Was it possible to predict the areas most vulnerable to the negative impacts of Covid-19?
Stuart Redding, Catia Nicodemo
22 October 2020
COVID-19 Policy & health systems Research methods & EBM
Stuart Redding and Catia Nicodemo, from the Centre for Health Service Economics and Organisation, describe a simple metric that predicts which English CCG regions are most vulnerable.
The Covid-19 pandemic is affecting the work, leisure activities and health of people across the country in all sorts of ways, but you can’t turn on the news without seeing that some parts of the country are more affected than others. Was it possible to predict that certain areas might have been more vulnerable to the spread and consequences of Covid-19?
The short answer is yes. In our recently published BMJ Open paper, co-authored by members of the CHSEO group with scholars from Oxford, Brunel, Birmingham, Berlin and Venezia, we create a simple metric that can be used to show how.
There is a large amount of publicly available data for local areas in England. Our study includes 15 variables containing information on population characteristics, the local prevalence of specific health conditions, and local health care resources, which we combined into one measure that allows Clinical Commissioning Groups (CCGs) to be compared with each other and their relative vulnerability identified. It is important to note that none of the variables we include had any information about Covid-19 or previous infection rates, and all data were from pre-pandemic times so our measure was not influenced by the disease or any actions undertaken by policymakers early in the crisis.
The most notable result from our study is that of the 30 English CCGs that we deemed most “vulnerable”, 24 are in the North of England. This is mainly due to these areas having older populations and populations with more underlying health conditions. Their hospitals also appear to be less-well resourced to cope with the extra pressures that will occur as a result of Covid-19, with tighter labour markets meaning they have less staff on average than hospitals elsewhere in England.
We deliberately didn’t include social deprivation amongst the variables included because we didn’t want to prejudge whether wealth or income would have a positive or negative impact on vulnerability. However, when plotting our index against the Index of Multiple Deprivation, it is clear that the areas we consider to be “vulnerable” to the effects of the pandemic are also much more likely to be deprived rather than affluent. This is obviously a concern, because it means that the worse off communities are likely to take a worse “hit” as a result of Covid-19 and this will inevitably cause inequalities to grow even greater.
One key experiment we performed was to check whether our measure was a good predictor of mortality associated with Covid-19. Our paper was published before even the first wave of infections was over, but studying mortality until the end of May 2020, we find that our measure was positively correlated with Covid-19 mortality. More than that, it was a better predictor of mortality than the Index of Multiple Deprivation, so there is more to our results than just replicating a measure that already exists. We plan to perform more analysis using our variable to predict further outcome measures as the pandemic continues, so keep an eye on this blog if you are interested in this work.
With Covid-19 cases on the rise again, consulting the Index of Vulnerability would help policy-makers to focus their efforts on helping the areas which most need support. We hope that our work can be useful in helping to give policymakers a quick and easy method of highlighting at-risk communities in the current crisis and when other undesirable health shocks occur in the future.