Introduction

Tuberculosis (TB), an infectious disease derived from various species of the genus Mycobacteria, remains a global threat to public health [1]. Although preventable and treatable, with 1.5 million deaths every year, TB is still on the world’s top list of infectious killer diseases [1]. Furthermore, with the advent of multidrug-resistant tuberculosis (MDR-TB), first-line drugs have been rendered ineffective [2,3]. An estimated 10 million active TB cases were recorded in 2020, which consisted of 5.6 million men, 3.3 million women, and 1.1 million children [4]. More than 95% of TB deaths occurred in low- and middle-income countries (LMICs), with 8 LMICs accounting for two-thirds of the total cases, with India having the highest number of cases, followed by Indonesia, China, the Philippines, Pakistan, Nigeria, Bangladesh, and South Africa [1].

As the global population ages, LMICs will face an increasing brunt derived from noncommunicable diseases (NCDs). As defined by the World Health Organization (WHO), NCDs are chronic diseases lasting for a long duration and are a result of a combination of genetic, physiological, environmental, and behavioural factors [5]. NCDs, such as diabetes mellitus (DM), cardiovascular disease (CVD), chronic lung disease, cancer, and mental health illness (MHI) among others, are increasing in prevalence [6]. Each year, 41 million people die from NCDs, with 85% of these deaths taking place in LMICs [5]. In particular, CVDs account for 17.9 million deaths, followed by cancer at 9.3 million and DM at 1.5 million [5].

Global population health data posit that progress towards meeting TB control targets, including the expansion of TB intervention programmes, is hindered by the synergistic relationship between TB and NCDs [7,8]. This is strongly attributed to NCDs and their associated risk factors such as CVD, DM, and smoking and alcohol abuse being risk factors for TB, consequently impacting the burden of TB at the population level [7,9].

At the societal level, the TB and NCDs share numerous underlying social determinants, such as high levels of poverty, poor health literacy, social and cultural values, and lack of social protection [10–13]. Additionally, coexisting communicable and NCDs increases the risk or effect of each other [14]. However, the current healthcare systems of LMICs might not have sufficient resources to respond to this combined burden of diseases derived from communicable and NCDs [15]. At the same time, disease-specific interventions are not the most effective approach for tackling these afflictions [16,17].

A known example is DM, which significantly increases the risk of TB and vice versa. The tight interplay between disease conditions at the biological level complicates the management of TB and contributes to poor TB treatment outcomes [18]. Concomitantly, TB can increase the risk of or aggravate DM status [19]. Other known linkages between TB and NCD typologies include that of CVD and TB [20], chronic lung disease and TB [21], MHI and TB [22,23], and cancer and TB [24]. These mutual predispositions necessitate better integration of health services to place TB and NCDs under control in LMICs after accounting for specific disease burdens within each country context.

Therefore, there are multiple ways in which service integration can be manifested. Integration covers disease-specific (vertical) programmes and system-wide (horizontal) pillars and initiatives, overlaying intersecting disease services (e.g., DM and TB treatment services), across community and tertiary level interventions, and between health and nonhealth sectors. This review focuses on integrating service delivery for diseases that are usually delivered separately but often affect the same types of end users. We also acknowledge that service integration goes across and beyond the status quo of a single healthcare entity or stakeholder and have therefore not only consolidated information on the different models of service integration but also provide practical steps (incorporating facilitators and avoiding barriers) when implementing TB and NCD integrated programmes.

Despite the increasing burden of disease derived from TB and NCDs in LMICs and the need for TB and NCD integrated programmes, most systematic reviews on TB service integration investigate the integration of TB- and HIV-related services or the health outcomes of TB and NCD comorbidities, their syndemic interactions, and prevalence [25–34]. Therefore, there remains a dearth of information regarding the mechanism of integration of TB and NCDs related services in LMICs, which this systematic review aims to address. Hereafter, this systematic review will present integration models that bring together the different services for TB and NCDs; points of service delivery; process modifications; and change management strategies through a quantitative and qualitative lens.

Methods

A protocol for this study has been registered on the PROSPERO international prospective register of systematic reviews (registration number: PROSPERO CRD42020202745). This study is reported as per the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (S1 and S2 Checklists) and illustrated in a flowchart below.

First, the findings from the literature review will be summarised based on its depth of integration, from least to most integrated, with each being delved into detail. A framework that illustrates the models of service integration will be developed from this step. Second, using our organised findings from the first step, WHO’s 6 operational “building blocks” of a health system, a well-established framework that identifies health system facilitators and constraints will be used to understand the facilitators and barriers for each integration model [35]. This step involves mapping various integration characteristics derived from the literature review to each dimension of the framework.

Drawing on the definitions proposed by Atun and colleagues [36], Dudley and Garner [16], and Legido-Quigley and colleagues [37], we operationally define programmatic integration as managerial or operational reorganising of health systems components to consolidate inputs, delivery, management, and organisation of particular service functions as a means to improve coverage, access, quality, acceptability, and (cost)-effectiveness [38]. For the purposes of this review, we consider the various tenets of integration that encompass the integration of service provision for both TB- and NCD-related services in terms of physical point-of-care delivery and types of services that include screening, testing or referral, and nonmedically related services like financial and social support.

Results

Overview of included studies

The screening process of this review adapted the PRISMA flowchart (Fig 1). Our database search yielded 4,906 articles. After the removal of duplicates, 3,484 articles were screened for titles and abstracts. A total of 184 full articles on integration of care of TB and NCDs were further assessed for eligibility, and 22 articles met the inclusion criteria for this review.

Download:

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

Fig 1. PRISMA 2020 flow diagram of the study selection process.

LMIC, low- and middle-income country; NCD, noncommunicable disease; PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-Analyses; TB, tuberculosis.

https://doi.org/10.1371/journal.pmed.1003899.g001

Characteristics of included studies

Of the 22 articles included, 13 were cross-sectional studies, 3 were cohort studies, 1 was a case–control study, 1 prospective interventional study, and 4 were mixed methods studies. A brief summary of the key characteristics of each study is provided in Table 1 below. Seven studies had a low risk of bias, 11 studies had a moderate risk of bias, 3 studies had a high risk of bias, and 1 study did not have enough data for risk bias analysis (elaborated in Tables 2–6).

Download:

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

Table 1. Description of the studies included.

https://doi.org/10.1371/journal.pmed.1003899.t001

Download:

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

Table 2. Summary of level 1 integration, entry via TB.

https://doi.org/10.1371/journal.pmed.1003899.t002

Download:

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

Table 3. Summary of level 2 integration, entry via TB.

https://doi.org/10.1371/journal.pmed.1003899.t003

Download:

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

Table 4. Summary of level 2 integration, entry via NCDs.

https://doi.org/10.1371/journal.pmed.1003899.t004

Download:

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

Table 5. Summary of level 2 integration, entry via TB and NCDs.

https://doi.org/10.1371/journal.pmed.1003899.t005

Download:

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

Table 6. Summary of level 3 integration, entry via TB.

https://doi.org/10.1371/journal.pmed.1003899.t006

Seven studies were conducted in Asia, 11 in Africa, and 4 in the Americas (Fig 2). Studies conducted in Africa and Asia focused more on the integration of mental health care in TB patients, whereas studies from South America explored the integration of services for NCDs such as DM and hypertension in TB patients.

Download:

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

Fig 2. Map of study sites.

Map shows the countries and the respective number of studies conducted in each country. The base layer of the map is taken from http://www.openstreetmap.org.

https://doi.org/10.1371/journal.pmed.1003899.g002

Synthesised model of integration based on literature review

A new framework of integrated care for TB and NCDs was developed where the depth of integration is stratified according to distinct levels 1, 2, and 3 in order to distinguish the integration of an additional service from the ones that are preexisting and/or stand-alone. The higher the numerical value, the deeper the level of service integration. Models of integration ranged from the type of services provided and point of entry into the healthcare system via TB or NCDs or both. We identified 7 models of integrated care and illustrated them in the framework below (Fig 3).

Download:

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

Fig 3. Framework for integration of TB and NCDs services.

Illustrated summary of different models of integration using either TB or NCD as points of entry into the health system. Numbers in brackets represent the number of studies reporting the respective model of integration. NCD, noncommunicable disease; TB, tuberculosis.

https://doi.org/10.1371/journal.pmed.1003899.g003

At level 1, the least integrated model, (i) patients entering the system via TB service are offered the additional service for screening for NCDs, and likewise (ii) the patients entering via NCDs service are screened for TB.

At level 2, moderate level of integration, (iii) patients entering from TB services are screened for NCDs and patients with potential or positive diagnosis are referred for care at appropriate healthcare facilities, and likewise (iv) patients entering the system from NCD services are screened and referred for TB treatment. Another model is also identified at this level, where (v) patients enter the system via both TB and NCDs services and are referred for the necessary additional TB- or NCD-related care after bidirectional screening is performed.

At level 3, the highest level of integration, (vi) patients entering the system via TB services are screened for NCDs, and (vii) patients entering the system via NCDs services are screened for TB and both models provide treatment or care management for both TB and NCDs at the same location (colocation).

The number of articles found at each level of integration is expressed in brackets, and entry points differentiated based on underlying known condition, be it TB or NCDs.

Facilitators and barriers to integration of TB and NCD services

Salient and overarching themes regarding facilitators and barriers facing integrated service delivery models were mapped to the WHO health systems framework which comprises six dimensions are summarised in Table 7 below.

Download:

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

Table 7. Summary of facilitators and barriers mapped into WHO health system framework.

https://doi.org/10.1371/journal.pmed.1003899.t007

Discussion

This review identified the potential applicability of integrating TB and NCD services in LMICs. Using the data extracted from the included studies, we derived key elements of the different levels of integration for TB and NCD services, followed by categorising them into 3 distinct levels, differentiated according to entry points into the health system. Our findings suggest that higher levels of integration conferred more benefits to patients in terms of managing TB and NCDs. However, we are also cognizant that numerous aspects spanning the 3 levels of integration are overlapping yet differ in certain operational areas depending on the study setting. Due to the pluralistic approaches taken by integration models that vary from how they exploit opportunistic screenings at either point of disease entry to onwards care treatment, the application of a complex systems approach to integrated services planning, delivery, and evaluation is, therefore, central to deciphering intervention implementation and impacts in real-world settings. We accomplished this by mapping the operational features of the various models of integration to WHO health systems framework, which provided a validated structure to organise our findings. As such, we elucidated the overarching facilitators and barriers that fall within the dimensions of service delivery, human resources, sustainable financing and social protection, information, and leadership and governance.

Our findings suggest that care integration is more common with entry via TB to the system, with 19 studies supporting this. Eight studies were conducted at level 1 integration, 8 studies at level 2 integration, and 6 at level 3 integration. A higher level of integration that requires longer chains of care as compared to a lower level of integration, whereby less services or services catering to one disease only is provided, comes with increasing challenges that need to be iteratively surmounted at all levels of the health system. More integrated models need to be mindful that additional integrated services do not compromise the functionality of existing TB or NCD programmes or clinical operations.

Integration aims to address fragmentation in services, enabling better coordinated and more continuous care. In short, integration is achieved by careful planning and financing, with a shared vision centred on a target patient population [62]. All studies pointed to integration as the genesis of structural links between previously separate services, encompassing aspects from organisational modifications, resource planning to physical colocation. At the lowest level of integration, only screening for either TB or NCDs is provided. In most studies, low-cost and straightforward screening techniques for TB and NCDs are used by trained staff. Level 1 integration studies had illuminated that providing point-of-care screening by either disease entry point, but without receiving the needed treatment through onwards referral or in-house treatment is insufficient to mitigate the increase in prevalence of either TB or NCDs effectively. The lack of providing referral to onwards care postscreening could be attributable to reasons such as the lack of comprehensive operational planning since such integrated programmes are relatively new in LMIC settings and cost concerns whereby the funds allocated for such programmes might have limited the prospects of linked care which may be considered unethical and a wasteful expenditure of scarce resources for diagnostic tests without any impact on patient management or outcomes.

Therefore, level 2 integration is more optimal in that regard but has its own challenges. Despite having referrals made to appropriate providers for both TB and NCD treatment postscreening, level 2 integration models face new difficulties that the level 1 integration models did not operationally face. This includes lost to follow-up due to loss of referral documents and patients facing the inconvenience of attending another visit, which might be constrained by work commitments and physical distance to the referred clinic [55].

Hence, full integration at level 3, whereby screening and treatment occur at the same location, has been shown in our reviewed studies to provide the best point of care for both disease groups and promotes, to an extent, care integration and continuity of care. Several studies have shown that colocation of services derives better health outcomes, follow-up care adherence, and overall higher patient satisfaction [63,64]. High levels of integration have also been shown to reduce fragmentation of care, thus minimising resource wastage and inefficiency by avoiding duplication of services and contradictory decisions by providers [65]. However, as direct outcome measures were not reported in the reviewed studies, we can only surmise that having the deepest levels of service integration confer the most benefits to both patients and the health system.

TB and NCDs control programmes must touch every aspect of a nation’s health system to manage the diseases comprehensively. Herein, we will discuss the facilitators and barriers with the health systems components in mind to further unpack the operational features of a high functioning integration model embedded in the health system of LMICs. Notably, 2 studies highlighted the need to factor in the social and financial elements for integrated models of care. Both factors fall beyond the remit of traditional TB control programmes. As sustainable financing and social protection is one of the dimensions of WHO health systems framework, we will use this as the entry point to delve deeper into other health systems dimensions using the facilitators and barriers summarised in Table 7 above.

First, sustainable financing and social protection is central to the management of TB and NCD integrated programmes as TB and NCDs have close links to poverty and social activation [66]. Hence, integration models will require more from health systems than the traditional medically oriented interventions that do not focus on the financial and social aspects of patients. OOP payment is shown to reduce the uptake of essential health services, where patients in LMICs might prioritise nonmedical yet essential aspects of their lives such as the purchasing of food and shelter. This will, in turn, become a missed opportunity if costs are the limiting factor to the uptake of TB and NCDs services. Therefore, health systems must explore robust financing and social protection mechanisms to promote the uptake of services at all levels of integration to ensure maximum reach and impact of integrated programmes [67]. From the providers and policymakers point of departure, the terms universal health coverage (UHC) and limited fiscal space lie in a luminal space, whereby finite resources need to be allocated to various aspects of the health system [68]. The study by Contreras and colleagues conducted in Peru showed that meeting the high demands for screening services necessitated the engagement of private laboratories where the costs were covered by the programme, while socioeconomic assistance such as transportation reimbursement and food vouchers were rendered to the patients by an NGO [49]. Therefore, having TB and NCD services provided at one facility can potentially reduce transport costs due to multiple visits. However, we are unclear of how the medical costs incurred by the patients will be divided in an integrated programme since certain medical conditions such as TB might already be covered by an existing programme but not others.

More importantly, risk factors for TB and NCDs are often multifaceted, and the most cost-effective way to mitigate both epidemics might be through early detection. Thus, financial attributes of screening services must consider the socioeconomic status of the targeted population and the ability of the country to finance it sustainably. This was further emphasised in another study by Byrne and colleagues in Peru, whereby glycosuria was used as a biomarker for undiagnosed DM and inadequate glycemic control for DM and CKD patients. Despite having moderate sensitivity, the study failed to employ more specific assessment tools for renal function impairment such as serum creatinine, due to cost and implementation feasibility concerns [45].

Second, at the service delivery level, contextual nuances such as cultural beliefs and preconceptions about a disease condition pose huge hurdles at the personal level and also impede the full functionality of an integrated programme. Poor or lack of public education and awareness of TB and NCDs can create misconceptions about the diseases such as “leaving it to fate.” This holds true not only for patients but also for policymakers, as seen in the disconnect from TB and NCDs agendas mentioned in the study by Lovero and colleagues [57]. A facilitator that can be employed to overcome this is to increase health literacy through community health workers [69]. The deployment of validated guidelines ensures that evidence-based process and outcome indicators are being monitored and fulfilled as cornerstones for effective patient-centred care, especially for TB and NCDs, which require prolonged management [70]. Furthermore, accessibility to care can also pose a barrier due to transportation costs or heightened opportunity costs for informal labourers when attending follow-up services. These challenges can be surmounted by deploying mobile testing units as used in the study by Govindasamy and colleagues in South Africa to meet patients where they are [55]. The lack of conducive space to screen for NCDs and provide counselling services might cause privacy concerns. As a result, patients might be less willing to take up these additional services, be it for mental health treatment or screening for other NCDs where anthropometric measurements are taken. Although not reported in the studies reviewed, colocation with TB services necessitates extra infectious disease precautions to reduce the nosocomial spread of TB when non-TB patients present themselves for NCDs [71].

Third, at the human resources level, it was apparent in nearly all reviewed studies that the increased workload posed a considerable roadblock to the operations of integrated programmes. This is especially salient in LMIC settings whereby training of health workers typically involve only a particular disease group such as maternal and child health or specific infectious diseases, while cross-training (i.e., TB and NCD services) for more than 1 disease condition is rare [72]. Additionally, the lack of manpower and training or expanded job scope places more stress on existing providers, who are already stretched by their normal clinical operations [73]. The overextending of these providers might lead to screening misses and treatment errors. As such, providers must not only be well trained but also motivated to provide the services. Attitude, professionalism, and commitment create stronger patient–provider relationships, which confer better patient outcomes and, hence, essential elements to any integration model [74].

Fourth, the availability of medical products and vaccines is essential to the implementation of integration programmes. However, aside from Bacillus Calmette–Guérin (BCG) vaccine, whose effect wanes after childhood, there are effectively no viable vaccines for TB. NCDs are also primarily developed due to genetic or lifestyle predisposition, which can only be prevented through behavioural modifications and medications [75]. Thus, the availability of efficacious medicines for TB, administered mainly through DOT, is an integral element to any TB programme. A break in the medication regime or shortage of medicines at all levels of integration can lead to the generation of MDR-TB [76]. Likewise, the provision of high-quality medication for NCDs is required to ensure optimal management of NCDs, which are often prolonged and require extended medication regimes and monitoring. Additionally, there is also a need for medical resources that aid in screening of patients for TB and NCDs. For example, the availability of handheld glucometers used for screening DM and electronic sphygmomanometers for hypertension as explicated by Govindasamy and colleagues provided a practical way to screen patients for NCDs [55]. Without the provision of medical equipment for screening, linkage to onwards care and treatment will be rendered ineffective in an integrated service delivery model. Although not discussed in the articles reviewed, having integration models that are equipped with adequate medical resources for both TB and NCDs is essential to managing both disease conditions concurrently. This is because the inability to control one might lead to the exacerbation of the other. For example, uncontrolled DM is associated with worsening of TB control as these 2 conditions are closely associated biologically [19].

Fifth, the information provided within and across healthcare entities is an essential tenet for the smooth running of any health programme, let alone one that integrates services catering to multiple disease conditions across different providers. Many of the studies reviewed highlighted the lack of coordination across providers, especially across different healthcare institutions, which hindered integration on many fronts. There is often a lack of common electronic medical record system or standardised medical information storage platform that tracks process and clinical outcome parameters of the patients and data access to providers within and across echelons of care was also reported to be limited. For example, Gnanasan and colleagues had expressed that certain clinical parameters such as blood pressure were not checked at the TB clinic but could have been extracted from DM or cardiovascular clinic’s records instead in order to give providers a more comprehensive medical history of the patient. However, the lack of coordination across provider settings stymied the exchange of patient data [59]. Furthermore, the information tracked must be evidence informed and relevant for clinical management of the patients’ condition. To this end, validated guidance principles and reporting measures/frameworks need to be deployed by providers. However, many reviewed studies had reported the lack of a former reporting system. Many had to turn to operationalised indicators to track the patient pathway, which might not be optimal. An example is using a modified WHO STEPwise Approach to NCD Risk Factor Surveillance (STEPS) framework that Anand and colleagues reported [50]. Although only pilot tested once in the study’s context, the operationalised questionnaire that screens for NCD predisposition was deemed quick and easy by both patients and providers. Therefore, despite the absence of thorough validation, rapid screening procedures can be deployed in resource constraint settings to streamline protocols to refer patients for onwards care.

Last, leadership and governance are central to implementing any integrated programme vis-à-vis other core functions that ensure the programme’s sustainability and credibility within a health system. Governance also cuts across all other aforementioned health system functions, making transparent, accountable and committed stewardship paramount in TB and NCD control. Good governance will also require taking the lead in upholding newly minted partnerships that include formal and informal productive relationships that span within and beyond the health sector and between the health system, its intermediaries and end users (patient and providers) through the accountable provision of evidence-driven guidelines and adequate and responsible incentivisation. An example of active collaboration is displayed in the study by Contreras and colleagues whereby the TB Cero programme was implemented as an integrative partnership model between public hospitals and an NGO in Peru [49]. At the national level, having a platform to bring together intersectoral actions driven by strong political commitment coupled with a definition of integration that captures an explicit agenda and a clear roadmap for implementation undergirds organisational change towards more assimilated TB and NCD service models. Specifically, governments can also promulgate national TB and NCD plans/guidelines/frameworks and earmark the needed funds to run plans into fruition. This was alluded to in the study by Ekeke and colleagues in Nigeria, whereby patients with DM were screened for TB, in line with the national TB guidelines [51]. Thus, there is a need to update these guidelines to include NCDs to prevent any missed opportunities in identifying NCD cases. Lönnroth and colleagues had also proposed a TB elimination framework to draw key takeaways from prospective TB and NCD integration models. This encompasses universal access to high-quality TB and NCD service with a lens that places groups at the highest risk first and reducing underlying vulnerabilities by incorporating health lenses in all policies, bolstering long-term political will for locally spearheaded integrated programmes, and garnering cross-sectorial and multilevel buy-in for the implementation of global TB and NCD action plans [77]. Finally, governance bodies must be mindful of the interconnectedness of interventions along the “cascade of care” from detection of TB or NCDs to successful completion of treatment or long-term management. This implies that governance bodies must be able to evaluate programmes using both process and clinical outcome measurements in a joined-up rather than a standalone manner to truly ascertain the effectiveness of the integration of TB and NCD services within one programme and at a common locale. As the effectiveness of the integrated programmes go beyond a single entity and across different sectors outside the health system, the need to incorporate systems thinking concepts as well as adopting relevant disciplines such as social sciences and health policy analysis will serve as cornerstones for a joined-up systems approach to evaluation [78].

This systematic review foregrounds the gaps in both policy and research in the areas of TB and NCD service integration. Policymakers and researchers have focused mainly on integrating TB and HIV services, as seen from the various existing and well-evaluated programmes and systematic reviews performed [33,37,79–82].

The research gaps can be plugged with more monitoring and evaluation works that track the process and clinical outcomes through validated indicators using a prospective or retrospective cohort study, case–control, or randomised control study approaches. Furthermore, the experiences and expectations of patients and providers can be explored through qualitative studies to confer appropriate models of care integration that meet the needs of all stakeholders. The urgency is now more than ever in the face of a swelling TB and NCD burden, and prospective research findings must be able to empower policymakers to drive integration models using implementation science and sustained political will.

Policymakers and programme implementers must recognise that the inability to attain sustainable and effective integration might be less attributed to ineffective programmes per se but rather due to the impediments underlying the health systems building blocks. Hence, integration necessitates stakeholders to envision the channelising of resources and agendas beyond individual programmes and to an extent sectors, while concomitantly investing in all health systems building blocks.

Although there is no magic bullet to achieve a fully integrated configuration for all settings, some actionable steps surfaced. We have summarised the key facilitators to be accounted for and barriers to be avoided during the implementation of integration models in Table 7 above. The high start-up costs to initiate such programmes require policymakers to place long-term returns at the forefront when deciding to allot resources into these models as return on investment in terms of better population health outcomes can only be seen in the longer term. Spillover effects such as better infrastructure and trained manpower can also enhance the health sector and improve access to care for the general population. Uptake must also be sustainable through carefully tailored integration models that account for the needs and preferences of the local population. A salient point that did not appear from our review but is crucial for the uptake of integrated services is the reduction of stigma when seeking care for certain services associated with TB and MHIs, which might affect the extent to which such models of integration are palatable.

This systematic review was limited by the dearth of research exploring the experiences of patients and providers, and specific process indicators that evaluate the feasibility of integrated interventions, as most studies focused on reporting outcomes at a particular time period. Furthermore, the main bibliographic databases used in this review may have had a limited coverage of material from LMICs regarding TB and NCD integrated programmes, and we acknowledge that the databases might have lacked coverage of material not published English [83–85]. In addition, the heterogeneity of the study designs, contexts of study, participants, sample sizes, outcomes, interventions, and diseases studied can potentially limit the generalisability of the study findings. Importantly, this review also revealed that most integration models were contextually bound, with each LMIC having their very own set of healthcare and political agendas writ large. Therefore, our findings have extracted the overarching themes that are cross-cutting and salient to LMICs, particularly in Asia, Africa, and the Americas.

One strength of this review is the usage of a widely deployed health system framework to guide the data analysis and reorganisation of themes for the facilitators and barriers, reducing the variability across research team members. However, we must be cognizant of the oversimplification of intricate process pathways and organisational culture. In practice, many of the key issues span across multiple building blocks, as to be anticipated when employing a health systems lens. However, we managed to utilise this framework for mapping integrated service options to each health system building block to derive high-quality and pragmatic design options for integration elaborated above.

Another strength is the development of a new framework that categorises models of integration according to their depth of integration. This will enable both policymakers and researchers to understand where a country’s TB and NCD programmes lie on the spectrum of integration and, with that knowledge, strive towards the deepest levels of integration by avoiding the barriers and incorporating the facilitators into their programmes.

Conclusions

The burgeoning number of TB and NCD cases has made a strong case for tackling this combined burden of TB and NCD prevalence with a comprehensive and integrated approach commensurate to its disease burden. Limitations notwithstanding, this systematic review provides valuable insights into integrating TB- and NCD-related services into one common programme in LMICs. This has the potential to improve health service delivery across disease conditions and levels of care that can lead to population health benefits.

Supporting information

Acknowledgments

We would like to thank Nima Asgari, Director of the Asia Pacific Observatory (APO) on Health Systems and Policies at WHO Regional Office for Southeast Asia, for his support throughout the project and Mr Tan Wei Kang, who kindly provided us with the statistical expertise in the conduct of this systematic review.