Forecasting imbalances of human resources for health in the Thailand health service system: application of a health demand method

Background For an effective health system, human resources for health (HRH) planning should be aligned with health system needs. To provide evidence-based information to support HRH plan and policy, we should develop strategies to quantify health workforce requirements and supply. The aim of this study is to project HRH requirements for the Thai health service system in 2026. HRH included in this study were doctors, dentists, nurses, pharmacists, medical technicians (MTs), physiotherapists (PTs), and Thai traditional medicine (TTM) practitioners. Methods and results The study mainly relied on the secondary data in relation to service utilization and population projection together with expert opinions. Health demand method was employed to forecast the HRH requirements based on the forecasted service utilizations. The results were then converted into HRH requirements using the staffing norm and productivity. The HRH supply projection was based on the stock and flow approach in which current stock and the flow in and out were taken into account in the projection. The results showed that in 2026, nurses are likely to be in critical shortages. The supply of doctors, pharmacists, and PTs is likely to be surplus. The HRH requirements are likely to match with the supply in cases of dentists, MTs, and TTM practitioners. Conclusion In 2026, the supply of key professionals is likely to be sufficient except nurses who will be in critical shortages. The health demand method, although facing some limitations, is useful to project HRH requirements in such a situation that people are accessible to health services and future service utilizations are closely linked to current utilization rates.


Introduction
Human resources for health (HRH) availability has improved in some countries. Still, shortages, skill-mix imbalances, maldistribution, barriers to inter-professional collaboration, rural turn over, and limited availability of health workforce data persist in many countries [1]. Recent efforts by the World Health Organization (WHO), the Global Health Workforce Alliance (GHWA), and partner organizations to facilitate the development of a global HRH strategy for the period 2016-2030 reflect the growing recognition of the importance of HRH planning. They also highlighted that "countries should build planning capacity to develop or improve HRH policy and strategies that quantify health workforce needs, demands and supply under different future scenarios" [1].
Human resources for health (HRH) planning is the process in which an organization attempts to estimate the HRH supplies and requirements; determine the most appropriate balance among the skills, distribution, and number of health workers; and make a plan to redress HRH challenges [2]. Effective HRH planning thus needs evidence-based information to support, so as HRH requirement and HRH supply projection can contribute quantitatively for the resilience HRH plan. Planning for HRH thus has a central role in order to continue progressing towards the realization of the sustainable development goals (SDGs).
Ranges of HRH projection approaches were made available; population ratio, health need, health demand, and service target methods and each of these approaches have their advantages and limitations [3]. The simple projection of population ratio method, although require less data, does not take into account the effects of changes in health service utilization and HRH performance [4]. Health need method explores changes in population needs for health services base on changes of patterns of health problems [3]. Compare to other methods, it is considered a comprehensive approach for HRH projection [4]; however, drawbacks include challenges in defining needs in terms of coverage and quality, and the projection may generate unrealistic number of HRH requirements without adjusting to mal-distribution [5]. Health demand method draws on observed service utilization rate, applies to future population profile, and converts to HRH requirements. The approach is appropriate for such health system that the population is accessible to a suitable mix of services. However, the approach requires considerable sets of data and the projection trends to base on status quo situation [5]. The service target approach sets targets for specific services using current service provision, other factors, and expert opinions [4]. This approach may be useful in planning critical health care services or service for specific population; however, the approach may depend on unreliable assumptions [4]. Taking all these aspects into consideration, this study employed health demand method to forecast HRH requirement for Thailand health service system because of the high health service coverage of Thailand health service [6] and the availability of administrative data system.

Thailand health service system
Thailand is currently an upper middle-income country and is making progress towards meeting the SDGs. Although Thailand has made significant success in some health indicators (SDG 3), some other health challenges persist and need crucial attentions. The unfinished agenda of infectious diseases such as tuberculosis and HIV/AIDS coupling with non-communicable disease, such as cancer, cardiovascular diseases, diabetes mellitus, and hypertension, and traffic accident have threatened Thailand health system [7]. What is more, the increase of elderly health needs has prompted Thailand to improve health services to be in line with people needs.
With regard to the Thai health service system, virtually, everyone (99.84%) was covered by health insurance while other forms of social security have expanded [8]. Available health facilities delivering services to people can be classified according to three-tier service systems: primary, secondary, and tertiary care facilities [9]. The primary health care facilities are the first contacts which individuals, families, and communities have with the health care system. Services provided incorporate common illness treatment, health promotion, disease prevention, rehabilitations, and community health interventions. This type of health facilities includes health centers and private clinics without bed. The secondary health care facilities provide curative care with the initial referral being made by the primary care professionals. Care provision is more complex and need higher medical expertise. Such health facilities include the Ministry of Public Health (MoPH) community hospitals, other public hospitals, and private hospitals. The tertiary health care facilities provide specialized care, usually on referral from primary or secondary medical care personnel, by specialists working in a facility that has personnel and facilities for special investigation and treatment. These facilities include MoPH general and regional hospitals, university hospitals, other public hospitals, and private hospitals [9].
Approximately 54% of Thai population resides in rural areas where they are mainly served by health facilities responsible to MoPH [10]. These include 9777 health centers which cover 100% of all sub-districts and 780 community hospitals (range from 30 to 120 bed hospitals) covering 88.8% of all districts [8]. The primary health care facilities located at urban areas comprise MoPH community medical centers, municipality health centers, private medical clinics, and drug stores. In urban areas of 77 provinces, the secondary and tertiary medical services are provided by 177 MoPH hospitals, 11 medical school hospitals, 105 other public hospitals, and 249 private hospitals [8].

HRH planning challenges
The density of doctors, nurses, and midwives of Thailand was 3.1/1000 population [11], well below the recommendation threshold of 4.45/1000 population [1], indicating the shortages of HRH. In addition, the maldistribution of HRH, particularly between rural and urban areas, hindered the progress of the Thai health system. It is well recognized that, to address the HRH challenges, the HRH requirements should be aligned with the health system needs in terms of the number, skills, and distribution. The evidence-based quantification of HRH requirements is thus required. Already several attempts have been made to estimate and forecast the HRH requirements for doctors [12], dentists [13], nurses [14], and pharmacists [15]; some challenges have been elucidated. The requirements were dated and each professional was projected individually without taking into account for team-based planning. Moreover, the requirements were mainly focused on public sectors which were likely to underestimate the country requirements. This study is therefore attempted to fill the gaps by forecasting HRH team in 2026 which comprise doctors, dentists, nurses, pharmacists, MTs, PTs, and TTM practitioners in primary health care, secondary health care, and tertiary health care services in Thailand.

Methods
To forecast the HRH requirements, the approach mainly relied on the secondary data and stakeholder committee opinions. Since the available secondary data were mostly related to curative and rehabilitative service utilization, the health demand method or service utilization method was then used to project HRH requirements. It is suggested that this method is appropriate for the countries where service utilization and related data are available and people are accessible to health services [5]. The method was considered appropriate as this study focused mainly on HRH requirements relating to service utilization at 3 health facility settings: primary, secondary, and tertiary health facilities. The high health insurance coverage (99.84% of population) and low unmet health needs (1.5% of sick patients) [6] have justified the utilization of the health demand method. HRH professionals included in the forecast are HRH whose main functions are related to curative and rehabilitative services including doctors, dentists, nurses, pharmacists, MTs, PTs, and TTM practitioners.
Health demand method forecasts the HRH requirements based on the forecasted service utilizations and then converts to HRH requirements by the staffing norm and productivity as the following steps. Pharmacists dispensed at about 20% of outpatients at primary care, and 70% of outpatient visits and inpatient days at secondary and tertiary care levels required drug dispensing service. All dental services at all 3 levels of care were provided by dentists. Due to unavailability of PTs and MTs at primary care, physiotherapy and laboratory services were only provided at secondary and tertiary care facilities. According to the Department of Thai Traditional Medicine [16], TTM service clients were estimated to 17.5% of general outpatients and 70% of this figure needed Thai traditional interventions (see details in Table 1). 3. Collect the service utilizations and related data to project the service utilizations in the next 10 years (2026), and these steps were followed: 1) Collect and collate the MoPH administration data relating to service utilizations at 3 levels of care. Since the 2013 data provided details of annual volume of main services provided at each level of care and was verified by MoPH committee [17], the TWG decided to use the service utilization and related data produced in 2013 for the purpose of HRH projection. 2) MoPH service utilization proportion by age and sex was used to compare with the country population of the same age and sex in 2013 [18].
The derived proportion was then used to multiply the projected population in 2026 of the same age and sex in order to obtain the projected service utilization at each level of care in 2026. By this approach, it was under the assumption that the pattern of service utilizations would remain constant in the future. 3) Use the percentage of service utilization by health facility types (MoPH, other public and private facilities) obtained from the survey report of the National Statistical Office [6] to add on the volume of the MoPH service utilization. Of all outpatient visits, 70, 24, and 6% of them used the services at MoPH facilities, private facilities, and other public facilities, respectively. For in-patient utilization, 76, 16, and 8% of all inpatients used the services at MoPH facilities, private facilities, and other public facilities, respectively. Regarding dental services, the unmet need was 1.05%, and service Ultrasound services In the projected scenario in 2026, it was presumed that all planning variables except population are held constant at current levels. 8. Match the projected HRH requirements with the projected HRH supply. The supply projections were reviewed from the 7 professional requirement and supply studies [20][21][22][23][24][25][26]. The stock and flow model was used for each study in order to project the HRH supply in 2026 [11]. The current stock of each professional was obtained from the licensed HRH who were actively practicing. The potential inflow into the HRH stock was based on the production plan of each professional, while the outflow against the stock was defined according to the potential number of retirements, death, resignation from the health system, and drop out from colleges/ universities. As the emigration and immigration of HRH was minimal, so this information was not taken into account for the forecast. The projected HRH supply was then calculated and matched with the projected requirements to identify the HRH imbalance.

HRH requirements in 2013
At primary care, only outpatient services were included in the HRH workload. The total outpatient services were 210 876 061 visits. Of this figure, 74 587 669 Table 2.
The services provided and HRH requirements at secondary and tertiary care facilities are shown in Tables 3  and 4. The quantity of services provided by doctors and nurses at secondary care were dominated by OP services, IP services, A&E services, ultrasound services, and normal and abnormal delivery services at approximately 110 290 828 visits, 12 949 845 patient days, 5 974 225 visits, 902 553 visits, and 281 048 cases and 15 664 cases, respectively. Comparing to those of the secondary care, although OP services at tertiary care facilities were about a half, the IP services was double, particularly those of the ICU service which were about 19 times more. The number of surgical cases was almost 13 times higher, and the abnormal delivery cases were about 3 times higher. All other services were comparable. Drug dispensing at tertiary care was slightly higher than those at secondary care. Laboratory services provided were about 3 times as high as those at secondary care. Clinical chemistry, microbiology, molecular biology, and blood banking tests/services were greatly higher at tertiary care. Details are shown in Table 3.
Dental services, PT services, and TTM services were higher at secondary care facilities, compared to those at tertiary care facilities. Dental services at secondary care services were 3 times as high as those of tertiary care services and the PT services at secondary care were slightly higher than those at tertiary care facilities. TTM services provided at secondary care facilities were about 1.8 times higher than those at the tertiary care. Details are shown in Table 4.

HRH requirements in 2026
The projection of service utilization and HRH requirements in 2026 are shown in Tables 5 and 6. The amount of OP services in 2016 will be the highest at the primary care compared to the secondary and the tertiary care. IP services are highest at the tertiary care; particularly, ICU services will be approximately 19 times of those of the secondary care. Surgery, abnormal delivery and ultrasound services will be also performed highest at the tertiary care. In contrast, A&E services will be higher at the secondary care facilities. Accordingly, laboratory services will be higher at the tertiary care. The volume of drug dispensing at the secondary care will be higher compared to that at the tertiary or the primary care. Details are given in Table 5.
Dental services will be highest at the secondary care compared to the primary or tertiary care. PT services at the secondary care will be close to those at the tertiary care. TTM services will be high at the primary and the secondary care compared to the tertiary care. Details are given in Table 6.
The projection of HRH requirement in 2026, after inclusion of allowance percentage, revealed that the health service system at the primary care facilities will require 8535 doctors, 29 691 nurses, 3118 pharmacists, 3575 dentists, and 6534 TTM practitioners. The secondary health facilities will require 14 998 doctors, 59 489 nurses, 7713 pharmacists, 5901 MT, 8416 dentists, 2254 PTs, and 9096 TTM practitioners. The projected tertiary

Sensitivity analysis
The effective HRH requirements is dependent not only on the number of service utilization but also on the average time spent on each service, working hour, proportion of time spent on other activities, etc. Time spent on each service or productivity of each professional varies across health facilities including public and private sectors or even within health facilities. The range of time spent on each service, i.e., the baseline time and the baseline time ± 10%, were chosen to perform sensitivity of various scenarios, while other factors held constant. Table 8 demonstrates the sensitivity of HRH requirements to different level of time spent per service. The higher the time spent per service, the higher the number of HRH requirements.
Matching HRH requirements and supplies in 2026 Table 8 illustrates the comparison between HRH requirements and supplies, and the gap identification in

Discussion
Although various approaches have been carried out to increase the accuracy in projecting the future HRH requirements and supplies, this kind of study often faces some limitations. First, several factors critically affecting demand of health care are not taken into account in the projection. Challenges facing Thai health system including rapid increase of aging population, increase of people living with chronic conditions, and growing demand of services provided in home and community settings should be included in the projection to show the direction towards which the health service system and HRH situation may be heading. Second, service utilization data used were focused mainly on curative and rehabilitation services. Health services in relation to health promotion, disease prevention, and home health care were not taken into account. Third, the HRH requirements did not cover HRH working as administrators, lecturers, researchers, and business related to health. The projected HRH requirements in the study therefore were likely to be underestimated. Fourth, the study mainly used the secondary data from various sources in projecting HRH requirements and supplies. Since the quality of data was varied among the sources, the results have some risks in the data accuracy and integrity. Fifth, the HRH requirements varied upon the volume of service utilization, time spent on each service, working hour, and proportion of time spent on other activities. The fact that the changes of these figures critically affect the number of HRH requirement, the results should be viewed with considerations. Sixth, the TWG play important role on crucial figures affecting HRH requirements. Despite the fact that several key stakeholders involved in decision making, the results risk towards professional bias and likely to require more HRH than the health system can afford. Despite some limitations, the present results demonstrated the potentials to project HRH requirements in a multi-professional manner rather than the single professional projection to consider the possible interaction and team work, particularly task sharing of doctors and nurses at the primary care level. The results support the recommendation made by Ono et al. [27] in that the HRH projection should shift from uni-professional to multi-professional approach. Moreover, in this study, multi-professional committees were involved at all important process of the HRH projection. This multi-professional projection approach has been carried out also in other studies, i.e., the calculation of family doctors, nurses, public health workers, and others at community health centers in China [28]; the projections of doctor, nurse, and midwife requirements in OEDC [29]; and the projection of skill-mix among doctors, advanced practice nurses, and physician assistants in Finland, the Netherland, Switzerland, and the United States of America [27].
The present results illustrated that, in 2026, doctors, pharmacists, and PTs were likely to be surplus, if the service system and HRH system be carried out as usual. On the other hand, the number of nurses will be in extremely shortage. The supply and requirements of dentists, MT, and TTM practitioners will be in balance. Comparing with the threshold of 4.45 (doctors + nurses + midwives)/1000 population to achieve the median score (25%) of attainment of 80% coverage for the 12 selected SDG tracer conditions [1], the requirement of doctors and nurses in this study was about 3.64/1000 population, slightly below the recommendation.
However, the HRH requirements in this study were likely to be underestimated. First, doctors and pharmacists provided care at about 20% of outpatient visits at public primary care facilities, and thus, the requirements of doctors and pharmacists were lower than what should be. Second, home health care for aged and chronically ill patients were not included in the scenario, so that the requirement of home health care providers, i.e., PTs, doctors, and nurses, were thus underestimated. Third, if drug industries and other drug-related functions were included in the projection, the requirements of pharmacists would be much higher. HRH requirements applying the service utilization for multi-professional approach in this study were lower than those used in different methods for single professional projections in Thailand. Dentist requirements in 2026 by the service target method by Jaichuen [18] showed the needs with the range of 19 677-20 955 which were higher than the supply. Pharmacist projection by the system dynamic approach required 47 787 pharmacists in 2026 [19]. Sawaengdee [17] combined service utilization and health need methods to project the nurse requirement which showed similar results to this study. At the same time, PT requirement by the combination of service utilization and service target approaches showed the requirements of 7855-8378 [21] which was higher than that of the results of this study.
Although the results of this study indicated that the production of key professional is likely to be sufficient, the concerned issues are the absorption, distribution, and retention of HRH which were not under the scope of this study.
There are some advantages of the use of health demand methods in projecting HRH requirements. This approach is appropriate for the countries where the population is accessible to a suitable mix of health services and future service utilization rates are closely linked to current utilization rates. The approach used in this study has already been used in several studies to project HRH future requirements, particularly for curative services. The health demand method was used to project for future doctor requirements in Mexico [30], Australia, Belgium, and Canada [27] and dentist requirements in Australia [31]. Moreover, the approach has been used at sub-national level, for example, there was an attempt to calculate the requirements for doctors, nurses, public health workers, and allied health workers working at community health center in a district in Beijing, China [28].

Conclusion
The results of this study implied that the production of key professional is likely to be sufficient in the future, except nurses which are likely to be in critical shortages. Health demand method used to project HRH requirements in this study, though there are some limitations, was likely to be applicable for countries where health service utilization constantly changed and was useful for micro-planning at the facility level. However, HRH planning should be reviewed and updated often as the context affecting health service is dynamic.