The implementation of community-based diabetes and hypertension management care program in Indonesia

Since 2010, Indonesian government has initiated a chronic disease management program, Prolanis (Program Pengendalian Penyakit Kronis) targeted for diabetes and hypertension. The program is continued at the commencement of universal health coverage (UHC) in 2014. “This study aimed to report the utilization and cost of the implementation of Prolanis in Indonesia from 2014 to 2016, or two years since the commencement of Indonesian universal health coverage.” Secondary data analysis was performed using publicly available data and data obtained from the national health insurance agency (BPJS); while data on disease prevalence were collected from basic national health survey. There was an increase trend of Prolanis participants, from around 11,000 participants in 2014 to more than 250,000 in 2016. More than 70% of participants were adults living in Java, however, the acceptance rate was very low in other area. Across different activities in Prolanis, physical activity was the most participated ones. In comparison to other regions, regions in Java were the most active area. The total expenditure for Prolanis program in 2016 increased almost triple from the annual cost in 2014. However, the cost per person was actually decreased more than 50%. Within two years of UHC implementation, there were increase covered participants and total costs, but cost per individual was decreased and there was significant difference in of cost between Java and outside Java. Further study and routine monitoring-evaluation process by health authority is needed to assess whether the cost difference would affect the service quality.


Background
Chronic diseases prevalence has been rising rapidly, especially in middle and low-middle income countries. Four of the most prominent chronic diseases-cardiovascular diseases (CVD), cancer, chronic obstructive pulmonary disease and diabetes mellitus-are linked by PLOS ONE | https://doi.org/10.1371/journal.pone.0227806 January 14, 2020 1 / 9 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 common and preventable physiological risk factors, notably high blood pressure, high blood cholesterol and overweight, and by related major behavioral risk factors such as unhealthy diet, physical inactivity and tobacco use. Efforts to reduce these chronic diseases should focus on controlling the key risk factors in a well-integrated manner. [1] Since 2010, Indonesian government has initiated a chronic disease management program (Program Pengendalian Penyakit Kronis-Prolanis) targeted for patients with diabetes and hypertension patients, under the civil service health insurance scheme. The program aims to prevent chronic disease patients fall into severe complications. Hence, the program was still limited to civil servants. With the commencement of universal health coverage in 2014, Prolanis has since been implemented to all patients under the national health insurance scheme. Patients with chronic diseases are able to register to this program and receive the following services: 1) medical consultation/health education; 2) regular health status monitoring; 3) home visit; 4) reminder through mobile short message service (SMS) gateway; 5) club activity; and 6) monthly routine drugs. [2] This chronic disease management program is a multi-faceted interventions that involve both, pharmacologic and non-pharmacologic strategies (e.g. patient education, physical activity, monitoring, and reminder system). The program is held by primary care centers, which already in collaboration with the national health insurance agency. The primary care centers include community health care centers (Puskesmas), primary care clinics, and private doctors.
Previously, some evaluation of Prolanis in Palembang showed that implementation of the program is still suboptimal. [3] One of the issues highlighted in the study was the inability to organize activity clubs due to lack of manpower and facilities. Furthermore, despite of the program, based on the national health survey data, there was an increase of diabetes prevalence diagnosed by medical doctors from 1.5 in 2013 to 2.0 in 2018. [4]This study aimed to report the utilization and cost of Prolanis in Indonesia from 2014 to 2016, or three years since the commencement of Indonesian universal health coverage. a mixed combination of a province in Java and two provinces outside Java, thus we categorized it as a separate region (mixed region).
We estimate the number of predicted populations of risk of diabetes and hypertension by multiplying the regional prevalence with the projected number of populations in the region at the same year. For example, to calculate the predicted population at risk of Diabetes in region 6 (Central Java and Yogyakarta) at 2014, we multiply the prevalence of the region of 3% with projected number of populations in region 6 in 2014 (27,993,000), resulted in 839,790 persons. (See Table 1).
We took the regional prevalence values from the highest disease prevalence of provinces located in specific BPJS region (based on data from Indonesian National Basic Health Survey 2013 (RISKESDAS) [5]); while the projected number of population in the region is the sum of projected population of the provinces included in the specific BPJS region (based on the Indonesian population projection 2010-2035 released by Indonesian Central Statistics Agency [6]). As there are two provinces, namely Banten and West Java that were split by BPJS within two different regions; we included half of total predicted population of Banten region 4, and another half in region 13. Subsequently, we included half of total predicted population of West Java in region 4, and another half in region 5.
RISKESDAS is a nationwide health survey that is conducted every 5-6 years by the Ministry of Health of Indonesia since 2007. The 2013 RISKESDAS was a household survey that covers 33 provinces of Indonesia. The report was recently updated in 2018 (at the time of this article is written, complete official report is not yet released). The survey described the situation at the national and provincial levels, and also at the district / city level. In the RISKESDAS, history of diabetes mellitus was asked to subjects aged 15 years and over. Prevalence of diabetes mellitus are presented in two ways: first, based on respondents' acknowledgment; second based on the results of interviews in the form of prevalence based on cases of diseases diagnosed by doctors / health workers. [5] In this study, we took the latter.
Hypertension was assessed through 2 methods, interview and blood pressure measurement. We took the hypertension prevalence per province based on the latter. Blood pressure measurements are performed using a standardized digital measuring device by trained enumerators. Each respondent is measured at least 2 times. If the results of the second measurement differ from �10 mmHg compared to the first measurement, then the third measurement is carried out. The two-measurement data with the smallest difference with the last measurement are calculated as a mean as a measure of tension. [5] For the projected number of population at the year 2014 to 2016, we used the Indonesian population projection 2010-2035 released by Central Statistics Agency. [6] The projection was made base on the results of the 2010 Indonesian Population Census. This projection is made based on assumptions about birth trends, death, and migration inter-provincial populations are most likely to occur over the next 25 years. First, the projected population of Indonesia were calculated, then projection was made per province. The projection results are discussed in the technical team formed by Central Statistics Agency, and then discussed in a team meeting consisting of officials from National agencies, academics and other relevant agencies.
To see whether the program covers a substantial part of those who already at risk, we calculated the percentage of coverage by dividing the number of registered participants with the sum predicted populations at risk for diabetes and hypertension at the same year. For example, to calculate the percentage of coverage in region 6 (Central Java and Yogyakarta) at 2014, we divide the registered participants (3,810) with the sum of predicted populations at risk for diabetes (839,790) and the predicted populations at risk for hypertension (7,390,152) in the region at 2014; resulted in 0.05% of coverage (see Table 1).  we included half of total predicted population of Banten region 4, and another half in region 13. We also included half of total predicted population of West Java in region 4, and another half in region 5. 1 At 2013 this province has not yet been established ; The cost incurred from Prolanis include the cost for screening for health status monitoring (for diabetes and hypertension), activity club and health education. The costs for home visit was already included under the capitation fund for primary healthcare centers, hence was excluded in this study. Health status monitoring included examination for random and fasting blood glucose test once per month, HbA1c test every 6 months, and blood chemistry test every year. Monitoring for hypertension was conducted regularly in each health status monitoring visits and activity club, thus was not expressed in costs incurred by BPJS. Health education/ medical counselling was provided every month by general physicians at primary health care center. Physical-activity clubs are group physical activities conducted regularly by the primary health care centers for the Prolanis participants. In the physical-activity club, cost may incur from sports coach fees and operational costs. We categorized the costs by regions and locations (Java/Outside Java and Mixed regions) and calculate the cost per capita for each activity by year. The cost per capita was calculated by dividing the cost of the activity by the number of participants in each activity. All costs were in Indonesian rupiahs.
Results were presented in tables and analyze descriptively.

Results
Characteristics of regions based on population at risk of diabetes and hypertension are described in Table 1. In general, data from BPJS showed an increase trend number of participants from around 11,000 in 2014 to more than 250,000 in 2016. About 70% of participants were living in Java island. The national prevalence of diabetes diagnosed by medical doctors in 2013 was 2.4%, while the national prevalence of hypertension was 25.8%. [5] Majority of provinces located outside Java had diabetes and hypertension prevalence lower than national prevalence.
From Table 1, we can see that in 2014, regions 6 and 7 had the highest percentage of coverage (both 0.05%). In 2015, region 2 (0.14%), 5 (0.15%), and 6 (0.38%) had the highest percentage of coverage and had a higher percentage coverage than the national coverage. In 2016, region 2 (0.57%), 6, (1.33%) and 7 (0.52%) had the highest percentage of coverage. Except for region 2, all the mentioned above are located in Java.
In Table 2, we presented the utilization of Prolanis activities per region by year. The number of each activities are rising from year 2014 to 2015, however, since the number of Prolanis members are also increasing greatly, the utilization rate of all activities seems lower in 2016 compared to 2015. The utilization rate for activity club is always higher compared to other activities in all year for all regions.
In Table 3, we showed the total expenditure for Prolanis program had increased from 29,8 trillion IDR (2,1 million USD) in 2014 to 83,5 trillion IDR (5,8 million USD) in 2016. This was in line with the increase number of Prolanis participants more than twenty times from year 2014 (11,079 persons) to 2016 (260,364 persons). However, if we calculated the individual cost, the cost per capita was actually decreased from 475,033 to 215,089.5 IDR per capita (54.7% decrease).

Discussion
In this study we found an increasing number of participants and activities of Prolanis within the year 2014 to 2016. Activity clubs' activities were favorited, especially for regions located  outside Java. We also observed almost triple increased of total expenditure for Prolanis program in 2016 from the annual cost in 2014 but the individual cost spent per patient has decrease for more than 50%. The increase in the number of Prolanis participants to more than 20 folds within the year of 2014 to 2016, can be translated as something fascinating, but expected. Even after a year of implementation, the coverage of Prolanis has reached all its predicted potential population, especially in Java. Since its introduction in 2014, the popularity of Prolanis and BPJS are rising because it was promoted as free-health care service by the Indonesian government. This might cause the massive hike of utilization and claims of any service release by BPJS right after its first year of implementation. However, we can see that after 2 years of implementation, the acceptance rate was still very low for regions located outside Java island compare to regions located in Java. This might be related with the lower rate of PHCs availability outside Java. Primary health clinics are the main driver of community participations in Indonesia. In 2015, the Indonesian Ministry of Health introduced the concept of "Healthy Community Movement (Gerakan Masyarakat Sehat -GERMAS) that focus more heavily on non-communicable disease prevention.
The number of Prolanis claims were increase greatly for activity clubs, especially from 2014 to 2015 (33,580 vs. 66,100). This increase number of activity clubs was especially shown for regions outside Java. We did not find the right reason for this finding, expect that it the cost for activity clubs outside Java might be much cheaper than regions in Java, thus making it interesting and easier to implement by PHC as part of implementation of Prolanis in their area.
We also observed the high increase of total expenditure for Prolanis program and decrease individual cost. This might have been caused by the increase number of participants. According to BPJS expenditure report, the allocated budget for Prolanis program per total national health budget is decreasing every year from 0.54% (2014) to 0.13% (2016). This was in contrary to the cost for chronic diseases, which was the biggest contributor the total national health expenditure. It shows how concern on disease treatment grabs more attention than investing in promotive and prevention program. This certainly gives more reasons for MOH and BPJS to increase the budget of promotive and preventive programs for non-communicable disease prevention, such as Prolanis, in Indonesia. Previous small study show the effectivity of Prolanis in controlling the progression of hypertension [7]. Our analysis had not look at clinical outcome of these patients and we acknowledged this as one of this study weakness. In order to evaluate the clinical impact of this program, more studies that look at the association of the program to clinical outcome of chronic diseases may need to be conducted. However, this was a first study who look at national data claim of Prolanis and includes the cost of the program. Conducting this research could also be beneficial for the payers as they can evaluate and monitor the effectiveness of their program. Other weakness of this study is we relied mainly on secondary data from national health security agency; while reliability on data collection remains questionable. Moreover, the claim data we have obtained from BPJS did not have more detail component of costs and some components of regions data are missing; for example, the health status monitoring claims for in 2015 and 2016. This might have been cause by underreporting by the regions. Nonetheless, we did not believe this under reporting was selective, thus is potential for selection bias. Furthermore, as we cannot get individual patients claim data due to confidentiality issues, we could not identify whether the population, within years are overlapping. Despite study weaknesses, this study provides evidence that can help implementation and expansion of similar chronic disease management, in low-middle income countries.
As conclusion, within two years of UHC implementation, there were increase covered participants and total costs, but cost per individual was decreased and there was significant difference in of cost between Java and outside Java. Further study and routine monitoringevaluation process by health authority are needed to assess whether the cost difference would affect the service quality.