This systematic review will include RCTs that used PROMs with feedback provided to clinicians to support their decision making in the clinical setting. The control will be standard clinical practice or usual care. RCTs have been chosen because their design ascertains the enrollment of a relatively homogenous group of patients that are balanced between the intervention and control groups and control for the impact of confounders. A review of RCTs offers an opportunity to study the fidelity of implementing the PROM intervention and factors associated with better implementation in an optimized setting.

The focus of this review is the use of PROMs as an intervention for patient care in clinical practice, and any study that does not use a PROM intervention will be excluded. RCTs that use proxy PROM measures (ie, those completed by a carer) will also be excluded, as this review focuses only on reports collected directly from the patient.

Participants will be any patients attending health care facilities, including hospital outpatients, specialist clinics, and health centers as well as the staff caring for them in these facilities, that is, patients completing PROMs reports and the clinicians responding to PROMs information.

The intervention will be the completion of PROMs by patients, with the results fed back to clinical staff to use in their ongoing clinical care.

The primary outcome is intervention implementation fidelity, that is, whether the intervention was delivered as it was intended or not. Intervention fidelity will be assessed using the criteria described in the systematic review by Proctor et al [13]: (1) adherence to the program protocol, (2) dose or amount of program delivered, and (3) quality of the program delivery.

Searches of databases to be accessed for this step will include MEDLINE, EMBASE, and the Cumulative Index to Nursing and Allied Health Literature via the OvidSP website. All searches will be saved using an account established in each search database. Databases such as ClinicalTrials.gov will be searched for ongoing studies. Other sources of grey literature will be limited to clinical trial results reported in theses, dissertations, and conference papers. Field experts will be contacted.

Field experts will be contacted for recommendations, and resources from the ISOQOL will be accessed. Reference lists of identified systematic reviews and studies will be searched to ensure a comprehensive list.

Studies will be screened initially by retrieving abstracts to identify whether they meet the eligibility criteria. Screening will be done independently by NR and MJ. If a study is identified as eligible for inclusion, the full-text version will be retrieved. After review of the full text, if a study is subsequently decided as not eligible, it will be excluded. A PRISMA flowchart will be used to present this information [15].

During the screening process, any discrepancies in the NR’s and MJ’s assessment of eligibility of the studies will be resolved by discussion. If this is not possible, the study will be referred to KA and DW for further input until a consensus is reached.

Endnote bibliographic software will be used to manage the literature by allocating studies using folders of included, excluded, and for discussion. Included studies will be entered for meta-analysis in RevMan 5.0. A database of the implementation data of included studies will be created during data extraction for narrative synthesis using Excel.

Data extraction will be done by NR and verified by MJ in 10% of studies. A data extraction table will be developed that, for each included study, states the research question, methodology, number of participants, primary outcome, secondary outcomes, tools used, and main findings to extract the demographics of the studies included.

Implementation data, including implementation strategies, process measurements, and outcome measurements will be extracted. The iPARIHS framework will be used to identify any implementation factors reported by study authors [19]. This information will be extracted using a content analysis approach. The key elements of iPARIHS are presented in Figure 1. The framework identifies what is present in the context for those exposed to the intervention and within the intervention itself. This framework also acknowledges the role of a facilitator, which has been identified as important [1]. RCTs often have research staff facilitating the study, so the impact of this can also be measured. This framework has been successfully used in a number of clinical studies, such as the prevention of Methicillin-resistant Staphylococcus aureus [20] and the Eat Walk Engage protocol [21]. If reported, these factors will be defined as impacting outcomes [13] or processes of implementation [14].

All eligible studies will be included in the narrative synthesis irrespective of the dose, delivery, duration, or intensity of the intervention, that is, independent of how often the PROM is administered, how it is administered, the duration of the PROM intervention and follow-up, and the number of PROMs administered.

If a meta-analysis is possible, it will be performed through a standard approach, that is, dichotomous data effect sizes will be calculated as odds ratios with 95% confidence interval, continuous data will be converted into standardized mean differences (SMDs), and time to event data will be calculated as hazard ratios. SMDs will be used to accommodate differences among PROM types reported by studies. If necessary, authors will be contacted if the report does not provide sufficient information. If the data cannot be obtained, the study will be excluded from the meta-analysis but still be included in the narrative synthesis. Similarly, studies with incomplete outcome data can still be included in the narrative synthesis if they report implementation factors.

A narrative data analysis approach [17,22] will be applied to the primary outcomes data collected based on the iPARIHS framework in Figure 1. This data will then be synthesized to describe intervention fidelity guided by the concepts presented in the study by Proctor et al [13]. The processes of implementation will be guided by the concepts described in the study by Powell et al [14]. Relationship within and between studies will be collated using mapping and tabulation of data [23]. The robustness of the synthesis will be assessed by comparing with previous systematic reviews, the scoping review, and study findings [18].

A meta-analysis will be performed to assess the impact of intervention fidelity on quality of life outcomes, provided there are sufficient data.

Stratified analysis of patient populations, for example, oncology, mental health, chronic diseases, will be attempted if there is heterogeneity in the data. Statistical heterogeneity will be assessed using I-square statistic to describe the variation across studies. There is likely to be heterogeneity in the evidence base because of enrolled patient populations or intervention design, and for that reason, a random effects model will be needed. The variability in effect estimates because of heterogeneity, rather than sampling error, will be classified as the proportion of observed effects [23]. This process is expected to allow quantitative consolidation of findings across studies.

A sensitivity analysis will be done by removing studies at high risk of bias as assessed by RoB 2.0, and the impact on study outcomes will be reported [23].