Introduction

Tuberculosis (TB) is a leading infectious cause of death, although rivalled in recent years by SARS-CoV-2 (COVID-19). Indonesia has the second highest TB burden globally with an estimated 845,000 TB cases and 98,000 deaths each year [1]. There have been improvements in TB case detection in recent decades, especially since a major case detection gap was identified by the National TB Prevalence Survey in 2014–2015. In Indonesia, overall TB case notifications rose by 69% from 331,703 in 2015 to 562,049 in 2019 but a wide case detection gap remains due to a combination of underdiagnosis and underreporting of people diagnosed with TB. Indonesia accounted for 10% of the estimated 2.9 million TB cases not detected or notified globally in 2019 [1]. In 2021, only 47% of the estimated TB cases in Indonesia were notified [2]. Within Indonesia, there are diverse determinants of TB epidemiology, and variation in performance of regional TB programs. Mimika district in Indonesia’s easternmost province of Papua is a multi-cultural mining district with an annual TB case notification rate of 818 per 100,000 population in 2019 [2]. Details of TB epidemiology and program performance in Mimika have been scant, although past reports have indicated a very high burden of disease [3, 4].

In 2017, in a collaboration between the Mimika District TB Program, University of Gadjah Mada, Timika Research Facility and Menzies School of Health Research (Australia), we commenced a TB health-systems strengthening project, the Stronger Health Systems for multidrug-resistant tuberculosis and malaria (STRATUM) initiative and a follow-up initiative called PRIME-TB (Papua New Guinea and Indonesia for the Micro Elimination of TB). The focus of this project was on prevention of TB through implementation and scale-up of screening household contacts of people with TB and use of TB preventive treatment (TPT) among young (<5 years) child contacts without TB. However, a broader set of health system strengthening activities were also implemented, including improvements to TB program information management systems, and use of a Continuous Quality Improvement (CQI) framework to identify gaps and opportunities for improving the clinical and public health management of TB [5] Mimika district has also received TB program support through United States Agency for International Development (USAID) funded initiatives, including promotion of active case finding and technical assistance to develop a district action plan for TB control [6].

Since early 2020, the COVID-19 pandemic has impacted the delivery of TB services. Staff and resources were diverted to the pandemic response and TB patients have avoided health care contact to limit their exposure to COVID-19. Utilization of primary and tertiary health care decreased dramatically [7] Indonesian national TB case notifications dropped to 393,323 in 2020, a 31% reduction compared to 2019 and less than half the estimated caseload [1]. In response to escalating COVID-19 cases in Mimika from March 2020 the local government implemented large-scale social restrictions [8]. Unintended impacts on access of the population to healthcare, including TB care, quickly become evident.

We aimed to evaluate the impact of strengthening of TB services on TB case notifications, diagnostic practices and treatment outcomes in this high-incidence setting of Indonesia, as well as the impact of the first two years of the COVID-19 pandemic on any progress made.

Methods

Interventions to improve TB case finding

To understand local contextual factors impacting TB program performance in addition to our own health system strengthening work [5], we sought information from the Mimika District Health Office about TB activities led by government and non-government organisations during the years of the study. These are shown in Table 1. In summary, in 2016 the Community Empowerment of People Against Tuberculosis (CEPAT) project was introduced to improve TB case finding, monitoring and evaluation, and TB surveillance [6]. This project recruited and trained local community health workers to do door-to-door TB symptom screening and to facilitate attendance of symptomatic individuals at a Primary Health Centre (PHC, known as Puskesmas) for evaluation and laboratory testing. In 2017, the TB Challenge project provided technical assistance to the district TB coordinator and staff in health facilities to optimize TB reporting, public-private mix in TB management, and develop a district action plan for TB control [6]. The CEPAT and TB Challenge projects ended in 2018 and 2019, respectively.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Health system activities introduced over time in Mimika District, 2014–2021.

https://doi.org/10.1371/journal.pgph.0001114.t001

From 2017, our research team implemented a multi-component intervention funded by the Australian government to initiate and strengthen household contact screening and management [5]. This project chiefly aimed to increase TB case finding among contacts and to provide TB preventive treatment (TPT) to young child contacts. Screening for co-morbidities was also supported with the distribution of blood glucose testing kits in 2018 to screen for diabetes and encouragement to offer HIV screening. Our project’s interventions have included group training, developing and providing educational materials, technical assistance and mentoring. Regular CQI meetings have been held to review progress and discuss achievements, barriers and challenges. All TB staff participate in the CQI meetings, raise their concerns and propose ideas to solve the identified problems.

A four-module GeneXpert (Cepheid, USA) machine was procured by the NTP for the district hospital in mid-2018 to detect Mycobacterium tuberculosis and rifampicin resistance (using ‘Xpert MTB/RIF’ cartridges) (Table 1). A second GeneXpert machine was installed in March 2020 at the Puskesmas with the highest TB caseload and three additional machines were installed in 2021, two at the district hospital and one at a PHC facility.

Results

Investigation of people with presumptive TB

Data on screening for TB were available for 2019–2021. In 2019, 7745 people were screened for TB. This represented three quarters of the annual target of 10,314 set by the national TB program [17]. The impact of COVID-19 is shown as an annotated plot in Fig 1 and S1 Table; there was a 38% decrease in numbers of people screened for TB to 4808 in 2020, attributable to lockdown, drops in presentations to clinics and a fall in contact screening activities to identify potential cases. This coincided with an increase in the proportion of Xpert assays that tested positive, from 14.8% of community members with presumptive TB in 2019 to 19.5% in 2020 and 20.5% in 2021.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Monthly number of TB suspects and newly diagnosed patients with TB before and during the COVID-19 pandemic.

https://doi.org/10.1371/journal.pgph.0001114.g001

TB notifications and characteristics of people with TB

During the eight-year study period, 11,803 TB cases were notified in Mimika District. Table 2 lists characteristics of TB cases by year. Overall, 69.1% were of Papuan ethnicity and 82.1% were from an urban setting, reflecting the general Mimika population [18]. The prevalence of co-morbidities among those tested were 10.5% with HIV infection and 5.4% with diabetes which compares to a prevalence of 2.4% and 1.1% in the general population respectively [18]. Overall, 22% of TB cases were EPTB, higher than the national average of 9% [1] and more common in children (27.3%), people living with HIV (26.2%) and Papuan people (24.7%).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Characteristics and treatment outcomes of TB cases notified in Mimika District, Papua, Indonesia, 2014–2021.

https://doi.org/10.1371/journal.pgph.0001114.t002

TB case detection and reporting increased annually until the onset of the COVID-19 pandemic in 2020 (Fig 2). Between 2015 and 2019 there was a 67% increase in notified TB cases (1078 to 1796 cases), achieving 94% of the estimated annual target case detection rate of 1910 cases for 2019 as set by the health authority. This was followed in 2020, coinciding with COVID-19 spread and lockdown measures, by a 18.7% decrease in case notification from 1796 in 2019 to 1461 in 2020, which then increased to 1716 in 2021.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Number of newly diagnosed patients with tuberculosis in Mimika district from 2014 to 2021.

https://doi.org/10.1371/journal.pgph.0001114.g002

Diagnosis and quality of care among people diagnosed with TB

The proportion of cases diagnosed in primary care increased over time from around a third to half of all cases (Fig 2). Although the proportion with bacteriological confirmation remained relatively constant at 35.8% overall, there was an 89% increase in the number of individuals with bacteriologically confirmed TB from 361 in 2015 to 682 in 2019 (Tables 3 and S1). All samples tested were sputum samples. The mode of diagnosis shifted from smear microscopy as the sole diagnostic option to Xpert being used in more than half the cases during the study period (Tables 3 and S1 and Fig 3). Bacteriological confirmation was uncommon in children (<15 years of age) but there was an increase each year in the proportion of children with laboratory confirmation from 1.5% in 2016 to 8.6% in 2020 (Table 3). This occurred as the proportion of children tested with Xpert increased; with diagnostic yield in children of 32.1% (62/193) positive on Xpert compared with 26.5% (56/211) for smear microscopy. Bacteriological confirmation in adults was 55.0% overall, again with Xpert having a higher diagnostic yield than smear microscopy.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3.

Smear microscopy and GeneXpert testing showing number of cases with testing performed (left y axis), and proportion of tests done which tested positive (right y axis).

https://doi.org/10.1371/journal.pgph.0001114.g003

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Bacteriological confirmation in people diagnosed with tuberculosis, 2014–2021.

https://doi.org/10.1371/journal.pgph.0001114.t003

Clinical diagnosis was made in 64.2% of total cases. Of 7,573 with a clinical diagnosis, 43.7% were adults with pulmonary TB, 23.7% were adults with EPTB and 32.6% were children. Of adults with clinically diagnosed pulmonary TB, only 4.7% had no microbiological testing undertaken; the rest (95.3% had a negative laboratory test (Xpert or smear). Among adults with presumptive pulmonary TB who seeks care at the hospital, 75.6% had chest x-ray. Clinically diagnosed TB cases were treated as per standard treatment regimens for drug-susceptible TB. The Indonesian pediatric scoring system was used to support clinical diagnosis in children without bacteriological confirmation with 53.3% recording a score and the majority (95.4%) of the scores recorded being 6 or more.

Xpert identified 107 cases of rifampicin-resistant (likely multidrug-resistant) TB (S1 Table). Prior to the introduction of Xpert into District facilities, only samples from TB cases in which there was a high likelihood of multidrug-resistant TB on clinical suspicion were sent for Xpert testing at the distant reference laboratory situated in the provincial capital. Once Xpert testing became available locally and indications for Xpert testing broadened, the prevalence of rifampicin resistance among TB cases detected by Xpert was 5–7% between 2018 and 2021. The proportion of TB patients tested for HIV and diabetes improved during the study period and mostly remained strong during the pandemic years (Fig 4).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Proportion of people with TB in whom HIV and diabetes were tested.

https://doi.org/10.1371/journal.pgph.0001114.g004

Treatment outcomes

Documentation of treatment outcomes improved after 2014 and treatment outcomes up until the 2020 cohort are listed in Table 2. Treatment success did not significantly change over time, ranging from 71.1% to 74.6% between 2015 and 2020 (χ² test for trend p = 0.21). A significant increase in the proportion of deaths was seen over time from 3% in 2015 to 6% in 2020 (p<0.001) and a fall in the proportion reported as lost to follow-up in recent years (Table 2). A rise in deaths of older individuals (>64 years) is noted in 2020 coinciding with COVID (S2 Table and S2 and S3 Figs). Treatment success rates were significantly higher in children (78.1%) compared with adults (69.2%), and in people treated in primary care (77.5%) compared with hospitals (67.1%) (p<0.001) (Table 4 and S1 Fig). One in five TB cases who initiated treatment in hospital were lost to follow-up, compared to one in ten whose treatment was initiated at a PHC (Table 2). People with TB/HIV co-infection had significantly lower treatment success (65.0%) than people without HIV (74.4%) (p<0.001) but ART status of those with TB/HIV was not routinely recorded. People with past TB had lower treatment success than those without (p = 0.001) (Table 4). Ethnicity was not associated with treatment outcome.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Tuberculosis treatment outcome by subgroups, Mimika district, 2014–2020.

https://doi.org/10.1371/journal.pgph.0001114.t004

Discussion

This study provides important insights into the epidemiology of TB in Mimika district, Papua province, revealing major improvements in TB case detection over an eight-year period. The 67% increase in TB case notifications exceeded population growth of approximately 35% over the same period [9]. Substantial downturns in TB detection occurred during the COVID-19 pandemic, but the TB program showed greater resilience than reported nationally: in 2020 compared with 2019, case notifications in Mimika district decreased by 18.7% which compared to a 30.9% decline in TB case notifications in Indonesia nationally [19]. Gains made in TB program quality indicators prior to the pandemic, such as proportion of patients accessing bacteriological TB diagnostics and HIV testing, as well as the proportion achieving treatment success, were largely maintained during the first two years of the pandemic. Health system strengthening activities underway in Mimika district may have contributed to these positive performance indicators. ‘TB CEPAT’, an active case finding initiative that conducted door-to-door symptom screening in 2016–2017, is likely to have contributed to increased TB notifications whilst our CQI projects (‘STRATUM’ and ‘PRIME-TB’) have provided a package of health system strengthening activities, with a strong focus on regularly engaging, training and motivating TB program staff. The Mimika district TB program has been recognised nationally for its successes, being awarded the Best TB Program in Papua Province in 2018 and 2019 and celebrated at a national World TB Day celebration in 2019 and 2021 [20–22].

Damaging impacts of the COVID-19 pandemic on TB program performance have been reported nationally and internationally [1, 23]. Indonesian national TB case notifications dropped to 393,323 in 2020, a 31% reduction compared to 2019 and less than half the estimated caseload. While case detection rates in Mimika dropped during the first year of the pandemic, the Mimika TB program successfully returned to close to its performance before the pandemic. The case detection rate dropped from 94% in 2019 to 76% in 2020, but recovered quickly and reached above 90% in 2021. During the same time, the case detection rate at province level only increased by 2.3%, and by 6.1% at national level [17], highlighting the stronger recovery in Mimika compared to other parts of Indonesia. The findings we report from Mimika suggest that investment in program strengthening activities, including health care provider education and CQI may mitigate external impacts on the TB program. Detailed evaluation of these interventions is underway. However, further improvements are needed with regards to screening for case detection. In Mimika district, reallocation of diagnostic equipment (GeneXpert machine) and human resources to COVID-19 management, including to support a National Sports Event week in October 2021, had demonstrable negative consequences for TB screening (Fig 2). This illustrates the importance of developing strategies to build surge capacity to protect routine service delivery while responding to health emergencies.

Our findings highlight that treatment outcomes and bacteriological confirmation are key areas for further improvement. An overall treatment success of 71.2% does not meet the target of 85% set by the Indonesian NTP or WHO [1, 24]. Furthermore, the treatment success rate did not improve from 2015 to 2020. The number of people with TB who are reported as lost to follow-up is high, especially among those receiving treatment through a hospital. Hospitals are often far from patient’s homes and have no outreach services, compared with primary care facilities which are better equipped to provide follow-up care in the community. In recognition of the need to decentralise care for treatment support, we worked with local healthcare providers to encourage treatment in primary care by implementing a referral process from hospitals to clinics. While the numbers of cases detected and treated at PHC level increased (Fig 1), the proportion of those lost to follow-up remained relatively unchanged until 2019 (Fig 5 and Table 2). A higher caseload provides additional strain on the health services to provide treatment support. Digital innovation to support active case finding, patient education, referrals, data recording and reporting is an urgent priority in settings where human resources are limited [25]. Specific challenges in Mimika district include a large and highly mobile mine worker population, and high language diversity among different ethnic groups. A large proportion of the population resident in Mimika are from other provinces in Indonesia and the high population mobility may also contribute to loss to follow-up and transfer out. More work is needed to address retention in care, such as through ensuring culturally appropriate approaches and outreach services. It is also noted that the proportion of deaths increased in 2018 and 2019. Specific reasons for this are not known, but notably, there were increases in MDR TB cases and in the proportion of TB cases being diagnosed in people aged 45–64 years. Deaths among people with TB increased further in 2020, mirroring WHO reporting on global TB-related deaths increasing by 7% during the pandemic [1]. Finally, poorer TB treatment outcomes were noted for people living with HIV including higher death rates but ART status of those with TB/HIV was not recorded highlighting the need for improved integrated care for those with comorbidities.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Treatment outcome among TB patients commenced on treatment, Mimika district, 2014–2020.

https://doi.org/10.1371/journal.pgph.0001114.g005

Bacteriological confirmation of TB was low at 35.8% compared with the national average in 2020 of 41% [17]. The only sample type reported in our study is sputum. No EPTB cases had bacteriological confirmation and hence any drug-resistant EPTB cases would have received incorrect treatment. Xpert is now increasingly used in preference to smear microscopy as per national guidelines and the use of Xpert increased during the study period. However, Xpert cartridges remain under-utilised while the analyser is required for COVID-19 testing and there was a lack of presumptive case finding. Only around 5000 of the annual allocation of more than 7000 cartridges distributed by the NTP to Mimika are being used annually. Xpert positivity among presumptive TB cases is high. Of all people tested with Xpert in 2020–2021, one in five tested positive which may suggest that not enough testing is being done (S3 Table) [26]. Programmatic planning to increase Xpert utilisation for TB testing is of particular importance while the competing need for COVID-19 testing with Xpert persists. Rifampicin resistance was 7.3% among those tested with Xpert in 2021, a concerning problem for Mimika given the high reliance on clinical diagnoses and empirical therapy with a standard drug-susceptible TB regimen. Xpert for EPTB diagnosis was unavailable in Mimika during this study due to lack of equipment (centrifuge and tubes), training or biosafety cabinet access for level 3 sample processing, and lack of finance from the NTP or health insurance to cover these costs.

Xpert clearly has high appeal given higher yield and lesser workload than microscopy and is now recommended in Indonesia for broad use. However, as utilisation of smear microscopy decreases and Xpert increases, there is a need to maintain microscopy expertise since this is still recommended for monitoring treatment response. By 2021, 27.7% of presumptive TB cases were diagnosed on microscopy and 63.5% using Xpert.

TB case finding in children is a marker of the quality of TB program performance. The proportion of TB cases in Mimika aged less than 15 years is consistently higher than national data. In 2020, 21.3% of TB cases in Mimika district were children aged 0–14 years, compared to only 9% nationally [2]. Infant BCG vaccination coverage in Mimika is low only 60.7% in 2020 and 64.5% in 2021. The high proportion of cases reported among children is similar to that reported from neighbouring Papua New Guinea (25% of cases aged less than 15 years), where TB is also highly endemic and where children make up a large proportion of the population [27]. Census data indicate that children comprise approximately 35% of the population in Mimika district compared to 23% in Indonesia overall. Only 5.6% of children were confirmed bacteriologically in this cohort highlighting the need for better TB diagnostics for children, especially young children with paucibacillary disease. Most (58%) of the bacteriologically confirmed TB in was diagnosed by Xpert and the proportion of bacteriologically confirmed child TB has increased since the introduction of Xpert, but remains low. Sputum induction or gastric aspirates for testing with Xpert can be done at the district hospital but these procedures require hospitalisation and are not routine for children with mild TB symptoms in the outpatient clinic. Furthermore, the WHO has recommended avoiding sputum induction to prevent airborne transmission during the COVID-19 pandemic. Recently updated WHO child TB guidelines (2022) recommend alternative samples for Xpert testing such as nasopharyngeal aspirates or stool [28]. The diagnosis of TB in children is usually clinical. The Indonesian pediatric TB scoring system is widely used and 52.1% of child TB patients in Mimika had a score ≥6, which indicates the need for TB treatment [17]. However, the scoring system includes tuberculin skin test results (largely unavailable at our study site) and chest X-ray also (limited in availability, only at hospitals requiring a referral from primary care) which limit its diagnostic value.

The main limitation of this study is incomplete data capture especially in the first year; however, we undertook stringent cross-checking to minimise missing data and validate entries, applied consistently across all years of the study. Ethnicity is not captured in the TB electronic records; some errors in assigning ethnicity could have occurred. We did not have access to directly comparable national TB datasets and therefore comparisons between program performance in Mimika district with other parts of Indonesia were only able to be descriptive. Compared to the other 28 districts in Papua, Mimika district ranked first in TB case finding in 2020 and 2021 [17]. While the population comprised 7.3% of the total population of Papua, it is estimated that about 10% of TB cases will be notified from Mimika district. However, in 2021, Mimika district was able to contribute to 18.4% of total TB case finding in Papua [17]. Strengths of this study include the highly comprehensive dataset from an under-reported, high burden setting, uniquely able to track program performance over a long period incorporating the unexpected disruption caused by COVID-19.

In summary, the implementation of several dedicated TB case finding and health system strengthening activities substantially improved TB case detection in Mimika District. Treatment success rate was sustained despite the increased burden that TB case numbers posed to the health services, but improvements in treatment outcome are needed. The analysis shows the negative impact of the COVID-19 pandemic on TB surveillance but there are already encouraging signs of recovery in case detection suggesting resilience in the TB services. There are clear areas for ongoing investment, including greater availability and uptake of rapid and accurate diagnostics, which international collaborative work in this study setting continues to address.

Supporting information

Acknowledgments

We thank the Mimika district TB program for access to district TB program data, Reynold Rizal Ubra, the head of Mimika District Health Office, for his consistent support to our research project, the Center for Tropical Medicine at the Gadjah Mada University for the technical assistance, and the National TB Program of Indonesia for access to national TB report. We would also like to thank our research assistants, Fransiska Nurul Rachmawati, Marwah Haji Gaffar, Christine Kaunang, Adrianus Ratu Nitit, and dr. Andrio Palayukan who have assisted in collecting and validating the data, as well as Christel van den Boogaard who were involved in designing the continuous quality improvement model for TB control program.