You are only as good as your data… or rather, your interpretation of available, quality data. Witness Donald Trump’s clever use of regional data by coming from “nowhere” to win the US election, then twisting and destroying Covid-19 data, while flouting public health rules and costing Lives. Such is the arrogance of politicians. Not all, mind you, just like theologians. Archbishop Desmond Tutu’s popularization of “ubuntu” – aided by the virus itself – reminded us of our inescapable human connection, and that a person is literally the sum of those around them. Here, four local academics with complementary skills and expertise, explain the collection, management and distribution of AS epidemiological data; in particular, how we have performed since the time Covid-19 landed. Hindsight reinforces the precision and relevance of this analysis, allowing a better response the next time. Data transparency and context are essential if we are to avoid the pervasive misinformation and manipulation we have experienced, they point out. Story courtesy of The conversation. –Chris Bateman
Covid-19 lessons: Scientists without quality data are like unarmed soldiers in a war zone
Some of the most important public health lessons from the Covid-19 pandemic relate to how government should share data with the public, how updates on responses should be clearly articulated, and the importance of sharing of knowledge, information and all relevant data are accessible to the public.
The pandemic has brought these issues to the fore. But the challenge extends beyond the borders of Covid-19 to all diseases.
Mistakes made during the pandemic in data collection, management and dissemination need to be acknowledged. And lessons need to be learned and shared about navigating public health data effectively.
We watched the effectiveness of containment in South Africa and how the data was used during the pandemic. We concluded that data collection and dissemination could have been much more efficient. And that if that had been the case, it would have determined better results.
For example, if more detailed localized data had been publicly available across the country, it would have been possible to quantify and compare the spread of the disease between cities, towns and rural areas. This, in turn, would have meant that those making political decisions were better informed.
Our analysis and conclusions underscored that quality data is the cornerstone of good science. Without it, scientists tasked with informing the public about vital public health issues are like unarmed soldiers in a war zone.
We cannot overemphasize the importance of epidemiological data and their relevance in managing the early stages of an outbreak. However, as a disease progresses, the underlying data and reporting must also improve to manage the progression of the epidemic.
Information sharing is not just about sharing data with the public.
Take the question of aggregated reports. Limited inferences can perpetuate public biases. Aggregated reports present data in a way that shows a cumulative count or a chronological progression of the total sum of data. Those World Health Organization graphics are a good example of good and bad practices. Good because the data is shared, bad because only one variable perspective is shared at a time.
Another challenge is that the underlying data is not easily accessible to other scientists. Thus, even if comprehensive and well-presented epidemiological reports are published by South African authorities National Institute of Communicable Diseases (NICD) and it now has a very usable dashboardthe underlying data is not available for further visualization or analysis by others.
Another problem with aggregate reports is that they summarize nuances and public health interventions and changes over time. This includes things like changing patient monitoring guidelines, introducing a new treatment regimen (as was the case with HIV/TB) and innovative clinical monitoring strategies.
Members of the public should have comparisons of the state of the current outbreak to previous outbreaks of a similar nature. This would be contextually relevant and could help people evaluate information as well as data and move towards evidence-based decision-making.
Timelines can be adjusted from these dashboards. But the way the data is presented means it’s difficult to contextually compare different infectious disease outbreaks (or clusters of outbreaks of a specific disease) and the impact on the healthcare system.
Reflect changing realities
Epidemics are not static. A sickness may lose epidemic status and become endemic, as it becomes a constant and more predictable presence in a particular location. For example, the contagiousness and harmfulness of a disease can change as a result of an actual intervention, such as an effective vaccine or effective non-pharmaceutical interventions.
In the early stages of an outbreak, there are three main data points that are useful to everyone and should be shared regularly: time, location and number.
Typically, after any outbreak, the government or health authorities take steps to share baseline data and infographics with the public that purport to support the interventions they may recommend.
This was the case during the Covid-19.
Corn We have identified some immediate problems with this approach.
First, much of the information is published only in formats such as infographics that are not computer readable. This makes further analysis impossible without research groups and members of society manually transcribing, collecting and sharing data. This causes a trust issue with the data: there can be multiple sources of the same information and the process is error prone.
Second, data shared over time and subsequent visualizations became less frequent (in the case of data sharing) and remained aggregated (in the case of dashboards and infographics). An unfortunate consequence was that there was no transparency or clear correlation between the underlying evidence and the decisions made.
So how can public health decision-making stop being treated as a state secret? Aren’t there just ways to openly share the required data and create platforms to engage with the numbers?
We think it is indeed possible.
The path to follow
Disaggregated data. In a country with inequalities like South Africa, aggregate data can mask the disproportionate effects of an event on specific communities. Making disaggregated raw data available can enable evidence-based advocacy and interventions to more effectively address the needs of marginalized communities.
Accessible data. Information should be shared with the inclusion of indices, measurements and simplified machine-readable data types. This would allow for wider use and add a layer of transparency. It would also create an opportunity for community-led monitoring and evaluation outside of government.
Choose the appropriate visualizations. We strongly recommend representing data as a relative number (in other words such as percentages or by population size) in addition to absolute numbers. It would make it more accessible. Ordinary citizens could better understand where things stand and how they are developing. It would also help inform any changes they might choose to implement to keep them safe.
Additionally, previous outbreaks of a similar or identical pathogen should also be displayed. This would allow people to contextually assess similarities and differences at a glance. Here is a good example.
Flaws to overcome
Covid-19 has revealed the fragmented way in which data is published and how insufficient data sharing can be if it is not done locally.
In some cases, data quality issues also undermine public confidence in the system. Trust is also affected by how often data is shared. Time and date inconsistencies for data sharing seem to be a universal problem. This breeds public distrust.
Finally, the information shared should not only support “good news”. Negative data – such as side effects of a particular treatment regimen or medical intervention – should also be shared.
From Covid-19 we have learned that there are multiple opinions around a specific issue. Some of these opinions have been misinformed. But one cannot blame the uninformed when information important for decision-making is not freely and easily accessible. Without the required supporting information, citizens will continue to make assumptions or believe misinformation and disinformation that is not based on evidence. Their spread may be unavoidable. But lack of access to quality data is not.
Nompumelelo Mtsweni, Data Visualization Developer, Elizabeth Cornelia Greyling, Head of Strategy at Columbus Stainless, and Emmanuel A Simon, Digital Strategy Consultant, also contributed to this article.
(Visited 1 time, 1 visits today)