Skip main navigation

Data as a Public Good: building data systems and competency for AMR surveillance

Brad Cunningham presents how we can build data systems and competency for AMR surveillance.
BRAD CUNNINGHAM: So I really tried to break this into the five major challenges which we see. And they’re not limited specifically to TB, but for any digital health infrastructure. And I’ve tried to frame them from the most concrete and technical to the higher level and more abstract challenges that we find. The first one is really how do you get access to the information, and what information is being produced. With diseases like TB, and COVID, and many of the other AMR tests, generally you’re moving away from centralised testing platforms for various reasons, including a long turnaround time from centralised laboratories for these kinds of tests, especially in low to middle income countries.
And these decentralised instruments do require a different kind of user, not skilled in highly complex laboratory techniques and result interpretation. And so a lot of these instruments often simplify the presentation of the result to make the interpretation easier, which can lead to the loss of some of the meta information that’s useful for disease surveillance, assuming the test is digital at all. We’ve also got rapid diagnostic tests. A lot of these are manual. They’re hard digitised. And sometimes they do require manual intervention. But I think for the focus here, we’re looking mostly at instruments capable of producing in digital results.
The second part of that is really the connectivity infrastructure. So you’ve got your instrument. It’s capable of producing a result. How do you get that result from the instrument to somewhere else? And that’s really what we’re referring to with the infrastructure. And this varies, really, country by country. South Africa, for example, has a WAN, a wide area network, where all labs are connected under the same network to a centralised system. Some countries have a local laboratory network or a LAN local network at each laboratory. Most of the countries that we work in system one rely heavily on the mobile. Mobile penetration in East Africa and Asia has a significant advantage over a fixed line system, such as fibre.
But then you’ve really got– your last mile is really your satellite connections, in terms of where there’s no mobile signal available, you can switch over to satellites. And so relying heavily on the mobile 2G, 3G, 4G, 5G systems, you know, we’ve really found that global sons are making the complex network of the different bands, the different operators, and even selecting which one in the country is much simpler. So we’ll have a local– a bunch of local networks and a global support that can connect to all of those. Counting additional costs, the cost per created by it is significantly higher.
But the robustness and the redundancy that you get out of that solution has really solved a lot of the technical issues with implementing these solutions reliably across mobile networks.
The third one is really looking at the acceptance of the stakeholders. And this is really acceptance from the top to the bottom. So this isn’t just the National Ministry of Health. On the contrary. It’s to say we’re going to use the system. Typically, you’ve got to get a regional buy-in. As well, the National Ministry generally provides recommendations. And then, even below that, you can implement the system in a lab. But if you don’t have the lab staff that actually accepted the system and is using it, you’re not able to generate the right lab dates.
And even down to the clinic level, you could get it this data back to the health care workers and to the clinicians that are seeing patients. And they’re not using this data. The whole purpose of the system kind of grinds to a halt if the data doesn’t get used by the right stakeholders. Now, the fourth one is really one of the more nuanced arguments these days in terms of the hosting of the ownership of the underlying platform, and more importantly, the data. So you’re getting this diagnostic data being generated with patient information, with a bunch of special personal information according to the legal definitions, and really trying to figure out how to appropriately control and share that data.
And then, the last one actually is cost and sustainability. These systems do come with some kind of investment that you often need a Capex– capital investment upfront to get the hardware and get everything installed and running. And there’s ongoing costs to make sure that the system is maintained and constantly being updated.
And so the conversation around cloud security, specifically, really has transformed, lifts into a conversation about security and more into a conversation about data sovereignty. We’re nearing the apex of the market cycle, if we’re not there already, when we move from early adopters and innovators to acceptance of mature cloud based systems. And I think there really is sufficient evidence in the field already that show both the technical and cost advantages of cloud based systems. So I don’t think those are really too much in question anymore of are cloud systems secure and are they more cost effective. I think we are close to moving past that as an issue in general.
And the prevailing issue that we find is really around the control and the ownership of the data, and the implications of that data being shared. Local servers, given their disadvantages for countries that are sensitive to sharing and more sensitive to controlling their data, it does give you a higher perception of control since you can physically access the device. You can walk up to it. You can turn it off. You can remove a hard drive. And it’s a lot more abstract in the system with the cloud system. You’ve got your data sitting somewhere else on somebody else’s infrastructure that you don’t have physical access to. And so that’s where the perceived lack of control comes in with this type of data.
And, really, the most important step to mitigating this and getting past the control and the access issues really comes down to the legislation that needs to be implemented, I mean, that is being implemented globally. We’re really looking at digital privacy laws that are put in place to promote the sharing– the appropriate sharing of information. And this really comes down to different countries acting their own laws or why they cannot be created. So for example, in Europe you’ve got the General Data Protection Regulation. So instead of each country in Europe having their own digital privacy laws, they said, as the economic region they will all share these same privacy legislation.
And I think that would be really helpful in low to middle income settings. It will vary from country to country. And, actually, not everybody is going to accept all things depending on the country’s policy and how sensitive they are to this data being shared. So the legislation, to me, is the first component of that, is making sure there is appropriate legislation that enables the sharing of those data. And then the second part of that, and perhaps a more concrete one, is the data use agreements.
So when you are placing your data onto third party systems or third party infrastructure, your data use agreement really outlines the party’s rights and the ownership, and what they can do with the data, and probably more importantly, what they can’t do with the data. What are the restrictions? What are the limitations? What data do they not have access to, and what data cannot be shared? I think those two components, in terms of the legislation and various agreements, are really key to overcoming these barriers. The other part of that is the sharing of the surveillance data often has significant impacts for the countries.
So if you’re looking at data like COVID or data like Ebolas, once this data gets published that there is an outbreak in the region, you know, this has a pretty significant impact across many industries. You have airlines that shut down, your extraction industries, your oil and gas, your mining operations shut down during these times. And it really impacts the lives of the citizens, the government, and there’s a massive economic impact on the country.
And that’s one of the barriers that really does prevail in these countries, that make it difficult for them to openly share a lot of the information just based on the consequences of that data being automated and public, and without them being able to control what gets sent out.
I actually agree with this completely. I think there’s a definite gap in the industry and in the market for available resources to help programmes understand and develop best practises for data analysis and data sharing. Digital health really has unlocked a wealth of data, which countries have never had access to– not just data at this level, but this amount of data at this level. Once you get one of these systems set up in a country, you can present the Ministry with a spreadsheet with 2,000 columns of different data fields.
Without any guidance on how to interpret the data, and how the data can be useful, it does become very difficult for countries to figure out where do they even start, what is the roadmap for them to follow. So I think we definitely are left with some best practises, and some standards. This is relatively still early in the lifecycle of the field. And so I think best practises on what countries should be doing with the data– what data should they be looking at on a weekly basis, on a monthly basis, on a quarterly basis. What is useful to report on?
Reporting on instruments utilisation, for example, on a daily basis is not nearly as useful as on a quarterly or annual basis. Similarly, you don’t want to wait 12 months to report on resistance rates for diseases like TB. You want to look at them on a much more frequent basis. Then, the second part of that, is because I don’t think people know which parts of that they should be reporting, should be, and how. And I also think maybe this is one of the fields that we’re actually– we’re innovating a little bit too quickly and too fast, without establishing best practises first.
You’ll see in the industry now we’re adding things like AI, and machine learning, and neural networks to those data, looking to collate as much disparate information as possible, and get these systems to figure out what’s useful. And so I think that adds a little bit more confusion to that. So one of the things that I can share is SystemOne, what you do in providing our system to the countries and trying to help countries understand this data. We launched a programme called The Data Fellows. And, basically, the inaugural session of this was we invited 10 officers from the monitoring and evaluation teams of five different countries that we provided the system to.
And it was a pretty intensive 6-month mentorship and course, in partnership with MSH and the Tableau Foundation, where we really took these everyday officers into a fundamental understanding of the data generated by these diagnostic instruments, what the data means, not only to how to interpret that and visualise it, but to then also interrogate that and get some additional learnings from that. And one of the things that we really tried to do under that is not just say here’s a bunch of data. Here’s a bunch of tools. It’s to actually get these representatives from the five different countries to share their best practises and say, now, this is really what our country does with drug resistant– drug resistant microbial data.
This is how we process it. This is a way we’ve processed it. This is what we’ve found useful. So trying to create a platform where countries actually have the ability to share this information and to share these learnings between themselves and each other really became helpful. And the whole purpose of this fellowship is to develop the capacity for a training of trainers models. So all of the 10 M&E officers that we brought to the programme, the intention was that each of them would now go back into their countries and continue to train, and continue the mentorship of local staff, to have a better understanding in terms of the processing of how this data can be useful for the countries.
I think we’ve proven, especially with COVID being on everybody’s minds, we’ve proven that we’re not ready to deal with these type of disease outbreaks sufficiently. And we know there will be huge potential pandemic pathogens that we do need to prepare for. And I think one could even argue that COVID kind of benefited from existing systems that are in the field, many of them established AMR systems. We’re really setting the foundation, and we’ve already been through the pilots on digital health not saying, can we do this. We’ve shown it can be done, and we’ve kind of already laid the foundations for that.
So I think you COVID kind of really took that and the massive investment globally to try and handle the COVID outbreak. It really catalysed that into a lot more innovative systems and a lot more real time systems being employed in the field that can now be utilised again for other infectious diseases. And I think one of the things that really has been fundamental since the COVID outbreak is prior to COVID it was– you have to pay for this type of system. You have to pay for digital health system. You have to pay for a surveillance system. And I think the argument now really has changed too you have to invest in a digital health system.
You have to invest in digital surveillance, just for the cost and the cost benefit that these systems provide. The World Bank shows that the GDP in 2020 has shrunk by $4.4 trillion since 2019, largely due to COVID and the subsequent recessions that the disease has caused. And so I really think that has helped catalyse the need for these systems to be placed into the field.
And I think public private partnerships really are a must for this infrastructure to be sustainable. You can build– you’ve got huge financial incentives with the COVID, with a lot of tech companies and biotech companies providing cutting edge solutions. And these innovations are being deployed to the field as quickly as possible. And these technologies will all be consumed by the public sector. Now, you can generate this fantastic disease surveillance programme. But if not being controlled by the public sector, the programme really won’t survive. And I also think, similarly, you’ve got these tech companies that are building these solutions, and they’re investing a lot of money in developing these.
And they need to find some kind of way to generate a revenue stream to survive as a company. So I think when we start talking about private public partnerships and sustainability, it really does come down to the argument of it needs to be sustainable for both parties. It needs to be sustainable not only for the public health system, which needs to be able to afford the solution and make sure the solution can be maintained and run, but it also needs to be sustainable for these companies that are providing the solution.
In absence of some kind of impact funding and donor funding, which has a limited lifespan, there does need to be this ecosystem that’s created where both parties are able to benefit from these kind of solutions in the field. And I think that’s, to me, the true definition of sustainability is it’s not just sustainable for one party. It has to be sustainable for both.

For most countries that would like to invest in data and information and overall digital health for their health care system, it is likely that they may not have the expertise within the MOH to adequately budget for and drive such initiatives on an annual basis.

Ministries of Health are used to procuring consumables such as diagnostics instruments and assays – but informatics/surveillance/digital health require a different set of skills and expertise.

Mr Brad Cunningham, who has been instrumental in building digital data systems for the tuberculosis programme in South Africa speaks of the key challenges facing governments today with regard to data security and data governance and how countries can build the capacity and competency to have data systems that can be used for pandemic response and antimicrobial resistance.

This article is from the free online

Diagnostics for AMR: Building Back Better from the COVID-19 Pandemic

Created by
FutureLearn - Learning For Life

Our purpose is to transform access to education.

We offer a diverse selection of courses from leading universities and cultural institutions from around the world. These are delivered one step at a time, and are accessible on mobile, tablet and desktop, so you can fit learning around your life.

We believe learning should be an enjoyable, social experience, so our courses offer the opportunity to discuss what you’re learning with others as you go, helping you make fresh discoveries and form new ideas.
You can unlock new opportunities with unlimited access to hundreds of online short courses for a year by subscribing to our Unlimited package. Build your knowledge with top universities and organisations.

Learn more about how FutureLearn is transforming access to education