Page 9 - EngineerIT September 2021

ICT DATA

Modelling a pandemic shows the

importance of data from multiple sources

By Dr Mark Lambrecht, Director of Global Health and Life Sciences Practice at SAS

t has become evident that one data set If anything, the pandemic has
often reveals a piece of the puzzle, but highlighted the need to have access
Iwhen linking multiple data sources, a to even more data and advanced
much bigger narrative becomes visible. technologies that can analyse that data
According to media reports, as effectively and quickly as possible.
analysed data for South Africa suggests The reason for this is clear even as
that the current wave of infections (wave South Africa has recently seen the
three) will have ended around the end of decline of the third wave of infections –
August 2021, but that estimates indicate there remains continued concern about
the fourth wave could start around 2 shortages in hospital resources including
December – comparable to wave three hospital beds, ventilators and personal
with an expected new variant of the protective equipment.
virus by then and lasting about 75 days. Data science and the associated
This kind of reporting demonstrates how modelling can assist government health
data science has played a crucial role in organisations and hospitals predict patient
managing the impact of the COVID-19 hospitalisation rates and plan resources Dr Mark Lambrecht
pandemic. accordingly. Throughout this, partnerships
The practical application of between public and private sector entities about how COVID-19 spreads, and
theoretical knowledge during these are crucially important. The former can harnessing insights to prevent or limit the
times of crisis has resulted in everything provide an enabling environment to effects of future pandemics.
from improving situational awareness to effect change, while the latter can deliver Epidemiologists track disease
conducting epidemiological modelling, on the technology and data analytics outbreaks using statistics. The likes
performing network analytics, and even requirements critically important to inform of cumulative frequency graphs and
holding randomised controlled trials. decision-making processes. exponential growth curves have been
It remains important to keep using Yes, governments hold much of shared widely to help visualise the
different approaches and to constantly the critical data needed to understand growth of the disease and understand
incorporate new information so that current conditions during an outbreak. when the growth might be peaking.
forecasts are always as accurate as However, it is analytics that provide them And then mapping dashboards show
possible. The appearance of new with the ability to synthesise this data where cluster outbreaks are occurring.
mutation virus variants is notably with other non-health (social indicators) This empowers government and health
difficult to predict and could impact the and non-governmental data to get the officials to develop area-specific models
assumptions of each model – pointing most insights from this unified data. and dashboards that help allocate critical
to the need for an agile data science Additionally, analytics can deliver insights healthcare resources. This can also
environment that can robustly work with about the spread of a disease and the significantly assist when it comes to
new information. effectiveness of public health action, region-specific lockdown regulations and
which can improve the response. highlighting hot zones where infections
Linking multiple data sources Of course, analysing data is a are high.
When linking multiple data sources, complex process that is significantly
the data linkage becomes critical impacted by the rate at which it [the data] Common goal
to also understanding the social is growing, especially given how quickly The pandemic has shown how real-
determinants of health, amongst other the pandemic spread globally. Given the world data science entails being part of
things, which inform effective policy life and death scenario that has played interdisciplinary teams. It is not about the
and intervention. Simply put, data and itself out numerous times over the past beauty of an algorithm or the choice of
its analysis have made a discernible +/- 18 months, data analysis must be programming language. It is about having
difference in how society reacts to this held to a higher standard to decipher it a concerted will, a unified mission and
crisis. It is helping governments and and glean insights to stem the flow of the a common goal to analyse data across
healthcare organisations around the pandemic. Therefore, to overcome the environments to help improve responses
world to find more effective ways to crisis, much comes down to collecting and provide guidance on where best to
deliver better outcomes. data, using it to help understand more allocate resources. n

EngineerIT | September 2021 | 7