Using the Theoretical Domains Framework to investigate clinicians’ behavioural determinants of antimicrobial prescribing in Qatar

Abstract

Introduction: A recent systematic review by Talkhanet al demonstrated the need for theoretically based behaviour change interventions in this area. [1] For developmentof such complex interventions, emphasis should be placed onusing theory to systematically identify behavioural determinants of antimicrobial prescribing.Aim: To identify and quantify clinicians' behavioural determinants of antimicrobial prescribing in Qatar.Methods: This cross-sectional survey is part of a multiphase explanatory, sequential mixed methods PhD project in Qatar. Questionnaires were distributed (onlineand paper based) to all doctors (~4,000) and pharmacists(~400) within Hamad Medical Corporation (HMC, themain healthcare provider). The questionnaire was developed with reference to the Determinants of ImplementationBehaviour Questionnaire (a generic questionnaire derived from the 14 theoretical domains of the TheoreticalDomains Framework, TDF). [2] Each item was presentedas a 5-point Likert scale (scored 5=Strongly agree to1=Strongly disagree). Personal and practice demographics were also collected for data contextualisation. The draftquestionnaire was reviewed for face/content validity by anexpert panel of six researchers in Qatar and the UK withexperience in the use of the TDF, followed by 'Think aloud'testing and piloting. Analysis investigated the behaviouraldeterminants and influential factors through descriptive,principal component analysis (PCA) and inferential analysis. Ethics approval was granted from a UK universityand HMC.Results: In total, 535 responses were received, 339(63.4%) from doctors and 196 (36.6%) from pharmacists. Respondents were predominantly male, 346 (64.7%).Just over half (n=285, 53.3%) had = 5 years' experience ashealth professionals. PCA showed a three component (C)solution with components incorporating a number of questionnaire items labelled: 'Guidelines compliance' (C1 with8 items), 'Influences on prescribing' (C2 with 7 items) and'Self-efficacy' (C3 with 5 items) in prescribing/recommendation activity. A scale score for each respondent was calculated through summation of Likert scores for the relevantquestionnaire items within each component. These scaleshad high internal reliability (Cronbach's alpha all >0.7)showing consistency in response between component itemsindicating statistical appropriateness for developing scales.The median score (possible scale range, midpoint) for eachscale was C1, 32 (8 to 40, 24), C2, 26 (7 to 35, 21) and C3,20, (5 to 25, 15). By way of example Table 1 shows levels ofagreement for items in C2. This shows lower levels of agreement than C1 scale with the median scale score (26) closerto the midpoint (21) indicating that respondents had lesspositive views. Inferential analysis using these scale scoresand free text analysis is in progress.Conclusion: A theoretical basis was used throughoutproviding insights to behavioural determinants for the development of a theory-based behaviour change intervention. Preliminary results suggest that social influences, staffdevelopment and quality monitoring may be useful targetsfor behaviour change interventions to improve antimicrobial prescribing practice. Limitations include potential socialdesirability bias and focus on one healthcare organisation/country in the Middle East which may limit generalisabilityof findings. More in-depth exploration is required to selectand test appropriate linked theory-based behaviour changetechniques.