For those who have followed my posts, I have previously referenced how mortality rates have been decreasing in LA County in spite of a rise in cases. Now public health officials, including Public Health chief Barbara Ferrer are grappling for an explanation.
Before discussing such, however, it is critical to point out that case rates in Los Angeles continue to rise, now reaching a new record of 19.67 daily cases per 100,000 population (Figure 1). Testing rates remain essentially flat with 155.71 daily tests performed per 100,000 this week and 151.69 the week before. An 11.3% increase in number of cases but only a 2.6% increase in testing. Not a good trend.
Coronavirus has led to 3,534 deaths and infected more than a confirmed 116,000 people in Los Angeles County.
As cases increase, why are we not also seeing increased rates of mortality? As seen in Figure 2, the most recent mortality rate is 42% lower than its maximum seen the week ending 4/24/2020. (But, before celebrating this welcome trend – it is critical to note that the mortality rate in Los Angeles remains 19 times higher than that of San Francisco).
Dr. Ferrer trots out the usual explanations for the decreasing death rate – namely that the newly infectious are younger and therefore less likely to succumb to the virus, that spread in nursing homes has stopped (there have been no new cases at skilled nursing facilities in the last 14 days), and that mortality is a ‘lagging indicator’.
“The real factor that is playing into the numbers we are seeing now is that we often have a lag time in deaths, from when we start seeing increases and hospitalizations,” Ferrer said. “Deaths will often lag behind slightly as much as one, two, three weeks.”
But data in Los Angeles directly contradict her point about lag time in mortality. Death rates have been in steep decline for at least 8 weeks, case rates have increased in every week save one (5/8/2020) since March.
More likely what is occurring (and is covered in introductory Statistics) is Simpson’s paradox. This phenomenon occurs when pooled data cause an apparent reversal of the true data trend. The classic example is one of apparent gender bias among graduate school admissions to University of California, Berkeley showing that men were more likely than women to be admitted. However, when examining the individual departments, it was shown that women tended to apply to competitive departments with low rates of admission whereas men tended to apply to less-competitive departments with high rates of admission.
So in the case of COVID-19 deaths, it is possible that mortality rates are indeed rising in every age group – but the overall death rate can appear to be decreasing as the proportion of cases in young people rises and falls among older people.
Los Angeles County does not publish age-specific details of each case, opting instead to tabulate totals by age group to date. With more granular data, it would be possible to better understand the true risk to Los Angeles County.
But, nevertheless, mortality rates are still 19 fold higher in LA than in San Francisco.