Global COVID-19 Death Rates & Access to Healthcare
April 2020 Data Visualization
Do more doctors mean decreasing death rates?
Disparities in healthcare within the USA and across the world have been of interest to me for a long time, and especially so in the COVID-19 era. Ventilators, hospital beds, and front-line healthcare workers are all hot topics in the news about the COVID-19 pandemic and how it’s being fought around the world. However, there is little evidence that going to the doctor can help you survive a COVID-19 related illness. With so many uncertainties at play, this article seeks to depict a relationship between access to healthcare and the change in COVID-19 death rates over time.
My hypothesis is that the higher the access to healthcare in a country, the more the death rates will decrease over the course of April 2020, and visa versa (the lower the access, the more the death rates will increase).
How do you measure this?
In order to test this hypothesis, I collected data from the daily case reports published by Johns Hopkins University (JHU) and calculated the death rates as the number of deaths / the number of confirmed cases in that country. I want to acknowledge at the outset that there’s a margin of error due to the fact that not everyone is tested, so it’s likely that there are more cases than are confirmed, and thus lower death rates across the board. However, this article charts the change in death rates over time.
“Access to healthcare” is measured as medical doctors per 10,000 people based on 2016 World Health Organization (WHO) data. Because people aged 65 and older are at much higher risk of dying from COVID-19, I have also included the percentage of the population that is 65 or older from 2018 World Bank data. I selected the countries based on a matrix of factors analyzed in a March 26, 2020 snapshot. Low, medium, and high were chosen for access to healthcare, cases overall, death rate, and recovery rate.
The Big Picture: 13 Countries, 1 Month
Here you can see the way the rates changed over time for all 13 countries, next to a table with the numbers for healthcare access, at-risk population, and death rate. The range of death rates is between almost 0 and 14%, with most death rates falling at an under 6% death rate. Below, I zoom in on low, medium, and high access groups and then compiled one comparison of a country from each group.
Group 1: Low Access & Risk
Death rates here dropped in Ghana’s case, rose slightly in Burkina Faso, and stayed mostly consistent in India and Peru from the beginning to the end of April. Based on this group, you might be able to say that less access does NOT mean rising death rates, but this could also be explained by the smaller at-risk population.
Group 2: Medium Access & Risk
Compared with group 1, more access to care did not correlate with more decrease in death rates. Both Poland and the US have rising death rates, although both have substantial access to healthcare, but this could be explained by the higher percentage of at-risk population. The other three countries stay consistent. Anecdotally, it is interesting to compare Saudi Arabia — 3.3% at-risk population — and Singapore —11.5% — because they both have around 23 doctors per 10k and consistently low death rates, but Singapore has almost 4 times the at-risk population.
Group 3: High Access & Risk
Compared to group 2, the death rates are higher overall and all of them increase over the course of the month. The death rate rises most in Sweden, where access to healthcare is the highest in the world; that 20.1% of their population is at risk might be an explanatory factor. This group is the clearest indicator that my hypothesis is wrong.
Group 4: Comparison
Here, I compare the lowest, middle, and highest access to healthcare of the 13 countries. The death rate increased the most where there was the highest access to healthcare, in Sweden. South Korea has 7x more people at risk than Burkina Faso, and 40x more doctors, and yet both show death rates staying consistent throughout the month. Again, the hypothesis seems wrong.
Conclusions and Future Considerations
In conclusion, I could not prove my hypothesis that death rates and access to healthcare are correlated by looking at the changes over the course of April. To do a more comprehensive analysis, I would need 6 months to a year of case reports, as well as data on other variables such as diet, population density, pre-existing health conditions, use of public transportation, smoking, etc. It may just be, as one doctor commented upon reading this article, that COVID-19 is not “doctor dependent” because there’s no set medical treatment. As with most things COVID-19, the real answer is to wait it out. So, stay tuned for the May 2021 version of this article.
