Integrating NCD into Primary Health Care in Thailand: a Mix Method Study

Background In response to an increased burden from non-communicable diseases (NCDs), primary health care (PHC) is advocated as an effective platform to support NCD prevention and control. This study aims to assess Thailand’s PHC capacity in providing NCD services, identify enabling factors and challenges and provide policy recommendations for improvement. This cross-sectional mixed-method study was conducted between October 2019 and May 2020. Two provinces, one rich and one poor were randomly selected and then a city and rural district from each province were randomly selected. From these four sites in the two provinces, 56 ocers from PHC centres were sampled purposively for a self-administrative questionnaire survey on their capacities and practices related to NCD. A total of 79 participants from and Oces, provincial and district and PHC centres who involved with NCD participated in focus group discussions or in-depth interviews.

Various global health forums reiterate and foster commitment on prevention and control of NCDs, for example: an Independent high-level commission on NCDs in 2018 3 ; the global action plan for the prevention and control of NCDs 2013-2020 4 ; and the Montevideo Roadmap 2018-2030 5 . SDG target 3.4 aims to reduce premature NCD deaths by one third by 2030 6 .
The global action plan on NCDs calls upon strengthening primary health care (PHC) for effective prevention and control of NCDs 4 , emphasizing early diagnosis, treatment and management of complications. Evidence shows positive health outcomes from integrating NCD at PHC level 8,9 , and reduction in hospital admissions 10 .

De ning PHC
In line with the principle of rst contact of citizens to health service, Thailand de nes PHC as a network of 10-15 PHC centres and a district hospital of 30-120 beds providing services to 50,000 population in the catchment area. PHC are close to where people live. PHC centres provide a comprehensive range of services including health promotion, disease surveillance, home healthcare, out-patient services with supervision and support by medical doctors from district hospitals.
At the community level, staff at PHC centres work closely with local government, other government sectors such as school and agriculture and community leaders. Over one million village health volunteers (VHV) play critical roles to support PHC.
Three decades of PHC infrastructure development since the 1970s reached full geographical coverage in all sub-districts and district hospitals nationwide 11 . In 2018, 9,806 PHC centres in 7,255 sub-districts provided a range of health services to an average ve thousand citizens in the sub-district catchment area 12,13 . Since 2002, Thailand achieved full population coverage through nancial risk protection schemes which offered a comprehensive bene t package (including the whole range of NCD interventions) free at point of services 11 . A study shows that the Chronic Diseases Clinic Model for diabetes and hypertension at PHC level has signi cantly shifted NCD patients from hospitals to PHC centres, minimising congestion in hospitals while maintaining good clinical outcomes 14 .
This study assessed PHC capacities in addressing NCDs, identi ed enabling factors and challenges, and provides policy recommendations for improvement.

Methods
This study was conducted between October 2019 and May 2020; it applied mixed-methods, using both qualitative and quantitative approaches. A literature review focused on NCD management at PHC centres, essential resources such as staff, essential medicines and diagnostic equipment, and referral systems to secondary and tertiary facilities. Findings from literature reviews were used to design self-administered questionnaire and interview guidelines for key informants who are healthcare professionals in selected PHC facilities.

Study sites
The study sites were selected based on provincial economic status and NCD mortality rate. Firstly, gross provincial product per capita, a proxy of socioeconomic status, was utilised. High and low socioeconomic provinces were de ned as the top ten richest and poorest provinces in Thailand. Secondly, the provincial speci c NCD mortality rate from the Department of Disease Control was retrieved. The average NCDs mortality rate in Thailand in 2018 was 114.28 per 100,000 population (min: 64.56, max: 199.49 ) 15 . In these ten richest and ten poorest provinces, out of the provinces which had higher than national average NCD mortality, one province was randomly selected. Saraburi province in the Central Region represented the highest socioeconomic province with an NCD mortality rate of 146.42 per 100,000 population, whereas, Phrae province in the Northern Region represented the lowest socioeconomic province with an NCD mortality rate of 176.33 per 100,000 population. Both provinces have higher than national average NCD mortality rate.
One city and one rural district of the two provinces were randomly selected to participate in this study. (See Table 1). In the four districts, all PHC centres including district (in rural area) or provincial hospital (in city), district public health o ces were invited and participated in this study. Data collection Quantitative data was collected by self-administered questionnaire surveys, one survey form for each PHC centre. In total 56 PHC facilities responded to the survey (38 in Phrae and 18 in Saraburi). The questionnaire comprised three parts: facility characteristics, responsibilities, and availability of essential resources.
Qualitative information was collected from in-depth interviews and focus group discussions (FGD) conducted among 79 participants purposively chosen from healthcare professionals who are responsible for NCDs in the PHC facilities, district hospitals staff who provide technical support to PHC and treatment of referral cases, and o cers at District Health O ces and Provincial Health O ces. Ten sessions of FGD were convened prior to 18 sessions of in-depth interviews to solicit key information for further assessment. (Table 2) Each interview lasted 40-60 minutes. Interviews were audio-recorded after written consent, and transcribed for qualitative analysis. Key ndings were triangulated by related documents, interviews and observations by the research team.

Data analysis
Quantitative ndings from questionnaires were analysed using descriptive statistics, namely mean, standard deviation and percentage to describe size, distribution, and pro le of PHC. For qualitative data, thematic analysis was applied, grounded by deductive and inductive approaches. Deductive methods were used to draft questionnaires and interview guidelines. Inductive analysis was performed when new information emerged after completion of eldwork.

Ethical approval
Ethical approval was granted by the Institute for Human Research Protection, Thailand (Reference: IHRP 096/2562). All data are kept anonymous and dissemination of data is for academic purpose without individual attribution.

Results
All 56 PHC facilities provided complete survey questionnaires (100% response rate), while 79 key informants participated in sessions of in-depth interviews and FGD.
PHC stakeholders from the two selected provinces fully engaged in ten sessions of FGD, which followed by 18 sessions of in-depth interviews relating to PHC's role in managing NCDs. Interviewees were selected based on their responsibility in managing or supporting NCD work at PHC level.
Three thematic areas emerged from key ndings from questionnaire survey, FGD, in-depth interviews and triangulation with literature reviews and other key informants. .

Theme 1 PHC foundation and enabling factors
Findings show that strong foundation for PHC is the result of continue policy and nancial support, improved management and human resources,

PHC functions
Self-administered questionnaire surveys found that PHC's key function is to provide a comprehensive range of health services such as health promotion and disease prevention, treatment, and rehabilitation. This accounts for 55% of the total workload, of which NCDs take a major share. Around one third of the workload contributes to community engagement such as support to disabled, home-bound and bedridden patients. Intersectoral collaboration with local government units, which address the social determinants of health and empower citizens, accounts for 18% of PHC centres' workload. (See Figure 1) From the survey, PHC centres provide all services such as diseases surveillance, environment health, mental health, home visits, NCD-related services and treatment, but dental health services are not provided by one third of 54 PHC centres due to the lack of dental personnel. (Figure 2) 2. PHC essential resources Self-administered questionnaire surveys also assessed essential resources for the functioning of PHC centres, include human resources and essential medicines.

Healthcare workers
This study categorizes PHC centres by the size of the catchment population: small (< 3,000 population), medium (3,000-8,000) and large size (> 8,000). The survey results revealed slight difference in numbers of staff by size. (See Table 3).
Total numbers of healthcare professional (including nurses, public health o cers and dental nurses) were three, ve, and eight in small, medium and large size PHC centres. There is no difference in the number of health care workers between richer (Saraburi) and poorer (Phrae) provinces. The number of registered nurses, mostly post-graduate trained as Nurse Practitioners increased by the size of catchment population to accommodate more curative service workloads. In contrast, there are, on average, two public health o cers who are four year trained regardless of size. A four-year trained dental nurse and dental unit are only available in medium and large PHC centres.  "NCDs quality standard comprises several indicators, but often without adequate budget allocation to ful l these mandates. As a result, our performance will be marked in the red zone due to budget shortfalls." [L7] Limited human resources for health Some NCDs services once provided by hospitals are increasingly shifted to PHC centres. Theme 3 Dynamic social context: an emerging challenge Urbanization and socialization rapidly transform local context and bring on board challenges including lifestyles diseases. This is another concern raised by local PHC and echoed throughout the study.

Urbanisation and social in uences
Rural PHC centres facilitate easy access to services and maintain good relationships with villagers and community leaders. Key informants con rm that patients prefer to seek health services from their local PHC centres rather than visiting over-crowded hospitals [L3, L4]. Trust and interpersonal relationships within the community, built over years, in uence people's decisions to visit PHC centres [L7, H6].
"People always choose the best option for themselves, therefore, easy access to PHC centres in their community are considered their best choice" [L7] Greater challenges are echoed by key informants from urban PHC centres. Patients in urban areas, with various choices of private and public clinics and hospitals, often bypass PHC centres [H3]. In addition, some private companies offer private insurance to employees who often use private hospital services.
Coverage of NCDs screening and treatment outcomes by the private sector are unknown, as this information is not captured by the MOPH information system. [H6].
Urban populations live in obesogenic environment, having more access to fast food, sweetened beverages and inadequate physical activity compared with rural counterparts. Energy-dense foods are key risks to obesity and NCDs [L4, L7, H2, H5]. Key informants also voiced that health promotion around smoking, alcohol, and physical activity is less effective if the population is not interested [H2, H3, H6]. Addressing commercial determinants of tobacco, alcohol and unhealthy diet through implementing WHO best buys measures is important, but beyond the capacity of PHC workers 20 .
Health literacy at the heart of NCD prevention and control Many key informants suggested promoting health literacy through mass or local media and that the MOPH should monitor and take legal actions against the promotion of falsely claimed products related to NCDs.
[L7] Promoting healthy diets through schools and community-based interventions requires parallel reforms for conducive food environments [L1, L7].
"We have been discussing 'Health Literacy' for years without applying it in context. We should apply these principles instead of repeating our talk." [L3]

Discussion
Thailand achieved UHC in 2002 with the whole population covered by nancial risk protection systems. The 9,806 PHC centres in 7,255 sub-districts nationwide provide the foundation for implementing UHC, as full geographical coverage contributes to equitable access to health services by all citizens 11,21 In past decades, PHC contributed to improved maternal and child health status, control of malaria and other infectious diseases and today PHC is also good for NCD epidemics. In 2017, government health spending was high at 15.03% of general government expenditure, and household out-of-pocket payments were 11.15% of health expenditure 22 . Spending on PHC as proportion of current health expenditure is unknown. Though the national health accounts have yet to estimate this gure, our eld assessment found it adequate as there is no access gap and no stock-outs of essential medicines.
Health workforce and nancial resources: key determinants for functioning PHC The cutting edge of PHC is the availability of quali ed nurses and other health professionals in every PHC centre, often recruited from local communities for training and home-town placement upon graduation. Implementing rural health workforce retention policies such as local recruitment and hometown placements, and nancial and non-nancial incentives result in higher rural retention 23 . Quali ed staff contribute to trust in quality PHC services by citizens 24 .
This study shows di culties in recruiting adequate numbers of dental nurses in small PHC centres due to limited posts and lack of career paths. The MOPH recognizes the problems but solutions have yet to be decided. A study shows 72.8% of dental nurses resign due to frustration and heavy workloads 25 .
VHVs (1.054 million in 2020) are lay people in communities who are recruited to support PHC functions.
Six three-hour modules are required for initial training of a new VHV 26  Findings from this study con rm PHC can ful l the rst mandate on service provision and PHC centre health workers are fully trained for this purpose including for NCDs. For example, 4 of 67 KPIs in 2020 are related to NCDs 29 and district, provincial and regional health authorities are required to report quarterly progress. Essential medicines for NCDs are available in more than 90% of PHC centres. Rotation of physicians to PHC centres boosts quality of care.
Challenges remain on intersectoral actions which address the determinants of health, as PHC is not designed, with inadequate competency to implement WHO best buy interventions 30 which can be most effectively implemented through national-level policy actions, effective law enforcement, such as increased tax and price, and control of alcohol availability. Despite these challenges, one study shows successful results in advocating zzy-drink-free schools by working closely with PHC centres, communities, local government and civil society organizations 31 . Scaling up these projects are underway.
Although PHC has empowered communities to optimize their health through working with community leaders, local government and VHV; challenges remain as these interventions are not effective compared with WHO best buys, which requires national synergies Figure 4 depicts the district health system, as a key PHC platform for integrating NCD prevention and control in Thailand. Different key actors provide full support for the functioning of PHC including NCD services.

Study limitations
A few limitations exist. First, the small sample size of PHC centres in two selected provinces limits generalizability of ndings. Second, the majority of participants were from PHC centres and in-depth interviews are dominated by PHC staff perspectives. Third, the COVID-19 situation in 2020, which enforced physical distancing and limit travel, did not allow interviews with the community representatives. Finally, there is an unavoidable re exivity bias of results as the researchers are also working in the public health sector.

Conclusion And Recommendations
From the 1970s it took three decades by successive governments to achieve a full geographical coverage of PHC centres and district hospitals in the 1990s. An average of three to eight staff members in small, medium and large size PHC centres can accomplish the PHC mandate in providing comprehensive health services throughout the life course, including NCDs interventions for the sub-district catchment population. PHC staffs are trained to perform these functions well, and are supported by adequate supplies of essential medicines. Over one million VHVs play critical roles in bridging PHC centres and communities and support diseases surveillance in times of public health emergencies.
Challenges remain to empower individuals and citizens to optimize their health, particularly in urban contexts. PHC has limited capacity concerning multi-sectoral collaboration to address social determinants of health as community-based interventions are not effective in implementing WHO best buy interventions.
To empower citizens and address social determinants through multi-sectoral action on NCDs, synergies and national-level support is needed for interventions such as tax and price policies on tobacco, alcohol and sweetened beverages, control of advertising, enforcing smoke free environments, and the availability and marketing of alcohol.

Declarations
Con ict of interests: The authors declare no con ict of interests.