Hack TB Blog Post #2
Big data has been compared to a holy grail, and like all important relics, Pakistan’s national TB data used in KIT’s Hack TB hackathon has a guardian. But unlike the fearsome dragons and greedy leprechauns in folk tales, the keeper of TB data is generous with his treasure, making sure that it is shared for an important quest.
In this blog post, Anny Huang, who took part in the Hack TB, talks to Abdullah Latif, the Data Manager at the National TB Control Program (NTP) Pakistan and custodian of this Big Data, and who has a riddle for you:
What is Pakistan’s sub-national TB burden?
Anny: KIT had already worked with the Pakistan NTP before they approached you with the Hack TB idea, what had you done together in the past?
Abdullah: Since 2017, the NTP in Pakistan and KIT have worked together on multiple research projects. In one study, we assessed access to TB care services, and in another, which is ongoing on the digitalisation of mobile chest X-ray vans, which are providing active TB case-finding in the community. We are looking at the data that is entered through digital means, in terms of how it can help to improve healthcare service planning and management. We are also going to start a MATCH approach to find missing cases.
An: When KIT first suggested the idea of Hack TB to you. What was your reaction?
Ab: We were very excited! The NTP was struggling to find the missing people with TB, and our last prevalence survey was done back in 2010/2011, which provided us with national estimates, but we badly needed sub-national estimates, to help us find more of the missing TB cases. When KIT approached us about Hack TB, we thought it was a great idea.
An: How exactly did you think that Hack TB would help your programme?
Ab: Pakistan is a high TB burden country, and the NTP has done a tremendous job controlling TB, but still we’re missing a lot of TB cases every year. Using the WHO estimates for modelling, and the prevalence survey, we know that we missed around 200,000 cases in 2018. Pakistan is a big country with a population of 212 million, so we needed some direction to find where these missing cases are. I’m very much hoping that with the results of Hack TB, we will have some idea where we can go and invest some money and find these missing cases.
An: During Hack TB, did everything go as expected? Were there any surprises?
Ab: Everything went very smoothly. We had a small delay in getting the prevalence survey data, because it was done eight years ago, and was all in the archives. It was a huge dataset, so it took some time, but otherwise, the rest of the whole process was perfect.
An: We were impressed with how good the data was. Did people have to put in a lot of extra work just to prepare the data for Hack TB?
Ab: Actually, thanks to the Global Fund, we had invested a lot in data management and surveillance. We have maintained TB surveillance data since 2000, and it is very easily accessible. For Hack TB particularly, there was no special request. It was just the ongoing surveillance data, which we had, and also the prevalence survey data. It was very easy for us to provide this.
An: I would imagine that when you did the prevalence survey, it was very challenging to collect all of the data?
Ab: I joined NTP back in 2011, at the end of the prevalence survey, which meant I had an idea about the challenges they had, especially in the collection of data. It was the second-largest survey ever done after China. The sample size was around 133,000 individuals, all of whom were screened for TB. It was a huge amount of data, all of which was collected on paper. Transferring these data into electronic records had its challenges, and it took a lot of time to clean that data.
An: What kind of modelling do you normally do, and did Hack TB add anything to this?
Ab: Yes, we have been using the TIME model since 2016 to see if our national strategic plan to control TB was having the desired impact, and we have had a very good experience working with this model. We are using some estimates from the WHO modelling exercise, which they do every year for their global TB report. We get the data of the incidence and mortality, especially from these models. But we’ve never used these models to look at the sub-national estimates, and Hack TB is the first opportunity we have had to do this, which is very useful for us.
This is something that we want to further improve and then maintain because new data from surveillance are coming in. Some new evidence is available from some other public health organisations and sectors. We are looking forward to updating it regularly, and to use it for the future.
An: Are there any other plans or policies that might be put into place with the results of the Hack TB?
Ab: Actually, we are very excited because Hack TB is something which is happening for the first time. The timing is perfect as the NTP is developing the National Strategic Plan for 2020-2023. We will be able to design our interventions and activities according to the sub-national level results which we have from Hack TB, which can help us find the missing cases, tell us where we need to invest more money, where we need to engage more partners, and where we need to conduct active case-finding activities. The programme is ready to use the results from the Hack TB for strategic planning.
An: There is going to be a new strategic plan, and KIT is going to be helping with new studies. Is there anything else planned for the future? Will there be another prevalence survey?
Ab: Yes, we are planning to conduct another prevalence survey in 2021, but at the moment, we are struggling with the funding. As you know, Pakistan is a low-income country. The population is huge, and the demand is to have more estimates, not only national estimates but also the sub-national level estimates. It will have a huge sample size and will need a lot of effort. With Hack TB, we will have the district-level estimates. We will adopt these estimates and see how they can help the programme because we don’t have the new prevalence survey, we don’t have the sub-national estimates. Hack TB is the only evidence we have at the moment, to plan sub-national activities. We are looking forward to having an active collaboration with KIT and together, we can see how best to use this evidence for future planning.
An: Apart from KIT, are there any other partners you collaborate with, especially if you are doing a big prevalence survey?
Ab: In Pakistan, we have a very strong private sector, they partner with the NTP, and do several activities which the NTP do not. For example, they carry out active case-finding in communities and some preventative treatment, they are very active in the TB care and prevention programme. We are going to share these results with them, and they can plan their activities accordingly. Then, we are also going to engage other organisations in the public sector, which can help us find these missing people with TB. They can also use the evidence for their planning, for example, infrastructure upgrading, and public health engineering.
An: Finally, if other governments were considering holding a similar event, would you encourage them? What advice would you give them?
Ab: Absolutely! Countries like Pakistan –meaning low-income countries– with limited resources and a high TB burden, should try the Hack TB approach. TB control programmes across the world have big datasets, for so many years, that can help them improve their programmes. They should learn from the experience of Pakistan. We have not invested anything. Everything happened free of cost for us. We didn’t put in much effort. We only provided this data to experts around the world to model and to help us find these missing TB cases. They have done a tremendous job! I would definitely and strongly suggest other countries to use this approach. I am sure they will have very good estimates available to them. They will have a very strong evidence-based practice in their programmes.