By Daniel Messer, Vice President, Technology Integration and Chief Information Officer, PSI
Gen Z might be the last generation to remember always seeing a healthcare provider in person for a diagnosis.
COVID-19 has accelerated the development and implementation of digital health solutions. Future generations might laugh about the way we used to access care.
“Since 2015, 1 billion people have gained access to the internet through a mobile phone – many for the first time. By the end of 2019, almost half the world’s population was using mobile internet (GSMA – The State of Mobile Internet Connectivity 2020).”
While not all communities across the globe experienced this rapid increase in internet access in the same way, digital health is now often accessible where in-person healthcare services might not be. Central to this development is the mobile phone that enables consumers with access to get healthcare services while generating data that allow us, public health implementers, to constantly improve the client’s experience. This includes:
- Service provision itself, that leverages familiar apps and offers signposting to providers
- Access to the patients’ medical records and thus improve the quality of care
- Using AI to power analytics to improve service provision in real time
SERVICE PROVISION: Digital health and the power of the mobile phone
With the spread of cell phones throughout emerging markets, it was only a question of time per when paper-based data collection for Health Information Systems (HIS) would be replaced by mobile devices. In the early 2000s, text message-based (SMS) solutions emerged and today global goods like DHIS2 collect data through Android based apps as a standard feature. While this is great for healthcare workers, healthcare consumers might not want to use their valuable bandwidth to download an app to access health information on topics such as COVID-19.
Leveraging familiar mobile apps — like Facebook and WhatsApp — can help to address this challenge.
Even data collection at the workforce level shows a clear preference for using platforms one already has and knows. A few years ago, PSI enhanced its well-established DHIS2 based malaria surveillance app to include Facebook Messenger to collect test results. The more the regions shifted from malaria surveillance to elimination, the less users remembered the initial, program specific Android app. In this case, Facebook and its messaging tool were already well established, and using a chatbot was an easier way to quickly transmit test results to centralized, government-run HIMS, as well as Ministries of Health, and the PSI programs supporting them.
Using chatbots through known platforms is also an effective solution for data collection and direct consumer engagements, such as counseling and training. Digital health programs developed by PSI during the pandemic use this kind of approach for e-learning (via WhatsApp) and social and behavior change (SBC) campaigns, including signposting to service providers.
However, this is just the first step.
The recent lockdowns accelerated the use of online diagnostic tools like those developed by Babylon and Ada. These tools let consumers directly access diagnostics and triaging by using AI. While digital health will not replace a doctor, the reality is that poorer countries do not have that many healthcare providers across their regions. As we have seen with fintech and mobile money (e.g. MPESA, Orange Money), countries in Africa now have an opportunity to leapfrog others in terms of digital health development. In addition, healthcare products are shifting toward a more consumer powered approach with self-care solutions, like Sayana Press as an injectable contraceptive option and HIV self-testing. In this way, the consumer becomes less dependent on traditional healthcare with in-person visits thanks to virtual assessments and self-care products and services.
ACCESS TO RECORDS: Tracking and serving clients through their health journey
Digital health tools that support diagnosing, triaging and signposting to services using a smart phone or feature phone give us the opportunity to address an old and critical problem. The client’s medical records can be seen by their provider, but often not by themselves. In addition, the medical records do not ‘travel’ well and doctors often may charge a fee to send the client’s records to another provider. Two key problems limiting easy transfer of data are:
- Uniquely identifying a client across time and space
- The wild mix or total lack of interoperable systems
Uniquely identifying clients can be done through manual or system generated codes. The latter might be unique, but cannot easily be re-generated without the original software, data and physical file. To avoid this problem, programs like DREAMS in Zimbabwe and Malawi implemented Unique Client Identifiers (UCI) that are linked to client’s personal data. A relatively high level of uniqueness could be achieved by just including the first two digits of the name coupled with year and town of birth and the mother’s name. This is not an unknown method and many states in the United States use this kind of approach to generate unique driver licenses IDs.
Why is it so important to uniquely identify the client?
The current project-driven funding model often results in ad-hoc based healthcare provision. The client gets a service and often has either no medical record or can only be recognized within the very same clinic and usually only for the same type of service. This is not only a problem from a quality-of-care point of view, but also inhibits follow-up for, say, a second vaccine shot. Some treatments rely on an accurate schedule—and an EMR system can help manage client history and visit reminders. Unfortunately, most clinics that have managed to invest in EMR systems do not all share the same solution. Even if they do, sharing and syncing patient records is still difficult. In addition, some donors insist on a specific EMR system that is often not a global good, like openMRS. The result is limited interoperability between systems and no access by the patient to their own medical history.
However, thanks to the (relatively new) FHIR (Fast Healthcare Interoperability Resource) standard, healthcare information can be easier exchanged between systems. Most systems that follow the Principles for Digital Development, have implemented FHIR and commercial companies like Apple use this standard (for its iPhone Health app). Still, that does not mean that client records can easily be transferred and stored while protecting personal identifiable information (PII) that are covered by various consumer data protection acts like GDPR. One solution might be to push data back to the smartphone of the client and ensure that all EMRs are compatible, supporting the FHIR standard.
USING AI TO POWER ANALYTICS: The power of the patterns
How can consumer-powered Digital Health influence analytics?
The client can now use a mobile phone to:
- Follow social media channels and be directed to a cyber educator or a chatbot
- Interact with a chatbot through a known messaging platform like WhatsApp to get health info
- Get an initial diagnosis or even have a remote session with a healthcare provider
- Get signposted to goods and services
- Access their medical data using a UCI that facilitates sharing of records between healthcare providers and allow smoother follow-up services that can be pushed through various platforms thanks to the FHIR standard
Just implementing some of these possible digital interactions will constantly create data and metadata (geolocation, time, etc.) to instantly analyze the success of your intervention. Traditional monitoring and evaluation methods collect data after the service/activity has taken place, sometimes with weeks and months of delays through manual data collection.
While insights and analytics as we know them will still be important, we now can review large datasets for patterns and causalities that go beyond the capacity of the human brain. A good example is COVID-19, which was already observed as an abnormality back in December 2019 using an AI-driven algorithm that scours foreign-language news reports. This is the power of analyzing patterns and triangulating datasets. The “collective intelligence” approach using the aggregation of highly anonymized search data, chat protocols and related metadata can help us to identify outbreaks and coordinate emergency response.
What does this mean for public health implementers?
Digital Health is not a niche development area anymore, and clearly can help public health implementers to have greater impact. The COVID-19 pandemic has shown how important digital is to reach large parts of the population and workforce during lockdowns and supply chain disruptions.
The opportunities and tools already available are exciting but, to drive health system change, we will have more success if we better coordinate our activities across the Global Health space. Thanks to guidelines like the Principles for Digital Development and standards like FHIR, we can solve challenges like interoperability and scale our interventions.