Developing a candidate competency framework for early-career pharmacists in Chinese public hospitals: A two-round modified Delphi study

Résumé fourni par la source

Background Hospital pharmacy in China is moving toward patient-centred, digitally enabled practice, and several competency frameworks already exist. The FIP Global Competency Framework explicitly addresses foundation and early-career pharmacists but is global and cross-sector, whereas a recent Chinese hospital framework was developed in grade A tertiary hospitals. Neither directly addresses a concise, role-spanning candidate common-core structure for early-career pharmacists across secondary and tertiary Chinese public hospitals. We developed and appraised a candidate framework for that purpose. Methods An evidence-informed framework of 6 primary domains and 20 secondary elements entered a two-round modified Delphi study; domains and elements both served as rating units, giving 26 units. Round 1 drew 54 consent-confirmed questionnaires, of which 51 met the analysability criterion. Panel members rated importance (IMP), contextual applicability (APP), assessability (ASM), and clarity (CLR) on 1–5 scales; option 6, ‘unable to judge’, was excluded itemwise. All 51 were invited to Round 2. Of 41 returned paper questionnaires, 38 identifiers matched the invitation set and one failed the analysability criterion, leaving 37 in the primary analysis set. A pre-data analysis-plan draft documented mixed high-rating thresholds of 80% for IMP/APP and 75% for ASM/CLR but was not formally locked. After the Round 2 distributions had been examined, a uniform 75% rule was first documented on 3 August 2026 during manuscript harmonisation to align reporting across dimensions and improve interpretability; it was therefore treated as post hoc. Classifications under the mixed rule and uniform 75%, 80%, and 85% cutoffs were reported transparently. Results Round 1 produced 7 unchanged units, 9 minor revisions, and 10 major revisions; none was deleted. In the Round 2 primary analysis set (n = 37), all 26 units met the earlier mixed rule and the post hoc uniform 75% and 80% analyses; at a uniform 85% cutoff, D2 alone did not meet the criteria. Proportions of ratings of 4–5 ranged from 91.9% to 100.0% (IMP), 81.1% to 97.3% (APP), 83.8% to 100.0% (ASM), and 86.5% to 100.0% (CLR). D1 and D2, both in the digital and AI domain, had the lowest relative support, whereas the domain header scored 97.3% on all four dimensions. D2 had no ‘unable to judge’ responses in any dimension. Common core was selected in 948/962 (98.5%) valid action recommendations. Kendall’s W remained low (0.062–0.101). Conclusions The unchanged 26/26 classification under the earlier mixed rule and post hoc uniform 75% and 80% analyses provides structured content-level support for the candidate framework, while the threshold chronology precludes a confirmatory interpretation. The study did not validate a measurement instrument or demonstrate patient or organisational outcomes. Differences in institutional digital maturity, system access, and role exposure are hypotheses for future investigation because they were not measured here. Behavioural anchoring and multicentre workplace validation are required before assessment use.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.