Could an average US person with Covid more likely die from medical error than from Covid?

[kcdmps]

For some diseases, the treatment is worse than the disease, itself. It looks like the death rates for people with Covid is about 3%. A study by Johns Hopkins suggests that the probability of dying as a result of medical errors is about 10%. Here’s the link to this info at Johns Hopkins website https://www.hopkinsmedicine.org/new...rs_now_third_leading_cause_of_death_in_the_us If (1) your chance of dying with Covid is 3% and (2) your chance of dying from medical errors is 10%, then the treatment is worse than the disease. The 10%, however, corresponds to the dead. For the living, the odds could be lower because about 1% of the population die yearly.

Below is a generalized evaluation of whether the treatment is worse than the disease for Covid. The assumption that medical error that causes death occurs in a hospital may be lacking because some could occur in a doctor's office, in an ambulance, etc. The fraction of deaths outside the hospital can be incorporated once it is known. Please comment.

A US person average about 40 days in hospital during lifetime (https://www.hcup-us.ahrq.gov/reports/statbriefs/sb246-Geographic-Variation-Hospital-Stays.jsp)

Covid patient average about 20 days in hospital (https://www.kpcnews.com/covid-19/article_8ab408ad-8fb0-5f74-8d57-11e586bd8a4f.html)

Covid patient uses 50% of average lifetime hospital stay (20 days/40 days)

Covid death rate is 3%

Medical error death rate is 10% over lifetime

Medical error death rate for Covid patients = (Covid patient uses 50% of average lifetime hospital stay) * (Medical error death rate is 10% over lifetime) = 5%

If medical error death rate for Covid patient (5%) is higher than Covid death rate (3%), an average US person with Covid will likely die from medical error than from Covid.

 

[kcdmps]

The Johns Hopkins report states "...based on a total of 35,416,020 hospitalizations, 251,454 deaths stemmed from a medical error, which the researchers say now translates to 9.5 percent of all deaths each year in the U.S."
This is consistent with considering only inpatient in my analysis.

 

[hlsmith]

I didn't review any of your links.

That COVID death rate seems high. Isn't it 0.1-0.3%? And that is disregarding all of the asymptomatic cases not diagnosed. Also, if I am in the hospital for COVID, isn't a medical error death a competing event, and are medical errors more likely in sicker patients? Also what are these medical errors and do they generalize to covid patients care.

 

[kcdmps]

That COVID death rate seems high. Isn't it 0.1-0.3%? No. It’s about 3% for people with Covid. Maybe your numbers are for the general population, including people who don’t have Covid. This model is only for people with Covid who are hospitalized. This is a back-of-an-envelope type model with generalized assumptions on whether it’s safe to seek treatment for Covid. Can you be more specific with links?

 

[hlsmith]

Yes, I was referencing public mortality. So I am just speaking from intuition in the following. There is a threshold for being admitted. A person with an oxygen saturation level under 90% isn't going to have a choice to be admitted or not. You have to think about things as conditional probabilties. Probability of dying given you are sick enough to be admitted is the setting.

Hospital treatments aren't that complex, maybe antiviral, anti-inflammatories, monoclonals, plasma, and supportive oxygen. The risks would come from individual risk factors (comorbidities, and immune responses [cytokine storms]) and nosocomial infections, but there are covid units so risk to other pathogens is low.

The risk of medical error shouldn't be on the mind since these people would have no other option.

 

[hlsmith]

I clicked on the Hopkins' link, it said that estimated 250k deaths from error. We know about over 400k Americans will be dead within the year, after closing down the economy. Not sure how many of those will be in the hospital, but similar studies state that some other deaths are being attributed to covid and many deaths aren't being attributed to covid.

Didn't open the BMJ article, but one would wonder what percent of deaths are related to comparable respiratory diseases and what the mechanisms are and if applicable (e.g. lack of safety net?)? So do respiratory based diseases have risk for ~10% error death rate?

 

[kcdmps]

The 10% medical error death rate could be compared to any disease, including respiratory.

 

[kcdmps]

You have to think about things as conditional probabilties. Probability of dying given you are sick enough to be admitted is the setting.

The question is, what is the probability of death for people with Covid who are sick enough to be admitted to a hospital? Are data available? My guess is that it’s higher than 3 percent used in the model, so medical errors may not be as significant when a person is sick enough to be admitted.

 

[AlexNillson89]

The 10% medical error death rate could be compared to any disease, including respiratory.

That is why I always try to confirm the diagnosis with several doctors. There it is less likely that they will all make an ode and the same medical error, unfortunately this method will not work if something is very urgent, for example, an operation

 

[hlsmith]

The 10% medical error death rate could be compared to any disease, including respiratory.

Well medical errors and COVID death are competing endpoints. If you die of medical error it prevents you from dying from COVID which could have been near deterministic given risk factors. Also, I would imagine medical errors are not completely stochastic.

 

[kcdmps]

According to JAVA Network (https://jamanetwork.com/journals/jamainternalmedicine/fullarticle/
2774572), mortality of Covid patients averaged about 8.5%, of which 5% is estimated to be caused by
medical errors (from above). That leaves 3.5% mortality from Covid. In this case, you are more likely
to die from medical errors (5%) than from Covid (3.5%) if hospitalized.

 

[hlsmith]

Ok I will keep this going. I will point out that I am not a 70+ year old. Also, covid mortality was worse initially when people were using Hydro chloroquine and not using prone positioning, plasma, cannulation O2, and other meds.

The Hopkins studies are older and some people have issues on what they called errors. Regardless of these things, How do you get 3.5% from 8.5% * 5%?

 

[kcdmps]

5% + 3.5% = 8.5% You have to carefully look at all assumptions and how they are derived before you can understand the results.

 

[noetsi]

I wonder how many die because they go to the hospital for covid and die of something they catch in the hospital. Sorry hlsmith, I have been told many times hospitals are dangerous places full of sick people you can catch diseases from.

I am far from sure we can calculate morbidity rates from Covid. We don't know how many got it and were asymptotic. Also people could have died from it and they not realized it, especially if they did not go to a hospital for it.

 

[kcdmps]

Median age of people who die from Covid is 71 according to cdc.

 

[kcdmps]

I wonder how many die because they go to the hospital for covid and die of something they catch in the hospital. Sorry hlsmith, I have been told many times hospitals are dangerous places full of sick people you can catch diseases from.

I am far from sure we can calculate morbidity rates from Covid. We don't know how many got it and were asymptotic. Also people could have died from it and they not realized it, especially if they did not go to a hospital for it.

This model only applies to Covid inpatients. As better assumptions become known, the model can be improved.

 

[hlsmith]

Clarify, please.

I havent gone to the source paper but the link states,

"Then, using hospital admission rates from 2013, they extrapolated that based on a total of 35,416,020 hospitalizations, 251,454 deaths stemmed from a medical error" =0.7%

But, per your calculation, a person hospitalized for 10 days has a 5% probability of dying.

So. 100 people are hospitalized for covid, their is an 8.5% probability they die related to disease and a 5% probability that death was actually related to an error. So you are trying to say there is a 5% probability of dying from a medical error and that it is not a 5% probability of the death being from error, which is different? How do these values jive with the above 0.7%?

 

[hlsmith]

PS, In US with the vaccination of group 1a and 1b, there will be fewer 'frail' persons with covid, lessening its death rate and making death by error 'possibly' more likely proportionally.

 

[kcdmps]

Clarify, please.

I havent gone to the source paper but the link states,

"Then, using hospital admission rates from 2013, they extrapolated that based on a total of 35,416,020 hospitalizations, 251,454 deaths stemmed from a medical error" =0.7%

But, per your calculation, a person hospitalized for 10 days has a 5% probability of dying.

So. 100 people are hospitalized for covid, their is an 8.5% probability they die related to disease and a 5% probability that death was actually related to an error. So you are trying to say there is a 5% probability of dying from a medical error and that it is not a 5% probability of the death being from error, which is different? How do these values jive with the above 0.7%?

Most of what you state are not consistent with the assumptions used in the model. The 0.7 % relates to deaths per hospitalizations (living people admitted to hospital). The number used in the model relates to the dead (not hospital admissions by the living) and 10% were by medical error. Can you see the difference?

 

[spunky]

Medical error death rate for Covid patients = (Covid patient uses 50% of average lifetime hospital stay) * (Medical error death rate is 10% over lifetime) = 5%

This calculation assumes stochastic independence of two events which are CLEARLY not independent.