- Split View
-
Views
-
Cite
Cite
Lydia Boudarene, Richard James, Richard Coker, Mishal S Khan, Are scientific research outputs aligned with national policy makers’ priorities? A case study of tuberculosis in Cambodia, Health Policy and Planning, Volume 32, Issue suppl_2, October 2017, Pages ii3–ii11, https://doi.org/10.1093/heapol/czx041
- Share Icon Share
Abstract
With funding for tuberculosis (TB) research decreasing, and the high global disease burden persisting, there are calls for increased investment in TB research. However, justification of such investments is questionable, when translation of research outputs into policy and health care improvements remains a challenge for TB and other diseases. Using TB in Cambodia as a case study, we investigate how evidence needs of national policy makers are addressed by topics covered in research publications. We first conducted a systematic review to compile all studies on TB in Cambodia published since 2000. We then identified priority areas in which evidence for policy and programme planning are required from the perspective of key national TB control stakeholders. Finally, results from the literature review were analysed in relation to the priority research areas for national policy makers to assess overlap and highlight gaps in evidence. Priority research areas were: TB-HIV co-infection; childhood TB; multidrug resistant TB (MDR-TB); and universal and equitable access to quality diagnosis and treatment. On screening 1687 unique papers retrieved from our literature search, 253 were eligible publications focusing on TB in Cambodia. Of these, only 73 (29%) addressed one of the four priority research areas. Overall, 30 (11%), five (2%), seven (2%) and 37 (14%) studies reported findings relevant to TB-HIV, childhood TB, MDR-TB and access to quality diagnosis and treatment respectively. Our analysis shows that a small proportion of the research outputs in Cambodia address priority areas for informing policy and programme planning. This case study illustrates that there is substantial room for improvement in alignment between research outputs and evidence gaps that national policy makers would like to see addressed; better coordination between researchers, funders and policy makers’ on identifying priority research topics may increase the relevance of research findings to health policies and programmes.
Using TB in Cambodia as a case study, we investigate how well evidence needs of national policy makers are addressed by topics covered in research publications.
Our systematic review of scientific research outputs on tuberculosis control from Cambodia revealed that only a small proportion of available evidence addresses priority areas identified by national policy makers.
In the context of substantial decrease in TB funding globally, improved coordination between funders, researchers and national policy makers is critical.
Introduction
There are strong calls for increased investment in research and development for tuberculosis (TB) (Marais et al. 2010; Walwyn 2013); a recent report on research funding for TB highlights that US$3.29 billion was disbursed between 2011 and 2015, which is only one third of the $9.84 billion of investment called for under the Global Plan to Stop TB (Frick 2016). However, there are questions about how justifiable continuing the current approach to investment in TB research and development is when translation of findings into policy and health care improvements continues to be a widely acknowledged challenge for TB and other diseases (Straus et al. 2011; Khan et al. 2016). Indeed, the historian Christian McMillan—who analysed responses to TB in the twentieth century—highlighted a pattern of ‘repetition and rediscovery’ among researchers and policy makers, owing to a tendency to seek new solutions while ignoring lessons that have been learnt (McMillen 2015).
Low and middle income countries (LMICs) are particularly lacking in governance and structures guiding research priority setting, with only half of countries surveyed in a 38-country study having any formal governance structures for health research (Ahmedov et al. 2008). A potential consequence of the absence of a clear research agenda driven by national policy makers is ‘a gulf between researchers and decision makers’, resulting in research outputs less useful for policy than they could be (Orton et al. 2011; Goyet et al. 2015).
In this study, we use TB research in Cambodia as a case study to investigate the level of alignment between evidence needs of national policy makers and publications generated by researchers. TB is a major public health challenge in Cambodia and there has been substantial investment along with a decline in prevalence since 2000 (CENAT 2002, 2011b). In parallel, there has been an increase in Cambodia-specific TB research; a recent scoping review showed that the number of publications relating to TB in Cambodia increased from 7 to 102 during the 15 years from 2000 to 2015 (Khan et al. 2016). To date, there has been no systematic review of published scientific literature—from Cambodia or from other countries in the region—to assess whether the increasing number of research outputs being generated are addressing information needs of policy makers and reducing high-priority gaps in evidence for policy and programme planning.
Methods
The study was conducted in three steps. First, a systematic literature review was conducted to identify relevant studies on TB in Cambodia published since 2000. In parallel, priority research areas for informing policy and programme planning were identified from the perspective of key national stakeholders. Finally, results from the literature search were analysed in relation to the priority research areas for policy to assess the level of alignment and identify priority areas that are not covered by current research.
Literature review
Three databases—Pubmed, EMBASE and IMSEAR—were searched using the following keywords: ‘Tuberculosis OR TB’ AND ‘Cambodia’, to identify any relevant studies published between January 2000 and the search date (20 October 2016). The same search was also conducted on Google Scholar and the first 100 hits were included. Grey literature on the World Health Organisation (WHO) website was searched from 2005 to 2016 (WHO—TB Publications). The most recent WHO publications reporting Cambodian data of any given series were included.
Based on a set of inclusion and exclusion criteria (Table 1), titles and abstracts of all publications identified through the search were screened. Publications that passed the first stage of title and abstract screening had a full text screening performed by two independent researchers (LB and RJ). Any disagreement regarding inclusion of an article was resolved through discussion.
Category . | Inclusion criteria . | Exclusion criteria . |
---|---|---|
Publication type |
|
|
Date of publication |
|
|
Subject of study |
|
|
Location |
|
|
Category . | Inclusion criteria . | Exclusion criteria . |
---|---|---|
Publication type |
|
|
Date of publication |
|
|
Subject of study |
|
|
Location |
|
|
Category . | Inclusion criteria . | Exclusion criteria . |
---|---|---|
Publication type |
|
|
Date of publication |
|
|
Subject of study |
|
|
Location |
|
|
Category . | Inclusion criteria . | Exclusion criteria . |
---|---|---|
Publication type |
|
|
Date of publication |
|
|
Subject of study |
|
|
Location |
|
|
Identifying priority research areas for informing policy
We shared a list of questions asking about research priorities in Cambodia, by email and in person, with two senior officials from the Cambodian National TB Control Programme (CENAT) and two WHO TB advisors to Cambodia in June 2016. We then held one to one semi-structured interviews in English with one senior official from each organization in order to gather more detail regarding the research priorities individuals perceived to be most important. During these interviews—which lasted between 25 and 45 min—we asked the key informants to describe the top priority areas in which research evidence to inform policy is required. We encouraged them to limit the number of research areas to five in order to ascertain the major priority areas. Detailed handwritten notes were taken in English and analysed following the discussions. In addition, we extracted information on stated research priorities from the most recent versions of two key CENAT documents: the national operational research concept note, generated from a collaboration between CENAT and WHO in 2016 (not published) and the TB strategic plans (CENAT 2011a). A thematic analysis was conducted to identify recurring topics covering the key priority areas for research from the policy makers’ perspective. We generated four broad research themes from our analysis which captured all of the main priority areas mentioned.
Data extraction and analysis
All eligible papers from the systematic search were included in a two-stage data extraction process, which was performed by two independent researchers (LB and RJ). First, papers were analysed to assess whether they addressed one or more of the priority research areas identified as described above. Standardized data extraction forms were then used to capture findings relevant to each priority area. Cambodia related data only were extracted from WHO reports. For each priority area, findings from the relevant literature were compiled and summarized, and evidence gaps identified. Data from a single publication were used in different priority area categories if appropriate.
Results
The database search returned 1687 unique papers. After screening of eligibility, 253 articles were included in the analysis. Figure 1 summarizes the search results.
The four overlapping key priority themes defined by national policy makers and WHO advisers, which were consistent with the most recent national operational research concept note and TB strategic plans were: TB-HIV co-infection (including epidemiology, prevention and management); childhood TB (including diagnosis, management and prevention); multidrug resistant TB (including epidemiology, prevention and management); universal and equitable access to quality diagnosis and treatment of TB. No discordance between the two organizations’ priorities was noted, which is likely due to their close collaboration. Moreover, the broad research priority themes in relevant reports did remain consistent between 2011 and 2016. Of the 253 eligible studies, 73 (29%) addressed at least one of the four priority research areas for policy and programme planning. Overall, 30 (11%), 5 (2%), 7 (2%) and 37 (14%) studies reported findings relevant to TB-HIV, childhood TB, MDR-TB; and access to quality diagnosis and treatment respectively (Figure 2). Fifty nine (23%) studies presented non relevant findings or a study design not meeting the inclusion criteria. Sixty eight (27%) studies reported lab-based work (genomics; immunology; molecular biology). Twenty two (9%), 14 (6%) and 14 (6%) of the studies reported a broad view of TB program management (policy, economics), descriptive epidemiological studies and TB associations and co-infection (other than TB-HIV), respectively. Two studies focused on vaccination and one study on extrapulmonary TB. A summary of evidence available and key evidence gaps for the identified research priorities in the literature is presented below.
TB-HIV
Much of the evidence on TB-HIV relates to epidemiology and treatment pathways. For example, reports estimate that 36 000 patients were co-infected by TB and HIV (WHO—Global TB Report 2015), and that 81% of TB notified patients knew their HIV status with 98% commencing antiretroviral therapy (ART). Epidemiological studies indicate a high incidence of TB-HIV co-infection (Bonnet et al., 2016), with TB identified as the first or second most frequent opportunistic infection amongst People Living With HIV/AIDS (PLWHA) (Pichith et al. 2001; Senya et al. 2003; Kong et al. 2007). A cross-sectional study with a retrospective cohort found 38% (216/574) of TB patients were co-infected with HIV and 24% (124/450) of the HIV patients with TB. The rate of death in the co-infected groups was much higher (27%) than TB patients with no HIV (5%) (Cain et al. 2007).
Recent studies provide evidence to suggest that there has been much progress on TB-HIV in the past decade. ART coverage has increased (Getahun et al. 2005) and HIV mortality has decreased (Eng et al. 2009). Strengthening of the TB-HIV management system between 2006 and 2008, with appointed TB-HIV coordinators to TB wards, is associated with improved HIV testing uptake and referral to HIV care and treatment services (Tsurugi et al. 2011). Two cross-sectional sero-prevalence surveys of TB patients registered by CENAT demonstrated that the HIV prevalence among TB patients had fallen from 11.8% to 9.9% between 2003 and 2005 (Tamura et al. 2008). The proportion of TB patients screened for HIV in Cambodia has also increased dramatically over the past 10–15 years with >80% and 81% of TB patients reported to be aware of their HIV status, in 2011 and 2014 respectively (WHO—Global TB Report 2015).
In terms of information to support programmatic guidelines, we identified 11 studies about symptoms associated with HIV-TB co-infections to assist with increasing detection rates (Lynen et al. 2007; Chheng et al. 2008; Kanara et al. 2008; Le Minor et al. 2008; Tamhane et al. 2009; Cain et al. 2010; Sculier et al. 2010; Chartier et al. 2011; Kim et al. 2011; Koole et al. 2011; Choun et al. 2013). In particular, national TB-HIV treatment guidelines were adapted after the results of the CAMELIA prospective randomized trial which demonstrated that initiation of ART 2 weeks after the start of TB treatment significantly decreased the risk of death (Blanc et al. 2007, 2011a,b).
Remaining barriers to HIV testing in TB patients identified by studies include AIDS-stigmatizing beliefs (Yi et al. 2009), distance to counselling and testing centres (Kanara et al. 2009), poor knowledge about TB/HIV, lack of communication skills among healthcare providers, the absence of protocols for HIV testing in TB patients and fear associated with informing patients of positive test results and owing to stigma (Samrith et al. 2015). We found little evidence from existing studies on interventions to address these barriers or about vulnerable groups that could be prioritized. One study showed that testing with Xpert MTB/RIF significantly decreased the turnaround time between specimen submission and diagnosis of TB-HIV patients (Auld et al. 2014), but the cost-effectiveness of this approach and challenges to widespread implementation have not been investigated.
There was also very little evidence available to assess treatment quality, issues of misdiagnosis and incorrect prescribing practices. Another area with little information for policy makers is factors driving drug resistance among TB-HIV patients; studies have estimated that between 5 and 8% have MDR-TB (Kimerling et al. 2002; Sar et al. 2009; Walls et al. 2015), but it is unclear whether MDR-TB is increasing among TB-HIV patients and what prevention strategies should be implemented.
Childhood TB
In contrast to TB-HIV, we found very few publications on childhood TB in Cambodia, despite this being a high priority area for national policy setting. Current prevalence estimates of TB in children aged under 15 are uncertain owing to a lack of studies on this topic. The national prevalence survey conducted in 2002 indicated that childhood TB is a major public health challenge in Cambodia, with only 50% of children aged between 1 and 14 showing a Bacillus Calmette–Guérin (BCG) vaccine scar and approximately 33% of children infected by age 12 (CENAT 2002). More recent estimates from a 2010 model indicated that BCG vaccination coverage among Cambodian children under 15 is over 90% (Dodd et al. 2014). It is difficult to draw firm conclusions about disease burden as the 2011 national prevalence survey did not include children under 15, resulting in a major gap in information about epidemiological changes in childhood TB and regional differences in disease burden, which is required to inform control strategies (CENAT 2011b).
The limited available evidence suggests that childhood TB remains an area that requires attention. A recent study in rural Cambodia found that among children under 15, TB was the sixth most frequent diagnosis, representing 1% of the household members screened (Ir et al. 2010). A 2012 study found that 7.7 per 1000 children admitted to a paediatric hospital had TB (Schopfer et al. 2012). In 2015, Morishita et al. reported that the number of TB notifications in children in 2013 was 6916, representing 18% of all notified TB cases. Wide variations in the notification rates between provinces was highlighted (Morishita et al. 2015).
While national policy makers recognize that tackling childhood TB is an urgent priority, they are hindered by a number of gaps in evidence to inform actions. Apart from a lack of nationally representative data on disease burden, research gaps span from studies on appropriate detection and diagnostic strategies for the local context to investigation of treatment quality and drug resistance. In particular, the issue of over diagnosis is of concern (Morishita et al. 2015), and the impact of childhood TB on the general TB epidemics should be assessed.
Multidrug resistant TB (MDR-TB)
With only 7 relevant studies on MDR-TB in Cambodia, information on this priority area is also lacking. The national drug resistance survey was conducted in 2001, before a new decentralized TB treatment programme was initiated in Cambodia. The survey indicated a low prevalence of MDR-TB, with approximately 10% of TB patients resistant to any one of the four first-line TB drugs (isoniazid, rifampicin, ethambutol and streptomycin) (Yamada et al. 2007). The 2015 Global TB report estimates the proportion of cases with MDR-TB to be 1.4% (95% CI 0.7–2.5) for new cases, and 11% (95% CI 4–22) in previously treated TB cases (WHO—Global Tuberculosis Report 2015), which is lower than other countries in the region such as Myanmar and Thailand. However, checks of standardized TB reporting records found that 27.3% of treatment failures were under-reported, and 22.9% of treatment successes were over-reported, indicating that accurate data for planning policy responses to MDR-TB is lacking (Hoa et al. 2010).
Between September 2011 and 2013, 20.8% of participants screened with Xpert MTB/RIF were found to be positive for TB, and 0.7% had rifampicin resistance (Creswell et al. 2014); this is consistent with the 2001 prevalence survey which found rifampicin resistance in 0.6% and may indicate that rifampicin resistance has not increased in the past decade (Yamada et al. 2007). A comparison with other resistance tests confirmed that Xpert MTB/RIF testing for rifampicin resistance was adequate to refer patients to second line treatment (Lorent et al. 2013, 2015b).
In general, information on barriers to identifying MDR-TB comes from a few studies, and further research is needed to aid early treatment and prevention on MDR-TB. A qualitative study indicated that the following barriers to identifying MDR-TB cases are important in Cambodia: low coverage and availability of diagnostics, long turn-around time for diagnostic results and inadequate access to testing for patients at high risk of MDR-TB (Royce et al. 2014). Detection of resistance by Xpert MTB/RIF improved the time to diagnosis from 97 to 2 days compared to standard drug susceptibility Tests (Lorent et al. 2015b). Decentralization of rapid diagnostic technologies has been recommended (Khann et al. 2013), but studies on challenges to implementing Xpert MTB/RIF and other new diagnostics in peripheral settings, and the cost-effectiveness of doing so, have not been conducted.
Universal and equitable access to quality diagnosis and treatment
Cambodia has a high rate of new case notifications (281 per 100 000 population) and 28% of new TB cases are estimated to remain undetected. Improving access to diagnosis and treatment is therefore a longstanding priority for national policy actors, and 46 studies were found addressing this area (CENAT 2011b).
In terms of access of TB patients to health facilities, studies have highlighted the importance of community referral. In a sample of 43 out of 93 basic TB management units, 73% of the total TB notifications (all forms) were attributed to community health workers or community volunteers (WHO—Global TB Report 2015). CENAT has been using active case finding (ACF) to increase case detection since 2005 (Eang et al. 2012; WHO—TB-Day-Brochure 2014). Several studies showed that ACF approaches lead to an increase in mean notification rates of all new TB cases per year (Thim et al. 2004), improved early case detection (Morishita et al. 2015), assisted in treatment initiation (Lorent et al. 2014) and, in a modelling study, was shown to reduce mortality (Yadav et al. 2014). Some researchers have recommended a systematic approach to implementing screening for TB involving: (1) setting clear goals after an assessment of the situation; (2) identifying and targeting high risk groups; (3) using an adapted screening algorithm; (4) planning context-specific implementation programs with relevant partners and (5) consideration of ethics and strong monitoring and evaluation as well as placing screening in the broader context of TB detection (Creswell et al. 2013). The authors stress that the need from detection at early stages of disease would be beneficial in reducing TB transmissions; however, evaluations of the impact of alternative screening approaches on TB transmission are needed and the cost-effectiveness of investments in alternative ACF strategies have not been established.
A study based on data from 2009 and 2010 routine TB registers concluded that, with a cost per case detected of USD $108, TB patient household contacts screening is cost-effective (Eang et al. 2012). Sutton et al. measured the costs of targeted case finding and isoniazid preventive therapy (IPT) services for HIV positive patients, and found an estimated costs of the two strategies per patient to be less than the reported cost of treating a new smear-positive TB case (Sutton et al. 2009).
In terms of diagnosis once patients have reached health facilities, key challenges include missed diagnosis during specimen testing as laboratory capacity remains low (WHO—Global TB Report 2015); there is a dearth of evidence on how to improve quality of sputum testing in health facilities, with intervention studies and health systems research needed urgently. Weaknesses in the linkage of patients with presumptive MDR-TB patients to culture centres has been documented, indicating the practices around initial diagnostic testing, and follow-up testing of patients that are not responding to treatment requires attention (Khann et al. 2013).
Several TB diagnostic techniques have been assessed in Cambodia, and these studies have provided important evidence for policy recommendations. The tuberculin skin test, produced variable results (Delgado et al. 2004) and a low and variable specificity and sensitivity, particularly in PLWHA (WHO—Publications on engaging all TB care providers); it is therefore used infrequently (WHO—International Standards for Tuberculosis Care-Diagnosis-Treatment-Public Health 2014). Serological tests, wand line Probe Assays, not recommended by CENAT, are rarely used (Grenier et al. 2011). Chest X-Rays, used >50% of the time as the only diagnostic tool in the public sector to detect pulmonary TB (Pande et al. 2015), were also found to be a significant predictor of smear-negative pulmonary TB (WHO—Priority Research Questions for TB-HIV in HIV-Prevalent and Resource-Limited Settings 2010). Xpert MTB/RIF presents a high sensitivity and specificity (Creswell et al. 2014); however, owing to resource constraints, smear microscopy was used 21 times more than Xpert MTB/RIF in 2012, allowing for possible improvement of diagnostics sensitivity and specificity (Qin et al. 2015). Alternative diagnostic algorithms using smear microscopy have been assessed in Cambodia (Lorent et al. 2015a), but further investigations are needed to make concrete guidelines. Particularly among HIV positive patients Xpert MTB/RIF provides better case detection than microscopy and contributes to a better detection of drug resistance detection, potentially decreasing the empirical treatments prescribed (Page et al. 2015; Auld et al. 2016). Schopfer et al. showed that rRNA detection is also useful for diagnosing children and shows a high specificity (Schopfer et al. 2012). While these studies provide information about effectiveness in terms of improved case finding, analyses of cost-effectiveness to inform investment and scale-up decisions are missing.
In light of the active private sector in Asia, a public-private mix program has been piloted, implemented and scaled up since 2005 in Cambodia, involving pharmacies, private clinics, hospitals and laboratories referring TB suspects to public sector clinics (Sheikh et al. 2007,WHO—Public Private Mix for DOTS—Towards Scaling Up 2005; WHO—Public-Private Mix for TB Care and Control 2007; Bell et al.2012, 2015, 2016). It is estimated that engaging the private sector can contribute to 17% of the total notifications (WHO—Public-Private Mix for TB Care Control 2007), indicating the importance of establishing a strategy for widespread and long-term engagement with private providers.
In line with the relatively low MDR-TB prevalence, studies in Cambodia show high treatment success rates among TB patients. Recent analyses indicate a treatment success rate of 93% (2013 cohort of 35 536 previously untreated TB patients) and a slightly lower treatment success rate for those TB patients previously treated (though excluding relapses) of 90% (2013 cohort of 1701) (WHO—Global TB Report 2015). Similarly, a low default rate of 1.6% and treatment failure rate of around 10.1% were found (Hoa et al.2012a,b). From 2012, data on MDR-TB, the treatment success rate was 79% in a cohort of 110 patients. As mentioned earlier, there is some evidence to indicate issues with accuracy of reporting (Hoa et al. 2010) and there has been one investigation of human resource requirements to implement quality-assured electronic data capture of the TB case registers (Hoa et al.2012a,b).
While prescribing practices of the TB healthcare providers were found to be good, TB drugs were also available at private pharmacies where access is largely uncontrolled (Uchiyama et al. 2006). Directly Observed Treatment (DOTS) and community DOTS (C-DOTS) were launched in 1999 and 2002, respectively (Khieu et al. 2007), and were shown to be highly effective in reducing delays to commencing TB treatment in Cambodia (Saly et al. 2006). There was no difference in treatment outcomes between the health centre and C-DOTS strategies (Khieu et al. 2007). A cost effective analysis of hospital versus C-DOTS (Pichenda et al. 2012) also found treatment success rates were not significantly different and that all costs among patients treated with hospital DOT were consistently higher than for those treated with non-hospital DOTs. An analysis of the TB programme concluded that three issues have been critical to its establishment: securing a stable and predictable funding base, adopting the DOTS treatment paradigm and integrating TB control into the health coverage plan (Hill and Tan Eang 2007).
Despite broad evidence availability on access to diagnosis and treatment of TB in Cambodia, specific gaps remain to be bridged. The lack of laboratory capacity and appropriately trained staff to perform quality analysis and run laboratories is an issue to be tackled. Evidence on cost-effective approaches for strengthening human resources and introducing new technologies for TB diagnosis is needed. Accessibility to TB screening and treatment is thought to have been improved by ACF and C-DOTS. However, the long-term impact of ACF on TB prevalence, as well as the sustainability of ACF implementation, has to be determined. Quality of diagnosis and treatment, as well as linkage between different health services has yet to be assessed. There are also notable gaps in the evidence base relating to TB prevention and about the growing issue of TB association with non-communicable diseases.
Discussion
There are two main findings from this case study of research outputs from Cambodia and their alignment with policy makers’ information needs. First, we found that a fairly small proportion of studies—less than a third—have any findings that relate to the main priority areas for policy and programme planning. Another third of the studies address more fundamental lab-based research issues. Second, there was large variation in the availability of evidence between the different priority areas. Taken together, the study findings indicate that there is much room for improvement in the coordination between scientific research and policy makers’ information needs, which may increase the relevance of research to policy setting and programme planning.
With 30 and 37 publications, respectively, TB-HIV co-infection and universal access to diagnosis treatment were the priority areas identified by policy makers with the greatest number of relevant publications. Direct links between the evidence produced and policy was observed for TB-HIV guidelines; for example, evidence from the multisite study trial named CAMELIA lead to a change of national and international guidelines for TB-HIV treatment (Blanc et al. 2011b) and research evidence also informed TB-HIV screening implementation (Eang et al. 2012). Childhood TB and MDR-TB have been studied far less, and our study suggests that more context specific evidence is required to inform policy decisions. Indeed, formation of national guidelines is being hindered by a lack of evidence to base them on; this is well documented for childhood TB in particular in LMICs broadly (Berti et al. 2014).
A previous study has already demonstrated that there is some misalignment between funding priorities and researchers’ agendas (Khan et al. 2016). The authors were not able to include national policy makers’ priorities in the analysis, and called for research investigating national policy makers’ views on research topics. The present case study thus provides important new information to show that there is substantial room for improvement in alignment between research outputs and evidence gaps that national policy makers would like to see addressed. A clear recommendation based on these findings is better coordination between researchers, funders and policy makers on identifying priority research topics. This could be achieved through the formation of national bodies for setting national or regional research agendas, which should include key national stakeholders and researchers engaged in the country, and feed into international research funders’ investment plans. To facilitate development of country plans and enhance global TB research, the WHO launched a Global Action Framework on TB Research in 2015 (WHO—A Global Action Framework for TB Research 2015). An exercise to engage national actors in identifying research priorities for malaria has recently been conducted, reflecting recognition of the importance of this other disease area as well (Canavati et al. 2016). However, setting of research agendas is not without challenges since the optimal process to follow for reaching a consensus, and the dimensions on which research priority areas are assessed (effectiveness, equity, cost etc.) is not clearly defined.
In addition to researchers, national policy makers can also play their part by ensuring that evidence gaps they would like to see addressed are clearly defined and communicated; often policy makers do not have a clearly articulated research agenda for researchers to engage with. In conducting this we had to undertake considerable work and required strong in country relationships to identify research priorities from different sources. Researcher’s limited awareness of policy maker’s research priorities exacerbates the misalignment. Another stage of engagement is in the evaluation of research proposals, where national policy makers should play a role judging applicability and potential impact on TB in high burden settings. Finally, in light of limitations in funding for TB research, and to avoid duplication, policy makers may be able to refer to studies conducted in similar settings to inform policy decisions. Facilitation of timely dissemination of regionally relevant research would be beneficial in addressing the gap in evidence for policy. Similarly, policy decisions can also be informed by analysis of good quality programmatic data; operational research using existing routine data can often be done at a relatively low cost, especially when implementing partners including non-governmental organizations—that are keen to impact on policy—play a role in maintaining high quality records.
A limitation of this study is that it is based on results from a single country and similar investigations in other settings both within the region and more broadly in LMICs would be useful to confirm whether our findings are generalizable. We acknowledge that the research reviewed did not include any studies in Khmer, which may have resulted in some publications being excluded, and that we only included CENAT and WHO stakeholders’ views when identifying priority research themes as focus was specifically on policy makers’ research priorities. Representatives of non-governmental organizations involved in service delivery, Cambodian academics and TB patients may have different opinions on priority research themes.
Conclusions
In the context of substantial decrease in TB funding globally, coordination between funders, researchers and national policy makers is critical (Frick 2016). Our case study of TB research in Cambodia indicates that only a small proportion of research outputs directly relate to priority areas for policy makers, and even those topics which are well covered by research studies have important gaps in evidence for policy setting. While ensuring that research diversity and independence is maintained, better synchronization of mutually agreed evidence gaps for policy with research agendas and research funding may result in greater impact of research on national policy.
Acknowledgements
This work was supported by the National University of Singapore, through the TB Cambodia grant R608-000-119-733.
Conflict of interest statement. None declared.