD > 0 favors N+C over P+A. CrI, credible interval; HR, hazard ratio; LSMD, least squares mean difference. Open table in a new tab aThe 95% CrI does not contain 1. HR < 1 favors N+C over P+A. bThe 95% CrI does not contain 0. LSMD > 0 favors N+C over P+A. CrI, credible interval; HR, hazard ratio; LSMD, least squares mean difference. In pts with 1L aRCC, MAIC analyses indicate that compared with P+A, N+C demonstrated a significant improvement in DRS and significantly delayed deterioration in HRQoL. These results, combined with the efficacy and favorable safety profile of N+C, may further inform treatment decisions of clinicians and pts with 1L aRCC.

BackgroundEMPOWER-Lung 3, a randomized 2:1 placebo-controlled phase 3 trial (NCT03409614), showed clinically meaningful and statistically significant overall survival improvement with cemiplimab, plus (+) platinum-doublet chemo (n=312), versus (vs) placebo (PBO)+chemo (n=154) among patients with aNSCLC (HR: 0.71; p=0.014). PROs were evaluated.MethodsPROs were assessed at day 1 (baseline), at start of each treatment cycle (every 3 weeks) for the first 6 doses, and then at day 1 of every 3rd cycle, per EORTC QLQ-C30 and lung cancer module, QLQ-LC13. Higher scores indicate better functioning or global health status/Quality of life (GHS/QOL) or worse symptom severity. Pre-specified analyses without multiplicity adjustments were conducted: 1) overall change from baseline per longitudinal repeated measures mixed-effects model; 2) time to definitive clinically meaningful deterioration (TTD) based on a 10-point threshold analyzed using a stratified log-rank test and a Cox proportional hazards model.ResultsBaseline scores were broadly similar between treatment arms. For pain symptoms (EORTC QLQ-C30), a statistically significant improvement in overall change from baseline (-4.98, 95% CI: [-8.36, -1.60]; p=0.004) and a statistically significant delay in TTD favoring cemiplimab+chemo (HR: 0.39, 95% CI: [0.26, 0.60]; medians not reached; p<0.0001) were observed. Overall change from baseline in GHS/QOL was seen in cemiplimab+chemo vs PBO+chemo (1.69, 95% CI: [0.20, 3.19] vs. 1.08, 95% CI: [-1.34, 3.51]); between-arms comparison was not significant (p= 0.673). Results showed a trend towards delay in TTD in GHS/QOL, favoring cemiplimab+chemo (HR: 0.78, 95% CI: [0.51, 1.19]; medians not reached; p=0.248). When comparing between arms, no analyses yielded statistically significant PRO results favoring PBO+chemo for any QLQ-C30 or QLQ-LC13 scales.ConclusionsCemiplimab+chemo demonstrated favorable PROs, with significant overall improvement and delayed TTD in pain symptoms. PRO results further support the favorable benefit-risk profile of 1L cemiplimab+chemo vs PBO+chemo in aNSCLC.Clinical trial identificationNCT03409614.Editorial acknowledgementMedical writing support was provided by Sumit Arora, of Prime, Knutsford, UK, funded by Regeneron Pharmaceuticals, Inc., and Sanofi.Legal entity responsible for the studyRegeneron Pharmaceuticals, Inc.FundingRegeneron Pharmaceuticals, Inc., and Sanofi.DisclosureR.G. Quek: Financial Interests, Personal, Full or part-time Employment: Regeneron Pharmaceuticals, Inc; Financial Interests, Personal, Stocks/Shares: Regeneron Pharmaceuticals, Inc. C. Ivanescu: Financial Interests, Personal, Full or part-time Employment: IQVIA; Financial Interests, Institutional, Funding: Regeneron Pharmaceuticals, Inc. K. Penkov: Financial Interests, Personal, Funding: AstraZeneca; Financial Interests, Personal, Funding: MSD; Financial Interests, Personal, Funding: Nektar; Financial Interests, Personal, Funding: Pfizer; Financial Interests, Personal, Funding: Regeneron Pharmaceuticals, Inc; Financial Interests, Personal, Funding: Roche; Financial Interests, Personal, Advisory Role: Nektar. E. Kalinka: Financial Interests, Personal, Funding: Pfizer; Financial Interests, Personal, Funding: AstraZeneca; Financial Interests, Personal, Funding: Roche; Financial Interests, Personal, Funding: Bristol-Myers Squibb; Financial Interests, Personal, Funding: MSD; Financial Interests, Personal, Funding: Nektar; Financial Interests, Personal, Funding: Amgen; Financial Interests, Personal, Funding: Regeneron Pharmaceuticals, Inc. C. Gessner: Financial Interests, Personal, Advisory Board: GlaxoSmithKline; Financial Interests, Personal, Advisory Board: Pfizer; Financial Interests, Personal, Advisory Board: AstraZeneca; Financial Interests, Personal, Advisory Board: Roche; Financial Interests, Personal, Advisory Board: Novartis; Financial Interests, Personal, Advisory Board: Bristol-Myers Squibb; Financial Interests, Personal, Advisory Board: Merck Sharp & Dohme; Financial Interests, Personal, Advisory Board: Berlin-Chemie; Financial Interests, Personal, Advisory Board: Chiesi; Financial Interests, Personal, Advisory Board: Boehringer Ingelheim; Financial Interests, Personal, Advisory Board: Sanofi. R. Passalacqua: Financial Interests, Personal, Advisory Board: Astellas; Financial Interests, Personal, Advisory Board: Bristol-Myers Squibb; Financial Interests, Personal, Advisory Board: Ipsen; Financial Interests, Personal, Advisory Board: Janssen; Financial Interests, Personal, Advisory Board: MSD; Financial Interests, Personal, Advisory Board: Roche; Financial Interests, Personal, Advisory Board: Sanofi-Aventis; Financial Interests, Institutional, Funding: Amgen; Financial Interests, Institutional, Funding: AstraZeneca; Financial Interests, Institutional, Funding: Novartis; Financial Interests, Institutional, Funding: Pierre-Fabre. G. Konidaris: Financial Interests, Personal, Full or part-time Employment: Sanofi; Financial Interests, Personal, Stocks/Shares: Sanofi. P. Rietschel: Financial Interests, Personal, Full or part-time Employment: Regeneron Pharmaceuticals, Inc; Financial Interests, Personal, Stocks/Shares: Regeneron Pharmaceuticals, Inc. G. Gullo: Financial Interests, Personal, Stocks/Shares: Regeneron Pharmaceuticals, Inc; Financial Interests, Personal, Full or part-time Employment: Regeneron Pharmaceuticals, Inc. All other authors have declared no conflicts of interest. BackgroundEMPOWER-Lung 3, a randomized 2:1 placebo-controlled phase 3 trial (NCT03409614), showed clinically meaningful and statistically significant overall survival improvement with cemiplimab, plus (+) platinum-doublet chemo (n=312), versus (vs) placebo (PBO)+chemo (n=154) among patients with aNSCLC (HR: 0.71; p=0.014). PROs were evaluated. EMPOWER-Lung 3, a randomized 2:1 placebo-controlled phase 3 trial (NCT03409614), showed clinically meaningful and statistically significant overall survival improvement with cemiplimab, plus (+) platinum-doublet chemo (n=312), versus (vs) placebo (PBO)+chemo (n=154) among patients with aNSCLC (HR: 0.71; p=0.014). PROs were evaluated. MethodsPROs were assessed at day 1 (baseline), at start of each treatment cycle (every 3 weeks) for the first 6 doses, and then at day 1 of every 3rd cycle, per EORTC QLQ-C30 and lung cancer module, QLQ-LC13. Higher scores indicate better functioning or global health status/Quality of life (GHS/QOL) or worse symptom severity. Pre-specified analyses without multiplicity adjustments were conducted: 1) overall change from baseline per longitudinal repeated measures mixed-effects model; 2) time to definitive clinically meaningful deterioration (TTD) based on a 10-point threshold analyzed using a stratified log-rank test and a Cox proportional hazards model. PROs were assessed at day 1 (baseline), at start of each treatment cycle (every 3 weeks) for the first 6 doses, and then at day 1 of every 3rd cycle, per EORTC QLQ-C30 and lung cancer module, QLQ-LC13. Higher scores indicate better functioning or global health status/Quality of life (GHS/QOL) or worse symptom severity. Pre-specified analyses without multiplicity adjustments were conducted: 1) overall change from baseline per longitudinal repeated measures mixed-effects model; 2) time to definitive clinically meaningful deterioration (TTD) based on a 10-point threshold analyzed using a stratified log-rank test and a Cox proportional hazards model. ResultsBaseline scores were broadly similar between treatment arms. For pain symptoms (EORTC QLQ-C30), a statistically significant improvement in overall change from baseline (-4.98, 95% CI: [-8.36, -1.60]; p=0.004) and a statistically significant delay in TTD favoring cemiplimab+chemo (HR: 0.39, 95% CI: [0.26, 0.60]; medians not reached; p<0.0001) were observed. Overall change from baseline in GHS/QOL was seen in cemiplimab+chemo vs PBO+chemo (1.69, 95% CI: [0.20, 3.19] vs. 1.08, 95% CI: [-1.34, 3.51]); between-arms comparison was not significant (p= 0.673). Results showed a trend towards delay in TTD in GHS/QOL, favoring cemiplimab+chemo (HR: 0.78, 95% CI: [0.51, 1.19]; medians not reached; p=0.248). When comparing between arms, no analyses yielded statistically significant PRO results favoring PBO+chemo for any QLQ-C30 or QLQ-LC13 scales. Baseline scores were broadly similar between treatment arms. For pain symptoms (EORTC QLQ-C30), a statistically significant improvement in overall change from baseline (-4.98, 95% CI: [-8.36, -1.60]; p=0.004) and a statistically significant delay in TTD favoring cemiplimab+chemo (HR: 0.39, 95% CI: [0.26, 0.60]; medians not reached; p<0.0001) were observed. Overall change from baseline in GHS/QOL was seen in cemiplimab+chemo vs PBO+chemo (1.69, 95% CI: [0.20, 3.19] vs. 1.08, 95% CI: [-1.34, 3.51]); between-arms comparison was not significant (p= 0.673). Results showed a trend towards delay in TTD in GHS/QOL, favoring cemiplimab+chemo (HR: 0.78, 95% CI: [0.51, 1.19]; medians not reached; p=0.248). When comparing between arms, no analyses yielded statistically significant PRO results favoring PBO+chemo for any QLQ-C30 or QLQ-LC13 scales. ConclusionsCemiplimab+chemo demonstrated favorable PROs, with significant overall improvement and delayed TTD in pain symptoms. PRO results further support the favorable benefit-risk profile of 1L cemiplimab+chemo vs PBO+chemo in aNSCLC. Cemiplimab+chemo demonstrated favorable PROs, with significant overall improvement and delayed TTD in pain symptoms. PRO results further support the favorable benefit-risk profile of 1L cemiplimab+chemo vs PBO+chemo in aNSCLC.

Cemiplimab significantly improved overall survival in pts with R/M cervical cancer after first-line (1L) platinum-based chemotherapy (chemo) (NCT03257267; ESMO-VP-2021). We now report patient-reported QoL, functioning and symptoms from the trial.

Determining a meaningful within-patient change (”responder”) threshold is key to interpret patient-reported outcome (PRO) data. There are numerous methods to estimate responder thresholds, including distribution-based methods (e.g., 0.5 standard deviation), anchor-based methods (using patient-reported and clinical biomarker anchors), and qualitative methods. Most authors use a combination of methods, which often results in several responder thresholds that may be similar but not numerically identical. We were interested in examining how authors synthesized thresholds derived using different sources into a single one. A targeted literature review of methods used to derive a single responder threshold derived from multiple methods was conducted using PubMed, Embase, and the websites of several congresses, and of regulatory and payer bodies. Searches were limited to the last 15 years with no geographic restrictions. The searches identified 596 unique publications; 22 were included in the review. There is no consensus on the optimal method or rationale to be applied to derive a single responder threshold. Three methods were proposed in the reviewed studies: (1) hierarchical integration of anchor-based and distribution-based approaches (with the former and ROC-derived thresholds given more weight and distribution-based thresholds used only to confirm the others), (2) graphical display of threshold estimates with the highest value often being selected for primary analysis and (3) triangulation methodology without weighting. The latter shares the same approach as the first method but is also applicable to qualitative-derived thresholds. A commonality across these three methods is their subjectivity and the need for critical review of the adequacy of the different threshold sources and of impact of false positives and false negatives before a threshold is selected. There is no established best practice to derive a single threshold. It would facilitate the field of PRO research greatly if a preferred method could be agreed.

The recent COVID-19 pandemic has caused several disruptions on some clinical trials. Although sponsors may have taken urgent operational measures to mitigate the risks, several unplanned intercurrent events may be identified. The ICH E9(R1) addendum offers an appropriate framework on which to build the right estimands to answer the study objectives. This work provides some considerations on potential strategies to handle intercurrent events caused by the pandemic. Considerations on potential changes in treatment regimens and populations will be described. Possible unplanned intercurrent events are listed and potentially appropriate strategies are proposed. Newly-identified intercurrent events include study-level holds, site-level closure/unavailability, patient-level treatment/study discontinuation due to patient's risk or unwillingness of patient to go to hospital or continue on an experimental drug, COVID-19 infection or death due to COVID-19. People with COVID-19 infection – even if continuing in the trial - may require hospitalization and receipt of non-permitted concomitant medication may cause intermittent missing data. Finally, pandemic-related measures may impact patient's psychological or physical functioning. While short treatment interruptions may be ignored, i.e. handled with a treatment policy strategy, most intercurrent events related to COVID-19 may be best handled with a hypothetical strategy evaluating the treatment effect "had the pandemic not occurred" allowing comparison to previous and future trials. Supplementary estimands ignoring certain events may help to assess the impact. Additional estimands where death can be considered an unfavourable outcome, i.e. composite strategy, or treatment effect is evaluated in patient stratums based on COVID-19 variables, i.e. principal stratification strategy, may be useful in some settings. Careful consideration of the intercurrent events that may occur in clinical trials during the COVID-19 pandemic will ensure studies are meeting their research objectives. Supplementary estimands and sensitivity analysis, as well as additional ad-hoc interactions with the relevant stakeholders, will need to be planned by sponsors.

Previous patient-reported outcome analyses in the phase 3 ADMIRAL study have provided an understanding of disease-related symptoms and health-related quality of life in FLT3-mutated relapsed or refractory acute myeloid leukemia (FLT3mut+ R/R AML). The objective of this analysis was to validate the Brief Fatigue Inventory (BFI), Functional Assessment of Cancer Therapy-Leukemia (FACT-Leu), and Functional Assessment of Chronic Illness Therapy-Dyspnea Short Form (FACIT-Dys-SF) instruments in FLT3mut+ R/R AML, using data from ADMIRAL. Analyses were performed using pooled baseline data from gilteritinib and chemotherapy arms. Individual item and domain characteristics were summarized by descriptive statistics, and floor and ceiling effects were examined. Reliability was assessed by Spearman rank correlations (rs) and Cronbach’s alpha (α) coefficient. Construct validity was evaluated by examining known-group validity (by Eastern Cooperative Oncology Group [ECOG] performance status, cytogenetic risk, and prior transplant status) and convergent validity (Spearman rank correlation coefficients). Clinically meaningful differences were calculated for BFI scores using distribution- and anchor-based approaches. Patients’ responses covered the full range of each item for all instruments, although all three showed some degree of ceiling effect. Moderate-to-strong correlations between all items within their subscale (rs≥0.3) were found without signs of redundancy. Moderate-to-high correlations for most items and corresponding scales (rs≥0.4) were observed; all instruments showed acceptable internal consistency (α>0.7). All instruments differentiated groups of patients by ECOG performance status at baseline, but not by cytogenetic risk or prior transplant status. Convergent validity was achieved with all instruments. Thresholds for clinically meaningful deterioration in BFI were 1-4 points, with the lowest thresholds obtained using the distribution-based method (1.27-1.32). In patients with FLT3mut+ R/R AML in the phase 3 ADMIRAL study, the BFI, FACT-Leu, and FACIT-Dys-SF showed good psychometric properties, although ceiling effects may limit the ability to detect treatment effects when high-functioning patients are enrolled at baseline.

This study examined the impact of enzalutamide (ENZA) vs. placebo (PBO) on self-reported fatigue across the prostate cancer disease continuum: metastatic [m]HSPC (ARCHES), nonmetastatic [nm]CRPC (PROSPER), mCRPC pre-chemo (PREVAIL), and mCRPC post-chemo (AFFIRM). Fatigue was self-assessed by patients using the item "I have a lack of energy" (0-not at all to 4-very much) on the FACT-P tool. Longitudinal change from baseline to week 13 and week 13 to end of study was analyzed (mixed models repeated measures). Responder analysis was conducted to examine the proportion of patients improving or worsening in fatigue. At baseline, patients across the disease continuum reported comparable levels of fatigue. Across studies, patients in ENZA and PBO arms experienced worsening of fatigue up to week 13, except in AFFIRM. A numerically larger proportion of patients experienced clinically meaningful (≥2 point) worsening of fatigue with ENZA vs. PBO with differences between groups ranging from 3-9%, particularly in trials involving asymptomatic to minimally symptomatic patients at baseline (Table). In AFFIRM, which included more symptomatic patients, fewer experienced fatigue worsening (8.5% ENZA vs. 16.3% PBO) and more patients experienced improvements with ENZA vs. PBO (5.7% vs. 1.1%). Differences were less apparent between groups in a mean change analysis; however, responder analysis showed the majority of patients in both arms reported stable or improved fatigue across all disease states.Table: 665PStudyEnd of study, weeksArmLeast squares mean changeResponder analysisBaseline to week 13*Week 13* to end of studyWorsen ≥2 points, %Improve ≥2 points, %ARCHES73ENZA0.310.1618.77.3PBO0.050.077.18.1PROSPER97ENZA0.640.2514.73.4PBO0.350.4910.04.4PREVAIL61ENZA0.420.3612.52.1PBO0.490.409.64.3AFFIRM25ENZA0.310.038.55.7PBO0.520.2916.31.1*PROSPER: week 17 was used instead of week 13 Open table in a new tab *PROSPER: week 17 was used instead of week 13 Across prostate cancer disease states, worsening of fatigue was generally observed in the first 13 weeks of treatment, particularly in trials of asymptomatic or minimally symptomatic patients. While ENZA was associated with increased fatigue in early disease states in some patients, most did not experience significant fatigue deterioration. In more advanced disease states, fatigue scores improved over time vs. PBO.

Fatigue is a major symptom experienced by mCRPC patients. Despite the availability of validated patient-reported outcome measures (PROMs) [e.g., BFI], patient-reported fatigue is not routinely or systematically collected in clinical trials (CTs). The purpose of this analysis was to assess the correlation of individual items from the FACT-P, a PROM used in prostate cancer CTs, and the BFI. Two items from the FACT-P (item 1 [“I have a lack of energy”] and item 7 [“I am forced to spend time in bed”]) were selected as proxies of fatigue. A combined subscale of items 1 and 7 was also created. Pearson correlation coefficients were calculated between items 1 and 7 and combined items 1 and 7 subscale and BFI scores using baseline data from the Phase 3 AFFIRM (NCT00974311) and PREVAIL (NCT01212991) enzalutamide CTs. Correlation effect sizes were presented as absolute values and classified as small (0.1), medium (0.3), large (0.5), or very large (0.7) using Rosenthal’s adjustment (1996) to Cohen’s conventions (1988). Item 1 showed large to very large correlations with BFI total score (r=0.7123, r=0.6468), fatigue interference (r=0.6720, r=0.5919), and fatigue severity (r=0.7090, r=0.6321) for both AFFIRM and PREVAIL, respectively. Item 7 showed large correlations in AFFIRM for total score and fatigue subscores (range, r=0.5264–0.5777) and medium correlations in PREVAIL (range, r=0.3528–0.4155). The combined subscale showed very large to large correlations with BFI total score (r=0.7522, r=0.6717), fatigue interference (r=0.7241, r=0.6275), and fatigue severity (r=0.7208, r=0.6346) for AFFIRM and PREVAIL. Item 1 and the combined subscale demonstrate strong correlations between the FACT-P and BFI in mCRPC patients. Using the single item “I have a lack of energy” or the combined subscale may help provide additional context when trying to understand and interpret the patient experience regarding fatigue. FUNDING: Astellas Pharma Inc.; Pfizer Inc. EDITORIAL: Complete HealthVizion

There is a growing literature on options for estimating a treatment effect for PROs, where an event censors the measure of the effect of interest. The censoring may stem from an event such as death that makes further collection impossible; censoring may also impinge directly on what it is desired to estimate, and may be required to be taken account of quantitatively. The objective of this project was to evaluate approaches to analysing PRO data in the presence of censoring due to death and other events. A targeted literature review was performed, with advice from subject matter experts (both statistical and PRO experts). Pubmed and Google were used to identify approaches that account for censoring of PRO data. Methods, and their attributes were abstracted and summarised. The approaches identified were further discussed with experts and a short list was prepared that was evaluated with respect to the assumptions required by each method, their limitations, ease of implementation and interpretability. We reviewed 26 publications and identified 8 approaches: win ratio for a ranked composite endpoint; continuation ratio modelling for a categorized composite endpoint; principal stratification with Rubin’s implementation; responder (binary) analysis; mixed model repeated measures (MMRM) with zero PRO imputation for death; area under the curve (AUC) as a patient-level summary measure; ordinal regression (proportional odds) modelling and joint modelling of survival and longitudinal PRO values. A short list of 4 approaches was assessed in depth and these methods were implemented using two publicly available clinical data sets as test cases. This review identified several analysis methods (estimators) for three different types of estimands for inference about PROs when an event impinges on the measurement of the PRO.