scroll

STRIATA by Macro-Eyes Maximizes Monitoring & Evaluation with Artificial Intelligence Innovations

Jun 4th, 2021
Jun 4th, 2021

Artificial Intelligence (AI) offers our world so much more than robots, google, AR goggles and robotic limbs, especially in the field of global health. Among the many of its applications, AI is poised to revolutionize the way we interpret health data, visually display those insights, and assist leaders in making better, data-driven decisions for maximized global health impact.

For me, nothing has made this clearer than working on the Macro-Eyes STRIATA initiative, hand-in-hand with dedicated leaders who support technological innovations from the Ministry of Health and Sanitation (MOHS) and the Directorate of Science, Technology and Innovation (DSTI) in Sierra Leone.

Today, many governments lack adequate real-time understanding of their health system readiness and what is currently at health facilities, such as infrastructure, human resource, materials or supplies. Dashboards and central repositories must be built and maintained; they often lack consistent or current data. Ministry leaders receive recommendations only to discover it is based on outdated sampling of data with little visibility into what is happening today. Data collected today rarely includes predictive analytics that offer a more comprehensive view of the whole health system.

STRIATA changes this.

Built as an interactive tool to measure health system capacity for COVID-19 response in Sierra Leone, STRIATA takes into account a multitude of different types of data: GoSL health, publicly available information/news and satellite imagery. STRIATA gives Ministry leaders deeper and more comprehensive visibility into their health systems to optimize scarce resources for the greatest impact, using a proprietary system of predictive analysis.

I believe AI — with STRIATA as a demonstrative and data-driven success story example — is primed to change the way global health data is collected, analyzed, visualized and utilized for empowered decision making. I see AI continuing to move to the forefront of data interpretation, because AI:

  1. Uses more types of data than traditional M&E.
  2. Analyses data faster than any team of M&E specialists.
  3. Processes imperfect data and large data sets to provide insights.
  4. Promotes an equitable data approach, for visibility into all levels of a system.
  5. Visualizes data insights in real-time.

To delve into more detail on these five functions of AI:

AI USES MORE TYPES OF DATA THAN TRADITIONAL M&E. THIS RESULTS IN DATA ANALYSIS THAT IS MORE COMPREHENSIVE AND DRIVES MORE INFORMED DECISION MAKING.

Data is critical to a continuously learning AI system. STRIATA incorporates more data than what is routinely collected, including:

  • Publicly available data. Macro-Eyes developed a web-scraping tool that continuously collects data about relevant communities
  • Satellite imagery in both low and high resolution
  • Natural language such as text messages from health workers
  • School and road data as well as other non-traditional data sources

AI’s ability to look across myriad data types impels new associations between seemingly disparate points, with a more precise context about operating environments. For instance, localized heavy rains knocked out a bridge crucial for community access to health services. This reduced access to a health facility and disrupted routine immunizations for a significant period.

STRIATA synthesized data from satellite images and the web to predict the readiness of that facility, within that timeframe and in real time: the regional office obtained data that might normally require a site visit or manual comparison of data points. AI synthesizes the complexity of the data to provide multi-layered insights. STRIATA becomes a trusted advisor that informs decision makers in cogently concise, understandable layers of data that encourage empowered decision making.

AI ANALYSES DATA FASTER THAN ANY TEAM OF M&E SPECIALISTS, RESULTING IN MORE TIME FOR DECISION MAKING, IMPLEMENTATION AND REACHING THE POPULATION IN GREATEST NEED.

In Sierra Leone, STRIATA was built in 6 months. It provides COVID-19 readiness scores for every health facility in the country; the current processes normally take at least a year. STRIATA provides Sierra Leone leaders with more complete data — and with more time. That means better insight for critical decisions. STRIATA quickly reveals blind spots where more data is needed, while using the available data to highlight how entities can improve COVID-19 readiness scores.

AI PROCESSES IMPERFECT DATA AND LARGE DATA SETS TO PROVIDE INSIGHTS, FOR A MORE COMPLETE OPERATING PICTURE EVEN WITH SMALLER DATA SETS.

Contrary to popular belief, AI does not not necessarily need large or perfect data sets. STRIATA in Sierra Leone proves that even the smallest data set has ‘signal’ to generate useful insights for government decision makers.

In Sierra Leone, STRIATA used 8 different data sets with over 600 data points to generate health facility readiness scores. The STRIATA system received small and incomplete facility data sets yet predicted COVID-19 readiness with less than 15% error for all health facilities in Sierra Leone, while generating insights on score improvement for infrastructure such as water supply, personnel, environmental disinfectant and number of beds.

AI is masterful at imputing missing values; this is invaluable for small or incomplete data sets. Imputed values ensure that analytics maximize insights for improved predictions.

AI PROMOTES EQUITABLE DATA APPROACHES, FOR VISIBILITY INTO ALL LEVELS OF A HEALTH SYSTEM. A BETTER UNDERSTANDING OF THE ENTIRE HEALTH SYSTEM ENABLES A MORE TARGETED NEEDS ASSESSMENT.

At both national and district levels, a health system often times visualizes aggregated data from multiple interfaces and software systems; it requires tedious integration with gaps that are murky and contradictory.

Correcting this deficit is how Ministry leaders can make the most impact.

STRIATA powers AI algorithms to synchronize this data into compelling visibility into all levels of the health system. This supports Ministry leaders to see, prioritize, and decide. The user easily zooms from country to facility, to focus on the specific aspect currently under review. With this comprehensive, multi-layered view, leaders can design and implement programs with an equitable lens; this ability to see blind spots clearly and understand priorities can drive the greatest health impact for their populations.

AI VISUALIZES DATA INSIGHTS IN REAL-TIME, TO PROVIDE REAL TIME VISIBILITY INTO YOUR HEALTH SYSTEM TODAY AND TOMORROW AS PREDICTIONS. DATA NO LONGER REFLECTS THE PAST.

Traditional assessments rely on data that reflects the past. Health system appraisals such as WHO SARA surveys take a long time and offer a static snap-shot, all while the world evolves. The act of publishing a report dates the data. STRIATA revolutionizes the lag time between data collection and visualization; it updates daily through APIs, web scrapers, text messages and other input methods for a real-time operating view.

STRIATA demonstrates how AI processes data to make the past relevant for today; today applicable to the future, and the future more predictable for maximal utilization of resources and optimal resolution of global health challenges. AI technologies for health are revolutionizing the way traditional M&E for global health is conducted and I am thrilled to be scratching the surface with STRIATA in Sierra Leone.

The STRIATA Dashboard delivers Health System Readiness Scores via powerfully striated views (Fig. 1). This interactive interface:

  1. Anchors the data with geographic representations that zoom from national and regional boundaries to site-level views. This encompasses satellite imagery, roads, borders, hospitals, labs and other infrastructure.
  2. Populates with data derived from a multitude of sources and categorized into the significant features that most impact the final readiness scores.
  3. Shows data in both tabular and graphical output for variable granularity and insight.
  4. Calculates a grade from ‘Most ready’ to ‘Least ready’ plus ‘More info needed.’ This gives eagle-eye accuracy and focus on capacity.
  5. Supports focused thinking through smart filters.
  6. Creates interactive overlays for varying health facilities in geographical context.

For a more detailed description of the Sierra Leone success story, download our newly-released STRIATA Sierra Leone Report.