MiM2M Delphi competency dataset: item-level consensus results, stability and attrition analyses, and standard rankings
Description
This dataset contains the complete item-level results of an international modified e-Delphi study (May-October 2025) that developed the Consensus-Based Competency Framework for Healthcare Professionals in Language-Discordant (Mental) Health Care for migrants and refugees. In Round 1, 101 experts from 24 countries rated 119 evidence-derived behavioural items for healthcare providers (HCP) and healthcare administrators (HA) on three-point Priority and Feasibility scales and gave open-text feedback. In Round 2, 51 panellists re-rated the 66 items that had not reached consensus and rated 28 newly generated items. In Round 3, 16 panellists deliberated and voted on 13 borderline items across three synchronous meetings. The consensus threshold was 75% of valid ratings in the Mandatory category; 76 items were ultimately ratified into the framework. The data folder holds nine CSV files and a combined Excel workbook, 00_combined_workbook.xlsx: Round 1 results for all 119 items with curated summaries of expert feedback (file 01); Round 2 results for the 66 re-rated items, including per-item stability lines (02); the 28 new Round 2 items, with feasibility distributions reported inline (03); the Round 3 deliberation and voting record with discussion summaries (04); the Round 1 consensus summary by domain (05); the attrition analysis comparing Round 2 completers and non-completers (06); Wilcoxon signed-rank stability statistics for all re-rated items (07); and end-of-Round-2 rankings of the overarching competency standards for both actor groups (08-09). Each table's notes, defining its abbreviations, thresholds and statistics, appear beneath the data on the corresponding sheet of the workbook and in the data dictionary (Data_dictionary_and_file_inventory.docx), which also explains the file structure, the helper columns (domain, domain section, actor) and the item-code scheme. Four related materials are published with the related research article rather than duplicated here: the survey instruments, rating scales and background questionnaire; the panel characteristics; the final 76-item framework; and the item-level feasibility results with the priority-versus-feasibility gap analysis. All data are aggregated at item level; no individual-level responses are included. Ethics approval: Scientific Council of Babeș-Bolyai University, Cluj-Napoca, Romania (Ref. Nr. 1821/09.02.2024; application reference 91/31.01.2024); all panellists provided electronic informed consent. Funder: Volkswagen Foundation, project "Multilingualism in Providing Quality Mental Health Care to Migrants: Needs, Resources and Practices" (MiM2M, Grant ID 9A675; www.MiM2M.net). The data support competency-framework adaptation and validation, methodological research on consensus processes, and health-workforce education and accreditation planning.
Files
Steps to reproduce
Item generation: The initial pool of 248 behavioural statements was generated by triangulating normative standards adapted from WHO competency instruments (the 2021 Global Competency Standards for health workers caring for refugees and migrants, and the Global Competency Framework for Universal Health Coverage), systematic and rapid reviews of communication strategies and multilingual-care provisions in migrant health care, and primary MiM2M qualitative studies with migrant service users and providers in five countries. Internal pilot testing by ten consortium researchers reduced the pool to 131 items; review by the multidisciplinary Steering Committee (13 members, without voting rights) yielded the 119-item Round 1 instrument (52 healthcare-provider and 67 healthcare-administrator items) organised in five domains: People-Centred Care, Communication, Collaboration, Evidence-Informed Practice and Personal Conduct. Panel and administration: Experts were recruited internationally through purposive and snowball sampling across clinical, academic and policy networks in migrant and refugee mental health, transcultural psychiatry, health communication, applied linguistics and interpreting studies. Three rounds were administered online via the European Commission's EUSurvey platform between May and October 2025, with controlled feedback of aggregated results between rounds. Round 1 was completed by 101 experts from 24 countries; 51 completed Round 2 (retention 50.5%); 16 panellists deliberated and voted on 13 borderline items in the three synchronous Round 3 meetings. Rating and consensus: Every item was rated on two three-point ordinal scales, Priority (Not Relevant / Optional / Mandatory) and Feasibility (Not at All / Moderately / Highly Feasible), with open-text comments throughout; the pre-specified consensus threshold was at least 75% "Mandatory" among valid ratings. Round 2 added 28 new items derived from thematic analysis of Round 1 feedback, plus a task ranking the overarching standards within each domain (weighted scoring algorithm). Analysis: Analyses were computed in IBM SPSS Statistics: frequencies and percentages with valid N per item; Wilcoxon matched-pairs signed-rank tests on the paired responses of the 51 panellists completing both rounds (file 07); Pearson chi-square and Mann-Whitney U attrition comparisons (file 06). Open-text feedback underwent inductive thematic content analysis; the curated per-item summaries are in files 01-04. Reporting follows the ACCORD checklist for consensus methods and the CONFERD-HP recommendations for competency-framework development. Data preparation: The result tables were exported from the analysis outputs to CSV; three helper columns identify the domain, the domain section and the actor of each row, and each table's notes accompany the data in the combined workbook and in the data dictionary.
Institutions
- Babeș-Bolyai UniversityCluj County, Cluj-Napoca