此例經常看見一些中醫愛好者用之指責現代醫學，請問這是真的嗎？

02-23

據說1973年，以色列全國醫生大罷工。為期長達一個月，根據耶路撒冷埋葬協會的統計指出該月的全國死亡人數下降了50%。
十年後，1983年，以色列醫生再次舉行全國總罷工，這次罷工的時間長，長達85天，神奇的事情發生了：過後，以色列耶路撒冷埋葬協會又做了一次統計，在這85天中，全國人口的死亡率又下降了50%。

其中一個原因就是做手術的人少了，短期內因手術引起的死亡也少了。但是這些做手術的患者本身就面臨著死亡風險，延期手術從長期統計上來說會造成更高的死亡率。

從統計學的觀點來說，這個例子的運用至少存在以下統計謬誤：

一、案例和統計結論的區別。

兩次罷工均引起死亡率下降，並不等於「每一次」罷工都導致。如果這十年期間發生了十起醫生罷工，但只有這兩起引起死亡率下降。那麼兩者之間的相關係數就很小了。這是樣本容量和總體之間的關係。

二、統計描述和偶然事件

對於不能用統計學進行描述的事件，理論上是不能用數學的方法來證明它的相關性的。從實踐的角度看，7個或以上的樣本才能引用它們的平均值，30個以上的樣本才能引用它的標準差。（這兩個數據在不同領域中也許有不同，但結論是通用的：樣本容量不夠的數據是無法進行統計描述的）。無法重複發生的事件，不能用統計學的角度來證明相關性，更加不能證明因果關係。

三、數據的相關性和數據的因果關係

假設有兩組數據，變數B隨著變數A的改變而改變，相關係數R平方超過0.8。那麼是不是能說明A就是引起B的原因呢？結論是不能。這個結果只能說明兩個組數據有相關性，不能證明數據之間的因果關係。典型的統計學謬誤如鸛鳥送來嬰兒。

四、能確定因果關係的只有設計實驗

當兩組數據高度相關時，必須進一步證明數據之間存在因果關係。唯一的辦法就是在受控的環境下，單獨改變變數A，同時觀察變數B有沒有跟隨變化。這是因為從第二點可知，變數A和變數B相關，有可能他們共同受第三個未知的變數C的改變而改變。因此必須設計一個實驗，讓所有未知變數保持不變，單獨改變變數A。如果此時變數B跟著改變，那麼因果關係成立。否則，因果關係不成立。

綜上所述，在沒有統計學證明變數A（醫生罷工）和變數B（死亡率）之間有因果關係之前，引用這個數據證明不了任何結論。

先不說兩個實例是否能說明相關性，相關性又是否說明因果性這種老生常談的問題了……

我就當它真有相關性，並且真有因果性好了（儘管我們原本應該另外找個工具變數來試著說明它的）……

短期影響和長期影響就這麼不做區分了？……

我是特別支持中醫愛好者不去看西醫的。

但是！

一般這種時候，中醫支持者馬上又裝回了中立派，表示不該打沒意義的賭。

怎麼辦？我也很難過啊……

只怪敵軍太狡猾

那麼問題來了。

題主既然問了，要不要站出來試一發？

順便為了防止缺心眼在評論區挑釁，自己先行立誓。

本人此生不尋求也不接受以陰陽五行等中醫理論為指導的治療方式。

Dear Cecil:

Hi, Cecil. I have come across a number of seemingly credible reports suggesting that every time doctors go on strike the overall death rate goes down, in some cases quite precipitously. Can you ascertain if this in fact is the case or if some other factors are at work here?

— Jacob

As well see below, Jacob, its not really so surprising that mortality statistics sometimes show a drop during a doctors strike. Whats staggering is that a reasonable person could see such stats and for even an instant think: Holy crap, those doctors are killing us. Sure, therell always be a few alternative-medicine fringe dwellers who genuinely see the medical establishment as some sinister cabal presiding over a high-density feedlot of human misery. But the way this "fact" about doctors strikes gets passed around suggests that a lot more people are a little more nuts than youd want to imagine.

No ones suggesting that changes in the availability of doctors wouldnt affect what happens to their patients in the short term - a study from last May, for instance, found that in both Australia and the U.S. birth rates decreased from 1 to 4 percent on days when OB/GYNs held their annual conferences. But looking over the data on doctors strikes, its hard to see where the MDs are saving many lives by staying out of the OR:

One example often cited is that of a monthlong strike by Los Angeles County physicians in 1976, during which the mortality rate for patients was seen to drop by 18 percent. But a 1979 study in the American Journal of Public Health showed that the overall area death rate remained unchanged, as enough personnel remained on duty to handle the real emergencies. Any seeming benefit to patients health likely resulted from about 11,000 fewer operations (presumably elective) being performed that month than was typical, meaning that an estimated 50 to 150 patients who could have been expected to die didnt.
A four-month Israeli doctors strike in 1983 was found to have some definable effects on public health - the percentage of cesarean sections increased somewhat, and one study suggested hypertension patients might have received worse treatment - but no observed impact on mortality. Nonetheless, the popular wisdom saw the work stoppage as a disaster: a detailed study of public perceptions afterward found that nearly one in four urgent-care patients (or their relatives) felt the strike had resulted in major health problems.
A 1984 doctors strike in Varkaus, Finland, mainly meant fewer visits for colds and stomachaches; no significant harmful effects to the public were seen, researchers found, and the uptick in visits after the doctors returned to work suggested that patients were glad to have them back.
Another key example used to support the proposition that fewer doctors means fewer deaths comes from a June 2000 article in the British Medical Journal written during another Israeli strike; the author reported that in the three months after doctors walked out death rates fell significantly in affected cities. However, her data was by no means the result of a scientific study but consisted mostly of anecdotal reports from funeral home directors, who claimed theyd seen "the same thing in 1983." What is known is that, as in the LA strike, many thousands of elective surgeries were postponed but emergency rooms and chronic care departments remained open.
In 2003 a SARS outbreak closed four hospitals in Toronto, and all non-emergency services were suspended. Among other things, this led to the canceling of a quarter to a half of joint-replacement surgeries, 40 percent of cardiac surgeries, and as much as 93 percent of some outpatient procedures. The result? The greater Toronto area did see a slight dip in mortality rate relative to the prior two years, but so did the rest of Ontario, and the decrease wasnt statistically significant anyway.

So despite media suggestions to the contrary, doctors going on strike doesnt seem to have much effect on the death rate one way or the other, and any reduction seen is probably the result of postponed or canceled nonemergency surgeries. And that figures: any surgery is risky, and some common procedures (like coronary bypass or aneurysm repair) have a death rate you just cant ignore. But leaving the tummy tucks out of it, most elective surgeries boast a pretty serious payoff, either in quality-of-life improvement right now or in medical trouble avoided down the line. If 600 people die each year as a result of hip-replacement surgery, does that mean the 200,000-plus patients that pulled through were fools to go under the knife? Youre welcome to calculate the odds however you like, and in certain cases it may well make sense to question the value of surgery. As a general proposition, though, if my health is on the line, Im glad to hear that the doctor is in.

— Cecil Adams

Death Rate Drops During Doctor Strike

看起來國外也一樣，替代醫學的信仰者也喜歡舉這個例子。

因為病人跑到其他國家看病了，

我認為這肯定是真的。

如果非典時期，醫生能全部罷工，後果不堪設想。

嗯…醫生罷工了，誰來call it ？

對啊，沒人開死亡證明了。

是真是假題主自己心裡沒點b數嗎？釣魚不要這麼明顯好不好啊。