Study design

To guide the assessment of readiness, this study will apply the Organizational Readiness for Change (ORC) model, adapted for AI integration in primary care. The framework considers two interrelated dimensions: Infrastructure Readiness (technical capacity, IT systems, and resource allocation) and Human Readiness (awareness, attitudes, and competency gaps among healthcare professionals). These dimensions will be evaluated through convergent mixed-methods design [8,9]. Quantitative and qualitative data will be collected during the same study period, analyzed separately, and integrated during interpretation using joint displays and narrative synthesis. The findings will be synthesized into an AI Readiness Score, which will inform an actionable roadmap for education and phased implementation aligned with national digital health priorities. The rationale for employing mixed methods design is justified by the complexity of the research objectives, which require both numerical data to quantify awareness and readiness levels, and narrative data to explore perceptions, attitudes, and contextual factors influencing AI education integration. This design allows for triangulation of findings, enhancing the validity and depth of the study.

Study setting

The study settings will involve all the Primary Health Care Centers (except AlJumailiya and Leghwairiya) and PHCC Headquarters AlJumailiya and Leghwairiya are excluded due to very small staff counts (<50) which would add logistical cost with little impact on representativeness relative to PHCC’s ~7,975 employees; the exclusion is therefore pragmatic and statistically proportionate.

Study population

All the PHCC healthcare professionals working in primary health care centers and Headquarters. The healthcare professionals will include Physicians, Dentists, Nurses, Pharmacist, Radiology technicians, Laboratory technicians, Physiotherapists, Dietitians, Health educators, administrative & support staff, and other paramedics.

Inclusion and exclusion criteria

All health care professionals in the selected health centers including Physicians, Dentists, Nurses, Pharmacist, Radiology technicians, Laboratory technicians, Physiotherapists, Dietitians, Health educators, administrative & support staff, and other paramedics. Two PHCC health centers, AlJumailiya and Leghwairiya—will be excluded from the study because each employ fewer than 50 healthcare professionals. Including these centers would add logistical complexity without significantly contributing to the representativeness of the sample, given their very small workforce relative to the overall PHCC population of approximately 7,975 employees. Exclusion ensures efficient resource utilization while maintaining the study’s focus on centers with sufficient staff to provide meaningful insights into AI awareness and educational needs across primary care settings.

Sampling

For the quantitative component, the study will adopt a census-style invitation. The HCPs will be approached initially through their official email with link to participate in the online survey. They will be sent a reminder in subsequent weeks. After two weeks if the response rate is low data collectors will approach HCPs with QR-code flyers which will be distributed at each center and PHCC Headquarters. In the qualitative component of this study, a maximum variation sampling strategy is employed to ensure a broad and diverse representation of healthcare professionals working in PHCC centers. This technique involves deliberately selecting participants who differ across key characteristics such as professional role (e.g., physicians, nurses, pharmacists, dentists), years of clinical experience, past exposure to AI tools, and geographic location of PHCC center.

Sample size calculation

The current submission targets the whole population (the Healthcare Professionals working in PHCC). That is why no sample size calculation is included. We hope to receive between 1,000 and 2,000 responses.

However, the minimum sample size of 367 is known to be sufficient for calculating a prevalence estimate of 50% with 95% confidence in a small finite population of 7975 (all PHCC employees, 2024), accepting an error of 5%. Considering the important scope of this study, laying the foundation for future interventions one should consider increasing the sample size to address a wider spectrum of prevalence estimates with the same precision (error margin of 10% of the estimate magnitude).

Download:

• PNG
  larger image
• TIFF
  original image

https://doi.org/10.1371/journal.pone.0345347.t001

If the email strategy succeeds in obtaining at least 1000 responses the data collection will stop by achieving a moderate study power (table above). If, the second step of personal invitation of targeted sample is used then the study will aim to the highest study power associated with 2000 sample size (table above). This sample size will used as a stopping criterion for the study.

The formula for sample size calculation estimating a single proportion:

Where:

• z = z score
• p̂ = population proportion
• n′ = sample size
• N = population size

For qualitative component, data will be collected till saturation of data is achieved [10]. Literature suggests that saturation is achieved normally between 12–20 interviews while conducting thematic analysis [10]. Two focus group discussions will be held with 10–15 participants in each session.

Data collection for Quantitative online survey

An online questionnaire form will be used for data collection (S1 File). The contents (the domains and items) of the online survey were constructed based on key themes and concepts discussed in recent published literature (and validated tools utilized) in this area of research [11–15]. The questionnaire form will start with a section about demographics (age, sex, professional role, and length of clinical experience). The next section will gather information measuring awareness and knowledge about AI applications in health care. A total of 7 questions will assess the self-rated AI knowledge, exposure to AI tools and identified sources of AI knowledge.

The form includes a section on training and educational needs. A total of 4 questions in this section is dedicated to enquiring about having formal AI training, the preferred training format Workshops, Online modules, CME), identifying possible barriers for not being trained in AI. The next section of the form is exploring perceptions and attitudes. It probes the study participants about their trust in AI tools (measured on a Likert scale), the perceived benefits of AI, and AI related concerns. The remaining sections of the questionnaire form are assigned to enquire about readiness, Impact of AI on clinical decision making and patient care, and the implementation and integration of these tolls. The tool was developed by reviewing the literature [16–20].

Tool validation plan

The face validity is defined as the extent to which a test or scale is subjectively viewed by the targeted study participants as covering the concept it purports to measure of the designed questionnaire. Face validity answers the question “Does it appear to the test taker that the questions on the test are related to the purpose for which the test is being given?”. This aspect of the tool validation will be assessed in a PILOT study. A total of 10 prospective participants (clinicians) will be enrolled for this pilot study. Their feedback on the questionnaire will be recorded and necessary amendments made (S2 File and S3 File).

The content validity is the extent to which the questions on a test are representative of the construct, trait, or attribute being measured. One way of obtaining a quantitative assessment of this type of validity is by examining the extent to which experts agree on the relevance of the content of the test items. A panel of 10 experts will be invited to review the study tool. For each questionnaire item, the panelist indicates whether the specific knowledge/ Attitude / Practice item measured is essential, useful but not essential, or not necessary as a contributor to the overall scale measure of that construct. In addition, the panelist indicates the reason for unfavorable (not essential) rating and writes any suggestion to improve the phrasing of the question. The reasons for unfavorable opinion on the side of the panelist were coded as 1 = irrelevant to the construct tested, 2 = too technical and difficult to the audience, 3 = repeated elsewhere and 4 = others (specify). The study tool will undergo iterative amendments and improvements based on the structured panelist review feedback.

The content validity ratio (CVR) for each item is computed. To determine whether an item is essential, its minimum value is compared with minimum values that depend on the number of experts who contributed ratings (a value of 0.62 for a panel of 10 experts) [[21]]. Changes required to the questionnaire will be submitted to the IRB for “Amendment Review”.

Where:

CVR_i = value for an item on the test

n_e = number of experts indicating that an item is essential

N = total number of experts in the panel

Data collection for qualitative component of the study (interviews and FGD with healthcare professionals)

Components of the quantitative study Instrument

The following variables will be extracted and analyzed from the survey responses:

Data analysis plan highlighting various components (variables) in the online survey

Prior to calculating the composite readiness measures, the psychometric properties of the instrument will be evaluated to ensure the validity and reliability of the derived scores. Internal consistency reliability will be assessed using appropriate reliability coefficients, and construct validity will be examined through relevant statistical procedures. Following confirmation of acceptable measurement properties, individual domain scores will be standardized to a common scale to facilitate comparison across constructs and ensure equitable contribution to the overall readiness assessment. The composite readiness measures will then be derived from the validated and standardized domain scores based on the predefined scoring framework.

The detailed procedures for analyzing and interpreting the AI awareness and knowledge scores, as well as the logistic regression model predicting AI readiness among clinicians in Qatar is described below. The analysis will be conducted after completion of the online survey distributed across PHCC centers

AI awareness and attitude scores

Statistical methods

Descriptive statistics: for quantitative data that is assumed to be normally distributed the mean, and standard error is calculated. The median and interquartile range will be used to describe the calculated composite index scores, which will depart from the assumption of normality. The Smirnov-Kolmogorov test for alignment of a distribution with normal assumption will be used.

Group comparisons: Independent t-tests or ANOVA to compare scores across demographic groups (e.g., profession, experience) for outcome variables that are assumed to be normally distributed. The non-parametric tests (Mann-Whitney U or Kruskal-Wallis) will be used if normality assumptions are violated. SPSS version 29 will be utilized for the analysis.

Qualitative data analysis

The qualitative data collected from the Focus Group Discussions and interviews will be analyzed utilizing thematic analysis. Qualitative data will be managed and analyzed with the support of NVivo software, which will facilitate the systematic organization, coding, retrieval, and interpretation of textual data obtained from interviews and focus group discussions. The software will be used to support thematic analysis by enabling the identification, categorization, and exploration of patterns, relationships, and emerging themes within the dataset. While NVivo will assist in data management and analytical processes, interpretation of the findings will remain the responsibility of the research team through an iterative and reflexive analytical approach.

The coding framework is derived from the themes outlined in the FGD and interview guides. Initial codes will be developed based on the following thematic areas:

1. Awareness and Knowledge of AI
2. Training and Educational Needs
3. Perceptions and Attitudes toward AI
4. Infrastructure and Readiness
5. Impact on Clinical Decision-Making and Patient Care
6. Recommendations for Enhancing AI Education

The qualitative data once recorded will be transcribed verbatim and then analyzed using Braun and Clarke’s (2006) Thematic Analysis. This approach encompasses (1) Familiarization with data; (2) Generating initial codes; (3) Searching for themes; (4) Reviewing themes; (5) Defining and naming themes; (6) Producing the report. Two researchers will independently code an initial subset (≈20%) and discuss discrepancies; a refined codebook will then be applied to the rest.

The study will report the results in accordance with ‘Consolidation criteria for reporting qualitative research (COREQ) and ‘Standards for reporting qualitative research ‘(SRQR) guidelines [24,25].

Quantitative results will identify patterns and gaps; interviews/FGDs will explain the “why” and co‑prioritize training content/delivery; integration products will include joint displays mapping survey findings to qualitative themes and the resulting education roadmap.

Integration of quantitative and qualitative findings

The quantitative survey findings will provide an overview of participants’ patterns of AI use, levels of readiness, and existing educational gaps related to artificial intelligence in clinical pharmacy practice. These findings will establish the breadth and distribution of AI-related knowledge, attitudes, competencies, and perceived training needs across the study population.

Subsequently, the qualitative interviews and focus group discussions (FGDs) will be used to explain, contextualize, and expand upon the quantitative results by exploring participants’ experiences, perceptions, and underlying reasons for the identified patterns. This qualitative component will provide a deeper understanding of the facilitators, barriers, and contextual factors influencing AI adoption and preparedness.

Integration of the quantitative and qualitative findings will be undertaken through the development of joint display tables, which will systematically map survey results against themes emerging from the qualitative analysis. This process will enable comparison, convergence, complementarity, and divergence between datasets, thereby generating a comprehensive understanding of AI readiness and educational needs.

The integrated findings will ultimately inform the development of a proposed AI education roadmap and readiness framework for clinical pharmacists, providing evidence-based recommendations to support workforce development, curriculum enhancement, and the effective implementation of AI-driven technologies in clinical pharmacy practice.

Quality control measures and good practices

The REPORT statement, which is an extension of the STROB statement checklist (international, collaborative initiative of epidemiologists, methodologists, statisticians, researchers and journal editors involved in the conduct and dissemination of observational studies, with the common aim of Strengthening the Reporting of Observational studies in Epidemiology) specially designed to assure the quality of reporting of secondary data analysis will be followed during analysis and writing of the report. For the qualitative component we aim to report the results in accordance with ‘Consolidation criteria for reporting qualitative research (COREQ) and ‘Standards for reporting qualitative research ‘(SRQR) guidelines.

To ensure the ethical and regulatory integrity of the study, oversight mechanisms will be implemented throughout its duration. A designated research monitor from the Clinical Research Department will conduct periodic monitoring and auditing activities. These will verify that the study is being conducted in accordance with the protocol approved by the Institutional Review Board (IRB) and outlined by Ministry of Public Health (MoPH) Qatar.

Timelines for study

The study is tentatively expected to be initiated in April 2026. The tentative timelines for the main study tasks are summarized in the table below:

Download:

• PNG
  larger image
• TIFF
  original image

https://doi.org/10.1371/journal.pone.0345347.t002

Ethical considerations

This study will be conducted in accordance with established ethical principles to ensure voluntary participation, minimize potential harm, and protect participant confidentiality. The principal investigator (PI), who will conduct the in‑depth interviews and facilitate the focus group discussions (FGDs), will not hold any supervisory, managerial, or evaluative role over the recruited healthcare professionals, thereby avoiding any power relationship that could influence participation or responses. Participation in the online survey will be anonymous, and the provision of any contact information for follow‑up communication will be entirely voluntary and at the discretion of the participant. The online questionnaire will explicitly state that participation is voluntary, that responses will not be linked to employment records, and that participation does not constitute an official Primary Health Care Corporation (PHCC) work task.

Specific measures will be implemented to protect potentially vulnerable participants, particularly given the workplace‑based nature of the study. Recruitment will be conducted through general announcements and institutional email communications, and line managers will not be involved in recruiting their staff. Participation will be non‑evaluative and will have no impact on employment status, professional standing, or performance appraisal. FGDs will be scheduled without the presence of supervisors or managers, and participants will be informed of their right to decline to answer any question and to withdraw from the study at any time without penalty or adverse consequences. All data will be deidentified prior to analysis, and study findings will be reported only in aggregate form to prevent individual identification.

Data confidentiality and security will be maintained through strict data management procedures. Survey data, audio recordings, and interview transcripts will be stored on PHCC’s secure network drive and will be accessible only to authorized members of the study team through role‑based permissions. No study data will be transmitted via email or shared outside PHCC systems. A fully de‑identified dataset will be used for all analyses and reporting. Study data will be retained for a period of three years following study completion, after which it will be securely deleted in accordance with PHCC data retention and disposal policies. To protect participant confidentiality and privacy, only fully de-identified datasets will be considered for external sharing. All direct and indirect identifiers will be removed prior to data release, and no identifiable participant information will be made publicly available. Any data sharing will be undertaken in accordance with the approved Institutional Review Board (IRB) requirements, PHCC data governance and information security policies

Ethics approval was received to conduct the study from ‘Institutional Review Board of Primary Health Care Corporation’ reference number BUHOOTH-25–00096 (supplementary document).

Dissemination plan of study findings

Upon completion of the study, findings will be disseminated through multiple channels to ensure broad and meaningful impact. Results will be submitted for publication in peer-reviewed journals and presented at relevant scientific conferences. In addition, tailored summaries will be shared with key stakeholders, including healthcare providers and policymakers, to inform practice and decision-making. Where appropriate, public-facing materials will be developed to communicate findings to the public, thereby enhancing transparency and promoting community engagement with the research outcomes.

Patient and public involvement

There was no Patient and Public Involvement (PPI) while formulating the study tools and designing the research methodology as the study mainly focuses on AI readiness among healthcare professionals providing primary care within the state of Qatar. However, a total of 10 prospective participants (clinicians) will be enrolled for content and face validity of the tool utilized in the study. Their feedback on the questionnaire will be recorded and necessary amendments made (supplementary document).