# How are epidemiologists calculating a 3% fatality rate for Wuhan Coronavirus? **URL:** \ **Category:** Factual Questions\ **Created:** [January 29, 2020, 7:52pm UTC](https://boards.straightdope.com/t/how-are-epidemiologists-calculating-a-3-fatality-rate-for-wuhan-coronavirus/847183 "2020-01-29T19:52:03Z")\ **Posts on this page:** 20\ **Page:** 1
**Author:** ![KidCharlemagne](https://avatars.discourse-cdn.com/v4/letter/k/ed8c4c/32.png) [@KidCharlemagne](https://boards.straightdope.com/u/KidCharlemagne)\ **Post date:** [January 29, 2020, 7:52pm UTC](https://boards.straightdope.com/t/how-are-epidemiologists-calculating-a-3-fatality-rate-for-wuhan-coronavirus/847183/1 "2020-01-29T19:52:03Z")
On this webpage: > **[ArcGIS Dashboards](https://gisanddata.maps.arcgis.com/apps/dashboards/index.html)** > > ArcGIS Dashboards stats show 132 deaths vs 126 totally recovered. That’s more than 50%. Is there a long post-peak phase where doctors know It won’t be fatal but can’t claim complete recovery? Because calculating by deaths/infected seems like a pretty shitty metric. ---
**Author:** ![Pantastic](https://avatars.discourse-cdn.com/v4/letter/p/3bc359/32.png) [@Pantastic](https://boards.straightdope.com/u/Pantastic)\ **Post date:** [January 29, 2020, 8:11pm UTC](https://boards.straightdope.com/t/how-are-epidemiologists-calculating-a-3-fatality-rate-for-wuhan-coronavirus/847183/2 "2020-01-29T20:11:27Z")
It’s some portion of the 6,165 ‘total confirmed cases’ at the upper left. What metric do you think should be used for ‘fatality rate’? Counting an infection that keeps someone feeling bad for a long time but doesn’t kill them as a ‘fatality’ doesn’t seem very accurate. I had pneumonia that lasted for like 3 months, and while I wouldn’t be at ‘complete recovery’ by the third month, it was pretty clear after the first month that it wasn’t going to kill me. The outbreak of coronavirus hasn’t even lasted an entire month yet, so there are a lot of people still suffering from it. I’m not even sure that ‘less than a month’ is a reasonable time to call a ‘long’ post-peak phase. The common cold often has symptoms (typically a cough) that last for 2-4 weeks after the major symptoms end, and that is much easier for the body to clear out than the heavy hitters like norovirus and coronavirus. ---
**Author:** ![KidCharlemagne](https://avatars.discourse-cdn.com/v4/letter/k/ed8c4c/32.png) [@KidCharlemagne](https://boards.straightdope.com/u/KidCharlemagne)\ **Post date:** [January 29, 2020, 8:28pm UTC](https://boards.straightdope.com/t/how-are-epidemiologists-calculating-a-3-fatality-rate-for-wuhan-coronavirus/847183/3 "2020-01-29T20:28:27Z")
> [@Pantastic](#): > > It’s some portion of the 6,165 ‘total confirmed cases’ at the upper left. What metric do you think should be used for ‘fatality rate’? Counting an infection that keeps someone feeling bad for a long time but doesn’t kill them as a ‘fatality’ doesn’t seem very accurate. I had pneumonia that lasted for like 3 months, and while I wouldn’t be at ‘complete recovery’ by the third month, it was pretty clear after the first month that it wasn’t going to kill me. The outbreak of coronavirus hasn’t even lasted an entire month yet, so there are a lot of people still suffering from it. I’m not even sure that ‘less than a month’ is a reasonable time to call a ‘long’ post-peak phase. The common cold often has symptoms (typically a cough) that last for 2-4 weeks after the major symptoms end, and that is much easier for the body to clear out than the heavy hitters like norovirus and coronavirus. That’s pretty much exactly why I asked if it was some sort of “past the peak” metric. Depending on the length of the illness, deaths/confirmed is not a helpful metric in determining how serious the illness is. I guess given the deaths/confirmed is about 2% I guess they’re using that, but calling it a fatality rate is misleading for comparing it to other diseases. ---
**Author:** ![Lord\_Feldon](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/lord_feldon/32/341_2.png) [@Lord\_Feldon](https://boards.straightdope.com/u/Lord_Feldon)\ **Post date:** [January 30, 2020, 1:56am UTC](https://boards.straightdope.com/t/how-are-epidemiologists-calculating-a-3-fatality-rate-for-wuhan-coronavirus/847183/4 "2020-01-30T01:56:49Z")
Doesn’t it go the other way too? The unconfirmed cases are presumably less severe than the confirmed cases. ---
**Author:** ![Chronos](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/chronos/32/134_2.png) [@Chronos](https://boards.straightdope.com/u/Chronos)\ **Post date:** [January 30, 2020, 3:31am UTC](https://boards.straightdope.com/t/how-are-epidemiologists-calculating-a-3-fatality-rate-for-wuhan-coronavirus/847183/5 "2020-01-30T03:31:09Z")
Sure, it’s not perfect, but how else could one possibly quantify a fatality rate? ---
**Author:** ![bob\_2](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/bob_2/32/3341_2.png) [@bob\_2](https://boards.straightdope.com/u/bob_2)\ **Post date:** [January 30, 2020, 10:16am UTC](https://boards.straightdope.com/t/how-are-epidemiologists-calculating-a-3-fatality-rate-for-wuhan-coronavirus/847183/6 "2020-01-30T10:16:01Z")
I would hazard a guess that it’s a combination of the above and past experience with similar viruses. ---
**Author:** ![KidCharlemagne](https://avatars.discourse-cdn.com/v4/letter/k/ed8c4c/32.png) [@KidCharlemagne](https://boards.straightdope.com/u/KidCharlemagne)\ **Post date:** [January 30, 2020, 10:43am UTC](https://boards.straightdope.com/t/how-are-epidemiologists-calculating-a-3-fatality-rate-for-wuhan-coronavirus/847183/7 "2020-01-30T10:43:49Z")
> [@Chronos](#): > > Sure, it’s not perfect, but how else could one possibly quantify a fatality rate? Say we don’t have enough data until they have enough data or create 2 different metrics. You’ve got reporters comparing Wuhan to the fatality rate of other diseases (like SARS at 10% (Confirmed)), and saying Wuhan is much less lethal by comparison. Well with only 2% recovered that’s seems unscientific. Doctors could make the reasonable case that a patient was past the danger zone and include them in the fatality statistic, but they don’t seem to be doing that. In a rapidly changing pandemic situation where transmission occurs exponentially and only 126 people recovered, it seems irresponsible to me. ---
**Author:** ![Chronos](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/chronos/32/134_2.png) [@Chronos](https://boards.straightdope.com/u/Chronos)\ **Post date:** [January 30, 2020, 11:47am UTC](https://boards.straightdope.com/t/how-are-epidemiologists-calculating-a-3-fatality-rate-for-wuhan-coronavirus/847183/8 "2020-01-30T11:47:57Z")
OK, so what’s the fatality rate from herpes? Nobody’s ever recovered from that, and I’m sure at least a few people somewhere, somewhen have died of it. ---
**Author:** ![Tired\_and\_Cranky](https://avatars.discourse-cdn.com/v4/letter/t/a87d85/32.png) [@Tired\_and\_Cranky](https://boards.straightdope.com/u/Tired_and_Cranky)\ **Post date:** [January 30, 2020, 2:13pm UTC](https://boards.straightdope.com/t/how-are-epidemiologists-calculating-a-3-fatality-rate-for-wuhan-coronavirus/847183/9 "2020-01-30T14:13:02Z")
> [@Chronos](#): > > OK, so what’s the fatality rate from herpes? Nobody’s ever recovered from that, and I’m sure at least a few people somewhere, somewhen have died of it. But the vast majority of people who had herpes lived long and fruitful lives until they died of something else so the fatality rate is near zero, which you already know. There is nothing insane about making the denominator the total of people who died of the disease, people who recovered, and people who had the disease but died of something else. In some ways, it would be a more informative metric with a fast-spreading outbreak like this, even though this isn’t actually how we do it. ---
**Author:** ![kayaker](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/kayaker/32/441_2.png) [@kayaker](https://boards.straightdope.com/u/kayaker)\ **Post date:** [January 30, 2020, 2:29pm UTC](https://boards.straightdope.com/t/how-are-epidemiologists-calculating-a-3-fatality-rate-for-wuhan-coronavirus/847183/10 "2020-01-30T14:29:11Z")
I’ve had influenza (presumed) a few times in my life, but never sought medical care. How do statisticians account for those like me who miss a few days of work but never see a doctor? ---
**Author:** ![KarlGauss](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/karlgauss/32/3713_2.png) [@KarlGauss](https://boards.straightdope.com/u/KarlGauss)\ **Post date:** [January 30, 2020, 3:20pm UTC](https://boards.straightdope.com/t/how-are-epidemiologists-calculating-a-3-fatality-rate-for-wuhan-coronavirus/847183/11 "2020-01-30T15:20:14Z")
> [@kayaker](#): > > I’ve had influenza (presumed) a few times in my life, but never sought medical care. How do statisticians account for those like me who miss a few days of work but never see a doctor? You may still have infected one or more people and there is some chance that one of them would be captured in the data if/when they sought medical care. Presumably, there are lots of other asymptomatic or nearly asymptomatic carriers where the same thing happens as caused by you. I imagine with a big enough data set you’d get a good estimate of the total burden of disease. ---
**Author:** ![Buck\_Godot](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/buck_godot/32/6573_2.png) [@Buck\_Godot](https://boards.straightdope.com/u/Buck_Godot)\ **Post date:** [January 30, 2020, 4:24pm UTC](https://boards.straightdope.com/t/how-are-epidemiologists-calculating-a-3-fatality-rate-for-wuhan-coronavirus/847183/12 "2020-01-30T16:24:57Z")
I would also question as to whether the record keeping regarding recoveries is as good as that regarding deaths. If someone dies of the disease it is very likely to be recorded. If someone starts feeling better and heads home there might not be any follow-up to confirm recovery. That the number of confirmed recoveries is so low would tend to support this. ---
**Author:** ![Kovitlac](https://avatars.discourse-cdn.com/v4/letter/k/2bfe46/32.png) [@Kovitlac](https://boards.straightdope.com/u/Kovitlac)\ **Post date:** [January 30, 2020, 4:31pm UTC](https://boards.straightdope.com/t/how-are-epidemiologists-calculating-a-3-fatality-rate-for-wuhan-coronavirus/847183/13 "2020-01-30T16:31:15Z")
> [@KidCharlemagne](#): > > Say we don’t have enough data until they have enough data or create 2 different metrics. You’ve got reporters comparing Wuhan to the fatality rate of other diseases (like SARS at 10% (Confirmed)), and saying Wuhan is much less lethal by comparison. Well with only 2% recovered that’s seems unscientific. Doctors could make the reasonable case that a patient was past the danger zone and include them in the fatality statistic, but they don’t seem to be doing that. In a rapidly changing pandemic situation where transmission occurs exponentially and only 126 people recovered, it seems irresponsible to me. What you provided was a lethal statistic, not a threat-o-meter. It’s entire purpose is to indicate how deadly the virus seems to be. And reports show it isn’t very, at least when compared to other illnesses. That isn’t to say it isn’t something you shouldn’t take seriously, particularly if you’ve been in contact with someone from China or have recently been there (or ARE there). But the risk of you actually dying is on the low side, and knowing that is \*important \*so that people don’t panic unnecessarily. ---
**Author:** ![naita](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/naita/32/5862_2.png) [@naita](https://boards.straightdope.com/u/naita)\ **Post date:** [January 30, 2020, 4:47pm UTC](https://boards.straightdope.com/t/how-are-epidemiologists-calculating-a-3-fatality-rate-for-wuhan-coronavirus/847183/14 "2020-01-30T16:47:05Z")
> [@KidCharlemagne](#): > > On this webpage: > > [ArcGIS Dashboards](https://gisanddata.maps.arcgis.com/apps/opsdashboard/index.html#/bda7594740fd40299423467b48e9ecf6) > > stats show 132 deaths vs 126 totally recovered. That’s more than 50%. Is there a long post-peak phase where doctors know It won’t be fatal but can’t claim complete recovery? Because calculating by deaths/infected seems like a pretty shitty metric. A better question would be where that “recovery” number comes from and what cases it involves. As for how the 3% rate is calculated and by who, you have to find an actual source and research who they got the number from, but this Guardian article from a couple of days ago indicates that it is indeed just a rough estimate based on reported cases and fatalities and likely an **overestimation**. > [@](#): > > **Why is this any worse than normal influenza?** > > We don’t yet know how dangerous the new coronavirus is – and we won’t know until more data comes in. Eighty-one deaths out of 2,827 reported cases would mean a 3% mortality rate. However, this is likely to be a overestimate since there may be a far larger pool of people who have been infected by the virus but who have not suffered severe enough symptoms to attend hospital and so have not been counted in the data. [https://www.theguardian.com/world/2020/jan/30/what-is-coronavirus-symptoms-from-wuhan-china-and-how-worried-should-we-be-sars](https://www.theguardian.com/world/2020/jan/30/what-is-coronavirus-symptoms-from-wuhan-china-and-how-worried-should-we-be-sars) ---
**Author:** ![Jasmine](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/jasmine/32/2964_2.png) [@Jasmine](https://boards.straightdope.com/u/Jasmine)\ **Post date:** [January 30, 2020, 4:50pm UTC](https://boards.straightdope.com/t/how-are-epidemiologists-calculating-a-3-fatality-rate-for-wuhan-coronavirus/847183/15 "2020-01-30T16:50:07Z")
I’m sorry, but I really can’t believe it is much more than a “guesstimate” at this point. I think you need more cases and more time to get a more accurate figure. Also, the figure would have to vary according to region because of differences in available and affordable medical care. Plus, I don’t think they’ve even firmly established a reliable R naught for this virus, and that’s an important factor because the faster it spreads, the more patients you have in a shorter time span. That affects the figures, too. ---
**Author:** ![KarlGauss](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/karlgauss/32/3713_2.png) [@KarlGauss](https://boards.straightdope.com/u/KarlGauss)\ **Post date:** [January 30, 2020, 4:55pm UTC](https://boards.straightdope.com/t/how-are-epidemiologists-calculating-a-3-fatality-rate-for-wuhan-coronavirus/847183/16 "2020-01-30T16:55:43Z")
> [@Jasmine](#): > > I’m sorry, but I really can’t believe it is much more than a “guesstimate” at this point. I think you need more cases and more time to get a more accurate figure. Also, the figure would have to vary according to region because of differences in available and affordable medical care. Plus, I don’t think they’ve even firmly established a reliable R naught for this virus, and that’s an important factor because the faster it spreads, the more patients you have in a shorter time span. That affects the figures, too. According to [this article](https://www.nejm.org/doi/full/10.1056/NEJMoa2001316) (free access), R is 2.2 ---
**Author:** ![KidCharlemagne](https://avatars.discourse-cdn.com/v4/letter/k/ed8c4c/32.png) [@KidCharlemagne](https://boards.straightdope.com/u/KidCharlemagne)\ **Post date:** [January 30, 2020, 5:51pm UTC](https://boards.straightdope.com/t/how-are-epidemiologists-calculating-a-3-fatality-rate-for-wuhan-coronavirus/847183/17 "2020-01-30T17:51:42Z")
> [@Kovitlac](#): > > What you provided was a lethal statistic, not a threat-o-meter. It’s entire purpose is to indicate how deadly the virus seems to be. And reports show it isn’t very, at least when compared to other illnesses. > > That isn’t to say it isn’t something you shouldn’t take seriously, particularly if you’ve been in contact with someone from China or have recently been there (or ARE there). But the risk of you actually dying is on the low side, and knowing that is \*important \*so that people don’t panic unnecessarily. A fatality rate is a lethality statistic. The reports of which you speak are generally referencing the fatality rate, the accuracy of which I’m questioning in this post. Note that this is really academic on my part. I’m not trying to work myself or anyone else into a panic. ---
**Author:** ![Buck\_Godot](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/buck_godot/32/6573_2.png) [@Buck\_Godot](https://boards.straightdope.com/u/Buck_Godot)\ **Post date:** [January 30, 2020, 5:53pm UTC](https://boards.straightdope.com/t/how-are-epidemiologists-calculating-a-3-fatality-rate-for-wuhan-coronavirus/847183/18 "2020-01-30T17:53:07Z")
> [@Jasmine](#): > > I’m sorry, but I really can’t believe it is much more than a “guesstimate” at this point. I think you need more cases and more time to get a more accurate figure. Also, the figure would have to vary according to region because of differences in available and affordable medical care. Plus, I don’t think they’ve even firmly established a reliable R naught for this virus, and that’s an important factor because the faster it spreads, the more patients you have in a shorter time span. That affects the figures, too. True, but it isn’t a complete shot in the dark. For example its pretty clear by now that it is less lethal than Ebola, and is probably less lethal than SARS. ---
**Author:** ![KidCharlemagne](https://avatars.discourse-cdn.com/v4/letter/k/ed8c4c/32.png) [@KidCharlemagne](https://boards.straightdope.com/u/KidCharlemagne)\ **Post date:** [January 30, 2020, 5:55pm UTC](https://boards.straightdope.com/t/how-are-epidemiologists-calculating-a-3-fatality-rate-for-wuhan-coronavirus/847183/19 "2020-01-30T17:55:37Z")
> [@naita](#): > > A better question would be where that “recovery” number comes from and what cases it involves. > > As for how the 3% rate is calculated and by who, you have to find an actual source and research who they got the number from, but this Guardian article from a couple of days ago indicates that it is indeed just a rough estimate based on reported cases and fatalities and likely an **overestimation**. > [https://www.theguardian.com/world/2020/jan/30/what-is-coronavirus-symptoms-from-wuhan-china-and-how-worried-should-we-be-sars](https://www.theguardian.com/world/2020/jan/30/what-is-coronavirus-symptoms-from-wuhan-china-and-how-worried-should-we-be-sars) I did ask about the nature of their definition of “recovery” in the OP. It would be perfectly reasonable for the purposes of determining a fatality rate in an emerging situation that doctors classify “recovery” as “probably past the peak of the illness and didn’t die.” ---
**Author:** ![KidCharlemagne](https://avatars.discourse-cdn.com/v4/letter/k/ed8c4c/32.png) [@KidCharlemagne](https://boards.straightdope.com/u/KidCharlemagne)\ **Post date:** [January 30, 2020, 6:05pm UTC](https://boards.straightdope.com/t/how-are-epidemiologists-calculating-a-3-fatality-rate-for-wuhan-coronavirus/847183/20 "2020-01-30T18:05:16Z")
> [@Chronos](#): > > OK, so what’s the fatality rate from herpes? Nobody’s ever recovered from that, and I’m sure at least a few people somewhere, somewhen have died of it. In most populations the stat isn’t relevant but for neonatal herpes, at least in one study, it was pegged at .82 per 100,000 live births. [Next page](https://boards.straightdope.com/t/how-are-epidemiologists-calculating-a-3-fatality-rate-for-wuhan-coronavirus/847183.md?page=2)