Protocol for a mixed-methods exploratory investigation into the role of health information technology for the provision of cancer care in Indonesia

Origin of research focus

In April 2019, a multi-stakeholder workshop was hosted by the study authors at Universitas Gadjah Mada (UGM), focusing on research priorities for cancer and digital technology research in Indonesia. The study team comprise applied health researchers with cancer and informatics expertise and experienced oncology clinical academics familiar with cancer provision in Yogyakarta. Attendees included representatives from the Ministry of Health (NCD Directorate), Ministry of Research, Technology and Higher Education, Ministry of Finance, Indonesian Cancer Foundation, independent commercial companies working in cancer care, academics from multiple departments in UGM including Internal Medicine, Surgery, Radiology, Public Health, and Nursing, and oncologists from hospitals in Yogyakarta, Indonesia. During discussions, it was highlighted that digital architecture in cancer care is not well understood in Indonesia. Research priorities highlighted during the workshop included determining available technologies used by cancer services, approaches to determining the quality of care delivery, patient-reported outcomes during and after cancer treatments, the development of palliative care approaches, and eliciting patient perspectives on the delivery of care. The proposed work is in response to the first priority, seeking to map the digital architecture of a defined region of Indonesia, to understand how it currently supports cancer care delivery and where there is scope for development.

Patient and public involvement

We consulted the Indonesian Cancer Foundation (ICF), Yogyakarta branch, on the design of the research study. The ICF is a non-governmental organization that aims to support the Government of Indonesia in its delivery of cancer care, by taking on promotional, preventive, as well as rehabilitative measures and supportive activities. The ICF also advocates on behalf of patients with cancer and their caregivers. We consulted the ICF about the questions for participants to make sure that the topic headings were exhaustive and capture the patient journey from the beginning of care and that the content of topic guides was acceptable to patients. The ICF supported engagement with patient groups and facilitated discussions on the types of study participants that were sought and the representativeness of the target population. We also conducted a preliminary discussion with healthcare professionals to inform the development of the topic guide for the study, ensuring the topic guide captures the flow of patients and the information processes undertaken by health facilities. During the dissemination of findings, we will consult with patient advocacy groups to i) guide key messages from findings of relevance to patients and members of the public; ii) support with developing plain language summaries of study findings for dissemination, and iii) identify which communication channels might be used to disseminate study findings for patient and public audiences.

Study design

The study will adopt an observational mixed-methods single case study design³² of health facilities involved in the delivery of cancer services at the primary, secondary, and tertiary levels in a sub-national health system. For a country implementing decentralization, health politics at a sub-national level is highly influenced by the policy at the national level. An observational mixed-methods single case study design is well-placed to support the development of new theory³². We will use the approach to support the development of an underpinning theory that reflects the current structure, content and use of IT infrastructure for referrals within cancer care in Yogyakarta, Indonesia.

Theoretical frameworks underpinning research

Multiple theoretical frameworks have been developed that may have relevance to exploring and understanding HIT adoption and implementation in the health sector^33–36. Within this study, we will draw on the socio-technical model³⁷, one of the models used to review HIT comprehensively from various dimensions in the health system. The socio-technical model frames the focus of project activities to focus on multiple, interacting and key components to better understand HIT for cancer care in Yogyakarta. Eight interrelated dimensions of the socio-technical model will be explored: 1) hardware and software computing infrastructure, 2) clinical content, 3) human-computer interface, 4) people, 5) workflow and communication, 6) internal organizational policies, procedures, and culture, 7) external rules, regulations, and pressures, and 8) system measurement and monitoring³⁷. While these are presented as single components, the eight specific dimensions are multiple, interacting components with non-linear, emergent, dynamic behaviour (i.e., small changes in one aspect of the system lead to small changes in other parts of the system under some conditions, but large changes at other times). The relationships and dependencies across the dimensions of the HIT for cancer referrals in Yogyakarta will be explored during the intervention development phase of the project. We will develop conceptual maps to depict identified relationships and relay these to system users and policymakers to validate their content and guide discussions around the prioritisation of intervention design to refine and develop existing HIT for cancer referrals.

A recent publication of the infrastructure supporting cancer referral systems in Indonesia has been published³⁸. The framework used is based on enterprise architecture (EA) which refers to the Open Group Architecture Framework (TOGAF). This will be used to provide a detailed explanation regarding the structure and key components of the referral system as it relates to cancer, alongside providing insight into the first of the eight dimensions of the socio-technical model to be explored; hardware and software computing. In general, cancer referrals in Indonesia are conducted according to hospital classification type within the national health insurance framework. Although the study will be limited to Yogyakarta, the system observed in this study should also be able to represent the national cancer referral process with slight differences between regions caused by the availability of services and providers in each area.

Health system context

The cancer control program in Indonesia predominantly focuses on promotion and preventative efforts (e.g. increasing knowledge for the public about approaches to cancer prevention and known risk factors) and early detection programs at the primary care level³⁹. Delivery of care is across general and specialty hospitals which the Indonesian government classifies into several classes (A, B, C, D) based on facilities and the services provided by the hospitals (Table 1)⁴⁰. Class A hospital is the highest level. This category was established for referral purposes⁴¹. For cancer service delivery, management is divided across primary, secondary, and tertiary referral facilities. Cancer services in Indonesia are classified based on the level of service, namely primary care (examination and or basic medical action in the health sector of care/clinic/health centre in the primary health care), secondary care (access to specialist doctors in D type/C type/B type hospitals), tertiary care (medical examination and or treatment performed by oncology subspecialist in A type hospitals)⁴².

The provision of services as part of cancer care varies across different levels⁴³. Services at the primary health care level involve education and counselling related to cancer, screening of cancer, providing first aid in cancer emergencies, case finding and passing referrals, and registering cancer data⁴⁴. Services at the secondary level involve further handling of patient referrals from primary health facilities, managing cancer emergencies, conducting special cancer medical examinations and actions, establishing cancer diagnoses, and multidisciplinary teams, providing health lifestyle advice, and conducting cancer data regimens. Services at tertiary level services include conducting primary and secondary checks related to cancer, following up on patient referrals from secondary health facilities and emergency cancer cases, advanced diagnostic investigations, follow-up examinations, integrated management, special medical actions, referrals, and registration of cancer data⁴⁵.

Using data from 2015, there were a total of 2,488 hospitals in Indonesia, comprising both general (n = 1,949) and speciality (n = 539) hospitals. As a first step to explore the role of HIT in cancer care, this study focuses on the Special Region of Yogyakarta, a region of Indonesia located on the island of Java. Reflective of population maldistribution in Indonesia, 70% of its inhabitants are condensed on the island of Java, which equals 11.5% of the land area of Indonesia, while the remainder is sparsely distributed across a further 17,000 islands⁴⁶. The current population of the Special Region of Yogyakarta was 3,656,108 in the most recent 2019 census⁴⁷.

Setting

The study will be conducted in the Special Region of Yogyakarta, Indonesia, across nine facilities supporting the delivery of cancer care. The facilities will be selected through stratified purposeful sampling intended to reflect variation among different levels of the health system (i.e. primary care, D type hospital, C type hospital, B type hospital, and A type hospital). The selected sample will be based on the maturity of the referral system within the healthcare center and whether the health facility is active in the national cancer registry network. A sampling frame will ensure the representation of health facilities at each level of care. This includes Primary Health Care Facilities (FKTP) and Advanced Referral Healthcare Facilities (FKRTL). All participating facilities will be based in the Yogyakarta special province, with an overview of facility groupings by health ministerial decree outlined in Table 2⁴⁸.

Participants

The study will recruit staff in healthcare facilities, both those involved in the management of cancer referral systems and those who are providing care to cancer patients (Figure 1). In this study, patients will also be recruited to provide insights into their experience engaging with the referral service system and maintaining communication with the service provider/health facilities.

Figure 1. Data collection process.

Sample selection

Following the selection of health facilities, interviews with staff and patients will be conducted within each of the participating health center sites. Participants will be purposively sampled to ensure the representation of perspectives to inform existing referral processes. This will include both health facility staff who have a direct role in cancer referral management and patients with cancer. Purposive sampling will ensure the selection of participants likely to provide richly-textured information relating to referral processes. The research team will liaise with the Head or Director of Patient Services at each facility to support the identification of eligible health professionals. The research will liaise with clinical staff supporting patients with cancer to identify and support the recruitment of patient participants.

Inclusion criteria for health facility participants include working in cancer care at any of the health facilities being recruited from and having worked in their role for 1 or more years. We will seek representation of participants across health facilities involved in the referral and management of patients with cancer, including representation across the range of staff types working within facilities (i.e. service directors, clinicians, information system managers and cancer program managers). A sampling frame will be developed for selecting patients at each facility, including characteristics of age (adults and children), sex (males and females), cancer type and service currently accessing (including inpatient, outpatient, one-day care, and radiotherapy). This will capture variation in the flow and processes of patients across clinical settings. For interviews with child participants, a mix of items will be included to elicit responses from both child participants and their caregivers who will be present during the interview. Only participants who are able to consent or assent to participation and have no communication impairment will be approached to participate. The criteria and sampling size of the patients are detailed in Table 3.

The sample size for health facility and patient participants to be recruited across sites was determined by the exploratory nature of the study and pragmatic considerations (i.e. resource constraints in terms of research budget and research team size). The selection of sites to ensure information-rich data collection was guided through consultation with the District Health Office. The adequacy of data quality and richness being gathered during interviews will be monitored through regular team meetings. A sampling frame for health facility staff is outlined in Table 4 and for patient participants in Table 3. We will consult with the clinical teams in each cancer unit in participation sites to support patient recruitment.

Qualitative data collection

Field observations

We will carry out participant observation in the field prior to interviews with health professionals to collect information on the routine method and variety of patient data collected in the health facilities. We will explore a range of forms used in healthcare centers (i.e. patient registration form, referral form, consent form, and medical resume). We will also gather data on the availability of radiotherapy and chemotherapy services, including details on the average daily visits, available beds, duration and frequency of the treatments, number of human resources, and the equipment used. The output from the observation will form the basis of interviews with the health personnel.

Health personnel interviews

Semi-structured face-to-face interviews will be conducted with staff involved in providing care for cancer patients. This includes hospital managers or directors of the patient service, clinicians, and health information system managers. Access to respondents will be through a written letter directed to the management of healthcare facilities. This interview will be conducted with staff from various areas because of the interrelation between units in conducting cancer care, including referral services. Data will be collected from 48 in-depth interviews. Where face-to-face interviews are not possible, interviews will be offered via telephone or video calls in place of meeting face-to-face.

Prior to the interview, written consent will be obtained to record and transcribe the interview. The one-hour semi-structured interviews will be conducted face-to-face by an interviewer; a note-taker will be present for the duration of the interview. Interviewers will record observations about the circumstances of the interview, summarise key opinions and factors that arise, and note considerations for modifying the topic guide for subsequent interviews.

A topic guide and questions have been developed to guide the interview (Table 5). Interviews will focus on the implementation of the existing HITs and cancer registry, the workflow and interoperability, current policies related to HIT, barriers in HIT operation and sharing across different units, and innovation strategies to address the challenges. We will use an interview topic guide to ensure the coverage of our research questions. Evidence from field observations will also be incorporated into interviews as a basis for discussion regarding treatment options and availability. The output of interviews will be centred around the existing and ideal cancer referral pathway, data management, and reporting mechanism to the central government. We anticipate a discussion on the possible gaps in the current pathway, including but not limited to the HIT architecture, digital literacy of the personnel, and human resources constraints. We acknowledge that only a small proportion of all general practices are sought for involvement in the study. This is an exploratory study, where we are seeking to understand a range of stakeholder experiences and ICT infrastructure being used. A small sample of general practice sites will enable the elicitation of views but may not fully capture the range of ICT approaches and current experiences in this setting.

Patient interviews

For adult patient participants, in-depth interviews will focus on an exploration of participants’ experiences of using a referral service system and how they maintain communication with the service provider, current use of digital technologies (such as smartphones), and gaps in provision that participants feel would improve or enhance existing service provision for patients with cancer. For child participants, the interview will be conducted with the caregiver to explore their experiences of using the referral systems. The interviews will be conducted following the topic guide detailed in Table 6.

We will conduct face-to-face interviews with cancer patients. Caregivers, typically family members, often accompany patients to clinical appointments and are involved in the decision-making process in Indonesia⁴⁹. As such we will be flexible about whether interviews are conducted with individual patients or with their caregivers present. All participants will be required to sign a consent form to participate in the study. Before the interview, where participant literacy levels are high, written consent will be obtained to record and transcribe the interview. Where literacy is low, verbal consent to record and transcribe the interview will be obtained from the participant. For those providing verbal consent, participants will be asked whether they would like to have a trusted relative or friend as their witness. The 30-minute-interviews will be conducted by an interviewer; As with health professional interviews, interviews will be held by telephone or video calls where face-to-face meetings are not possible.

Audio recording and electronic copies of consent forms will be stored in password-protected files using the Microsoft OneDrive platform. An electronic copy of consent materials will be made as soon as possible following an interview and once an electronic version has been created, paper copies will be destroyed. Only research team members will have access to the OneDrive platform folder set up for the study. On completion of this study, all electronic data on the OneDrive folder, (i.e. electronic copies of consent forms and audio recordings of interviews), will be moved to an electronic archive for 5 years prior to being permanently destroyed. Following completion of the study, we will assess the suitability of wider sharing of research data, once deidentified, via a research data repository. All study participants will be assigned an identification code, which will be delinked from their identity across all study documentation.

Quantitative data collection

We will collect the summary of the distribution and demographic data of cancer patients from hospitals and Puskesmas routine reports, along with data from the district health office; Data will be drawn from the report of the 24-month period prior to the study taking place in May 2020. We will collect the number of admitted and referred cancer patients in the last two years. Variables included will be the type of admission (emergency room, inpatient, outpatient), referral status (referred or not), and the outcome (discharged, died, referred). The data will be used to determine patterns of referral as reflected through routine data and explore whether these data can inform where gaps exist in the existing referral system.

Data analysis

Qualitative data. Interviews will be transcribed verbatim before being imported into NVivo 12 software for deductive framework analysis. An initial exploration and charting of pseudonymised transcripts will be undertaken (AE) through a process of line-by-line coding. Initial coding will be conducted using a random set of transcripts (n=3) for each of the three key stakeholder groups (i.e. patients, health professionals involved in the care of patients with cancer, health professionals and administrators involved in the management of cancer referral systems). A final framework will be agreed through discussion with the wider team. Once inconsistencies are resolved, all remaining interview transcripts will be coded within the joint coding frame. In order to explore common themes and divergences across the stakeholder groups, comparative analysis in the framework will be undertaken. A model of the coding frame will be developed, and each theme and subtheme defined to ensure the internal consistency of each code. Illustrative codes will be reported for each theme, with the study ID code to demonstrate reporting from across the sample breadth. Safeguards for rigour and quality of data analysis will be embedded into the analysis process, including regular team meetings to discuss transcripts, codes, and development of themes, alongside verification of accuracy of data through listening to audio recordings of interviews and assessing transcripts by team members (AF, MJA). Where possible, respondent validation will be incorporated into the analysis, sharing summaries of preliminary findings with research participants to provide an opportunity for them to engage with, and add to, interview and interpret data. The final output from the analysis will be used to inform the development of conceptual maps, ensuring the representation of data from all participant groups.

Quantitative data. Descriptive statistics will be generated to inform specific themes of interest (as outlined in Table 7). Exploration of routine datasets will provide insights into the current distribution and demographic data of patients accessing cancer services in participating sites. This will also be explored at a regional level, with descriptive analysis of regional data providing insights into the distribution and patterns of cancer prevalence in the Yogyakarta region. Secondary data from routine reports will be analysed using Stata/MP 14.1 (StataCorp LLC, College Station, USA). We will also apply network representation analysis to visualise the data. Pairwise relationships between health facilities across the levels of the health system will be derived from descriptive analysis and will be described as a collection of nodes and edges to depict relationships between the facilities. This is intended to provide visually informative data alongside guiding conceptual model development as part of subsequent Theory of Change workshops.

Intervention development and modelling

This study intends to understand and map existing referral processes as part of cancer service provision in Yogyakarta. However, findings are also intended to be used to guide prioritization of future intervention development to refine existing referral processes. The findings from the quantitative and qualitative components of the study will be fed back to key stakeholders in cancer service delivery within Yogyakarta and national representatives. This will be used to elicit feedback on study findings and derive priorities for future research and intervention development.

A narrative synthesis approach will be used as a framework to guide the describing, comparing and combining of heterogeneous qualitative findings and quantitative results from across the study. The synthesis of findings will be used to generate conceptual maps. We will use elements of Popay et al.’s methodology of narrative synthesis⁵⁰ which guides the bringing together of disparate literature from multiple sources. We will follow the steps outlined by Popay et al.⁵⁰, which will include: i) developing a theory of how the referral system currently works, why and for whom; ii) developing a preliminary synthesis of findings from across the study activities; iii) exploring relationships in the data, including exploration of alignment of findings against the TOGAF approach (e.g. components of business, architecture, technology and data) and eight interrelated dimensions of the socio-technical model; and; iv) assessing the robustness of the synthesis. Findings from qualitative and quantitative activities will be synthesised to guide the development of an underpinning theory of how referral systems currently work within Yogyakarta and generate conceptual maps for refinement through engagement with key stakeholders in the existing cancer referral system, including the Ministry of Health of the Republic of Indonesia, the Ministry of Information and Communication of the Republic of Indonesia, the Yogyakarta Special Region Health Office, the Association of Yogyakarta Special Region Hospitals, and BPJS as the national health coverage agency.

The use of a conceptual map (also referred to in the literature as logic models, analytical frameworks, concept maps, or influence diagrams) will provide a means of visualising the underlying theory of change beneath the existing referral system as determined through the study. The conceptual map presents a useful framework for describing how the referral system is believed to achieve its effects. By presenting this information graphically, rather than in narrative form, we hope to gain a clearer view of the relationship between constituent parts and how they interact. We will produce different types of conceptual maps for engagement with key stakeholders, including both activity maps (focusing on the implementation process and aiming to describe the intention of the referral system and the specific steps needed to achieve this intention) and outcomes maps (focusing on the connections between the resources, activities and the outcomes, which may be subdivided into short-term outcomes, long-term outcomes and impact). Discussions will also incorporate prioritisation of future efforts including around digital health technologies specifically for cancer, regulations that ensure the effectiveness and efficiency of cancer referrals (including financing, infrastructure, human resources, technology regulation, interoperability, and data standards), and potential partnerships between government and private sector (public-private mix).

Conceptual maps will be presented, reviewed, discussed and refined during stakeholder meetings and workshops. Two workshops will be conducted. The first will focus on Yogyakarta stakeholders (District Health Office, Provincial Health Office, National Insurance Agencies, Indonesian Cancer Foundation, and health facilities) to determine pathways to impact at the provincial level. The second meeting will focus on the national perspective (with representatives from The Ministry of Health, Yogyakarta Provincial Health Office, National Insurance Agency, and Hospital Organizations) to discuss the recommendations derived from the research and their relevance for national policy. Alongside enhancing and validating findings from the study, conceptual maps will be used formatively as a mechanism for engaging with stakeholders and stimulating debate about the future development of the cancer referral system in Yogyakarta. The data collected will be the basis for interventions related to improvement and innovation for cancer referral services in Yogyakarta and may have national relevance.

Ethics

The study involves the participation of human subjects requiring consideration of ethics issues relating to informed consent of participants, beneficence, ensuring participant confidentiality and protection of study data, and a respect for privacy. This research has received ethical approval from the joint Medical and Health Research Ethics Committee (MHREC) of Universitas Gadjah Mada and Dr. Sardjito Hospital (No. KE/FK/1075/EC/209). The MHREC has the rights to monitor the research activities at any time.

Dissemination of information

A number of activities will be undertaken to support the dissemination of project findings. These will include: (1) developing newsletters and press releases to communicate key project findings to the general public, (2) delivering presentations at local and national conferences in Indonesia, alongside presentations at international conferences, (3) publishing articles in peer-reviewed journals and (4) social media through research team member and institutional accounts. Participants will be anonymised in any dissemination activities. Only pseudonymised, non-identifiable characteristics and quotes will be used in dissemination.

Study status

The study is currently recruiting participants.

Research registration

This study has been registered with Research Registry (researchregistry5807) 15/07/2020.