At the Johnson & Johnson Center for Health Worker Innovation, we believe that in order to effectively strengthen community health systems, including closing the 18 million health worker shortage, we must prioritize the needs of individual health workers and what they need to thrive on the job.
Those needs include ensuring that health workers have access to reliable data to inform how they deliver care, and in turn are able to provide patient data through simple, intuitive, and effective data collection processes.
A recent survey shows that a majority (56%) of a frontline healthcare worker's time is spent accessing and updating patient records instead of caring for patients. Community health workers, who are often already undersupported in their primary caregiving roles, must take on collecting, inputting, and verifying patient data. This additional burden can impact health workers’ ability to fully deliver on their caregiving responsibility, the quality of services they provide, and ultimately, the reliability and accuracy of the data collected themselves.
To address this, we’ve partnered with DataKind, a global data science nonprofit organization providing pro-bono data science, machine learning, and artificial intelligence innovation to social organizations. Through its Impact Practice on Frontline Health Systems, DataKind gives organizations the tools to evaluate the integrity of the data collected by frontline health care workers, improve their data collection systems, and empower health workers to be part of this process.
To learn more about how increasing the accuracy and effectiveness of data collected on the front lines actually supports a thriving health workforce and increases trust in community-generated data, we asked Mitali Ayyangar, Portfolio Manager of Frontline Health Systems for DataKind, to answer a few questions.
Q: Why has DataKind prioritized assessment of data quality among frontline health care workers as a key target area in its Impact Practice?
A: Frontline health workers around the world provide critical services and care in areas where other health infrastructure is often lacking. Through the use of digital health tools and mobile applications, millions of patient interactions are logged every year and yield large volumes of data. This data could potentially drive more effective national health policies, differentiated health service delivery, and improve program management and evaluation.
A common pain point in the sector is that the data collected at the front lines of care are still considered unreliable for data-driven decision making. Without trust in frontline health data, how can health systems managers and policymakers make informed decisions about public health priorities and invest in them? And how can healthcare workers themselves grasp opportunities to optimize their care delivery and patient outcomes? We believe that tools and systems that promote responsible data collection with rapid remediation pathways to improve data quality offer a turnkey solution that will go a long way in creating trust and transparency across health systems. These solutions will benefit multiple frontline health organizations, giving our work a pathway to supporting genuine sector-wide change.
Q: How will strengthening data integrity processes benefit frontline health care workers and the patients they serve?
A: I’ll start with a quote from the UN Statistical Commission that stated unequivocally in March 2019 that “Every misclassified or unrecorded [maternal or neonatal] death is a lost opportunity to ensure other mothers and babies do not die in the same way…when it comes to health, better data can be a matter of life and death.”
Frontline health workers collect a wide range of data in the communities on maternal and child health, infectious diseases, family planning, noncommunicable diseases, and much more. Collecting data responsibly, surfacing problematic data, and following pathways for data remediation helps frontline health workers organize their work and monitor their patients more efficiently. This process makes the data that frontline health workers collect accessible, informative, and useful, enabling them to provide the right care or intervention at the right time, which is the most significant action health workers can take to serve their patients and reduce deaths from preventable causes.
Q: How does DataKind involve frontline health workers as part of the data integrity and assessment process, and why is this so important to do?
A: At DataKind, our primary principle is to design with, not for, and where possible we aim to work with the end users of our products throughout our design process. While it has been harder for us to work directly with health workers right now due to COVID-19, our partners use every opportunity to collect input and feedback from health workers. This enables us to develop data quality assessment tools that can be as easily and seamlessly integrated into health workers’ routine care as possible.
Q: How does data integrity strengthen community health and health systems overall?
A: High-quality and trusted data can increase the utility of that data and become a lever for sector-wide innovation to improve global health outcomes. For instance, high-quality patient data can be aggregated for community-level analysis and improve the practice of making decisions backed by trusted data. It can also open up pathways for collaboration between national, sub-national, and community health systems and provide insights into complex health service delivery questions. This creates continual opportunities for improvement and better chances of having public health policies and investments match actual needs of patients and communities.
And finally, frontier data science tools like machine learning and AI are essential for the transformation of health systems. To realize the potential of AI for precision public health and predictive analytics to anticipate, say, the next health emergency or to prevent bad outcomes, data integrity is a foundational pre-requisite.
Q: What key accomplishments have been achieved in assessing the quality of data being collected on the frontlines to date—and what impact have they had on health workers?
A: DataKind has achieved some key milestones in developing automated data quality tests. In a pilot exploration which was generously funded by the Johnson & Johnson Foundation, DataKind built an automated data testing pipeline on Medic Mobile’s platform called the Community Health Toolkit. We were able to scan maternal and child health data generated by over 2,000 community health workers and rigorously test for more than 160 different types of data quality problems. With this machine level intervention, Medic Mobile and their partner in Kenya, now require less human capacity to monitor those data streams while simultaneously gaining more trust in the data. Now, DataKind and Medic Mobile are moving forward to the next phase of development, which is a closed beta to test this toolkit in different contexts and on different types of health data. We’re also working to transfer this framework to our new partner Lwala Community Alliance, an organization that uses Dimagi’s CommCare platform. This is the most widely used platform by frontline health workers in the world.
For more information on the DataKind/Medic Mobile pilot, see their recently published case study: "Engineering Scalable Data Quality Assessments for Frontline Health with Medic Mobile."