Data Science x Public Health
This podcast discusses the concepts of data science and public health, and then delves into their intersection, exploring the connection between the two fields in greater detail.
Data Science x Public Health
Everyone Uses Case Fatality Rates… But They Fail When Detection Is Unequal
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
Case fatality rate is one of the most commonly cited numbers during outbreaks and health emergencies. It seems to offer a direct answer to a simple question: how deadly is this disease? But what if the rate is being shaped less by biology and more by who gets detected as a case?
In this episode, we break down why case fatality rates often fail when detection is unequal, how testing and surveillance distort the denominator, and why context matters when interpreting outbreak severity.
👉 Enjoyed the episode? Follow the show to get new episodes automatically.
If you found the content helpful, consider leaving a rating or review—it helps support the podcast.
For business and sponsorship inquiries, email us at: 📧 contact@bjanalytics.com
Youtube: https://www.youtube.com/@BJANALYTICS
Instagram: https://www.instagram.com/bjanalyticsconsulting/
Twitter/X: https://x.com/BJANALYTICS
Welcome to today's deep dive. We've got a stack of your sources today: medical articles, epidemiological data, research notes, and well, our mission is to unpack what sounds like the simplest math in public health, which is the case fatality rate or CFR.
SPEAKER_01Aaron Powell Yeah, it really sounds like the most straightforward metric you could have.
SPEAKER_00Aaron Powell Exactly. But before we get into the actual numbers, I want you to imagine you're trying to figure out if a a new restaurant in town is actually popular. But instead of looking inside at all the packed tables, you only taunt the people standing outside in the rain complaining about the weight.
SPEAKER_01Aaron Powell Well, yeah, you end up with a very loud but fundamentally incomplete picture of the situation.
SPEAKER_00Aaron Ross Powell Right, because you're completely missing everyone who is already inside eating happily. And I mean that brings us to the core of our sources today. The CFR, the proportion of identified cases that end in death is basically doing that exact same thing. So why does such an appealing, clear-cut metric fail us so badly in the real world?
SPEAKER_01Aaron Powell Well, the failure comes down to a misunderstanding of the bottom half of that fraction, the denominator. It is never all infected people, it is strictly people classified as cases. And getting classified as a case is entirely at the mercy of your access to testing, the local surveillance infrastructure, and how severe the symptoms are.
SPEAKER_00So, like if a health system is overwhelmed and they only have the capacity to swab the most severe cases.
SPEAKER_01The people rolling into the ER needing immediate oxygen.
SPEAKER_00Exactly. That denominator stays tiny, making the overall fatality rate look terrifyingly high.
SPEAKER_01Aaron Powell Right. And because that denominator is artificially small, the initial numbers cause just absolute panic. But watch what happens when the surveillance infrastructure actually catches up. Testing expands to the general public, capturing the people who just have a scratchy throat or a mild fever.
SPEAKER_00And suddenly the denominator balloons and the fatality rate plummet.
SPEAKER_01Exactly. Which could easily make you think the pathogen miraculously weakened. In reality, our detection simply got better.
SPEAKER_00We finally started counting the people eating inside the restaurant.
SPEAKER_01And beyond just who gets tested, the data gets really warped by when the testing happens. There was a massive mathematical delay built into outbreaks. Because I mean, it takes seconds to test positive for a virus, but it can take weeks for an illness to actually become fatal.
SPEAKER_00Oh, okay, wait, let me make sure I'm tracking the math here. Early in an outbreak, the denominator is filling up instantly with hundreds of newly diagnosed cases every day. Right. But the numerator, the deaths, hasn't matured yet because that physical process takes time.
SPEAKER_01Yes. And that lag causes the rate to swing wildly. Early on, the rate looks artificially low because the deaths haven't caught up to the diagnoses. But then a few weeks later, cases might drop off, but those previous patients start passing away, making the rate look artificially high. You get these massive statistical swings depending purely on which week you do the division.
SPEAKER_00Wait, so if this metric is so incredibly volatile based on testing capacity and timing lags, why is it the go-to number on every news network?
SPEAKER_01That is the big question.
SPEAKER_00Every time there's a global health event, we see leaderboards comparing different countries by their CFR to judge who is handling the crisis better.
SPEAKER_01Yeah, and that highlights a fundamental communication failure in how we handle public health data. Society tends to treat the case fatality rate as if it were a stable biological property of the pathogen itself. But the sources point out that CFR is actually a property of a region's health and surveillance system.
SPEAKER_00Oh, so it measures the strength of the radar, not just the size of the storm.
SPEAKER_01That is a great way to look at it. To get the real story of an outbreak's severity, you cannot rely on CFR alone.
SPEAKER_00You need other data.
SPEAKER_01Right. You have to look at a combination of metrics. You need infection fatality estimates, you need to track hospitalization patterns, and you need seroprevalence data.
SPEAKER_00Aaron Powell, which is essentially running blood tests on the general population to find the antibodies from past hidden infections.
SPEAKER_01Exactly, because that is how you find your true denominator.
SPEAKER_00Okay. So when you, the listener, are trying to make sense of the news during the next health event, you have to demand context. Remember that the CFR is entirely surveillance dependent. It only shows you what your flashlight is pointing at.
SPEAKER_01It forces us to really question a lot of the absolute certainty we see during the early days of any medical crisis.
SPEAKER_00Which leaves us with a final thought for you to ponder. If a disease's fatality rate is actually a mirror reflecting a society's testing and surveillance capacity, how many historical outbreaks have we completely misjudged simply because their radars were broken? Maybe the restaurant was always full? We just never knew how to look inside.