Why did you put "works" in quotes? Office 2003 still does, in fact, work. It works just fine.

A lot of people are still using Office 2003 because the number of new features that impact daily usage seems to shrink with every new release. Why upgrade when the version you have does everything you need it to, and the new version doesn't do anything you wish it did?

There's always someone who will benefit from [insert new feature here]. But for the rest of us, Office has suffered from a paucity of innovation since 1995. If anything, things have gotten worse -- e.g. they keep trying to make Microsoft Word "smart," but the result is a program that's too smart to be obedient and too stupid to do what you actually want it to do.

The writing's on the wall for Office. If the folks in Redmond don't figure out something reeeal soon, Office is toast.

This is only true in theory. The data is often not as rigorous as we would like it to be (e.g. ezetimibe, which was approved without any mortality data, whose efficacy is now being questioned). Meanwhile, many "off-label" uses are actually backed by very strong evidence, but no one [not even the FDA] bothers getting the "label" for the indication because the drug is already on the market. The various professional organizations that publish treatment guidelines tend to do a much better job of reviewing the evidence than the FDA does.

http://www.youtube.com/watch?v=VSkT5XykJzo

Well, every generation has its candle burners and its lazy souls. The good ones are selected to succeed, and the lazy ones get fired or demoted to lesser companies or positions. Unfortunately, twenty years later the ones who have succeeded have a tendency to look around, see the incoming generation with its unselected mix of the hardworking and the lazy, and then prophesy the end of the world.

Part of the problem is that Vicodin/Percocet/Lortab/etc. are not always even correctly prescribed. For example, when you look up Vicodin in Epocrates (the most commonly used electronic drug reference), the dosing regimen comes up as:

1-2 tabs every 4-6 hours
Not to exceed 8 tabs in 24 hours

Most physicians just write the first line in the script. Very few medications come with a "Not To Exceed" line, because the maximum is usually built into the dosing and frequency, so most physicians are not accustomed to having to write a "Not To Exceed" line.

The resulting script, with the NTE line omitted, allows the patient to take 2 tabs every 4 hours, for a total of 12 tabs in a 24-hour period. For Vicodin, that is 6 grams of acetaminophen over 24 hours, which can be toxic. Unfortunately, MANY scripts for Vicodin are written this way. Fortunately, most patients don't take the max dose/frequency (especially when sleeping), and most patients have sufficient reserve function in their liver to handle 6g/24hrs of acetaminophen for a few days.

Either way, "1-2 tabs every 4-6 hours" is already too many parameters to juggle; "1-2 tabs every 4-6 hours not to exceed 8 per day" is a little absurd. Truthfully, the average Slashdot reader might be able to figure it out just fine; whereas, the average USA Today reader might struggle.

I would kindly suggest that Epocrates (and others) change their formulary to say "1-2 tabs every 6 hours," which would avert this specific problem, and protect patients who are in fact following directions -- they were just given the wrong damned directions. It would not, however, address the larger issue surrounding educating one's patients about potential toxicities and interactions with Tylenol...

Paramedics work within pre-defined protocols; the planning has been done in advance, so that it doesn't need to be done at the scene. However, a paramedic's job is to stabilize the patient, provide initial treatment, and get them to the hospital; medics generally do not provide definitive care.

On the other hand, if you are sedated in the ICU with multiple organ failure, you had better hope that you have a team of doctors who know how to formulate an appropriate treatment plan. In a medical ICU, the bulk of this planning typically does NOT take place at the bedside; it happens "behind the scenes."

Quickly assessing the situation and applying the right techniques works for certain tasks, but there are many other tasks that require careful, meticulous planning.

While I do agree that we need to achieve a better separation between industry and research funding, it is also important to remember that industry-funded research does not automatically mean "it's all a bunch of lies!"

The volume of evidence pointing to the LDLmortality connection has been staggering. I wholeheartedly agree that there are flaws in these studies, just as there are flaws in any study. These flaws create "holes" in our knowledge that many people are hard at work trying to patch, and we learn more by doing it. However, finding flaw in a study does not prove the null hypothesis (i.e. does not demonstrate that the opposite conclusion must be true).

Ultimately, while there is much evidence to support the idea that cholesterol is NOT the ONLY cause of atherosclerosis (this is a very hot area of research at the moment), I have found very little evidence to support the notion that cholesterol has NO role in atherosclerosis. As usual, the truth probably lies somewhere in the middle.

A closer examination of the article you linked to reveals some critical flaws. I will address the most salient part here regarding cholesterol and progression of coronary artery disease.

Most of the studies cited in the article point to the poor correlation between serum cholesterol levels and progression of plaques as measured by degree of stenosis (degree of narrowing of the artery). It is now believed that a heart attack is not the result of progressive narrowing of the artery, but rather a cataclysmic "explosion" of clot arising from an area of prior damage to the vessel wall - damage that may not be seen on angiography, as angiography only examines the diameter of the vessel lumen. Evidence shows that the correlation between degree of stenosis and What We Care About (i.e. heart attacks, cardiac death) is poor. Hence, proving a poor correlation between serum cholesterol and progression of stenosis doesn't really mean much in terms of What We Care About.*

There are, on the other hand, large placebo-controlled studies that have clearly shown that (1) patients with high LDL levels have heart attacks earlier, and die earlier from cardiac causes; and, (2) that patients whose high LDL levels are treated with statins/fibrates/etc. can profoundly delay heart attacks and cardiac death. This is called a study with hard endpoints, and this is the sort of study that directly addresses What We Care About. Many scientists are working on exactly what happens between LDL and a heart attack.

One must examine ALL of the evidence before concluding that cholesterol is a "problem being created out of nothing in order to sell more drugs."

---
* [Note: We are talking about chronic stenosis here. This is an entirely different issue from the idea of placing stents, as stents are placed (i) to relieve symptoms resulting from reversible ischemia, or (ii) to restore blood flow immediately following a heart attack in order to "save the muscle" before it has a chance to die. As a side note, exactly when to stent and when not to stent is a huge controversy that the evidence has yet to resolve.]

One can call it a pseudoscience, but it is sometimes the best we can do. It is very hard to achieve scientific rigor in a field where it is so difficult to control your experiment.

Consider: In other fields, you can build chemical compounds that achieve absurd levels of purity. You can breed mice that are genetically identical to one another. You can place said mice in a controlled environment - you control their diet, their exposure to stimuli, their exposure to pathogens. You can set the protocol for how you do your experiment, when you do it, under what conditions. The entire experiment can take place in a locked lab.

On the other hand ... In medicine, you cannot breed humans that are genetically identical to one another - each one is genetically unique, and more and more research is beginning to link one's genotype with disease predisposition and therapeutic response. You cannot ensure that your humans have a homogenous set of environmental stimuli, because you can't exactly lock them up. You can't even be sure that they aren't cheating, and taking additional meds behind your back that you know nothing about. The majority of them will have a diverse array of comorbid conditions that pollute your data and make it difficult to draw conclusions. See the problem?

Ironically, if you were able to perform such controlled experiments, their conclusions would often be poorly applicable, because your studied population (genetically identical disease-free humans living in a cage) would bear NO resemblance to the patient in front of you. You can't study apples and apply the findings to oranges.

The article does not promote the "retirement" of beta-blockers -- in fact, it shows that beta-blockers reduce the risk of reinfarction and ventricular fibrillation, but increase the risk of cardiogenic shock.

Mind you, ventricular fibrillation is a fancy word for cardiac arrest. This brings us between a rock and a hard place. The good news is, there are many findings that can serve as (albeit imperfect) harbingers of conditions such as cardiogenic shock. If I were a patient who did not have such findings, I sure would want the beta-blockers, and the evidence would support it.

Contrary to your statement, there is no "strong evidence that [beta-blockers are] ineffective." The evidence simply shows that there are situations where they are indicated, and others where their use should be cautioned against. Calling for the "retirement" of beta-blockers ignores the evidence. This would not be evidence-based medicine.

Part of evidence-based medicine entails examining whether it is appropriate to apply the evidence to your particular patient (in EBM parlance, external validity). For example, a study done on hospitalized patients (who are sicker) may have very different findings from a study done on patients with the same illness that did not require hospitalization, and the conclusions drawn should be applied to different populations. In a field where we know so little in the way of how the body works, where we know so little about how disease works, and where each patient has a million confounding factors that makes him/her different from the next, achieving good external validity can be very difficult.

You are right that the formal definition of AIDS leads to circular logic (HIV causes AIDS, AIDS is defined by HIV). The fact remains, however, that the evidence shows that HIV leads to a syndrome of severe immunocompromise - we can call it AIDS, or we can call it severe HIV-associated immunocompromise, but whatever we call it, it's something and it's real and the evidence would strongly suggest that it is caused by HIV. Such "invalidation" of postulates 1 and 4 based on the formal definition of AIDS is an issue of semantics, not of clinical realities.

Strep throat is defined as a tonsillopharyngitis caused by Group A Streptococcus. One could make the similar argument that this cannot be proved with Koch's postulates, and is therefore unscientific. Does that argument stand to reason? Perhaps. Is it useful? Only marginally.

As for postulate 3, I agree that the evidence is not exquisitely compelling. However, is it ethical to prove postulate 3? No, it is not; the third postulate thus remains unproven for many diseases. We rely on other types of data to support (not prove) postulate 3. "Statistical controls" would not help us here -- there are many other causes of severe immunodeficiency that can cause syndromes very similar/identical to AIDS, but the existence of such etiologies neither proves nor disproves causality. I can prove that there are non-strep infections which can cause symptoms indistinguishable from strep throat, but that doesn't have much bearing on the causality of Group A Strep for strep throat.

Logically, it is ultimately possible that HIV is but a confounded variable with the true cause of AIDS. This is the centerpiece of Peter Duesberg's argument for drug abuse as the cause of AIDS (an argument that is strongly contradicted by the available evidence). However, no one has yet identified such a factor that is supported by any evidence, experimental or not. The bulk of the non-experimental evidence surrounding HIV suggests a causal relationship. Unscientific? Perhaps, but this is the best data we have. The challenges inherent in proving anything in a biological context are an unfortunate reality of biomedical research.

Sidestepping the arguments regarding causality for a moment, I would like to point out that anti-retroviral therapy has been demonstrated to be effective in placebo-controlled trials toward delaying both opportunistic infections and, ultimately, death, in HIV+ patients. It may not be perfect science - but in biology and medicine, our knowledge is often incomplete and our ability to perform controlled experiments is poor. You do the best you can, but ultimately you have to take a pragmatic approach and see if things work as you hope they do. One can call it unscientific, but the thousands of patients who are alive due to anti-retroviral therapy are probably grateful for it nonetheless.

Causality is different from virulence. You do not need modifiers like "always" or "never" to establish causality. A microorganism can cause disease in some individuals and not in others; it can cause disease sometimes, and other times not. This (somewhat frustrating) aspect of infectious disease results from the complexity of the interplay between the microbe, your body, your body's immune system, and the environment.

For example, you can have Hepatitis B virus floating around in your blood, but have no symptoms. That does not mean that there is not a distinct clinical syndrome caused by Hepatitis B; it just means that some people only "carry" the virus, and do not manifest the full-on disease.

As another example, the same bug that causes strep throat (Group A Streptococcus, S. pyogenes) is often found just chilling out in people's throats, causing no symptoms whatsoever. A lot of people who have gotten strep throat "a hundred times" are actually Strep carriers, and their repeated sore throats are often caused by viruses instead. However, that does not mean that Group A Strep doesn't cause strep throat. It just doesn't do it all the time.

Coming back to HIV, there are people with a genetic mutation in a chemokine receptor (CCR5) that prevents infection by HIV. It is believed that this receptor plays an important role in the HIV virus' ability to enter and infect a host cell. Similar phenomena of genetic protection from disease have been described for acute gastroenteritis ("stomach flu"), and a long list of other infectious diseases. It is speculated that the sickle cell trait (where you have one normal gene and one sickle cell gene, leading to a milder [often asymptomatic] form of the disease) is so prevalent among those of African descent owing to the partial protection that it provides against malaria.

Unfortunately, to date, no one has found an HIV-infected patient who, given enough time to live through the latent period (up to 10-20 years), did not eventually develop the clinical syndrome of AIDS. Part of the challenge in controlling the spread of HIV, and perhaps part of the reason behind the reluctance to accept causality between HIV and AIDS, is that the time between initial infection and clinically apparent disease is so long. The length of this latency period is, in part, a reflection of how much "safety factor" there is in the immune system. Your immune system has to suffer a pretty devastating insult before you begin to see symptoms.

HIV causing AIDS? It's a consensus that has an overwhelming amount of evidence to back it up. I won't even begin to try to summarize it all, but I will describe the gist of it.

Causality between a microorganism and disease is commonly established through the demonstration of Koch's Postulates. These are not hard-fast rules; some of Koch's Postulates are difficult to prove through ethical experiments. However, in the case of HIV, all of Koch's Postulates have been fulfilled:

1. The microorganism must be found in abundance in all organisms suffering from the disease. The virus has been isolated from every patient with AIDS. On top of that, people have sequenced its genome, elucidated its structure, and taken a picture of it.
2. The microorganism must be isolated from a diseased organism and grown in pure culture. The virus has been isolated from patients with AIDS, and grown in culture. Critics cite the fact that this is very difficult to do, and requires special conditions. Most scientists believe that such special conditions are necessary when you are trying to culture something like a retrovirus. Special requirements for growth are not unique to HIV; for example, no one has ever successfully cultured a pathogenic strain of Treponema pallidum (syphilis) in vitro, but anyone who has ever had syphilis will tell you it is a VERY real disease.
3. The cultured microorganism should cause disease when introduced into a healthy organism. When introduced into a healthy individual, the HIV virus has been found to cause disease. It should be noted that this has only happened a few times in monitored environments, through needle-stick exposures; however, it would not be ethical to experimentally inoculate a healthy person with HIV. There is an overwhelming body of non-experimental evidence to support bloodborne and sexual transmission of the HIV virus, and the evidence shows that everyone who contracts HIV eventually gets AIDS - with OR without therapy.
4. The microorganism must be reisolated from the inoculated, diseased experimental host and identified as being identical to the original specific causative agent. This has been demonstrated in the cases of needle-stick exposures.

Anti-retroviral therapy - while itself is quite dangerous and filled with side effects - has nevertheless been shown in numerous studies to reduce morbidity and mortality in HIV+ patients. Anti-retroviral therapy has also been compared to placebo, and its effects have been found to be beneficial over placebo. Other studies, mostly performed in Africa, have examined the "natural history" (i.e. the untreated progression) of HIV infection; such studies have shown that the natural history of HIV infection leads to the severe immunocompromise characteristic of the AIDS syndrome, followed by death.

Yes, there is plenty of money flowing into AIDS research and drugs. However, that fact fails to prove anything related to this discussion, one way or another. There was a point in time when the HIV-AIDS connection was, indeed, a hypothesis; many people cite evidence from that period of time in making the claim that HIV->AIDS is still a widely disputed theory. However, a careful examination of the current scientific evidence will reveal an overwhelming body of evidence supporting a causal relationship between HIV and AIDS.