Patient-Reported Outcome Measures in Cancer Care: An Updated Systematic Review and Meta-Analysis

Abstract

Importance: Patient-reported outcome measures (PROMs) come directly from the patient, without clinician interpretation, to provide a patient-centered perspective.

Objective: To understand the association of PROM integration into cancer care with patient-related, therapy-related, and health care utilization outcomes.

Data sources: Searches included MEDLINE and MEDLINE Epub ahead of print, in-process, and other nonindexed citations; Embase databases (OvidSP); PsychINFO; CENTRAL; and CINAHL from January 1, 2012 to September 26, 2022.

Study selection: Randomized clinical trials (RCTs) that enrolled adult patients (ages 18 years and older) with active cancer receiving anticancer therapy using a PROM as an intervention.

Data extraction and synthesis: Pairs of review authors, using prepiloted forms, independently extracted trial characteristics, disease characteristics, and intervention details. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses reporting guideline was followed. Random-effects analyses were conducted.

Main outcomes and measures: Overall mortality, health-related quality of life (HRQoL) measures, and hospital utilization outcomes.

Results: From 1996 to 2022, 45 RCTs including 13 661 participants addressed the association of PROMs with outcomes considered important to patients. The addition of a PROM likely reduced the risk of overall mortality (HR, 0.84; 95% CI, 0.72-0.98; moderate certainty), improved HRQoL (range 0-100) at 12 weeks (mean difference [MD], 2.45; 95% CI, 0.42-4.48; moderate certainty). Improvements of HRQoL at 24 weeks were not significant (MD, 1.87; 95% CI, -1.21 to 4.96; low certainty). There was no association between the addition of a PROM and HRQoL at 48 weeks. The addition of a PROM was not associated with reduced ED visits (OR, 0.74; 95% CI, 0.54-1.02; low certainty) or hospital admissions (OR, 0.86; 95% CI, 0.73-1.02; low certainty).

Conclusion and relevance: The findings of this study suggest that the integration of PROMs into cancer care may improve overall survival and quality of life.

Figure 1.. PRISMA Diagram of Article Selection for Updated Review

The original meta-analysis included randomized clinical trials (RCTs) and observational trials. We combined the 20 RCTs from the original review with the 25 RCTs from the updated search.

Figure 2.. Forest Plot and Risk of Bias for Overall Survival

Weights are from random-effects model. Risk of bias categories included: A, random sequence generation; B, allocation concealment; C, masking of participants and personnel; D, incomplete outcome bias; E, selective reporting. DL indicates DerSimonian-Laird random effects meta-analysis; HR, hazard ratio; PROM, patient-reported outcome measures; SOC, standard of care.

Figure 3.. Forest Plots and Risk of Bias for QLQ-C30

Weights are from random-effects model. Risk of bias categories included: A, random sequence generation; B, allocation concealment; C, masking of participants and personnel; D, incomplete outcome bias; E, selective reporting. DL indicates ; PROM, patient-reported outcome measures; QLQ-C30, European Organization for Research and Treatment of Cancer Core Quality of Life questionnaire; SOC, standard of care; WMD, weighted mean difference.

Figure 4.. Forest Plots and Risk of Bias for Emergency Department Visits and Hospitalizations

Weights are from random-effects model. Risk of bias categories included: A, random sequence generation; B, allocation concealment; C, masking of participants and personnel; D, incomplete outcome bias; E, selective reporting. DL indicates ; OR, odds ratio; PROM, patient-reported outcome measures; SOC, standard of care.