Pakistan has one of the highest burdens of tuberculosis (TB) in the world, a disease that infected as many as 10 million people and caused 1.4 million deaths globally in 2019.
Around a quarter of the world’s population are thought to be infected with latent TB. The disease can remain dormant for years, until, for example, the immune system is suppressed by diseases such as HIV or diabetes or with advancing age. One in ten of those infected will develop active TB disease during their lifetime.
This lengthy and unpredictable incubation period makes detection of TB difficult using conventional measures and is one of the reasons why the disease has been so successful in remaining hidden and widely distributed throughout communities in low- and middle-income countries, posing unique challenges for halting the transmission of the disease.
Modelling local TB positivity rates using active case-finding data
In November 2023, KIT presented a study comparing the predictive accuracy of a Bayesian and frequentist approach to modelling TB positivity rates which can be found in active case-finding (ACF) settings in Pakistan. Both models included the same regressors reflecting the MATCH conceptual framework, using Mercy Corps’s ACF data as an outcome variable.
The Bayesian model, the ‘MATCH AI’ model co-developed by KIT and our partner
EPCON, performed well and quite similarly to a negative binomial regression. The results of this study highlight the potential value of using ACF data to target which harder-to-reach communities/locations may most benefit from TB case finding in Pakistan. See the highlights here.
The Case for Community Case Finding
There is considerable evidence for the benefits of using active case finding (ACF) to locate people with TB in communities instead of waiting for people to seek healthcare in facilities. Using mobile diagnostics to find people with TB can lead to faster referral and treatment, which leads to better health outcomes, reduced transmission, and fewer new cases. ACF can fill the TB case detection gap which is often a challenge for hard-to-reach communities. However, resource and logistical constraints rule out screening them all, so the challenge is determining where and among which populations to look.
This challenge is compounded in Pakistan due to political instability, security risks, and a population of over 212 million spread out over a vast geographical area. To overcome these issues, a specific and targeted approach to community case finding is required.
Improving Case Finding
KIT Royal Tropical Institute’s electronic Case-Based Surveillance (eCBS) project aims to specifically address this challenge as well as to bridge the TB case detection gap by using real-time geospatial analysis of mobile chest camp and secondary data to more efficiently identify communities with a high likelihood of having undetected cases of TB. Using a Bayesian model to make prevalence predictions at the subnational level, the project seeks to identify these ‘hotspot’ communities at the most local level possible. The project also strives to improve resource and time efficiency by shifting from paper to electronic data collection.
Digitization of Data
A central component of the eCBS project is the timely digitization of chest camp screening data. The project is predicated on using real-time data to help steer programmatic activity and validate the model’s predictions. As MercyCorps’s data collection in Pakistan has been paper-based, project resources have been aimed at digitizing this process.
The research team conducted field visits to understand
MercyCorps’s workflow and data needs. EPCON developed a tailor-made data collection tool and worked with KIT to introduce a digital dashboard for real-time monitoring and validation of the data quality. Virtual training sessions and hands-on support from the National TB Program (NTP) have helped MercyCorps to make the transition to digital data collection. Not only fundamental for the project, efforts to digitize screening data in Pakistan also contribute to strengthening the resource constrained health systems, adding an element of sustainability to the eCBS project.
21st Century Case Finding
Digital data from community-based chest camps are now streaming into the eCBS server in real-time. The Bayesian model uses both chest camp screening data and a repository of TB, demographic and additional epidemiological data linked to geographic regions in Pakistan, to make predictions of locations where TB case finding yield is likely to be high. These predictions, along with chest camp data are visualized on the dashboard and geoportal (geographical visualizer), where clustered geographical areas with high case predictions are identified as good locations for screening.
While KIT will be initially supporting with the project set up and analysis of hotspots, the project will be gradually integrated in to public and private sector partnerships between the NTP and participating non-governmental organisations. The ultimate goal of this integration is for the partners in Pakistan to have complete autonomous control of eCBS.
Education, training and coaching
There is probably no better contribution towards sustainable development than investing in people and building capacities at all levels. KIT Royal Tropical Institute plays a major role in this by offering a range of education and capacity building services, from formal education at masters level to client-oriented training and coaching support. Our training and coaching programmes are tailor-made to address the unique objectives and capacities of the requesting organisation.
More about Education, training and coaching
The Match Approach
A cornerstone of the Centre for Applied Spatial Epidemiology is the MATCH Approach. This innovative geo-spatial analysis framework evaluates the effectiveness of interventions in the context of the local disease risk, burden and health efforts. This information is especially valuable in informing local interventions.
The power of the MATCH approach is twofold. First, it employs and analyses multiple — and often underused — sources of geographically, temporally and demographically disaggregated data on epidemiological risk factors. And finally, it links data to policy objectives, combines and simplifies complex data into a more intuitive format and builds capacity for data management and analysis. This enables decision makers to use their program data to make better informed decisions supported by local evidence. MATCH is currently being applied to support TB programs around the world to identify persistent gaps in program functioning which lead to over four million people with TB not being diagnosed or reported globally. MATCH consequently provides a valuable tool to identify geographic areas where TB case detection, diagnosis and reporting can be strengthened.