Naveen Gattu, co-founder at Gramener, explains the role data can play when managing capacity during a pandemic.
Healthcare systems across the world have been taken to breaking point as a result of the fast-paced spread of COVID-19. The virus has been contracted by millions of people and has taken hundreds of thousands of lives, putting healthcare professionals and hospitals in an unprecedented situation.
Overcoming crises involves well thought-out strategies that are backed by science in order to make tough decisions in a short span of time. With such novel data on the virus and the nature of its spread that changes at breakneck speed, the decision-making process should not be driven by past experiences, nor a “gut feeling,” but by data.
It’s not yet clear how long it will take to find a vaccine – if at all. Limiting the spread of this highly infectious virus will help to ensure that hospitals don’t become completely overwhelmed. Achieving this will require an understanding of demand and capacity management for hospitals, and the availability of personal protective equipment (PPE), ICU beds, ambulances, and numerous other vital elements of the fight against COVID-19.
Armed with data, hospital administrators can make evidence-based decisions around capacity management. However, a wholly data-driven way of doing things will come only when data as a culture has been achieved throughout the entire healthcare system. Let’s dive deep into how data can help hospitals understand their demand and capacity, and how nurturing a data culture is essential to achieving this.
Supply and demand management tools
By using simulation and modelling tools, healthcare professionals can estimate the best possible outcomes through to the worst-case scenarios, of a given situation.
This illustrative simulation model has been generated using population statistics and bed availability by region in the UK. With this, healthcare providers can determine which regions are ready to handle outbreaks and redistribute vital resources like hospital beds, ventilators, and PPE kits to the areas that have insufficient supplies.
Gramener’s visualisation shows the availability of hospital beds across the UK: The NHS has 96,166 available beds across the country. This kind of management tool is especially useful for meta use on the part of the government because not all regions are uniformly equipped to handle a surge in cases. By plotting out hospital capacity on a map, healthcare and government leaders can immediately spot the regions which could face a crisis of bed shortages.
For example, if 1,500 people per million from the age groups specified in the visualisation contracted the virus, the East Midlands and London would be left with a shortage of beds. This number can be adjusted according to calculated trajectories, and the shortage of beds can be estimated across regions.
In another scenario, if people aged over 50 contract the virus at thrice (3,000 per million) the rate than the under 50 age group (1,000 per million), there would be a total shortage of ~6000 beds in the UK – highest in the East Midlands (~1750 beds).
It’s not just important to do analysis, but equally important to present this in a consumable manner for non-technical users to use at ease and to make decisions. Improving the data literacy of hospitals will enable smarter data decisions and realising the value of model outputs. According to Gartner, by 2023, data literacy will become an explicit and necessary driver of business value, demonstrated by its formal inclusion in over 80% of data and analytics.
Implementing a data-driven culture for decision making
Collecting the data necessary to make informed supply and demand decisions often means going beyond the data that’s obtained internally, within hospitals. For example, this information would need to come from third-party actors such as logistics companies, manufacturers, and government agencies. When an organisation has integrated data as a culture, there is a widespread understanding for the necessity of this data in order to power decisions.
Keep data centralised, and decisions localised
Over the last couple of decades, chief medical information officers (CMIOs) and healthcare chief information officers (CIOs) have been reacting to the growing need for a data supply. By investing in data warehouses, data lakes, data transformation, and data cleaning, these healthcare leaders have laid the crucial foundations.
The primary challenge they face at this moment is driving data value. To achieve this, healthcare IT professionals must create a data infrastructure where multiple analytical models are run on past historical data with minimal latency. This is as well as providing an intuitive user experience to healthcare users and an environment where decisions are driven by local teams working within a data culture.
The importance of visual narratives
The cohesive consumption of data, even by the least data-literate members of staff, is needed in order to firmly embed data culture within healthcare institutions. Humans relate to storytelling and visual narratives much better than lines of numbers, so in order to make the insights derived from AI-driven analytical models accessible for everyone across an organisation, they must be delivered in the form of well-crafted, visual stories.
In fact, human brains process visual information 60,000 faster than they do text, and studies have shown that stories stimulate more parts of the brain than statistics do. Data narratives allow non-technical users to leverage data when making important decisions, as it becomes something they can understand, appreciate, and action without requiring knowledge of data science.
Data scientists should focus on presenting model outputs in strong narratives to clinicians that need simple narratives to make decisions. These narratives help clinicians to understand patterns from data in a more simplified manner and thereby convince others of their decisions, as well as providing other benefits like shortening discussions, making data more memorable, and speeding up processes.
Data tools can simplify user experience
It’s not only decision-makers that need digestible insights from data-crunching. For example, front-line doctors could use AI-powered algorithms to help them detect COVID-19 cases based on past data, or allow them to better differentiate between the coronavirus, pneumonia, or tuberculosis.
In addition to leveraging AI and visualisations to make data accessible, healthcare teams with a strong data culture can also begin to explore the possibilities of new tools that provide accurate results and save them time. However, in order for clinical staff to adopt these tools and use them to their full potential, the interfaces must be user-friendly and intuitive, so as not to add extra effort to their already high workload.
While the worst of the crisis may be over in some countries, we – and healthcare teams across the globe – are far from out of the woods. With a strong chance that COVID-19 will remain a concern for populations the world over until a vaccine becomes available, it’s vital that hospitals and healthcare decision-makers champion data culture and leverage data insights to best manage the situation.