This study protocol is part of a large multi-method and multi-step project: “Comparing the cost-effectiveness of self-management interventions in four high priority chronic diseases in Europe: COMPAR-EU”. This is a European Union-funded project aiming to identify the most effective and cost-effective SM interventions for patients living with chronic obstructive pulmonary disease (COPD), heart failure (HF), obesity or type 2 diabetes mellitus (T2DM).

Our main objective is to identify, appraise, and analyse the perspectives of patients and their caregivers on outcomes of SM interventions for the selected chronic conditions. To provide a more detailed assessment, we will conduct four overviews (one per chronic condition). Our specific objectives are 1) to explore how patients with one of the four chronic conditions, and their caregivers, value the importance of different outcomes of SM interventions, and 2) to identify the factors that might impact the acceptability and feasibility of SM interventions based on their preferences and experiences with SM through the disease trajectory and when taking part in SM interventions.

Type of reviews. We will consider for inclusion: 1) quantitative SRs 2) qualitative evidence syntheses QES; and 3) mixed methods research synthesis (MMRS). To be eligible for inclusion, the SRs should report as a minimum: 1) the search strategy for at least one database; 2) the list of included studies; and 3) the methods applied for synthesis ³⁶ .

Context/setting. We will include SRs of studies without geographical restriction and will include all settings, except those confined to inpatient care.

Population/perspectives. We will limit inclusion to SRs of studies conducted in adult patients (18 or older) with one or more of the chronic conditions of interest (COPD, HF, obesity and T2DM), without restricting by disease severity. We will consider SRs assessing multiple chronic conditions (e.g., patients with more than one of the chronic conditions of interest or patients with chronic conditions of interest plus other chronic conditions) We will exclude SRs focused on patients with the chronic conditions of interest and cancer comorbidity or severe mental disorders. We will include SRs exploring the perspectives of patients or their informal caregivers. We will include SRs assessing both perspectives if these are reported separately. We will exclude SRs addressing only healthcare providers’ or healthy population perspectives.

The phenomenon of interest/intervention. We will include SRs aimed at exploring the perspectives of patients and their caregivers on the importance of outcomes of SM interventions. Their perspectives are based on their values, preferences and experiences. As described by Zhang et al. ¹⁴ , preferences for or against intervention are implicitly related to the relative importance people place (or value) on the expected or definite health outcomes connected to a specific intervention. We define SM intervention as supportive interventions systematically provided by healthcare staff, peers or laypersons to increase patients’ skills and confidence in their ability to manage a chronic condition. SM interventions aim to prepare patients (and, where appropriate, informal caregivers ³⁷ ) to actively participate in the management of their disease ^{37, 38} .

We will apply the COMPAR-EU taxonomy for the domain of outcomes of SM interventions, with 35 elements categorised in seven subdomains ³⁹ : 1) empowerment, 2) level of fulfilment of the expected SM behaviours, 3) clinical outcomes (progression of disease, complications, adverse events, and mortality), 4) quality of life of patients and caregivers, 5) perceptions and satisfaction with care, 6) health care use, and 7) cost. We will not exclude SRs addressing other outcomes not listed in the COMPAR-EU taxonomy. Instead, these will be incorporated as a new category if applicable.

Comparison. While not addressed as an objective in these overviews, if reported, we will consider usual care as the comparison.

Outcomes. We will include SRs evaluating studies reporting any of the following outcomes:

We will design and execute a literature search strategy in MEDLINE (accessed through PubMed), the Cumulative Index of Nursing and Allied Health Literature (CINAHL), and PsycINFO. The leading search concepts will be: 1) the perspective of patients, for which we will include a sensitive content search strategy previously published ⁴² ; and 2) specific terms for each chronic condition. We will apply methodological filters limiting to SRs, retrieving references using the filters available in each database, and to qualitative research, synthesis using non-controlled terms. We will consider English-language studies for inclusion without restriction on publication date. We will hand-search the reference list of overviews identified through our search strategy and conduct a forward citation search of the included studies in Scopus ⁴³ . We will establish regular search update alerts. Studies identified through these alerts and meeting the eligibility criteria will be incorporated until the stakeholders’ consultation starts.

We will manage references with Endnote X9. Pairs of authors will conduct the screening of references and selection of studies. After initial calibration with 10% of retrieved references, one author will independently select references to be included, and a second author will check the final list of included and excluded studies with reasons. Disagreements will be solved by discussion or with the help of a third author.

We will extract data using Nvivo 12 Pro ⁴⁴ . We will design and pilot-test one form for each type of SR (quantitative, qualitative or mixed methods). After initial calibration between pairs of authors, one author will extract data from the included reviews, and a second author will check the data against the full text for accuracy. We will solve disagreements by consensus or with the help of a third author. We will collect the following information from each type of SR:

In the extraction process, we will record the extent of relevant information not reported (e.g. information about age), but we will not take additional steps to get further information from primary studies ^{29, 45} . For discrepant data (two or more SRs reporting different data for the same primary study), we will extract data from the SR with the highest methodological quality, and if similar methodological quality, from the review with the most updated search ⁴⁵ . For discrepant methodological quality assessment (two or more SRs using different tools or reporting different results for the same primary study), we will report both results.

We will assess the overlapping primary studies (studies appearing in more than one review) using a citation matrix ⁴⁷ . We will extract the references of the primary studies answering our research question within all selected reviews, calculate the “corrected covered area” (CCA), and describe the extent of overlap (slight [0 to 5], moderate [6 to 10], high [11 to 15], or very high [over 15]) ⁴⁸ . We will include all SRs that meet the eligibility criteria regardless of overlap ⁴⁹ . We will conduct this assessment using Excel files.

For duplicated data (two or more SRs reporting the same data from the same primary study or with over 25% of overlap within primary studies), we will include data only once, selecting the one with the most complete and detailed data ⁴⁹ .

We will use the JBI Critical Appraisal Checklist for Systematic Reviews and Research Syntheses to test the quality of the included reviews ³⁰ . The JBI Checklist, which comprises 11 items, is a tool to assess the methodological quality of both quantitative SRs and QES ³⁰ . After initial calibration, one author will independently apply this checklist, and a second author will check results to validate scores. We will solve disagreements through consensus or with the help of a third author. We will adopt a previously published scoring method ⁵⁰ that allocates “one point” to positive responses and “zero points” to negative or unclear responses, and we will calculate the percentage of positive responses for each review (sum of positive responses divided by total items tested). The quality of each SR will be ranked based on the following criteria: 0–33% of criteria met (low quality), 34–66% of criteria met (medium quality), and 67% or more of criteria met (high quality) ⁵⁰ . We will not exclude SRs based on their quality scores.

To synthesise findings from selected SRs, we will use a mixed-methods three-step approach, with a convergent parallel analysis of quantitative and qualitative evidence ⁵¹ .

Step 1: Quantitative evidence. To analyse and synthesise quantitative findings (utility and non-utility data), we will use narrative synthesis ⁵² . We will calculate descriptive statistics (mean, SD for variables with a normal distribution and the median, interquartile range, for other distributions).

For reporting, we will present results in summarised tables (grouping by relevant categories of outcomes). We will plot reported results, e.g. mean values, 95% confidence intervals and I ² statistic for heterogeneity (if pooled), or the range of mean values or other measures (if not pooled) across SRs, the number of studies and the number of respondents for each outcome. Where possible we will present results by country of study, location, type of respondent (patient or caregiver), respondents’ characteristics (sample size, the severity of disease), and methods used. We will perform statistical analyses using STATA software V.14 ⁵³ .

Step 2: Qualitative evidence. We will analyse and synthesise qualitative data using a thematic synthesis ⁴⁶ approach using NVivo 12 Pro ⁴⁴ . The levels of analysis will comprise second-order (primary studies) and third-order SRs constructs ^{24, 54} ( Figure 1). We will not specify any a priori theme.

The synthesis process will follow the three stages of thematic synthesis ⁴⁶ , the first two taking place concurrently: 1) text coding - line-by-line coding of the text of each SR; 2) developing descriptive themes; 3) re-interpretation and synthesis of this newly organised information, to produce analytical themes that go beyond the findings of the SRs authors (overview of SRs) ( Figure 1). For this step, we will conduct a collaborative analysis with at least three review authors to consider alternative interpretations and ensure that fourth-order constructs remain grounded in the primary studies.

Step 3: Integration of qualitative and quantitative evidence. We will integrate the results of these two sources of evidence in the interpretation phase of results ⁵⁵ . For findings informing outcomes, we will organise them according to the COMPAR-EU taxonomy (or to additional outcomes, if applicable) ³⁹ . For emerging themes informing about feasibility and acceptability of SM interventions, we will develop a framework, according to the classification of analytical themes. The goal of the integration phase is to categorise the importance of outcomes and provide a framework to identify the factors that influence acceptability and feasibility of SM interventions, always from the perspectives of patients and their caregivers. We will present findings using tabular formats, which will inform the development of decision-making tools for the interactive platform of the COMPAR-EU project.

Sub-group analysis. For quantitative evidence and depending on the sufficiency of SRs, we plan to analyse the primary outcomes according to the patient population (e.g. age, the severity of the disease, number and types of chronic conditions, and type of respondent), intervention (e.g.,. type of SM intervention) and setting (e.g., country). For qualitative evidence, it is not possible to specify the groups in advance, due to the nature of this synthesis.

Sensitivity analysis. If applicable, we will conduct a sensitivity analysis exploring the impact of the quality of the included SRs (low, medium and high) in the final synthesis of findings.

After the integration of quantitative and qualitative findings, we will conduct an online survey with patients’ and caregivers’ representatives from each chronic condition. We will ask for their feedback on the draft of summaries of perspectives of patients and their caregivers derived from the SRs. We will design a tabular format for the seven subdomains of outcomes with close and open-ended questions assessing: 1) to what extent the overview findings resound with their preferences and experiences; and 2) the relevance of findings for healthcare decision-making on SM interventions from their point of view.

For the recruitment of participants, we will invite patients’ and caregivers’ representatives through the patient’s networks collaborating in the COMPAR-EU project. We aim to include at least five patients with different profiles of gender, age group, and severity of disease and one informal caregiver for each condition. After giving their informed consent, patients will receive the link to complete the survey. Their responses will be analysed based on frequencies, and thematic analysis for the open-ended questions. After discussion within the review working group, we will include their perspectives in the discussion section of each chronic condition manuscript.

Ethical approval was obtained by the Clinical Research Ethics Committee of the Avedis Donabedian Research Institute, the coordinator partner of the COMPAR-EU Project (EU 754936), the University Institute for Primary Care Research (IDIAP Jordi Gol) and was signed on March 2018. We will disseminate our findings as conference abstracts, peer-review manuscripts, and social community media. Patients’ and caregivers’ representatives participating in the draft summaries assessment will receive feedback based on their responses. Our results will inform the development of decision-making tools for the COMPAR-EU Project.

Since we planned to conduct the overviews of reviews for the four conditions consecutively, we are in different points on each one: 1) we conducted the searches in the selected databases and selected the initial list of included reviews for the four conditions on March 2019; 2) we finalised the extraction of the quantitative data of SRs reporting health utilities for the four conditions on September 2019, updated until November 2020; 3) we started the extraction of qualitative data, and thematic synthesis for reviews of T2DM on November 2019 and this process is still in progress; 4) in the upcoming months we will follow with obesity, HF and COPD. Finally, we plan to publish results also in a consecutive manner, one manuscript per condition, during the second semester of 2021 and the first of 2022.

No underlying data are associated with this article.

OSF: The perspectives of patients and their caregivers on self-management interventions for chronic conditions. https://doi.org/10.17605/OSF.IO/GFSA5 ⁵⁶

OSF: The perspectives of patients and their caregivers on self-management interventions for chronic conditions. https://doi.org/10.17605/OSF.IO/GFSA5 ⁵⁶ .

Data are available under the terms of the Creative Commons Attribution 4.0 International license (CC-BY 4.0).