Doctors Kill
On reflection, I suspect the main thing keeping me from convincing people that the second half of medical spending hurts as often as it helps is a reluctance to believe that "doctors kill." I should argue more like the October AEI Health Policy Outlook:
Hospital-acquired infections1 (HAIs) are estimated by the World Health Organization (WHO) to kill between 1.5 and 3 million people every year ... Even in the United States, nearly 100,000 people die from HAIs every year. Someone who is already sick - that is, a hospital patient - is especially susceptible to new infection, since his immunity is compromised. Whatever bug is going around is likely to flourish. Hospitals can be dangerous places. ... A recent study ... at fifteen hospitals in Ontario showed that less than one third of doctors and nurses washed their hands between patients as required by good practice ...
A recent investigation into how long nosocomial pathogens can survive on dry surfaces [found that] ... the most common - Methicillin-resistant Staphylococcus aureus (MRSA), Clostridium difficile, E. coli, and tuberculosis lasted for months on surfaces and remained continuous sources of infection if no regular disinfection was performed. The study also flatly contradicts the common belief that HIV can last only a short time outside the body, stating that "blood-borne viruses, such as hepatitis B and HIV, can last over a week on dry surfaces." ...
WHO estimates that at least 16 billion injections are given every year in developing and transitional countries. Less than 5 percent of these are immunizations; over 70 percent are unnecessary or could be given orally. Patients believe injections deliver faster, stronger medicines, and doctors over-prescribe injections to satisfy them. In some cases, nine out of ten patients receive an injection at every visit. ... Assessments carried out by WHO in numerous countries have revealed that syringes and needles are often just rinsed in a pot of tepid water between injections. Worldwide, up to 40 percent of injections are given with syringes and needles reused without sterilization. In some countries, this proportion is as high as 70 percent. ...
A recent mathematical model suggested that unsafe injections may cause 8-16 million cases of hepatitis B, 2.3-4.5 million cases of hepatitis C, and 80-160,000 cases of HIV annually worldwide ... In the past year, a study in Zambia led by the CDC found that medical injections - whether into muscle or veins - "were overwhelmingly correlated with HIV prevalence, exceeding the contribution of sexual behaviours in a multivariable logistic regression." ... On demographic, medical, cultural, and sexual behavior ... criteria the women were in the low-risk category for HIV... and yet 30.3 percent were found to be HIV positive. ... the most significant risk factor to which these women had been exposed was going to a clinic and having an injection. As the authors conclude, "medical injection history made an overwhelming contribution to explaining prevalent HIV infection, even after demographic variables, sexual behaviours, and substance abuse were already parcelled out of . . . the . . . equation." In this study, all risk factors were compared with each other for perhaps the first time, and the results seriously undermine current public health messages on HIV.
A colleague of my wife was a nurse at a local hospital, and was assigned to see if doctors were washing their hands enough. She identified and reported the worst offender, whose patients were suffering as a result. That doctor had her fired; he still works there not washing his hands. Presumably other nurses assigned afterward learned their lesson. Hat tip to Michael Cannon.
Posted by Robin Hanson at 06:00 AM
overcomingbias.com
Another major cause of death is medications. There are 180,000 deaths annually from medication errors and adverse reactions. An Australian study found 2.4-3.6 % of hospital admissions to be due to prescription medication events. Of these I would expect medication errors to be of less significance than adverse reactions, which tend to occur because we currently are not very good at tailoring the medication to the patient.
Phillips D, Christenfeld N, Glynn L. "Increase in US medication-error
deaths between 1983 and 1993." Lancet. 1998;351:643-644.
Barbara Starfield, MD, MPH, Is US Health Really the Best in the World?, JAMA, Volume 284, No. 4, July 26, 2000
Institute of Medicine (IOM), "To Err Is Human: Building a Safer Health System", 2000
nap.edu
I guess one could lump the iatrogenic death and injuries into three main categories. One is simply the unavoidable risks: some treatments simply are dangerous, hospital are risky places because there are sick people around. The second one is incompetence: not washing hands, not checking drug dosages carefully, all the things that could have been avoided by paying more attention. People tend to become emotional about this one. The third one is the knowledge category: if people had just known about new research or successful practice, if they had a particular piece of information about the patient, then they would have acted differently.
Category #1 is hard to get rid of, except by reducing overall overtreatment (which is also going to help the other two). #2 is sticky, since much of it is caused more by organisational problems. But since people are emotional about it, it is easier to get support for interventions to reduce this category of causes. There is probably a lot that can be done, but I suspect the costs of doing it can become arbitrarily high. #3 has some components that can be fixed very cheaply and effectively; there was a project a while ago that demonstrated that just by teaching how the most successful hospitals handled various conditions treatment success at other hospitals rose significantly (can't find the paper for the moment, though). But personalized medicine on the other hand is likely to start out quite expensive and would require expensive changes in how the medical system works.
Meanwhile the alternative therapy people are gleefully rubbing their hands since they look so good in comparision. After all, they don't keep *any* statistics of how many they kill.
Posted by: Anders Sandberg | October 16, 2007 at 07:20 AM
overcomingbias.com |
