Original Research••

Tuberculous meningitis patient pathways and delays to diagnosis in Indonesia: a retrospective cohort study

Abstract

Introduction Delays in diagnosis and treatment contribute to high mortality of tuberculous meningitis (TBM). We studied TBM patient pathways including delays to diagnosis, and their alignment with available diagnostic services in Indonesia.

Methods We recruited patients admitted to two tertiary hospitals who started TBM treatment. Participants or their relatives were interviewed to recall healthcare visits preceding TBM treatment. We also surveyed available diagnostic capacity for TBM at hospitals that had been visited by at least two patients preceding their study enrolment.

Results Of 175 participants (median age 31 years, 57.1% men), 85.1% had reduced consciousness or coma, and 46.9% had motor deficits including hemiparesis. Patients attended a first healthcare provider, most often private clinics (38.3%) or informal healthcare providers (22.3%), at a median 14 days (IQR 1–34) after symptom onset. They visited multiple providers (median 5, IQR 3–8) over a prolonged time period (median 31 days, IQR 10–79) preceding TBM diagnosis. Of 40 surveyed hospitals, 52.5% could not or not always perform lumbar puncture, 22.5% lacked cerebral imaging facilities and 31.6% and 84.2%, respectively, could not provide routine microscopy or GeneXpert MTB/RIF on cerebrospinal fluid.

Conclusion In these urban settings in Indonesia, pathways to TBM diagnosis are complex and lengthy, and patients often visit healthcare providers with limited capacity to diagnose TBM. There is an urgent need for interventions to strengthen health literacy and diagnostic and referral processes in public and private health sectors for complex patient groups like TBM.

What is already known on this topic

  • In Indonesia, as in many other parts of the world, tuberculous meningitis (TBM) has a very high morbidity and mortality. Studies have mainly focused on clinical characteristics, diagnosis and antibiotic or immunomodulatory treatment, but not on patients’ access to healthcare.

What this study adds

  • Patients with TBM experience complex and lengthy pathways before TBM is diagnosed, mostly at an advanced stage. Diagnostic capacity for TBM and other brain infections is very limited outside top-referral hospitals.

How this study might affect research, practice or policy

  • In a country like Indonesia, public health interventions to improve patients’ access to high-quality healthcare services will be crucial for improving outcomes of TBM and other brain infections.

Introduction

Tuberculous meningitis (TBM) is the most devastating manifestation of tuberculosis (TB). It is one of the most common causes of brain infections, with an estimated 164 000 adult incident TBM cases globally in 2019.1 Based on this study and the Global TB report 2022, we estimate that there were at least 13 400 incident TBM cases in Indonesia in 2021, which is 8% of the global estimate. In a recent Indonesian cohort among 793 patients with central nervous system infections, 44% were caused by Mycobacterium tuberculosis (Maharani, unpublished). TBM is fatal if untreated, but even with treatment, patients have an estimated mortality of 23–25% globally and up to 67% in studies from Asia,2 with at least one-third of survivors suffering from neurological sequelae.2

Efforts to improve outcomes of patients with TBM have mainly focused on improving sensitivity of diagnostic tests and optimising antimicrobial or adjuvant therapy. However, few studies have addressed whether patients with suspected TBM have access to high-quality diagnostic and therapeutic services in a timely manner, and what barriers they may experience in this regard. Delays in TBM diagnosis and treatment have been reported from different countries, and have shown strong associations with mortality.3 Accelerating diagnosis and treatment is therefore a key objective in improving patient outcomes for TBM. To do so, more insight is needed into ‘patient pathways’, concerning where and when patients with TBM seek care prior to diagnosis, and whether the right diagnostic services are available, and in such a way that patients’ needs and preferences are met.4 For pulmonary TB, previous studies have described patient pathways and their alignment with diagnostic services.5 6 However, patient pathways for TBM are likely different as TBM is more rare, more rapidly progressive and severe at later stages, but most importantly much more difficult to diagnose, due to its initial non-specific symptoms, and the need for lumbar puncture, cerebral imaging and more advanced microbiological tests, none of which are sufficiently sensitive.7 Better understanding of individual-level TBM patient pathways could therefore help to identify barriers to diagnosis and improve allocation of resources to areas where they are needed most.

Therefore, the aim of this study was first to characterise TBM patients’ pathways, including delays, prior to hospital admission in two tertiary hospitals in Indonesia, which has the second highest number of TB cases worldwide, and a broad range of health system challenges, including a complex network of public and private healthcare providers. Second, for the first time we aligned patient trajectories with the availability of TBM-related diagnostic services using facility-level data. Findings in this study about TBM may share features with patient pathways for other severe diseases in high-burden settings, and could provide insights into how to improve efficiencies in the delivery of wider healthcare services.

Methods

Study setting

This retrospective cohort study was conducted from January 2020 to May 2022 in Cipto Mangunkusumo hospital (RSCM) in Jakarta and in Hasan Sadikin hospital (RSHS) in Bandung. Both hospitals are government-owned university hospitals and serve as sole tertiary referral centres for brain infections in Jakarta (~11 million inhabitants) and West Java province (~50 million inhabitants). Indonesia’s public healthcare system is decentralised, and the district level government has the main responsibility for provision and management of healthcare service in city hospitals and primary care centres (pusat kesehatan masyarakat or ‘Puskesmas’). Also, the private sector plays an important role in healthcare provision in Indonesia, with ~63% of hospitals managed by the private sector in 2021,8 and a large number of private primary care providers. Many Indonesian patients seek care in private sector facilities or at informal care providers, such as drug stores, traditional spiritual healers or alternative practitioners.6 9 To achieve universal health coverage, the government launched Jaminan Kesehatan Nasional in 2014, Indonesia’s mandatory social health insurance scheme. By 2021 it covered 86% of the population, and included almost all public hospitals, the majority of Puskesmas and private hospitals, but only a limited number of private clinics.8

Study population

Patients were included in the study if they were ≥18 years old, admitted to the emergency room or neurological ward in one of the two study hospitals, and started on TBM treatment for a presumed diagnosis of TBM. Patients were excluded if they (or their relative for patients that were not fully conscious or had cognitive impairment) did not provide written informed consent, if they returned to the hospital with sequelae of previous episodes of TBM (e.g. seizure) without current TBM, if the patients or their relatives could not be contacted, if patients died or went home before recruitment, or if they were discharged with a diagnosis other than TBM.

Procedures and definitions

Patients were interviewed in their native language using a structured questionnaire (online supplemental material 1). The questionnaire was developed based on the ‘Tool to estimate patients’ cost’ from the Dutch Tuberculosis Foundation (KNCV TB Foundation), and adapted to TBM and the Indonesian context through rounds of discussion with experts and through pilot testing. If patients were not fully conscious or had cognitive impairment, close relatives accompanying the patients in the hospital and familiar with their medical history were interviewed instead. The interviewee was asked to recall in chronological order the dates of onset of symptoms and subsequent visits to healthcare providers until inclusion in the tertiary study hospitals. If patients could not remember exact dates, they were asked to relate their experiences to memorable events, such as Islamic Eid al-Fitr. Subsequently, the first day of the week, month or year that the interviewee was able to recall was recorded.

Symptoms were categorised as neurological and general. Neurological symptoms included altered consciousness, headache, behavioural changes, vomiting (often occurring as a result of raised intracranial pressure), motor and sensory abnormalities, cranial nerve palsy, seizures and other symptoms mentioned by patients, such as visual impairment or vertigo. General symptoms included fever, lethargy, weight loss, night sweats, cough, haemoptysis and other symptoms mentioned by patients, such as gastrointestinal symptoms and lumps in the neck.

Subsequently, for each healthcare visit, the interviewee was asked to recall the type of healthcare provider, reasons for visiting that provider, tests performed, potential diagnoses made and the treatment provided. Provider types were defined based on Hanson et al,4 and adapted to the Indonesian context (table 1).9

Table 1
|
Definitions of healthcare providers and their diagnostic and treatment capacity

Dates of hospital admission, diagnosis and treatment initiation were verified using patients’ medical records. Date of diagnosis was defined as the day of presumptive or confirmed diagnosis of TBM (whichever came first) by an experienced team of senior neurologists in both sites, based on clinical assessment, cerebrospinal fluid analysis and/or cerebral imaging. As Indonesian guidelines recommend identical treatment regimens for pulmonary and meningitis TB, date of treatment initiation was defined as the day in which anti-TB drugs for either pulmonary TB or TBM were initiated. Treatment initiated for pulmonary TB but interrupted or completed before TBM diagnosis was considered ‘history of TB treatment’, and not considered in the date of TBM treatment initiation. If patients explicitly reported that doctors mentioned a TBM diagnosis and/or started anti-TB treatment during a previous visit in their trajectory, this was considered the date of diagnosis or treatment, respectively, and was checked in the medical records.

Time delays were summarised based on World Health Organization (WHO) definitions10: patient delay was defined as the number of days between onset of symptoms and the patient’s first visit to any healthcare provider; diagnostic delay was defined as the number of days from symptom onset to TBM diagnosis; and treatment delay as the number of days between TBM diagnosis and initiation of anti-TB drugs.

To evaluate how patient pathways align with diagnostic services in the healthcare system, hospitals that were visited by two or more patients prior to their study enrolment were assessed. Through an online survey or a phone interview, we asked neurologists employed in these hospitals to provide information about availability of TBM-related diagnostic resources, that is, the number of neurologists, lumbar puncture equipment, cerebral imaging and laboratory analyses on cerebrospinal fluid (online supplemental material 2). For hospitals without a neurologist, we consulted the hospital’s website or other physicians or administrators employed in the concerning hospital. Information about the hospital’s type and size was retrieved from the online hospital information dashboard.11

Analysis

We performed descriptive analyses to summarise data on patient pathways and delays. Furthermore, the type of providers visited by patients over sequential healthcare visits are presented in a bar graph. We also visualised the type of providers visited by patients over time since symptom onset in a bar graph. If patients visited multiple healthcare providers in a single time frame, we selected the highest-level provider for that time frame according to table 1. The association between HIV status and diagnostic delay was evaluated using the Mann-Whitney U test, and the effect of diagnostic delay on the British Medical Research Council (MRC) grade at recruitment was assessed with ordinal logistic regression.

Ethical considerations

Patients or their family members, if the patient was not fully conscious or had cognitive impairment, provided written informed consent prior to study enrolment.

Patient and public involvement

To our knowledge, this is one of very few studies within the TBM field that is guided by patient’s perspectives and that highlights their experiences and challenges. During pilot-testing of the questionnaire, participants provided feedback about the burden and time to participate.

Results

From January 2020 to May 2022, 175 patients with TBM were included; 99 from the study hospital in Jakarta and 76 from Bandung (online supplemental material 3). A close relative accompanying the patient was interviewed for 114 (65.1%) patients with reduced consciousness.

Patients were generally young, with a median age of 31 years (IQR 24–46), and 57.1% were man (table 2). Furthermore, 58.3% had other forms of TB besides TBM at the time of admission, mostly pulmonary TB. Patients presented with severe signs and symptoms, including reduced consciousness (85.1%), motor abnormalities (46.9%) and a history of seizures (25.7%), as well as frequent headache (84.0%), fatigue (78.9%) and fever (74.9%). Most patients (91.4%) self-reported that this resulted in difficulty with their daily activities. Most patients were diagnosed with advanced disease, TBM grade II (79.1%) or III (16.9%).

Table 2
|
Patient characteristics

Patient pathways and delays

Patient pathways to diagnosis were characterised by numerous healthcare visits to different providers and long delays. The median time between symptom onset and patients’ first encounter with a healthcare provider, that is, patient delay, was 14 days (IQR 1–34). Patients mostly visited private clinics (34.7%), informal care providers (21.2%) and community health centres (12.4%) after onset of general symptoms, mainly because of proximity to their home or easy accessibility (36.6%), and familiarity with the provider (14.9%). After neurological symptoms, patients mostly initially visited private clinics (30.3%), tertiary referral hospitals (17.1%) and informal care providers (16.6%). Patients had a median of 5 healthcare visits (IQR 3–8), with a range of 1–24 visits, preceding TBM diagnosis (figure 1A). The time from first healthcare visit to TBM diagnosis was median 31 days (IQR 10–79). The median diagnostic delay, that is, time from symptom onset to TBM diagnosis, was 66 days (IQR 31–128) (figure 1B); 74 days (IQR 32–172) in Jakarta and 61 days (IQR 29–97) in Bandung, possibly due to difference in case mix. For those with other forms of TB (58.3%), time from general symptoms to neurological symptoms was 20 days (IQR 0–74) and time to TBM diagnosis was 65 days (IQR 31–128). HIV positive patients did not have significantly different median patient delay compared with HIV negative patients (23 days (IQR 6–39) vs 14 days (IQR 1–33), p=0.12), which was similar for time from first healthcare visit to TBM diagnosis (32 days (IQR 14–89) vs 31 days (IQR 9–73), p=0.45). With every 10 days increase in diagnostic delay, the likelihood of a higher MRC grade at recruitment increased with 3% (OR 1.03, 95% CI 1.00 to 1.05). Finally, the median treatment delay was 1 day (IQR 0–1).

Figure 1
Request permissionrequest permission icon
Figure 1

Type of healthcare providers and visits made before TBM diagnosis. The types of healthcare providers visited prior to diagnosis, with the cumulative proportion diagnosed with TBM in green. In (A) the x-axis indicates the number of sequential healthcare visits leading up to a diagnosis, and the y-axis indicates the proportion of patients with TBM that visit different types of healthcare providers in the concerning visit. In (B) the x-axis indicates the time after onset of first symptoms, regardless of general or neurological nature. The different colours distributed over the y-axis indicate the highest level of provider visited by patients during that time frame, according to the order in table 1. = has not sought care yet, = has visited a healthcare provider in a previous time frame, but not in this time frame, = visited an informal provider but not diagnosed yet, = visited a private clinic but not diagnosed yet, = visited a community health centre but not diagnosed yet, = visited a private secondary level hospital but not diagnosed yet, = visited a public secondary level hospital but not diagnosed yet, = visited a public tertiary level hospital but not diagnosed yet, = diagnosed in a private secondary level hospital, = diagnosed in a public secondary level hospital, = diagnosed in a public tertiary level hospital. TBM, tuberculous meningitis.

Patients’ pathways were variable, but many patients had a similar sequence of events (figure 2). Most patients developed neurological symptoms about 2 weeks after onset of general symptoms, without getting a diagnosis of another form of TB prior to TBM diagnosis (figure 2, pathway 1). Many of them (66.7%, n=78) only started seeking care after onset of neurological symptoms, with a median of 5 visits (IQR 2–6) over 23 days (IQR 7–60) until a TBM diagnosis was made. They most frequently visited private clinics (37.6%) and informal care providers (23.1%) after general symptoms, and private clinics (26.5%) and tertiary referral hospitals (20.5%) after neurological symptoms. In an alternative scenario, some patients recorded first having neurological symptoms (figure 2, pathway 2), followed by more general symptoms after a median of 38 days (IQR 29–90) for 13 patients (72.2%). Patients with this scenario needed a median of 6 visits (IQR 4–7) over 58 days (IQR 15–96) prior to TBM diagnosis. They most often visited private clinics (38.9%), informal care providers (16.7%) and community health centres (16.7%) after neurological symptoms, and private hospitals (61.5%) and public secondary hospitals (15.4%) after subsequent general symptoms. Finally, some patients (22.9%, n=40) were first diagnosed with another form of TB, a median of 45 days (IQR 27–145) after onset of general symptoms and 31 days (IQR 15–51) prior to TBM diagnosis (figure 2, pathway 3–8). Of those, 23 started TB treatment a median of 24 days (IQR 15–42) prior to TBM diagnosis (figure 2, pathway 3, 5, 7). There was no clear relation between particular pathway groups and disease severity at time of recruitment. In the first group, 4.3% had MRC grade 1 at recruitment, 76.7% grade 2 and 19.0% grade 3, and for the second group 0.0%, 83.3% and 16.7%, respectively. Individual patient pathways are presented in online supplemental material 4.

Figure 2
Request permissionrequest permission icon
Figure 2

Sequence of events for subgroups of patients with TBM. The different orders of events for subgroups of patients with TBM with regards to symptoms, diagnosis of TBM and other forms of TB, as well as TB treatment. TB regimens that were initiated because of other forms of TB, but that were discontinued prior to TBM diagnosis are not considered in this graph as TB treatment. TB, tuberculosis; TBM, tuberculous meningitis.

Barriers to diagnosis

From patients’ narratives of the four most common pathways, some challenges that caused diagnostic delays were identified (panel 1). For instance, involvement of informal and private care providers likely led to significant delays. This is shown by the fact that median time from first healthcare visit to TBM diagnosis was significantly longer for those who initially visited informal or private providers compared with those who initially visited public community health centres and public secondary level hospitals (39 days (IQR 14–89) vs 24 days (IQR 10–58), p=0.04). Moreover, doctors often suspected alternative diagnoses, such as typhoid fever (n=30), gastritis (n=17), ‘regular headache’ (n=14) or stroke (n=14). In six patients, traditional or spiritual healers falsely diagnosed them as having a curse or disturbance of spirits.

Panel 1 Patients’ narratives of their pathways until tuberculous meningitis (TBM) diagnosis

Pathway 1: involvement of midwife and cleric

A 19-year-old pregnant woman with no comorbidities and no history of tuberculosis (TB) developed a headache, dizziness and lethargy. Four times she went for regular check-ups to her midwife, who attributed her symptoms to pregnancy. After 2.5 months, the midwife referred the patient to a private hospital, where she was admitted and diagnosed with a urinary tract infection. Two weeks later, the patient had fever, vomiting, altered consciousness and behavioural changes for which she visited a religious healer twice, because she thought her symptoms were due to supernatural causes. The religious healer diagnosed the patient as having a ‘trance’. Two weeks later, the patient visited her midwife again for a check-up, who now referred her to a different private hospital, leading to a diagnosis of typhoid fever and antibiotic treatment. Eventually, 3.5 months after onset of initial symptoms, the patient developed motor abnormalities and cranial nerve palsy and presented to the emergency room of the tertiary referral hospital the next day, where she was diagnosed with TBM and commenced on treatment.

Pathway 2: prolonged patient delay and ophthalmology

A 31-year-old HIV-positive man without a history of TB initially developed visual impairment. Around a month later, the patient bought eye drops at a pharmacy, but after a while he also had a persistent headache. Three months after his initial symptoms the patient visited a general practitioner at a private clinic, who referred him to an ear-nose-throat specialist in a private hospital. A diagnosis of sinusitis was made, and the patient was treated with antibiotics. The following 2 weeks, the patient twice visited a private ophthalmologist, who diagnosed the patient with optic nerve damage. A few days later, the patient also started vomiting and became lethargic, and visited two private hospitals that suspected the patient of glaucoma and papilloedema. Again a week later, the patient had reduced consciousness, behavioural changes and cranial nerve palsy. A general government hospital referred him immediately to an ophthalmologist in the tertiary referral hospital, who suspected a brain tumour and referred the patient to the neurology department, where he was diagnosed with TBM and commenced on treatment.

Pathway 3: TBM despite ongoing pulmonary TB treatment

A 24-year-old woman without any comorbidities or history of TB developed a fever, cough and night sweats. In the following 2 weeks, she twice visited a pharmacy to buy vitamins. About 2.5 months later, she also developed a headache, vomiting and a seizure, so she twice visited a private general practitioner, who diagnosed her with dyspepsia and dehydration. The general practitioner at the private clinic referred her to a private hospital where the patient was admitted, diagnosed with pulmonary TB and treated for 14 days. However, in the following week, despite TB treatment, she developed altered consciousness, behavioural changes, cranial nerve palsy and motor abnormalities. She again visited the private hospital, which referred her to the emergency ward of the tertiary referral hospital, where she was diagnosed with TBM and started on optimal treatment.

Pathway 4: dyspepsia and discontinued TB treatment

A 24-year-old man without comorbidities or a history of TB developed weight loss, cough and a headache. In 2 months the patient visited four drug stores and three different private clinics, where his symptoms were attributed to typhoid fever and dyspepsia. Two months after onset of the initial symptoms, the patient also developed lethargy, fever and behavioural changes. A primary healthcare clinic referred him to a secondary government hospital and subsequently a tertiary hospital for diagnosis. There, the patient was admitted, diagnosed with pulmonary TB and treated for 7 days. The patient discontinued his treatment, and twice visited a private clinic where dyspepsia was suspected and treated accordingly. Around 1.5 month later, the patient developed altered consciousness, haemoptysis and vomiting, for which he visited a secondary public hospital. The doctors suspected TBM, and referred him to the tertiary referral hospital, where this was confirmed, and appropriate treatment was started immediately.

Alignment of patient pathways with availability of TBM diagnostic services

Of 49 hospitals that were visited by two or more patients prior to enrolment, 40 (81.6%) provided information about availability of TBM-related diagnostic capacity and services (online supplemental material 5). Even though most hospitals (92.5%) had a neurologist, 52.5% did not always have access to supplies for lumbar puncture, and 22.5% lacked cerebral imaging. Finally, many had no facilities for routine cerebrospinal fluid analysis (31.6%), cerebrospinal fluid smear microscopy (47.4%), GeneXpert MTB/RIF (84.2%), or TB culture (65.8%).

Discussion

In this study, we found that patients, almost all of whom were young, had experienced complex, variable and lengthy pathways before TBM was diagnosed, mostly at an advanced stage. Moreover, patients often visited healthcare providers with limited capacity to diagnose TBM. It took on average two weeks for patients to seek care for their symptoms, and on average five healthcare visits over about a month’s time until a diagnosis of TBM was made. Especially in the early stages of the diagnostic pathway, patients visited informal and private healthcare providers, which typically lack the necessary diagnostic capacity for TB or for brain infections. Similarly, hospitals visited by patients with TBM often lacked capacity for diagnosis of brain infections, such as cerebral imaging, lumbar puncture facilities and cerebrospinal fluid analysis.

There are several possible reasons for delay in TBM patients’ pathways to diagnosis. First, patients may not (immediately) seek care after onset of symptoms. Patient delays reported in our study were shorter than reported for pulmonary TB in the same setting in Indonesia,6 likely because TBM symptoms can be more severe. Comparable to our study, studies from Taiwan and China reported a median time of 1–4 weeks from onset of TBM symptoms to presentation to a tertiary hospital.12–14 Care-seeking of patients with TB may be hindered by patient-related barriers, such as limited knowledge about TB, gradual onset of symptoms, stigma associated with the disease, attribution of symptoms to other causes (e.g. HIV), negative perceptions and attitudes towards healthcare providers, and health-system related factors such as geographical and financial barriers to access care.15–17

Second, patients who do seek care for their symptoms, may not go to caregivers or facilities that have the clinical knowledge, experience and capacity to diagnose and treat TBM. In our study, the majority of patients with TBM first went to an informal provider or private clinic, similar to pulmonary patients with TB in Indonesia.5 6 9 These entry-points have a key role in timely diagnostic-workup and efficient referral but are also characterised by provider-related barriers. For instance, in high TB burden counties like Indonesia and India, informal and private providers, especially in outpatient primary care but also in high-end specialised hospitals, are known to largely lack or underutilize diagnostic services for TB, to not treat TB according to local guidelines, to poorly refer to national TB programmes and to have large gaps between knowledge and practice.18

Third, those who eventually reach higher level, specialised hospitals still experience complex and lengthy pathways to diagnosis and treatment. Studies from China, India, South Africa and Taiwan have reported delay between 9 days to 6 weeks from TBM symptom onset to treatment initiation.12 19–21 Furthermore, studies from China, India and South Africa have reported a median 2.4 to 4 healthcare visits preceding TBM diagnosis,19 21 22 and a mean 1.7 hospitalisations and 4.7 outpatient clinic visits prior to treatment initiation.14 Health system delays in TBM are clearly universal, and caused by challenges both from demand side, that is, the patient, and from the supply side, that is, the health system. Health-system barriers to pulmonary TB diagnosis and treatment besides previously mentioned factors are for instance scarcity of trained doctors, lack of laboratory facilities, shortages in TB medication and catastrophic household costs for care.15 16 For TBM these are further complicated by atypical symptom presentation, low sensitivity of diagnostic tests, hesitancy among clinicians to perform lumbar puncture, limited availability of diagnostic resources and inadequate TBM treatment regimens.7 17 23 Limited availability of diagnostic resources in hospitals reported in our study was consistent with reports from a recent large survey, where 73.3% of African inhabitants had access to routine lumbar puncture services, but often only in teaching hospitals and not regional or local hospitals.24

Although Indonesia has made significant progress in moving towards universal health coverage,25 the findings of this paper call for further strengthening of care cascades for patients with complex diseases like TBM. This requires further investments in all levels of the healthcare system, including both the public and the private sector, to achieve strong patient-centred care with universally accessible, high-quality and equitable diagnostic and therapeutic health services. Although further research should be aimed at systematically identifying and addressing modifiable factors related to patients, healthcare providers and the wider health system, we can make some preliminary recommendations based on the findings of this study. First, community-based interventions could be aimed at improving health literacy. Health literacy interventions have been shown to promote awareness and knowledge about particular health-related issues, and to enhance patients’ ability to find, understand and use information and healthcare services,26 which could reduce patient delays.27 Second, at the provider-level, interventions could focus on encouraging and facilitating prompt use of lumbar puncture,23 28 which has been associated with reduced mortality in acute bacterial meningitis,28 and better triage of TB and brain infections at lower level public and private healthcare facilities,29 including timely notification and referral. Finally, health-system wide efforts should be aimed at further engaging the private sector in provision of TB care,29 30 decentralising TB diagnostic services29 and increasing efficiency of referral processes. Some of the issues raised in this paper are likely to be similar across different settings, such as mild symptom onset or difficulty with recognising symptoms as TBM. However, the structure and organisation of the health system, resource capacity and sociocultural context, that also greatly influence patient pathways, can vary substantially between settings.

This is the largest study to examine diagnostic pathways for TBM. In one of the most populous parts of Indonesia and for the first time, we also aligned TBM pathways with diagnostic services of 40 hospitals visited by patients prior to diagnosis. There are some study limitations. First, because of the study design we could not include patients who did not access a tertiary hospital or who were never diagnosed, including those who died prior to diagnosis. This may have resulted in recruitment bias towards more severe presentation of TBM, although severe TBM may also be present but under-recognised, underdiagnosed or die at other facilities (such as informal or private healthcare providers). In our study, those with longer diagnostic delays had more severe disease at time of recruitment. However, we do not have data about disease severity or course during the patient pathway and its effect on recognition of TBM by health providers, or on delays. Those with mild presentation may be harder to recognise as possible central nervous system infection, and lumbar puncture is performed infrequently23 especially for those with mild disease. Patients experiencing long delays and complex diagnostic pathways in our study may serve as a proxy and could help to understand pathways of those who are not appropriately diagnosed with TBM. Second, diagnostic pathways in this study were assessed using patients’ and families’ recall, noting that it is difficult to determine the exact time of onset of TBM. TB and TBM symptoms can have an insidious start, which can form difficulty in exactly remembering dates of symptom onset. Although we asked patients to relate their experiences to memorable events, some recall bias likely exists. Finally, healthcare may have been affected by surges of COVID-19 infections during the study period. Hospitals were frequently overburdened, with limited human and material resources, and TB symptoms may have been attributed to COVID-19, which may have caused further delays.31

Conclusion

In conclusion, pathways of patients with TBM in this high-burden setting in Indonesia are complex and lengthy and patients often visit healthcare providers with limited diagnostic capacity for TBM. This highlights the need to look beyond improving diagnostic tests and therapeutic strategies for TBM, and to improve care continuity for complex diseases like brain infections, for instance, by improving health literacy in the community, further decentralisation of diagnostic services and strengthening of referral processes including guidelines and insurance regulations, in both the public and private sector. This should shorten patient’s journeys and increase the proportion of patients receiving appropriate treatment in time, and surviving this most deadly manifestation of TB.

Overview

Article metrics
Altmetric data not available for this article.
Dimensionsopen-url