Protocol for a mixed-methods study to explore implementation outcomes of the Phone-based Interventions under Nurse Guidance after Stroke (PINGS-II) across 10 hospitals in Ghana

Design

This mixed-methods descriptive exploratory study assesses specific implementation outcomes and contextual factors affecting the PINGS-II intervention.

Study setting and participants

This research will be conducted across 10 district-level and teaching hospitals where the PINGS-II intervention is being implemented. The PINGS II study is a phase III randomised controlled trial involving 500 recent stroke survivors, seeking to assess the efficacy of an mHealth technology-centred, nurse-led, intervention on BP control among stroke survivors. It consists of the following components: (1) home BP monitoring, (2) use of phone alerts to set medication reminders and (3) patient education on hypertension, cardiovascular risk management and stroke, all over 12 months of follow-up. The patients will also have access to a nurse navigator, who is the trial study nurse, the participants can call for medical advice regarding their blood pressure if needed.8

This implementation research protocol is however a separate observational enquiry which will be deployed in the 10 study sites where the PINGS-II intervention is being implemented. Three population groups are involved: (1) hypertensive patients who are using the PINGS-II intervention will be engaged in in-depth interviews, focus group discussions (FGDs) and will respond to a structured questionnaire, (2) health workers to be surveyed using an online questionnaire, (3) medical doctors involved in PINGS-II at each implementing site of PINGS-II for key informant interviews and (4) actors in the health policy domain such as the focal person for the non-communicable diseases in the Ghana Health Service, other relevant Ministry of Health agencies, academic research institutions and non-governmental organisations. The sampling is purposive although participants will be selected to maximise variation across demographics, severity of hypertension, facility location, socioeconomic factors and level of the healthcare facility as far as feasible.

To identify relevant questions and the most appropriate and feasible strategies for participant recruitment into this implementation study, we worked with the site investigators for the PINGS-II trial and engaged participants who had a stroke and their caregivers at the various study sites. The insights generated from these background activities were further reviewed by the research team to transform them into relevant questions to investigate strategies for recruitment, data collection and potential analyses. These actions were deemed important to ensure alignment of the study aims to the interests of patients as well as health workers involved in the care of patients who had a stroke, without compromising the scientific rigour of the study, while also complying meaningfully with the need to involve patients in study design and execution. Findings will be communicated to relevant stakeholders such as clinicians, policymakers and patient groups via appropriate forums such as dissemination workshops. Table 1 shows the inclusion and exclusion criteria for the study.

Study timelines

The online questionnaire with health workers was conducted in August 2021 as a preimplementation study to gauge feasibility, acceptability and appropriateness of the PINGS II intervention from the perspective of health workers. Results of this baseline assessment were published in August 2022.21 Data collection for the assessment of fidelity, costs to patients, and interviews with policymakers, interviews and FGDs with patients who had a stroke have been on-going since June 2022. Transcription of all qualitative data collection finally concluded in June 2024. Quantitative data collection on ‘costs’ and ‘fidelity’ from patients is expected to conclude in September 2024, with write up of all results expected to be completed by December 2024, for onward peer-review submission.

Sample size and recruitment for the FGDs with patients

Patients enrolled on the PINGS-II intervention will be invited to voluntarily participate in the FGDs. One FGD will be conducted in each of the 10 health facilities with each discussion involving 6–10 participants.

Sample size for the quantitative data collection from patients

An anticipated 250 patients will be enrolled on the PINGS-II intervention within the PINGS-II clinical trial with sample size calculation and rationale already provided for the trial.8 These patients will be invited to respond to the structured questionnaire to collect data on measures including implementation fidelity and costs. We will also include approximately 250 patients who were assigned to the control group to serve as a comparator population for analysis.

Sample size for the interviews with clinicians and policymakers

Creswell22 indicates that between 5 and 25 interviews are sufficient for enquiries where participants have experienced the phenomenon under study. Thus, 10 key informant interviews are planned, comprising one clinician from each of the 10 health facilities. In addition, five health policy actors would also be interviewed.

Sample size for the online health worker survey

The survey will be deployed via various social media groups for health workers in Ghana, including medical doctors, nurses, pharmacists and physiotherapists. The hyperlink to the survey will be shared widely to maximise participation and representation of the different health worker groups.

We used the Cochran formula at the 0.95 CI (Z=1.96) and 5% level of precision (e). Since the population proportion was unknown for any of the measures that is, ‘Feasibility’, ‘Acceptability’ and ‘Appropriateness’, a proportion of p=0.5 was chosen assuming maximum heterogeneity, to calculate the minimum sample size (n).

,adjusted by an additional 10% to a final sample of 423, for non-response.

Assessing feasibility, acceptability and appropriateness outcomes

Feasibility is the extent to which an intervention can be deployed successfully in a given setting while appropriateness refers to the perceived fit, compatibility or relevance of the intervention for implementation.23 Acceptability, also referred to as adoptability, is concerned with how well the target population will receive the intervention and how their needs might be met by the intervention.24 Weiner and colleagues developed and validated three measures: Acceptability of Intervention Measure, Intervention Appropriateness Measure and Feasibility of Intervention Measure. These measures will be adapted for assessing these outcomes under this protocol, and were designed to be pragmatic, and demonstrated validity across several domains, including content, structure and responsiveness to change.25 Table 2 summarises the relevant data types and study populations. The adapted tools for assessing feasibility. acceptability and appropriateness have been added as online supplemental information.

Assessing fidelity

Fidelity is the extent to which an intervention is implemented as intended.26 This will be assessed using data from the structured patient questionnaire supported by the FGD with patients, and the routine administrative and operational data from the project implementation. The instruments for data collection will capture different dimensions of fidelity such as content, process, exposure and responsiveness.27 28 A summary of the relevant data strategy and study population is presented in table 1. The set of items to assess fidelity are shown in the online supplemental information.

Assessing costs

Under this protocol, downstream costs such as self-reported additional personal costs incurred by patients who had a stroke partaking in the PINGS-II intervention will be assessed. This includes data on expenses made for related laboratory tests and other investigations, therapeutic agents, transportation, BP monitoring devices, among others. Downstream costs are those arising further after the implementation. Many interventions measure intervention and implementation costs but not costs incurred further downstream. These additional costs are an important factor when evaluating the overall intervention for cost-effectiveness.29 This protocol will assess some of these additional costs to patients such as transportation costs to the hospital for follow-up, costs of purchasing items like batteries for home BP measurements, phone call credit to contact the nurse navigator, and so on. The complete set of cost items of interest is shown in the online supplemental information.

Assessing facilitators and barriers to PINGS-II implementation

Contextual factors that may impact PINGS-II implementation will be explored from the perspective of patients, clinicians and other relevant stakeholders in the health policy space. This will be done via interviews and FGDs with these target populations. It is necessary to account for the influence of context to explain why implementation outcomes are achieved. Failure to do this may result in limited generalisability of the findings to different circumstances.30–32 Data to be gathered will include patients’ demographic details, health status and accessibility issues, among others. Clinicians’ professional backgrounds and attitudes towards technology that provide insights into individual and organisational challenges will be explored. Meanwhile, stakeholders’ perspectives on the care delivery ecosystem and healthcare priorities will offer a broad contextual understanding. Integrating these diverse viewpoints through interviews and focus groups is expected to enhance implementation strategies. The guides for the interviews and FGDs are provided in supplementary materials.

Data collection, management and analysis

The data collection and analysis team comprises five research scientists (including two lead physician researchers) with training and experience in mixed-methods research who are not directly involved in the PINGS-II clinical trial, and supported by other research physicians who are directly involved in the PINGS-II trial. The quantitative data will be collected using the REDCap software. All data will be kept password-secure and confidential with access restricted only to the investigators.

Descriptive summaries will be obtained for sociodemographic characteristics and other relevant domains using means, medians and proportions. Where appropriate, measures of dispersion such as SD and IQRs will be shown. The Likert type responses obtained for the feasibility, appropriateness and acceptability measures as described will be used to calculate a composite score for each measure. Relevant techniques such as χ² analysis will be used to explore associations between various physician characteristics and their ratings of the PINGS II initiative, regarding feasibility, acceptability or appropriateness. Cronbach’s alpha will be calculated to determine the internal consistency of the data collection tools. The different dimensions of fidelity measured in PINGS implementation will also be analysed descriptively to show how PINGS II was implemented with reference to the planned procedures.

The qualitative data will be collected by digital audiorecording and stored on a password-protected personal computer with access available only to the investigators. Qualitative data will be analysed by thematic analysis, which is a process of identifying themes or patterns within qualitative data,33 and an appropriate method for seeking to understand experiences, thoughts and behaviours across a dataset.34 By this approach, the transcripts will be read and re-read to familiarise with the data and then labels or codes assigned to segments of text that represent the key concepts. The codes will then be iteratively refined and grouped into broader themes and mapped onto the CFIR domains,18 for example, themes related to the Intervention Characteristics domain may include ease of use and perceived effectiveness. The Inner Setting domain may include resources, culture, leadership support while the Outer Setting will include community resources, patient needs and policy context. The Individual Characteristics domain will also include factors like patient preferences and clinician attitudes to the intervention, while themes affecting the implementation process for example, barriers and facilitators will be classified under the Process domain of the CFIR. The codes will then be iteratively refined and grouped into broader themes related to implementation outcomes (feasibility, acceptability, appropriateness, fidelity and cost) and contextual factors. Once the data have been coded according to CFIR constructs, they will then be iteratively organised under the specific implementation outcomes of interest. This will allow exploration of the contextual factors identified through CFIR analysis and how they influence the achievement of these outcomes, that is, feasibility, acceptability, appropriateness, fidelity and costs, to provide valuable insights into the implementation. Such deeper and integrated use of the CFIR has been advocated in a systematic review that considered the application of the framework across diverse settings and objectives.35

Ethics and dissemination

Ethics approval for the study has been granted by the Committee for Human Research and Publication Ethics at the Kwame Nkrumah University of Science and Technology, Ghana (CHRPE/AP/117/22). Voluntary written informed consent will be obtained from the participants. All the rights of the participants and ethical principles guiding scientific research shall be adhered to. Findings from the study will be presented in scientific conferences and published in a peer-reviewed scientific journal. A dissemination meeting will be held with relevant agencies of the Ghana Ministry of Health, clinicians, patient group representatives, and non-governmental organisations.

Patient and public involvement

The study was developed with consideration of patients’ priorities, experiences and preferences. Through informal consultation sessions, stroke survivors provided invaluable insights that guided the formulation of the research question and the selection of meaningful outcome measures.

Patients’ input influenced the design to ensure that the research approach resonates with their convenience, needs and preferences to enhance the study’s effectiveness. While not directly involved in recruiting other patients, stroke survivors will continue to influence the implementation and evaluation of PINGS-II through ongoing feedback sessions.

The study results will be disseminated to participants through clear, accessible means such as stakeholders’ engagements with patient groups. Participants will also have opportunities to engage with researchers to discuss implications for their care and future interventions.