« Another Call to End Aid to Africa | Main | Choke To Submit? »

April 04, 2009


In medical school, we learn that beta blockers are ineffective during an acute coronary attack. They (as well as nitrates, which also are ineffective) are used to help combat the agonizing anginal pain.

Thrombolytic agents, such as tpa, and anti-platelet medications, such as aspirin, however, are effective in preventing death. Many doctors may be unaware of what works and what doesn't, but we medical students are usually shown what is true and what is wishful thinking. Perhaps some people stop caring, once they leave school, and others forget that not everything that doctors do is intended to confer a mortality benefit.

I'm a swedish med student and I want to point out at least one thing where I disagree. You write "Patients with ear infections are more likely to be harmed by antibiotics than helped. While the pills may cause a small decrease in symptoms (for which ear drops work better), the infections typically recede within days regardless of treatment." I don't know how this works in America but in Sweden at least every doctor knows that most ear infections do recede without treatment. However, the problem with it is that if you don't treat it while the infection is still confined to the ear is that if it doesn't go away by it self, there is a great risk of the infection spreading both to the bone and to the brain. These infections are very hard to treat and especially brain infections are very serious. The same goes for sinusitis in some cases (depending on in which sinus it is located).

The risk is basically deemed too great, better treat too many than too few in this case. As long as you don't give the patient broad spectrum or other very strong antibiotics the side effects are also quite mild (mostly mild diarrea and an occasional exanthema). Of course you should not prescribe antibiotics if a patient only have had the infection for a day or two but you can't wait indefinitely before treating.

Bjorn, I think the doctors' problem relating to ear infections is they fail to determine whether the cause is viral or bacterial. Most of the infections are viral, and would be best left to heal on their own. The few that are bacterial (and which are generally much more dangerous) usually ought to be treated with the appropriate antibiotic (another mistake is using one that is inappropriate toward the particular organism), which would have great benefit for the patient. Some docs probably figure it's easier to send people home with pills than do proper workup (including a throat culture, which usually requires over a day to see the result). As we see, they are often mistaken, when they rid themselves of the extra work.

Well, most often you can get pretty sure about if an otitis is viral or bacterial just by doing a CRP (a basic test of the patients immune reaction) since you have to have a really bad viral infection to get up to the levels of a bacterial infection. This test take about 10 minutes total and the doctor doesn't have to do it himself so it seems pretty stupid not to do it in the case of a suspected otitis. Anyway, it doesn't really matter if your point is true since in that case it still doesn't show bias among doctors, only laziness.

Note that Medicare didn't say they wouldn't pay for beta-blockers

This statement reflects lack of knowledge about the Medicare system. For hospitals, Medicare reimbursements are by "diagnosis related groups". Medicare pays the hospital a set amount for the treatment of the condition, regardless of what drugs or care techniques are involved. An exception is certain "durable goods", such as pacemakers, for which there are extra pre-set reimbursements, regardless of the cost. Interestingly, the reimbursement scales frequently are lower than the actual costs of the care or the durable items. The hospitals are expected to eat the loss, and in return, they gouge the private payers.

It'd be great if all such 'myths' about not-so-serious health care issues (such as cough syrups helping cough, antibiotics helping ear-infections) were aggregated at one place. Some sort of 'Encyclopedia of myths related to common health care issues'. What do you think?

Is it reasonable to expect Obama to expend all his political capital on this issue while (a) the economy is melting down and (b) scientists haven't come together one side of it? Doctors can't be expected to know such things, but wouldn't it be more reasonable to blame economists for not being able to speak with a near-consensus voice on the issue, than to ask politicians to do what economists cannot?

retired, if Medicare had no financial influence over treatments, how could a federal treatment evaluation group influence practice?

Eliezer, I wasn't saying how Obama should have spent his capital.

I agree Harry, I would love to see some kind of Snopes + Wikipedia focused on comparing medical/nutrition/health studies to various health procedures and actions.

The potential for related advertising would be immense as well if someone wants to pursue this as a commercial venture. Surprised it doesn't already exist.

adina, how can the doctor tell if its viral or bacterial? Is a culture necissary? If the infection is dangerous, I could understand why a doctor would want to prescribe anti-biotics, just in case (especially given our system of medical insurance and tort law).

Recent press reports detailing the dangers of cough syrup for children have noted that cough syrup doesn’t work. True: No cough remedies have ever been proven better than a placebo, either for adults or children. Yet their use is common.
I would think the existence of cough syrup would go along with it being an effective placebo. True you might be able to fool a child with anything, but an adult who actually goes out and buys the product himself needs to think it actually works.
Back surgeries to relieve pain are, in the majority of cases, no better than nonsurgical treatment. Yet doctors perform 600,000 of these surgeries each year, at a cost of over $20 billion.
I had a rather painful and persistent neck injury (C5-C6 herniated and C4-C5 bulging discs hitting a brachial nerve and the spine) and found it very difficult to find any good information about what treatments really work. I didn't want to do under the knife unless I had to, but of course all the doctors I went to recommended their own pet procedure. Eventually I found statistics more or less repeating what you just wrote, that cervical fusion very often does not relieve pain. I ended up getting a cortisol injection and all was well, but I would have very much liked to have had a resource where empirical evidence of treatment results could be found. I really think there might be a demand for that sort of thing (I'm wondering if a wiki could work here?). I know people less studious than myself would have simply gone with what their doctor recommended (a cervical fusion in my case).

Grant, I share your skepticism about spinal surgery. I also feel that we ought to have a greater degree of certainty about its benefits than we should about beta blockers, given the enormous physical and financial costs.

For people wondering if there is an organization that does meta-analyses on studies, and makes recommendations based on degree of evidence, I believe the Cochrane Review is a very good organization that fulfills that purpose.

adina, I didn't know about the Cochrane Review; thanks. However, its not really marketed towards the patient, which is what I think is needed. Patients certainly don't care about using their insurer's (or the government's) money for simple procedures or safe drugs, but most don't want to undergo surgery or take medications with strong side effects if its not going to do them any good (e.g., a friend of mine being on prednisone after a bone marrow transplant, only to find out later there were alternative medications with more agreeable side effects).

Keep in mind that when one new study contradicts multiple old studies you don't automatically throw out all the old ones, especially when it is long-standing clinical practice. I'm not saying this particular new study isn't good and it probably does mean a change in beta-blocker administration is warranted, but it's not always as simple and "doctors are ignoring medical evidence".

Further, studies can only answer questions that are capable of being asked about large numbers of homogeneous patients. Doctors often look closer at specific subgroups of patients where clinical experience leads them to favor a treatment that varies from what the big study says that the "average" patient should be getting. Yes, doctors often ignore valid evidence, but evidence is also not always as convincing to an expert as it may seem to a layman.

It's certainly the case that most of that list is perfectly correct in criticizing clinical practice that doesn't match up to good evidence but I felt like there's also some nuance that people should be aware of.

Brent brings up a good point, about patient stratification. For some reason, in treatment of hypertension (not post-MI), beta blockers, for seem to confer disproportionate benefit to white and young patients, while diuretics confer disproportionate benefit to African-American and elderly patients. I have no clue why, but may have to do with the fact that diuretics mainly target the kidneys, while beta blockers target the heart.

Robin -- I don't want to break this to you but Beta-blockers are pretty cheap drugs. As other people have pointed out, multiple older studies seem to show a significant benefit. The problem is medical practice has changed to include a slew of new interventions -- and whether beta blockers still confer an advantage when you add all the other treatments, perhaps in the ACUTE setting there is a risk. For long term treatment, they work. Obviously if you weren't lazy you would have deciphered this yourself by trying to read a review first.

In terms of antibiotics for common URIs, that are VIRAL, -- most physicians know antibiotics only reduce symptoms by a day or two at best (if its bacterial) -- if you have ever talked to a doctor you would know that some patients DEMAND antibiotics.

@Hanson: retired, if Medicare had no financial influence over treatments, how could a federal treatment evaluation group influence practice?

Medicare has completely different policies for hospitals versus doctors and other fee-for-service providers. This site (and the links contained there) describes the determination of hospital reimbursement for patient treatments based on the patients' diagnoses: http://www.cms.hhs.gov/AcuteInpatientPPS/.

This site describes the role of the Medicare Quality Improvement Organization: http://www.commonwealthfund.org/Content/Performance-Snapshots/Variations-in-Care/Quality-of-Care-and-Medicare-Spending-at-the-State-Level.aspx .

Acute-care hospitals are not governed by fee-for-service regulations, but rather are motivated by the fact that they will receive the amount determined to be appropriate for care of a certain disorder, with modifiers that may qualify for extra reimbursement. Among those modifiers are extra amounts for teaching hospitals based on several factors, including numbers of doctors in training. This is the main reason that such hospitals as Mass General may receive more money for the same service than Beth Israel, which has a smaller in-training force (your earlier post about these two institutions suggested the increased cost was because of patients' willingness to pay for "prestige"). In the Medicare system, no matter what the patient is willing to pay, the price is fixed, and nothing higher can be collected. Voluntary over-payment is illegal. Medicare regulatory bodies can see the statistics of all participants, review the costs of the actual services performed, and assign what they consider to be a fair compensation nationwide or regionally. Hospitals that have high rates of complications and unnecessary services lose money (I should more accurately say "lose more money", since all acute-care hospitals lose money on Medicare in-patients; it's a part of the federal design).

Doctors are regulated by fee-for-service (Part B) rules. CPT codes for every imaginable service are published with the reimbursement for that service. Doctors who have excessive complaints or who are caught in fraudulent practices lose their permission to treat Medicare patients. The feds regulate the type of services allowed through the use of CPT coding: if there's a CPT number, it's considered a legitimate procedure/service. Unlisted services are considered on an individual basis, and frequently disallowed or priced so low that the physician has no financial incentive to perform the service. If the service is not approved by Medicare, the participating physician can perform it on a Medicare patient, but he cannot charge anyone at all. The physician can withdraw from participation with Medicare, and treat Medicare patients for cash (except in some states), but he cannot cherry-pick; his participation must be for all Medicare patients or no Medicare patients. Also regulated under fee-for-service rules are Home Health Agencies (HHAs), Skilled Nursing Facilities (SNFs), Comprehensive Outpatient Rehabilitation Facilities (CORFs), and Hospices, and the like.

As the site about the Medicare Quality Improvement Organization describes, it is more a gatherer of info that influences general reimbursement trends, a recommender of quality practice policies, and a source of comparative performance among hospitals that can shame them into doing better, or award "attaboys". Of course, a horrible provider can be excluded from the program.

While this long answer barely scratches the surface of your question, perhaps only those who have dealt extensively with Medicare can imagine how truly ridiculous and wasteful the system is.


For some reason, in treatment of hypertension (not post-MI), beta blockers, for seem to confer disproportionate benefit to white and young patients, while diuretics confer disproportionate benefit to African-American and elderly patients.

I've heard that blacks are more prone to low-renin hypertension than whites, though apparently the whole thing is controversial:

"It has been well documented that, as monotherapy or in the absence of a diuretic, {beta}-blockers, ACE inhibitors, and ARBs do not lower blood pressure to the same extent in African American patients that they do in white patients with hypertension. It has also been reported that, as monotherapy, thiazide diuretics and CCBs have greater blood pressure–lowering efficacy than do other drug classes in African Americans. However, studies reporting these types of data have certain common limitations: (1) they generally do not report SBP responses; (2) they generally reported response rates based on a reduction of 10 mm Hg or more from baseline DBP rather than achievement of target blood pressure; (3) individual agents cannot be used as a proxy for class effect; and (4) conclusions cannot be drawn regarding the best course of treatment for patients for whom antihypertensive treatment was not efficacious in these studies."

With regards to the efficacy of diuretics, specifically thiazides, it's not completely clear how they work.

Thanks for the info, Tom!

Again you inaccurately report research results. You write: "No cough remedies have ever been proven better than a placebo".

Yet the linked article:

  • reports only on studies of over-the-counter cough medicines, not all cough medicines;
  • reports only on studies of children, not all users of cough medicines; and
  • reports only on coughing due to upper respiratory tract infection, not all causes.

You can do better.

doctors are in the doctoring business to save money. CYA applies here. why risk your neck to save a life when you can plausibly claim you did everything you could by following the tried and true procedures to the letter? malpractice insurance etc.

Now if we had a system where doctors are paid not by large insurance corporations but by their patients...

>Recent press reports detailing the dangers of cough syrup for children have noted that cough syrup doesn’t work. True: No cough remedies have ever been proven better than a placebo, either for adults or children. Yet their use is common.

Was the placebo syrupy too? My recollections of cough medicine involve a decrease in coughing right after consumption--suggesting that either the consistency is what does it, or I'm easy to brainwash.

"Why? Because according to the theory of the straining heart, the treatment makes sense. It should work, even though it doesn’t. Ideology trumps evidence."

Are there any important differences between these medics and economists, who believe in free trade because of theory, despite the lack of empirical evidence that it is beneficial?

The comments to this entry are closed.

Less Wrong (sister site)

May 2009

Sun Mon Tue Wed Thu Fri Sat
          1 2
3 4 5 6 7 8 9
10 11 12 13 14 15 16
17 18 19 20 21 22 23
24 25 26 27 28 29 30