4 ms·

former pathology resident here- disagree. The main reason we do fewer autopsies is because of better diagnostics while the patient is alive. The numbers they pr

by gmarx 11y ago

former pathology resident here- disagree. The main reason we do fewer autopsies is because of better diagnostics while the patient is alive. The numbers they provide for significant changes in diagnosis based on autopsy seem pretty high (based only on my own experience). Even the examples they pick for the article are kind of lame- it wasn't St. louis encephalitis, it was west nile! Um...so? Do we have a treatment now for West Nile? Would they have done something different knowing that it was west nile (honestly I don't know; I've been out of the medical biz for years)

niels_olson 11y ago
> Would they have done something different knowing that it was west nile Considering it reinforced the value proposition of county, state, and federal epidemiology units, all of whom are now monitoring crow populations, and it has launched a significant amount of research, and every primary care and ER doctor now looks for symptoms, yes, I do believe something is different.
CatDevURandom 11y ago
>> former pathology resident here The implication being that you did not complete your training. I'm curious -- if you don't mind the derailment -- why not?
- gmarx 11y ago
  Completed it. Never practiced as an attending. Went into the computers instead
sampo 11y ago
> The main reason we do fewer autopsies is because of better diagnostics while the patient is alive. The article claims that "About 15 to 30 percent of the time, the diagnoses at time of death are wrong".
- gmarx 11y ago
  Maybe, I haven't read the study, but there is a huge selective pressure here. That is, these days you tend to order autopsies only when there was something mysterious about the death.
  - jacquesm 11y ago
    > Maybe, I haven't read the study I don't get that. If you haven't read the study, how can you comment on it? I understand that you'd have your opinion canned and ready in the normal case but this study supposedly moves the needle and I'd be very interested on what it does to your position after you read the study, but without that it's as if you are rejecting out of hand any evidence the study might provide.
    - gmarx 11y ago
      I can comment based on my observation (lately shared by many) that most of the medical literature is bunk and my experience participating in more than 100 autopsies which tells me that 30% is absurdly high, 15% is too high and the range reported is bizarrely large. In those over 100 autopsies I recall only two surprises. Mostly we learned nothing the clinicians didn't know. Occasionally we found something unknown that wouldn't have altered treatment or course. My experience was that the autopsies were overwhelmingly uninformative.
  - sampo 11y ago
    As far as I understood, an "autopsy study" means a study where the researchers decide to perform an autopsy to all of patients in their sample. A quick googling seems to support this meaning, links below. If this is the case, then the results do not contain the bias you are suggesting. http://www.ncbi.nlm.nih.gov/pubmed/10877491 http://www.ncbi.nlm.nih.gov/pubmed/10877491 http://www.ncbi.nlm.nih.gov/pubmed/18288887 http://www.ncbi.nlm.nih.gov/pubmed/18288887 http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0033685 http://journals.plos.org/plosone/article?id=10.1371/journal....

by gmarx 11y ago

niels_olson 11y ago

> Would they have done something different knowing that it was west nile Considering it reinforced the value proposition of county, state, and federal epidemiology units, all of whom are now monitoring crow populations, and it has launched a significant amount of research, and every primary care and ER doctor now looks for symptoms, yes, I do believe something is different.

CatDevURandom 11y ago

>> former pathology resident here The implication being that you did not complete your training. I'm curious -- if you don't mind the derailment -- why not?

gmarx 11y ago
Completed it. Never practiced as an attending. Went into the computers instead

sampo 11y ago

> The main reason we do fewer autopsies is because of better diagnostics while the patient is alive. The article claims that "About 15 to 30 percent of the time, the diagnoses at time of death are wrong".

gmarx 11y ago
Maybe, I haven't read the study, but there is a huge selective pressure here. That is, these days you tend to order autopsies only when there was something mysterious about the death.
- jacquesm 11y ago
  > Maybe, I haven't read the study I don't get that. If you haven't read the study, how can you comment on it? I understand that you'd have your opinion canned and ready in the normal case but this study supposedly moves the needle and I'd be very interested on what it does to your position after you read the study, but without that it's as if you are rejecting out of hand any evidence the study might provide.
  - gmarx 11y ago
    I can comment based on my observation (lately shared by many) that most of the medical literature is bunk and my experience participating in more than 100 autopsies which tells me that 30% is absurdly high, 15% is too high and the range reported is bizarrely large. In those over 100 autopsies I recall only two surprises. Mostly we learned nothing the clinicians didn't know. Occasionally we found something unknown that wouldn't have altered treatment or course. My experience was that the autopsies were overwhelmingly uninformative.
- sampo 11y ago
  As far as I understood, an "autopsy study" means a study where the researchers decide to perform an autopsy to all of patients in their sample. A quick googling seems to support this meaning, links below. If this is the case, then the results do not contain the bias you are suggesting. http://www.ncbi.nlm.nih.gov/pubmed/10877491 http://www.ncbi.nlm.nih.gov/pubmed/10877491 http://www.ncbi.nlm.nih.gov/pubmed/18288887 http://www.ncbi.nlm.nih.gov/pubmed/18288887 http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0033685 http://journals.plos.org/plosone/article?id=10.1371/journal....