4 ms·

This. This is the cure for cancer we need: affordable routine scans. Though more scans means more results for doctors to analyze. Hopefully algorithms can mak

This. This is the cure for cancer we need: affordable routine scans. Though more scans means more results for doctors to analyze. Hopefully algorithms can make that part cheaper too.

tisme 14y ago
No. Affordable routine scans will result in massive overloading of the health system with unnecessary procedures, and the complications arising from those procedures will kill more people than would have died otherwise from whatever shows up on those scans. There is no such thing as an intervention without risk, and that starts with the risks associated with general anaesthesia. Scanning is one thing, interpreting a scan and deciding when and where to intervene is something that does not scale well at our current level of technology. Basically it needs a human to sit down and spend time analysing the scan to make sure that what is there warrants intervention. So even if the scan costs would drop to 0 you'd still have the component of the interpretation to deal with. The end result of overloading the system like this is that those with symptoms that need care are less likely to receive it or receive it timely, and that those that didn't need the care will suffer from complications. Knowing when to scan and when not to scan is a crucial ingredient in giving good care.
- lostlogin 14y ago
  Better than what I said. And your last point flows into socializes versus privatized care quite quickly, where fear of litigation drives further testing, which creates a positive feedback loop. The more you test for the more you need to test for. In an exceptionally cynical view of private healthcare by S. Sheem, House of God, he says Rule 10: If you you don't take a temperature you can't find a fever. There are quite a few gems in there for those with a dark sense of humour.
- mdonahoe 14y ago
  Ok. What if no one looks at the scans? What if I get scanned every month and no one looks at them until I start showing symptoms? My aunt got breast cancer recently. She got mammograms every year but the doctors never saw anything. Finally there was a mass discovered, and a when they went back to look at the old images, they could see it growing.
lostlogin 14y ago
No, not now, maybe in the future (with regards to MRI). Even with breast imaging where MRI provides great results, routine screening is only shown to be helpful in those genetically predisposed to disease. Generating masses of biopsy procedures, untold stress, misdiagnoses, botched procedures etc. The false positive rate for MRI is currently too high in breast imaging for it to be a routine tool, let alone whole body imaging with the pitfalls in the technique.
pja 14y ago
No. Many, many people have small, non-malignant cancers growing within us right now. Some people carrying such growths will develop life-threatening cancers, whilst the rest (probably the vast majority) of the population will die of something else at a ripe old age before the growths in question ever become malignant. Once you start screening everyone, suddenly you have this huge population of growths that you have to biopsy in order to determine whether or not they're malignant or not. Your false-positive rate means that you start giving extremely harmful treatments (chemotherapy is itself mutagenic & can cause cancer!) to individuals who if they'd never have been scanned would have remained blissfully unaware of the harmless growth within them & would have had productive, happy lives. Routine scans are not cost free: They put a huge diagnostic burden on the system & they inevitably mean that you deliver invasive, potentially harmful treatments to people that don't need them.
sskates 14y ago
I'm surprised there are three responses to this comment, all starting with the word "no". I'm not arguing for scanning irrespective of the cost. I'm saying that if the cost were 0 then you should always scan. More information should always lead to better decisions and better outcomes. If it's the case that more information leads to worse decisions then it's indicative of a problem in your rule set, not a problem with your information. Imagine if taking someone's blood pressure was a multi-thousand dollar test that required specialized expertise to interpret, and could easily lead to misdiagnoses of heart problems. Now imagine that we just got the technology to do it cheaply and easily. It wouldn't make anyway sense to be afraid of testing blood pressure- it's just more information that can lead to a more complete picture of a patient's condition. But that's exactly what my sibling comments are arguing for now with MRI scans. Also see number 7 in a list of ambitious startup ideas: http://paulgraham.com/ambitious.html http://paulgraham.com/ambitious.html
- aidenn0 14y ago
  The parent post wasn't some theoretical awesome scan with no false-positives; it was specifically whole-body MRI. Even if you could do a whole-body MRI scan for say a nickel, and have automated software mark all new tumors, more people would die sooner than if you hadn't done this. Once you find a tumor you need to biopsy it before you can treat it. There is a small, but non-zero, chance of death from a biopsy. The number of tumors that won't kill you right away is so much greater than the number of tumors that will kill you right away, that this small chance of death multiplied by the large number of people with non-threatening tumors is a larger number than the small number of people with highly malignant tumors.
  - sskates 14y ago
    So then it's a problem with the rule set- you shouldn't automatically biopsy every tumor. It's doesn't make sense that there's a rule in place (you must biopsy every tumor) such that people are afraid of getting more information because it will lead them to make worse decisions. More information should always lead to better decisions and better outcomes. Why are we forced to biopsy every tumor, even in the case where you agree it's suboptimal? Imagine if every time someone had really high blood pressure, we had a rule to perform bypass surgery. People would then avoid taking blood pressure because it might lead to surgery. That doesn't make sense- knowing someone's blood pressure isn't the problem, the problem is our suboptimal rule.

This. This is the cure for cancer we need: affordable routine scans. Though more scans means more results for doctors to analyze. Hopefully algorithms can make that part cheaper too.

tisme 14y ago

No. Affordable routine scans will result in massive overloading of the health system with unnecessary procedures, and the complications arising from those procedures will kill more people than would have died otherwise from whatever shows up on those scans. There is no such thing as an intervention without risk, and that starts with the risks associated with general anaesthesia. Scanning is one thing, interpreting a scan and deciding when and where to intervene is something that does not scale well at our current level of technology. Basically it needs a human to sit down and spend time analysing the scan to make sure that what is there warrants intervention. So even if the scan costs would drop to 0 you'd still have the component of the interpretation to deal with. The end result of overloading the system like this is that those with symptoms that need care are less likely to receive it or receive it timely, and that those that didn't need the care will suffer from complications. Knowing when to scan and when not to scan is a crucial ingredient in giving good care.

lostlogin 14y ago
Better than what I said. And your last point flows into socializes versus privatized care quite quickly, where fear of litigation drives further testing, which creates a positive feedback loop. The more you test for the more you need to test for. In an exceptionally cynical view of private healthcare by S. Sheem, House of God, he says Rule 10: If you you don't take a temperature you can't find a fever. There are quite a few gems in there for those with a dark sense of humour.
mdonahoe 14y ago
Ok. What if no one looks at the scans? What if I get scanned every month and no one looks at them until I start showing symptoms? My aunt got breast cancer recently. She got mammograms every year but the doctors never saw anything. Finally there was a mass discovered, and a when they went back to look at the old images, they could see it growing.

lostlogin 14y ago

No, not now, maybe in the future (with regards to MRI). Even with breast imaging where MRI provides great results, routine screening is only shown to be helpful in those genetically predisposed to disease. Generating masses of biopsy procedures, untold stress, misdiagnoses, botched procedures etc. The false positive rate for MRI is currently too high in breast imaging for it to be a routine tool, let alone whole body imaging with the pitfalls in the technique.

pja 14y ago

No. Many, many people have small, non-malignant cancers growing within us right now. Some people carrying such growths will develop life-threatening cancers, whilst the rest (probably the vast majority) of the population will die of something else at a ripe old age before the growths in question ever become malignant. Once you start screening everyone, suddenly you have this huge population of growths that you have to biopsy in order to determine whether or not they're malignant or not. Your false-positive rate means that you start giving extremely harmful treatments (chemotherapy is itself mutagenic & can cause cancer!) to individuals who if they'd never have been scanned would have remained blissfully unaware of the harmless growth within them & would have had productive, happy lives. Routine scans are not cost free: They put a huge diagnostic burden on the system & they inevitably mean that you deliver invasive, potentially harmful treatments to people that don't need them.

sskates 14y ago

I'm surprised there are three responses to this comment, all starting with the word "no". I'm not arguing for scanning irrespective of the cost. I'm saying that if the cost were 0 then you should always scan. More information should always lead to better decisions and better outcomes. If it's the case that more information leads to worse decisions then it's indicative of a problem in your rule set, not a problem with your information. Imagine if taking someone's blood pressure was a multi-thousand dollar test that required specialized expertise to interpret, and could easily lead to misdiagnoses of heart problems. Now imagine that we just got the technology to do it cheaply and easily. It wouldn't make anyway sense to be afraid of testing blood pressure- it's just more information that can lead to a more complete picture of a patient's condition. But that's exactly what my sibling comments are arguing for now with MRI scans. Also see number 7 in a list of ambitious startup ideas: http://paulgraham.com/ambitious.html http://paulgraham.com/ambitious.html

aidenn0 14y ago
The parent post wasn't some theoretical awesome scan with no false-positives; it was specifically whole-body MRI. Even if you could do a whole-body MRI scan for say a nickel, and have automated software mark all new tumors, more people would die sooner than if you hadn't done this. Once you find a tumor you need to biopsy it before you can treat it. There is a small, but non-zero, chance of death from a biopsy. The number of tumors that won't kill you right away is so much greater than the number of tumors that will kill you right away, that this small chance of death multiplied by the large number of people with non-threatening tumors is a larger number than the small number of people with highly malignant tumors.
- sskates 14y ago
  So then it's a problem with the rule set- you shouldn't automatically biopsy every tumor. It's doesn't make sense that there's a rule in place (you must biopsy every tumor) such that people are afraid of getting more information because it will lead them to make worse decisions. More information should always lead to better decisions and better outcomes. Why are we forced to biopsy every tumor, even in the case where you agree it's suboptimal? Imagine if every time someone had really high blood pressure, we had a rule to perform bypass surgery. People would then avoid taking blood pressure because it might lead to surgery. That doesn't make sense- knowing someone's blood pressure isn't the problem, the problem is our suboptimal rule.