N=1000 Res=replicate(N,F()) par(mfrow=c(2,1)) hist(Res[1,],xlab="P(winner)",ylab="",main="Change",col="gray") hist(Res[2,],xlab="P(Winner)",ylab="",main="No Change",col="gray") summary(Res[1,]) summary(Res[2,]) #http://i.imgur.com/XmQ82Lc.jpg0752卵の名無しさん2017/04/26(水) 15:25:22.24ID:vNHKSJLD ROLE AND LIMITATIONS OF STATISTICS Much of medicine is inherently probabilistic. Not everyone with hypercholesterolemia who is treated with a statin is prevented from having a myocardial infarction, and not everyone not treated does have one, but statins reduce the probability of a myocardial infarction in such patients. Because so much of medicine is based on probabilities, studies must be performed on groups of people to estimate these probabilities. Three component tasks of statistics are: selecting a sample of subjects for study, describing the data from that sample, and drawing inferences from that sample to a larger population of interest. 0753卵の名無しさん2017/04/29(土) 08:57:14.64ID:j6+oCrzG .stan_code = ' data { int n_obs; real[n_obs] x; } parameters { real mu; real<lower=0> sigma; } model { x ~ normal(mu, sigma); }'
model{ for(i in 1:2){//行 for(j in 1:2){//列 n[i,j]~ binomial(N[i],p[i][j]);//ex. n[1,2] ~ Binom(N[1],p[1][2]) }//行ごと(=介入群、対照群ごと）に二項分布 } /* n[1,1] ~ binomial(N[1],p[1][1]); n[1,2] ~ binomial(N[1],p[1][2]); n[2,1] ~ binomial(N[2],p[2][1]); n[2,2] ~ binomial(N[2],p[2][2]); */ } 0760卵の名無しさん2017/05/04(木) 18:33:21.26ID:6FvZGmBc generated quantities{ real d; real delta_over; real p11; real p10; real p01; real p00; real RR; real OR; p11 = p[1][1];//介入あり効果あり、(暴露あり病気あり) p10 = p[1][2];//介入あり効果なし、(暴露あり病気なし) p01 = p[2][1];//介入なし効果あり、(暴露なし病気あり) p00 = p[2][2];//介入なし効果なし、(暴露なし病気なし) d = p11 - p01;//暴露あり病気あり割合−暴露なし病気あり割合 delta_over = step(d);// ifelse(d>0,1,0)に相当 RR = p11/p01; OR = (p11/p10)/(p01/p00); } 0761卵の名無しさん2017/05/04(木) 18:33:49.02ID:6FvZGmBc library("rstan") rstan_options(auto_write = TRUE) options(mc.cores = parallel::detectCores())
options(scipen = 10)
# UC Berkeley gender bias # Men 8442 44% # Women 4321 35%
n = matrix(round(c(8442*.44,8442*.56,4321*.35,4321*.65)),ncol=2,byrow=TRUE) ; n N=apply(n,1,sum) ; N data <- list(n=n,N=N) model2x2=stan_model('2x2.stan') fit2x2 <- sampling(model2x2,data=data,seed=1234)
l.05=uniroot(function(x)rr2p(a,N1,b,N0,x)-0.05,c(0.5,1.0))$root points(l.05,0.05,pch=19) u.05=uniroot(function(x)rr2p(a,N1,b,N0,x)-0.05,c(5,10))$root points(u.05,0.05,pch=19) text(l.05,0.05+0.02,round(l.05,2)) text(u.05,0.05+0.02,round(u.05,2)) points(1,prr1) text(1,prr1+0.02,round(prr1,2)) text(rr,0,round(rr,3)) 0765卵の名無しさん2017/05/10(水) 22:14:20.37ID:trFWFCkB There is no reason for national medical school gradutates to envy uraguchi bona fide morons who bought their way into the bottom medical school.
No offense, but a grim reality. There is nothing to be more ashamed among doctors than buying their way into the bottom medical school. Owing to this original sin, they cannot even name their honorable alma mater.
In order to keep their self-esteem, these moronic graduates cannot help but call other genuine doctors charlatans against their better judgement. What a pity! The bottom line is that it sucks to be an uraguchi.
If any doctor tells that he will see a patient immediately when he cannot, the patient will label him a liar. The doctor of his word won't behave in such a dishonest way. If one insists that he can name ten of his classmates when he actually cannot, he will be a liar.
As for the uraguchi graduates, it seems to be intolerably dishonorable to uncover their alma mater. Far more dishonorable than being a liar.
There will be no English response from bona fide morons suffering EBMS（Expellee from Bottom Medical School) Syndrome.
Last but not least, it is not the bottom medical school but its enrollee that is despicable, which deserves to be called a bona fide moron beyond redemption. 0766卵の名無しさん2017/05/11(木) 19:03:05.80ID:/tTz9Crr 製薬会社の説明会ではよく大規模スタディの解析結果がよく提示される。 大規模試験でp<0.001が示されると、それだけ信頼できると騙される医者が多すぎ。 まあ、臨床に統計は必要ないというド底辺特殊シリツ医大卒は問題外だが。
大規模スタディとは数多く集めないと統計的有意差がでないような薬効の検証試験。 スカイダイビングにパラシュート着用が死亡率を低下させるかどうかを証明するのに大規模試験は不要ｗ ド底辺特殊シリツ医大卒が国立卒より低学力であるのは>29と>30の投稿を比較するだけでEvident. 0767卵の名無しさん2017/05/11(木) 20:50:15.65ID:/tTz9Crr # Normal Ratio Distribution
X=scale(rnorm(N))*sx+mx Y=scale(rnorm(N))*sy+my hist(X/Y, freq=FALSE,breaks=50,col='lightblue',main='Gaussian ratio distribution') curve(dGRD(x,mx,sx,my,sy,d),add=TRUE,lwd=2) 0768卵の名無しさん2017/05/12(金) 06:02:32.57ID:e0eXW0W5 データをプールして求めたサマリー感度は62.3％（95％信頼区間57.9-66.6％）で、サマリー特異度は98.2％（23.8-98.7％）。これらの値を基に計算した陽性尤度比は34.5（23.8-45.2）、陰性尤度比は0.38（0.34-0.43）になった。 0769卵の名無しさん2017/05/12(金) 06:32:32.16ID:e0eXW0W5 Data Synthesis:
159 studies evaluated 26 RIDTs, and 35% were conducted during the H1N1 pandemic. Failure to report whether results were assessed in a blinded manner and the basis for patient recruitment were important quality concerns. The pooled sensitivity and specificity were 62.3% (95% CI, 57.9% to 66.6%) and 98.2% (CI, 97.5% to 98.7%), respectively. The positive and negative likelihood ratios were 34.5 (CI, 23.8 to 45.2) and 0.38 (CI, 0.34 to 0.43), respectively.
Sensitivity estimates were highly heterogeneous, which was partially explained by lower sensitivity in adults (53.9% [CI, 47.9% to 59.8%]) than in children (66.6% [CI, 61.6% to 71.7%]) and a higher sensitivity for influenza A (64.6% [CI, 59.0% to 70.1%) than for influenza B (52.2% [CI, 45.0% to 59.3%). 0770卵の名無しさん2017/05/12(金) 06:51:38.84ID:e0eXW0W5 そういえば薬の説明会に某社のMRがマスクをしてきて社内でインフルエンザが流行っているといってたな。 先週の当直でも近くの老健でも入所者や職員にインフレンザがくすぶっていた。 0771卵の名無しさん2017/05/13(土) 09:05:01.10ID:QvHCxQMI ci.pLH <- function (a, b, c, d, cl = 0.95) { # a:TP b:FP c:FN d:TP LH = (a/(a+c))/(b/(b+d)) # TP/FP Z = qnorm(1 - (1 - cl)/2) # 1.96 RRL = LH * exp(-Z * sqrt(1/a - 1/(a+c) + 1/b - 1/(b+d))) RRU = LH * exp( Z * sqrt(1/a - 1/(a+c) + 1/b - 1/(b+d))) CI = c(LH,RRL, RRU) return(CI) }
http://www.asahi.com/articles/ASK3J3CHDK3JULBJ004.html0776卵の名無しさん2017/05/14(日) 12:39:20.08ID:75s6edjBhttp://i.imgur.com/ArPaux9.png0777卵の名無しさん2017/05/15(月) 19:25:44.29ID:k0A2P6EO If, however, one uses the limits only to determine whether the null point lies inside or outside the confidence interval, one is only performing a significance test. It is lamentable to go to the trouble to calculate confidence limits and then use them for nothing more than classifying the study finding as statistically significant or not. One should instead remember that the precise locations of confidence limits are not important for proper interpretation. Rather, the limits should serve to give one a mental picture of the location and spread of the entire P-value function. 0778卵の名無しさん2017/05/15(月) 19:28:09.11ID:k0A2P6EO The likelihood of a specified parameter value given observed data is defined as the probability of the observed data given that the true parameter equals the specified parameter value. 0779卵の名無しさん2017/05/18(木) 19:01:17.64ID:zeLyudIOhttps://www.ncbi.nlm.nih.gov/pmc/articles/PMC5147894/table/pone.0167480.t002/
ns=round(Stay0[1]/4) sty0=c(sample(Stay0[2]:Stay0[3],ns,replace=TRUE), sample(Stay0[3]:Stay0[4],ns,replace=TRUE), sample(Stay0[4]:Stay0[5],ns,replace=TRUE), sample(Stay0[5]:Stay0[6],ns,replace=TRUE)) # hist(sty0,breaks=100,main='control');quantile(sty0) wilcox.exact(sty0,sty1)$p.value } N=10000 Res=replicate(N,Sim_Stay()) hist(Res,freq=FALSE,col='lightgreen',breaks=100) lines(density(Res),add=TRUE) mean(Res<0.05) max(Res) 0784卵の名無しさん2017/05/20(土) 10:52:57.73ID:GAnNqfB7 Physician age and outcomes in elderly patients in hospital in the US: observational study http://www.bmj.com/content/357/bmj.j1797.full.pdf のTable 2のデータを用いて コクラン・アーミテージ検定をしてみる。
it is not the kickass bottom medical school but its enrollee that is despicable, which deserves to be called a bona fide moron beyond redemption. ド底辺シリツ医大が悪いんじゃない、本人の頭が悪いんだ。 0789卵の名無しさん2017/05/24(水) 14:21:19.79ID:olQQP7cX 女 男 若 １５ ５ 老 ３ １０ 0790卵の名無しさん2017/05/25(木) 13:31:59.33ID:SNaobD6X 死亡患者数 総患者数 医師数 女性医師 46007 415605 18751 男性医師 137811 1199399 39593
2-sample test for equality of proportions without continuity correction
data: c(M1[1, 1], M1[2, 1]) out of c(M1[1, 2], M1[2, 2]) X-squared = 54.01, df = 1, p-value = 0.0000000000001994 alternative hypothesis: two.sided 95 percent confidence interval: -0.005312783 -0.003089587 sample estimates: prop 1 prop 2 0.1106989 0.1149000 で有意となり、 生存率差の９５％信用区間もアブストラクトの95%CI, -0.57% to -0.28%;とほぼ一致。 0792卵の名無しさん2017/05/26(金) 16:45:01.86ID:g0v5lQqI 10 13 15 6 0793卵の名無しさん2017/05/28(日) 04:15:53.81ID:23SWWCfN A little old lady goes to the doctor and says, "Doctor I have this problem with gas, but it really doesn't bother me too much. They never smell and are always silent. As a matter of fact I've farted at least 20 times since I've been here in your office. You didn't know I was farting because they didn't smell and are silent". The doctor says, "I see. Take these pills and come back to see me next week." The next week the lady goes back, "Doctor" she says, "I don't know what the heck you gave me, but now my farts... although still silent they stink terribly." "Good" the doctor said, “now that we've cleared up your sinuses, let's work on your hearing!" 0794卵の名無しさん2017/05/28(日) 09:12:32.95ID:sdpbW2IZ これって名投稿だと感心する。
it is not the kickass bottom medical school but its enrollee that is despicable, which deserves to be called a bona fide moron beyond redemption. ド底辺シリツ医大が悪いんじゃない、本人の頭が悪いんだ。 0795卵の名無しさん2017/05/28(日) 09:12:56.33ID:sdpbW2IZ 10 5 6 8 0796卵の名無しさん2017/05/28(日) 18:57:09.64ID:1dIxNVDD ## RR=OR*(1-P1) + P1 OR=(RR-P1)/(1-P1) # P1:暴露群でのイベント発生率=a/(a+b) P1=(OR-RR)/(OR-1)
library(fmsb) fisher.test(.mm) pairwise.fisher.test(.mm,p.adjust='holm') pairwise.fisher.test(.m[,1],.m[,2],p.adjust='holm') pairwise.fisher.test(.m[,1],.m[,2],p.adjust='bon') pairwise.fisher.test(.m[,1],.m[,2],p.adjust='none') pairwise.fisher.test(.mm,p.adjust='holm') 0803卵の名無しさん2017/06/03(土) 06:14:04.74ID:fHjCdMkb 身体診察の教科書において最高峰として名高い『Sapira's Art and Science of Bedside Diagnosis』 にこんな記述がある。
More than 100 books and thousands of articles have been published on EBM. It has been subjected to major criticism, including the observation that it is not itself evidence based, in that there is no convincing evidence that physicians using it provide any better care than those who do not.
Perhaps the most balanced definition is that EBM constitutes “methods of incorporating epidemiologic evidence into clinical practice” (Cohen et al., 2004). 0804卵の名無しさん2017/06/06(火) 00:38:28.52ID:vZoASpBs Three business men were sitting in a bar, drinking and discussing how stupid their wives were. The first says, "I tell you, my wife is so stupid. Last week she went to the supermarket and bought $300 worth of meat because it was on sale, and we don't even have a fridge big enough to keep it in!" The second agrees that she sounds pretty thick, but says his wife is thicker. "Just last week, she went out and spent $17,000 on a new car," he laments, "and she doesn't even know how to drive!" The third, a blond male, nods sagely and agrees that these two women sound like they both walked through the stupid forest and got hit by every branch. However, he still thinks his wife is dumber. "I have to laugh when I think about it," he chuckles. "Last week my wife left on a vacation to Greece. I watched her packing her bags and she must have taken at least five boxes of condoms with her. She doesn't even have a penis!"
