General Practitioners

Starting from the conceptual model of health literacy,⁹ (link) a Delphi process among the HLS-EU Consortium members was conducted to generate items for assessing health literacy: the way people access, understand, appraise and apply information to make decisions regarding health care, disease prevention and health promotion. The resulting draft questionnaire was pre-tested for face validity in three focus groups (in Greece, Ireland and the Netherlands) and field-tested with 50 computer-assisted face-to-face interviews in two countries (n = 99 in Ireland and the Netherlands). Following the results of a principal component analysis and reliability analysis of the data, as well as inputs from consultations with external experts, a pre-final version of the questionnaire was achieved through a consensus-based item selection process. The pre-final version was subjected to a ‘plain language’ assessment by literacy experts to obtain the final version, which is known as HLS-EU-Q47 because it includes 47 items across 12 subdomains. For each item, respondents rated the perceived difficulty of a given task on a four-category Likert scale (i.e. very easy, easy, difficult and very difficult). This kind of operationalisation follows the tradition of subjective assessments of health literacy¹³ (link) and reflects the interactive or relational nature of health literacy by measuring the fit of personal competences with contextual or situational demands of social systems.¹⁴ (link) More details about the questionnaire’s development and the specific items of the HLS-EU-Q47 are presented in Sorensen et al.¹¹ (link)
For the purpose of the HLS-EU survey the HLS-EU-Q47 was supplemented with an additional section, which contained 39 items referring to antecedents and precedents outlined in the conceptual model.⁹ (link) They included inter alia indicators for the respondents’ health service use, health behaviour, subjective health status and socio-demographic and socio-economic situations supplemented with the Newest Vital Sign, which is a quick assessment of literacy.¹⁵ (link) Examples of items include gender, age, education (using score cards and answers were transformed to international standard classification of education (ISCED) levels, social status, financial deprivation, self-assessed health (SF-36), long term illness, visit to general practitioner, hospital admission, alcohol consumption, smoking, exercise, work experience in health sector and insurance coverage. The final version of the questionnaire for the HLS-EU survey included the 47 health literacy items and the additional 39 items and was named the HLS-EU-Q86.

The primary outcome measure was mental well-being as measured using the self-completion Warwick-Edinburgh Mental Well-being Scale (WEMWBS) [19 (link)]. This 14-item instrument has been validated for the UK population and adopted by the Scottish Health Survey and the Health Survey for England. It asks respondents to read statements about feelings and thoughts and to choose the response (a 5-point scale ranging from none of the time to all of the time) that best describes their experience over the previous two weeks. Example items are “I’ve been feeling optimistic about the future” and “I’ve been thinking clearly.” It has been shown to have good content validity and shows high correlations with other scales of mental health and well-being. It has a near-normal population distribution, with no ceiling effects.
Secondary outcomes were self-completed Center for Epidemiologic Studies Depression scale (CES-D) depression scores, Generalized Anxiety Disorder 7-item (GAD-7) anxiety scores, EuroQol Group 5-Dimension Self-Report Questionnaire (EQ-5D) quality of life scores, physical activity (self-reported frequency of exercise), and use of health services (self-reported general practitioner consultations or hospital visits). All outcomes were measured at the start of the trial (baseline before the intervention), immediately following the intervention (6 weeks after baseline), and 6 weeks after the intervention was finished (12 weeks after baseline).

Cognitive functions were assessed using the Wechsler Memory Scale – Revised (WMS-R) [12 ]. The test was administered during a routine visit to the general practitioner and it required around 1 h to complete. The WMS-R is a neuropsychological test designed to measure different memory functions such as verbal, visual memory, attention/concentration and delayed memory. The psychometric characteristics of the WMS-R are evaluated and related to its clinical utility in the Romanian population [13 , 14 (link)]. The weighted scores (weighted raw score summaries) were calculated according to the scoring system from in the WMS-R administrative and scoring manual, with a higher score indicating better functioning.

This paper draws on interview-based research conducted from May to June 2022, as part of a broader project investigating how information about medicines moves between patients, pharmacists, and general practitioners. To explore how GP’s approached prescribing decisions and understood their relationships with patients and pharmacists around medicines, we elected to use an interview methodology. Interview methodologies are well-suited to research that investigates people’s thoughts and experiences, and the semi-structured format accommodates emergent findings that might not be anticipated in the original research design. The project received ethics approval from the Victoria University of Wellington Human Ethics Committee (#28324).
Given our aim to solicit general practitioners’ views specifically, we followed a purposive sampling strategy and recruited from a national pool, so as to gain cross-sectional insights from different geographic, socioeconomic, and institutional settings. Using the Medidata database of general practitioners, we issued an invitation for interview participants that reached 1,331 recipients. Of these, 25 people registered their interest via the supplied link, and 16 followed through to an interview. Our only selection criterion was that participants must be currently practicing general practitioners, as all 25 initial respondents were. Recruitment stopped when participants stopped opting in to the study.
While most of our resulting sample worked full-time in general practice, some of our interviewees worked part-time, as locums, or combined their general practice work with, e.g., working for hospice. Of our sixteen participants, 10 were female and six male. Eight were based in one of New Zealand’s three major cities, six in a smaller city or large town, and two rurally. Thus, while our sample is not representative, it does encompass a range of professional and regional experiences and clear themes were evident across the data set.
Interview guides were developed from the overall project’s aims, and included questions about participants’ professional backgrounds and contexts, prescribing practices and views on medicines, relationships to pharmacists, and perspectives on the New Zealand health system. In keeping with the semi-structured interview process, we allowed these guides to steer our conversations without dictating or unnecessarily limiting their course. The interview procedure was not adjusted in light of emergent findings, so as to ensure maximum comparability across the data set. However, the interview guide was adapted in some cases to suit individual participants’ time availability.
The authors conducted all interviews via Zoom, between May 1 and June 30 2022. Interviews lasted from 24 to 68 min, with a mean duration of 46 min 30 s. Typically, shorter interviews were those conducted during participants’ lunch breaks, and longer ones were conducted on participants’ days off or in the evenings. All interviews were audio recorded using the computer’s inbuilt recording software, then professionally transcribed verbatim. The original audio files and resulting transcripts were allocated pseudonymised alphanumeric file names (linked to identifying information in one securely stored spreadsheet) and stored in the University’s secure cloud storage system.
Although Denise Taylor training as a pharmacist could potentially be expected to influence how interviewees spoke about the pharmacists they interacted with, we did not note any discernible differences in the results collected by each interviewer, and neither had any existing relationship with any participants. It is possible that conducting interviews via Zoom also mitigated the extent to which the researchers’ positionality shaped interviews. We were known to participants’ primarily by our qualifications, professional roles, and association with a respected funder, with few cues as to our respective physical presentations beyond the shoulders-up Zoom window.
Our data analysis was conducted by (Courtney Addison) June through October 2023, and checked and discussed with Denise Taylor. Our analytic process was grounded in the constructivist tradition of Corbin and Strauss [34 ], which acknowledges the role of the analyst in meaning-making, and followed a Reflexive Thematic Analysis process [35 (link)]. This began with familiarisation with the data set (all interview transcripts), followed by iterative, open coding of the corpus. Codes were then reviewed and grouped into themes, which were themselves reviewed against the transcripts. The review process enabled refinements to the theme, so that, for example, the theme ‘Doctor/pharmacist interactions’ became ‘Sharing information about patients’, ‘Seeking medicines information’, ‘Correcting mistakes’. To check the validity of findings, the authors discussed transcripts and codes from their distinct disciplinary perspectives (Anthropology and Pharmacy, respectively) and also compared findings against the data set from the other arm of this project, which consists of interviews and observations from a community pharmacy. This confirmed, for example, that doctors were using pharmacists in the ways they described in the interviews reported on here, and that pharmacists were indeed picking up mistakes as interviewees report.