{"id":741,"date":"2019-12-29T17:38:23","date_gmt":"2019-12-29T08:38:23","guid":{"rendered":"https:\/\/info.zanet.biz\/?p=741"},"modified":"2020-01-07T17:25:36","modified_gmt":"2020-01-07T08:25:36","slug":"network-meta-analysis%e3%82%92openbugs%e3%81%a7","status":"publish","type":"post","link":"https:\/\/info.zanet.biz\/?p=741","title":{"rendered":"Network Meta-analysis\u3092OpenBUGS\u3067"},"content":{"rendered":"\n
<\/div>

Network Meta-analysis\u3082\u3060\u3044\u3076\u666e\u53ca\u3057\u3066\u304d\u3066\u3044\u308b\u3051\u3069\u3001Arm-based model\u3068\u3044\u3046\u306e\u304c\u51fa\u3066\u304d\u3066\u3001\u4eca\u307e\u3067\u306e\u76f8\u5bfe\u7684\u52b9\u679c\u6307\u6a19\u3092\u7d71\u5408\u3059\u308b\u3068\u3044\u3046\u8003\u3048\u65b9\u304c\u30c1\u30e3\u30ec\u30f3\u30b8\u3092\u53d7\u3051\u305f\u3088\u3046\u306a\u72b6\u614b\u306b\u306a\u3063\u3066\u3044\u308b\u3068\u601d\u3046\u3002<\/p>

<\/div><\/div><\/div>\n\n\n\n
<\/div>

\u4eca\u307e\u3067\u306e\u4e3b\u6d41\u306fContrast-based model\u3067\u3059\u3088\u306d\u3002<\/p>

<\/div><\/div><\/div>\n\n\n\n
<\/div>

\u305d\u3046\u306a\u3093\u3060\u3002\u3057\u304b\u3082\u3001\u30d9\u30fc\u30b9\u30e9\u30a4\u30f3\u30ea\u30b9\u30af\u306b\u3064\u3044\u306f\u3001\u5916\u90e8\u306e\u30c7\u30fc\u30bf\u3092\u7528\u3044\u308b\u3053\u3068\u304c\u52e7\u3081\u3089\u308c\u3066\u3044\u305f\u308a\u3059\u308b\u3002Dias S\u305f\u3061\u306f\u305d\u308c\u3092Baseline model\u3068\u547c\u3093\u3067\u3044\u308b*\u3002

\u30e9\u30f3\u30c0\u30e0\u5316\u6bd4\u8f03\u8a66\u9a13\u306e\u5bfe\u7167\u7fa4\u306e\u30c7\u30fc\u30bf\u3092\u305d\u306e\u307e\u307e\u4f7f\u308f\u306a\u3044\u65b9\u304c\u3044\u3044\u3068\u3044\u3046\u8003\u3048\u65b9\u306b\u3082\u4e00\u7406\u3042\u308b\u3002\u30e9\u30f3\u30c0\u30e0\u5316\u6bd4\u8f03\u8a66\u9a13\u306f\u5bfe\u8c61\u8005\u304c\u30bb\u30ec\u30af\u30c8\u3055\u308c\u3066\u3044\u308b\u304b\u3089\u3001Real world\u3067\u306f\u9055\u3046\u7d76\u5bfe\u30ea\u30b9\u30af\u3067\u3042\u308b\u53ef\u80fd\u6027\u306f\u9ad8\u3044\u3068\u8003\u3048\u3089\u308c\u308b\u304b\u3089\u3002\u3057\u304b\u3082\u3001\u500b\u4eba\u500b\u4eba\u306f\u30ea\u30b9\u30af\u304c\u7570\u306a\u308b\u304b\u3089\u3001\u30ea\u30b9\u30af\u6bd4\u306e\u3088\u3046\u306a\u76f8\u5bfe\u7684\u52b9\u679c\u6307\u6a19\u3092\u30e1\u30bf\u30a2\u30ca\u30ea\u30b7\u30b9\u3067\u7d71\u5408\u3057\u3066\u3001\u305d\u308c\u3092\u9069\u7528\u3057\u3066\u7d76\u5bfe\u52b9\u679c\u3092\u8003\u3048\u305f\u65b9\u304c\u3044\u3044\u3068\u3044\u3046\u8003\u3048\u306f\u7406\u89e3\u3067\u304d\u308b\u3002\u3057\u304b\u3057\u3001\u4e00\u65b9\u3067\u7d76\u5bfe\u30ea\u30b9\u30af\u3092\u3069\u306e\u30c7\u30fc\u30bf\u304b\u3089\u6c42\u3081\u308b\u3079\u304d\u304b\uff1f\u306a\u304b\u306a\u304b\u7c21\u5358\u306b\u306f\u3044\u304b\u306a\u3044\u3002

\u30ea\u30b9\u30af\u6bd4\u306e\u3088\u3046\u306a\u76f8\u5bfe\u7684\u52b9\u679c\u6307\u6a19\u306e\u5024\u306f\u30d9\u30fc\u30b9\u30e9\u30a4\u30f3\u30ea\u30b9\u30af\u304c\u5909\u308f\u3063\u3066\u3082\u5927\u304d\u304f\u306f\u5909\u52d5\u3057\u306a\u3044\u3068\u3044\u3046\u3053\u3068\u304c\u7d4c\u9a13\u7684\u30c7\u30fc\u30bf\u3067\u3082\u793a\u3055\u308c\u3066\u3044\u3066\u3001\u30e1\u30bf\u30a2\u30ca\u30ea\u30b7\u30b9\u3067\u7d71\u5408\u3055\u308c\u308b\u305d\u308c\u305e\u308c\u306e\u7814\u7a76\u306e\u30d9\u30fc\u30b9\u30e9\u30a4\u30f3\u30ea\u30b9\u30af\u304c\u7570\u306a\u3063\u3066\u3044\u3066\u3082\u76f8\u5bfe\u7684\u52b9\u679c\u6307\u6a19\u306f\u307b\u307c\u540c\u3058\u306b\u306a\u308b\u306f\u305a\u3060\u304b\u3089\u305d\u308c\u3092\u7d71\u5408\u3057\u305f\u307b\u3046\u304c\u3044\u3044\u3068\u601d\u308f\u308c\u3066\u6765\u305f\u3002\u3000

\u5358\u7d14\u306b\u8ad6\u7406\u7684\u306b\u306e\u307f\u8003\u3048\u308b\u3068\u3001Arm-based model\u306f\u7814\u7a76\u3054\u3068\u306e\u30e9\u30f3\u30c0\u30e0\u5272\u308a\u4ed8\u3051\u3092\u7121\u8996\u3059\u308b\u3053\u3068\u306b\u306a\u308b\u3068\u3044\u3046\u6279\u5224\u3082\u6b63\u3057\u3044\u3088\u3046\u306b\u3082\u601d\u3048\u308b\u3051\u3069\u3001\u5b9f\u969b\u306b\u306f\u305d\u3046\u3067\u3082\u306a\u3044\u3068\u3044\u3046\u53cd\u8ad6\u3082\u3042\u3063\u3066\u4e00\u6982\u306b\u306f\u305d\u3046\u306f\u8a00\u3048\u306a\u3044\u3067\u3057\u3087\u3046**\u3002

\u307e\u305a\u306fZhang J, et al: Network meta-analysis of randomized clinical trials: reporting the proper summaries. Clin Trials 2014;11:246-62. PMID: 24096635<\/a>\u3067\u51fa\u3066\u304f\u308bArm-based model\u306e\u30cd\u30c3\u30c8\u30ef\u30fc\u30af\u30e1\u30bf\u30a2\u30ca\u30ea\u30b7\u30b9\u306e\u305f\u3081\u306eBUGS\u30b3\u30fc\u30c9\u3068\u30c7\u30fc\u30bf\u90e8\u5206\u3092\u793a\u3057\u307e\u3059\u3002(\u3053\u3061\u3089\u306e\u8ad6\u6587\u3082\u3088\u304f\u66f8\u304b\u308c\u3066\u3044\u307e\u3059\uff1aZhang J, et al: Bayesian hierarchical models for network meta-analysis incorporating nonignorable missingness. Stat Methods Med Res 2017;26:2227-2243. PMID: 26220535<\/a>)

\u4eca\u65e5\u306e\u76ee\u7684\u306fOpenBUGS<\/a>\u3092\u52d5\u304b\u3057\u3066\u30cd\u30c3\u30c8\u30ef\u30fc\u30af\u30e1\u30bf\u30a2\u30ca\u30ea\u30b7\u30b9\u3092\u5b9f\u884c\u3059\u308b\u3053\u3068\u306a\u3093\u3060\u3051\u3069\u3001\u3053\u306eArm-based model\u306e\u30cd\u30c3\u30c8\u30ef\u30fc\u30af\u30e1\u30bf\u30a2\u30ca\u30ea\u30b7\u30b9\u3092\u3084\u3063\u3066\u307f\u3088\u3046\u3068\u601d\u3046\u3002OpenBUGS\u306e\u4f7f\u3044\u65b9\u306e\u7b2c\u4e00\u6b69\u306f\u4ee5\u524d\u6295\u7a3f<\/a>\u3057\u307e\u3057\u305f\u3002

\uff08\u53c2\u8003\u6587\u732e\uff1a
*Dias S, Ades AE, Welton NJ, Jansen JP and Sutton AJ: Network Meta-Analysis for Decision Making. 2018, John Wiley & Sons Ltd.
**Hong H, Chu H, Zhang J, Carlin BP: Rejoinder to the discussion of “a Bayesian missing data framework for generalized multiple outcome mixed treatment comparisons,” by S. Dias and A. E. Ades. Res Synth Methods 2016;7:29-33. PMID: 26461816<\/a>\uff09<\/p>

<\/div><\/div><\/div>\n\n\n\n

model {
for(i in 1:tS) {
p[i]<-phi(mu[t[i]]+ vi[s[i], t[i]]) # model
r[i]~dbin(p[i], totaln[i]) # binomial likelihood
}
for(j in 1:sN){
vi[j, 1:tN]~dmnorm(mn[1:tN], T[1:tN,1:tN]) # multivariate normal distribution
}
invT[1:tN, 1:tN]<-inverse(T[,])
for (j in 1:tN){
mu[j]~dnorm(0, 0.001)
sigma[j]<-sqrt(invT[j, j])
probt[j]<-phi(mu[j]\/sqrt(1+invT[j, j]))
#population-averaged treatment specific event rate
}
T[1:tN,1:tN]~dwish(R[1:tN, 1:tN], tN) # Wishart prior
for (k in 1:tN) {
rk[k]<- tN+1-rank(probt[],k) # ranking
best[k]<-equals(rk[k],1) # prob {treatment k is best}
}
for (j in 1:tN){ # calculation of RR, RD and OR
for (k in (j+1):tN){
RR[j, k] <- probt[k]\/probt[j]
RD[j, k] <- probt[k]-probt[j]
OR[j, k] <- probt[k]\/(1-probt[k])\/probt[j]*(1-probt[j])
}
}
}
#DATA
list(tN=4,sN=24,tS=50,mn=c(0, 0, 0, 0),R=structure(.Data=c(1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1),.Dim=c(4,4)))
s[] t[] r[] totaln[]
1 1 9 140
1 3 23 140
1 4 10 138
2 2 11 78
2 3 12 85
2 4 29 170
3 1 75 731
3 3 363 714
4 1 2 106
4 3 9 205
5 1 58 549
5 3 237 1561
6 1 0 33
6 3 9 48
7 1 3 100
7 3 31 98
8 1 1 31
8 3 26 95
9 1 6 39
9 3 17 77
10 1 79 702
10 2 77 694
11 1 18 671
11 2 21 535
12 1 64 642
12 3 107 761
13 1 5 62
13 3 8 90
14 1 20 234
14 3 34 237
15 1 0 20
15 4 9 20
16 1 8 116
16 2 19 146
17 1 95 1107
17 3 134 1031
18 1 15 187
18 3 35 504
19 1 78 584
19 3 73 675
20 1 69 1177
20 3 54 888
21 2 20 49
21 3 16 43
22 2 7 66
22 4 32 127
23 3 12 76
23 4 20 74
24 3 9 55
24 4 3 26
END <\/p>\n\n\n\n

<\/div>

\u4e0a\u306e\u30d6\u30ed\u30c3\u30af\u3092\u3059\u3079\u3066\u9078\u629e\u3057\u3066\u3001\u30b3\u30d4\u30fc\u3057\u3066\u3001OpenBUGS\u3092\u7acb\u3061\u4e0a\u3052\u3066\u3001File\u30e1\u30cb\u30e5\u30fc\u304b\u3089New\u3092\u9078\u3093\u3067\u3001\u30a8\u30c7\u30a3\u30bf\u30fc\u753b\u9762\u306b\u8cbc\u308a\u4ed8\u3051\u307e\u3057\u305f\u3002<\/p>

<\/div><\/div><\/div>\n\n\n\n
<\/div>

\u305d\u3046\u305d\u3046\u3002\u305d\u3057\u305f\u3089\u3001\u4e0b\u306e\u56f3\u306e\u3088\u3046\u306b\u3001Model\u30e1\u30cb\u30e5\u30fc\u304b\u3089Specification…\u3092\u9078\u3093\u3067\u3001Specification Tool\u3092\u8868\u793a\u3055\u305b\u3001\u30b3\u30fc\u30c9\u306emodel\u306e\u6587\u5b57\u5217\u3092\u9078\u629e\u3057\u3066\u53cd\u8ee2\u3057\u305f\u72b6\u614b\u3067\u3001Specification Tool\u306echeck model\u30dc\u30bf\u30f3\u3092\u30af\u30ea\u30c3\u30af\u3057\u307e\u3059\u3002<\/p>

<\/div><\/div><\/div>\n\n\n\n
\"\"
check model<\/figcaption><\/figure>\n\n\n\n
<\/div>

\u30dc\u30c8\u30e0\u30d0\u30fc\u306bmodel is syntactically correct\u3068\u8868\u793a\u3055\u308c\u307e\u3057\u305f\u3002\u78ba\u304b\u6b21\u306e\u30b9\u30c6\u30c3\u30d7\u306f\u30c7\u30fc\u30bf\u306e\u8aad\u307f\u8fbc\u307f\u3001load data\u3067\u3057\u305f\u3088\u306d\u3002<\/p>

<\/div><\/div><\/div>\n\n\n\n
<\/div>

\u305d\u3046\u305d\u3046\u3002\uff03DATA\u306elist\u306e\u6587\u5b57\u5217\u3092\u9078\u629e\u3057\u3066\u53cd\u8ee2\u3055\u305b\u3066\u3001Specification Tool\u306eload data\u30dc\u30bf\u30f3\u3092\u30af\u30ea\u30c3\u30af\u3057\u307e\u3057\u3087\u3046\u3002<\/p>

<\/div><\/div><\/div>\n\n\n\n
<\/div>

list\u3092\u30c0\u30d6\u30eb\u30af\u30ea\u30c3\u30af\u3057\u305f\u3089\u3001Specification Tool\u306eload data\u30dc\u30bf\u30f3\u304c\u30a2\u30af\u30c6\u30a3\u30d6\u306b\u306a\u308a\u307e\u3057\u305f\u3002\u30af\u30ea\u30c3\u30af\u3057\u307e\u3057\u305f\u3002\u3042\u3063\u3001\u30dc\u30c8\u30e0\u30d0\u30fc\u306bdata loaded\u3068\u8868\u793a\u3055\u308c\u307e\u3057\u305f\u3002<\/p>

<\/div><\/div><\/div>\n\n\n\n
\"\"
load data<\/figcaption><\/figure>\n\n\n\n
<\/div>

\u305d\u308c\u3067\u306f\u6b21\u306b\u89e3\u6790\u5bfe\u8c61\u306e\u30c7\u30fc\u30bf\u3092\u8aad\u307f\u8fbc\u307e\u305b\u307e\u3059\u3002\u3053\u306e\u30c7\u30fc\u30bf\u306fR\u306e\u30d1\u30c3\u30b1\u30fc\u30b8\u3067\u3042\u308bpcnetmeta\u3084\u3044\u308d\u3093\u306a\u30d1\u30c3\u30b1\u30fc\u30b8\u306b\u4ed8\u5c5e\u3057\u3066\u3044\u308b\u7981\u7159\u6cbb\u7642\u306e\u30e9\u30f3\u30c0\u30e0\u5316\u6bd4\u8f03\u306e\u30c7\u30fc\u30bf\u3067\u3059\u3002
\u7814\u7a76\u6570\u306f24\u4ef6\u3001\u6cbb\u7642\u306e\u6570\u306f4\u7a2e\u985e\uff081) no contact (NC); 2) self-help (SH); 3) individual counselling (IC); 4) group counselling (GC)\u3067\u3001\u6cbb\u7642\u30a2\u30fc\u30e0\u306e\u6570\u306f50\u3067\u3059\u3002\u4e00\u884c\u304c\u3072\u3068\u3064\u306e\u6cbb\u7642\u30a2\u30fc\u30e0\u3067\u3059\u3002\u30b3\u30fc\u30c9\u306e#DATA\u4e2d\u306etN\u304c\u6cbb\u7642\u306e\u6570\u3001sN\u304c\u7814\u7a76\u6570\u3001\uff54S\u304c\u6bd4\u8f03\u30da\u30a2\u306e\u6570\u3001\u306b\u5bfe\u3059\u308b\u5909\u6570\u540d\u306b\u306a\u3063\u3066\u307e\u3059\u3002\u6cbb\u7642\u306e\u7a2e\u985e\u306f\u6b21\u306e\u30c7\u30fc\u30bf\u90e8\u5206\u306et[]\u3068\u3044\u3046\u5909\u6570\u306b1\uff5e4\u306e\u6574\u6570\u3067\u6307\u5b9a\u3055\u308c\u3066\u3044\u307e\u3059\u3002\u7814\u7a76\u756a\u53f7\u306f\u540c\u3058\u304fs[]\u3068\u3044\u3046\u5909\u6570\u3067\u8868\u3057\u3066\u307e\u3059\u3002
\u3053\u306e\u30c7\u30fc\u30bf\u30bd\u30fc\u30b9\u306fHasselblad V (1998) “Meta-analysis of multitreatment studies.” Med Decis Making 18(1), 37\u201343. \u3060\u305d\u3046\u3067\u3059\u3002 <\/p>

<\/div><\/div><\/div>\n\n\n\n
<\/div>

\u306a\u308b\u307b\u3069\u3002\u6b8b\u308a\u306e\u30c7\u30fc\u30bf\u306f\u8868\u5f62\u5f0f\u307f\u305f\u3044\u3067\u3059\u306d\u3002\u78ba\u304b\u3001Excel\u3067\u7528\u610f\u3057\u3066\u3001\u30b3\u30d4\u30fc\u3057\u3066\u3001OpenBUGS\u306eEdit\u30e1\u30cb\u30e5\u30fc\u304b\u3089Paste Special…\u3067Plain Text\u3067\u8cbc\u308a\u4ed8\u3051\u308b\u3093\u3067\u3057\u305f\u306d\u3002<\/p>

<\/div><\/div><\/div>\n\n\n\n
<\/div>

\u305d\u306e\u901a\u308a\u3002\u3067\u3001s[]\u304b\u3089\u4e00\u756a\u4e0b\u306eEND\u306e\u6700\u5f8c\u307e\u3067\u9078\u629e\u3057\u3066\u3001\u53cd\u8ee2\u3055\u305b\u3001Specification Tool\u306eload data\u30dc\u30bf\u30f3\u3092\u30af\u30ea\u30c3\u30af\u3057\u3066\u304f\u3060\u3055\u3044\u3002<\/p>

<\/div><\/div><\/div>\n\n\n\n
<\/div>

\u3084\u308a\u307e\u3057\u305f\u3002\u30dc\u30c8\u30e0\u30d0\u30fc\u306bdata loaded\u3068\u8868\u793a\u3055\u308c\u305f\u306e\u3067\u3001Specification Tool\u306ecompile\u30dc\u30bf\u30f3\u3092\u30af\u30ea\u30c3\u30af\u3057\u307e\u3059\u3002
\u30dc\u30c8\u30e0\u30d0\u30fc\u306bmodel compiled\u3068\u51fa\u307e\u3057\u305f\u3002
load inits\u3068gen inits\u30dc\u30bf\u30f3\u304c\u30a2\u30af\u30c6\u30a3\u30d6\u306b\u306a\u308a\u307e\u3057\u305f\u306d\u3002<\/p>

<\/div><\/div><\/div>\n\n\n\n
<\/div>

\u4eca\u56de\u306f\u3001\u521d\u671f\u5024\u306e\u8a2d\u5b9a\u306e\u305f\u3081\u306e\u30c7\u30fc\u30bf\u3092\u7279\u306b\u7528\u610f\u3057\u3066\u3044\u307e\u305b\u3093\u3002\u4e8b\u524d\u5206\u5e03\u306e\u5024prior\u3092\u30b5\u30f3\u30d7\u30ea\u30f3\u30b0\u3059\u308b\u5834\u5408\u3001\u5e45\u5e83\u3044\u5206\u5e03\u306e\u5834\u5408\u3001\u521d\u671f\u5024\u3092\u8a2d\u5b9a\u3057\u306a\u3044\u3068\u3001\u975e\u5e38\u306b\u306f\u3058\u3063\u3053\u306e\u65b9\u306e\u5024\u304c\u5f97\u3089\u308c\u305f\u5834\u5408\u3001Markov Chain Monte Carlo (MCMC)\u306b\u3088\u308bGibbs\u30b5\u30f3\u30d7\u30ea\u30f3\u30b0\u304c\u3046\u307e\u304f\u52d5\u304b\u306a\u304f\u306a\u308b\u3053\u3068\u304c\u3042\u308b\u306e\u3067\u3001\u305d\u306e\u3088\u3046\u306a\u5834\u5408\u306f\u3001\u521d\u671f\u5024\u306e\u8a2d\u5b9a\u7528\u306e\u30c7\u30fc\u30bf\u3092\u7528\u610f\u3057\u305f\u65b9\u304c\u3044\u3044\u3067\u3059\u3002\u66f8\u304d\u65b9\u306f\u3001\uff03DATA\u306elist()\u3068\u540c\u3058\u3088\u3046\u306a\u66f8\u304d\u65b9\u3067\u3059\u3002
\u4eca\u56de\u306f\u3001T[1:tN,1:tN]~dwish(R[1:tN, 1:tN], tN) # Wishart prior\u306e\u30b3\u30fc\u30c9\u306e\u90e8\u5206\u3067\u3001\u30de\u30c8\u30ea\u30c3\u30af\u30b9\u30c7\u30fc\u30bfR\u304b\u3089Wishart\u5206\u5e03\u306e\u30e9\u30f3\u30c0\u30e0\u30b5\u30f3\u30d7\u30ea\u30f3\u30b0\u3092\u884c\u3063\u3066\u3001\u30de\u30c8\u30ea\u30c3\u30af\u30b9T\u306b\u683c\u7d0d\u3059\u308b\u64cd\u4f5c\u3092\u3057\u3066\u3044\u308b\u306e\u3067\u3001\u3044\u305a\u308c\u306b\u305b\u3088gen inits\u306e\u5b9f\u884c\u304c\u5fc5\u8981\u306b\u306a\u308a\u307e\u3059\u3002load inits\u306a\u3057\u3067\u3001gen inits\u3092\u5b9f\u884c\u3059\u308b\u3068\u3001\u3059\u3079\u3066\u306e\u30e9\u30f3\u30c0\u30e0\u30b5\u30f3\u30d7\u30ea\u30f3\u30b0\u304c\u5b9f\u884c\u3055\u308c\u3001\u305d\u308c\u3089\u306e\u5024\u304b\u3089MCMC\u304c\u958b\u59cb\u3055\u308c\u308b\u3053\u3068\u306b\u306a\u308a\u307e\u3059\u3002<\/p>

<\/div><\/div><\/div>\n\n\n\n
<\/div>

gen inits\u30dc\u30bf\u30f3\u3092\u30af\u30ea\u30c3\u30af\u3057\u305f\u3089\u3001\u30dc\u30c8\u30e0\u30d0\u30fc\u306binitial values generated. model initialized\u3068\u51fa\u307e\u3057\u305f\u3002
\u6b21\u306f\u3001Model\u30e1\u30cb\u30e5\u30fc\u304b\u3089Update…\u3092\u9078\u3093\u3067\u3001Update Tool\u3067updates\u306e\u56de\u6570\u3092\u8a2d\u5b9a\u3059\u308b\u3093\u3067\u3057\u305f\u306d\u300210000\u56de\u306b\u8a2d\u5b9a\u3057\u3066\u307f\u307e\u3057\u305f\u3002\u305d\u308c\u3067\u3001update\u30dc\u30bf\u30f3\u3092\u30af\u30ea\u30c3\u30af\u3057\u3066\u30d0\u30fc\u30f3\u30a4\u30f3\u3092\u5b9f\u884c\u3057\u307e\u3059\u3002<\/p>

<\/div><\/div><\/div>\n\n\n\n
\"\"
Update Tool<\/figcaption><\/figure>\n\n\n\n
<\/div>

\u3042\u3063\u3068\u3044\u3046\u9593\u306b\u30d0\u30fc\u30f3\u30a4\u30f3\u304c\u7d42\u4e86\u3057\u307e\u3057\u305f\u3002

\u6b21\u306f\u3001\u30b5\u30f3\u30d7\u30ea\u30f3\u30b0\u3057\u305f\u5024\u3092\u8a18\u9332\u3059\u308b\u5909\u6570node\u3092\u8a2d\u5b9a\u3059\u308b\u3093\u3067\u3057\u305f\u306d\u3002Inference\u30e1\u30cb\u30e5\u30fc\u304b\u3089Samples…\u3092\u9078\u3093\u3067\u3001Sample Monitor Tool\u3092\u8868\u793a\u3057\u3066\u3001node\u306e\u53f3\u306e\u30d5\u30a3\u30fc\u30eb\u30c9\u306b\u3001probt\u3068\u5165\u529b\u3057\u3066\u3001set\u30dc\u30bf\u30f3\u3092\u30af\u30ea\u30c3\u30af\u3057\u307e\u3059\u3002\u7d9a\u3051\u3066\u3001RR\u3001RD\u3001OR\u3001rk\u3001best\u3068\u5165\u529b\u3057\u3066\u9806\u6b21set\u30dc\u30bf\u30f3\u3092\u30af\u30ea\u30c3\u30af\u3057\u307e\u3057\u305f\u3002\u5168\u90e8\u30676\u500b\u306e\u5909\u6570\u3067\u3059\u3002
probt\u306f\u5404\u6cbb\u7642\u7fa4\u306e\u7d76\u5bfe\u30ea\u30b9\u30afAbsolute risk\u3001RR\u306f\u5404\u6cbb\u7642\u30da\u30a2\u3067\u306e\u30ea\u30b9\u30af\u6bd4\u3001RD\u306f\u540c\u3058\u304f\u30ea\u30b9\u30af\u5dee\u3001OR\u306f\u30aa\u30c3\u30ba\u6bd4\u3001rk\u306f\u5404\u4ecb\u5165\u304c1\u30012\u4f4d…\u306e\u78ba\u7387\u3001best\u306f\u5404\u4ecb\u5165\u304c\u6700\u5584\u3067\u3042\u308b\u78ba\u7387\u3067\u3059\u306d\u3002<\/p>

<\/div><\/div><\/div>\n\n\n\n
<\/div>

\u305d\u306e\u901a\u308a\u3002\u305d\u308c\u3067\u306f\u3001\u30b5\u30f3\u30d7\u30ea\u30f3\u30b0\u306e\u500b\u6570\u309250000\u306b\u8a2d\u5b9a\u3057\u3066\u3001\u3064\u307e\u308a\u3001Update Tool\u306eupdates\u309250000\u306b\u5909\u3048\u3066\u3001update\u30dc\u30bf\u30f3\u3092\u30af\u30ea\u30c3\u30af\u3057\u307e\u3057\u3087\u3046\u3002<\/p>

<\/div><\/div><\/div>\n\n\n\n
<\/div>

\u4eca\u5ea6\u306f30\u79d2\u304f\u3089\u3044\u304b\u304b\u308a\u307e\u3057\u305f\u3002\u7d50\u679c\u3092\u898b\u308b\u306b\u306f\u3001Sample Monitor Tool\u3067node\u306e\u30d5\u30a3\u30fc\u30eb\u30c9\u306b*\u3092\u5165\u529b\u3057\u3066stats\u30dc\u30bf\u30f3\u3092\u30af\u30ea\u30c3\u30af\u3059\u308b\u3093\u3067\u3057\u305f\u306d\u3002\u3053\u308c\u3067\u3001\u8a18\u9332\u3055\u308c\u305f\u5168\u90e8\u306e\u5909\u6570node\u306e\u5e73\u5747\u5024\u3001\u4e2d\u592e\u5024\u300195\uff05\u78ba\u4fe1\u533a\u9593\u306a\u3069\u306e\u5024\u304c\u8868\u793a\u3055\u308c\u308b\u306f\u305a\u3067\u3059\u3088\u306d\u3002
\u51fa\u307e\u3057\u305f\u3002<\/p>

<\/div><\/div><\/div>\n\n\n\n
\"\"<\/figure>\n\n\n\n
<\/div>

\u3053\u306e\u7d50\u679c\u3067\u3001Arm-based model\u306e\u7279\u5fb4\u3068\u3057\u3066\u3001\u5404\u6cbb\u7642\u7fa4\u306e\u7d76\u5bfe\u30ea\u30b9\u30afAbsolute risk\u304c\u5f97\u3089\u308c\u308b\u3068\u3044\u3046\u3053\u3068\u304c\u3042\u3052\u3089\u308c\u307e\u3059\u3002\u3053\u306e\u30c7\u30fc\u30bf\u3067\u306f\u3001probt\u306e\u5024\u304c\u5927\u304d\u3044\u65b9\u304c\u3088\u3044\u7d50\u679c\u3067\u3059\u3002\u3064\u307e\u308a\u7981\u7159\u6210\u529f\u7387\u304c\u9ad8\u3044\u3068\u3044\u3046\u3053\u3068\u3067\u3059\u3002
\u6cbb\u76424\u304c\u4e00\u756a\u6210\u7e3e\u304c\u3044\u3044\u3053\u3068\u304c\u5206\u304b\u308a\u307e\u3059\u306d\u3002\u3053\u308c\u306fGroup counselling\u304c\u7981\u7159\u52b9\u679c\u304c\u6700\u3082\u9ad8\u3044\u3068\u3044\u3046\u3053\u3068\u3067\u3059\u306d\u3002best[4]\u306e\u5e73\u5747\u5024\u304c0.6667\u306a\u306e\u3067\u3001\u6cbb\u76424\u304c\u6700\u5584\u3067\u3042\u308b\u78ba\u7387\u304c\u304a\u3088\u305d0.67\u3067\u3042\u308b\u3053\u3068\u304c\u5206\u304b\u308a\u307e\u3059\u3002<\/p>

<\/div><\/div><\/div>\n\n\n\n
\"\"<\/figure>\n\n\n\n
<\/div>

\u306a\u308b\u307b\u3069\u3001RR, OR, RD\u306b\u3064\u3044\u3066\u306f\u3059\u3079\u3066\u306e\u30da\u30a2\u6bd4\u8f03\u306e\u5024\u304c\u51fa\u3066\u307e\u3059\u306d\u3002
Inference\u30e1\u30cb\u30e5\u30fc\u304b\u3089Comparison Tool…\u3092\u9078\u3093\u3067\u3001Comparison Tool\u3092\u8868\u793a\u3057\u3066\u3001node\u306bRR\u3068\u5165\u529b\u3057\u3066\u3001caterpillar\u30d7\u30ed\u30c3\u30c8\u3092\u8868\u793a\u3057\u3066\u307f\u307e\u3057\u305f\u3002\u30d7\u30ed\u30c3\u30c8\u3092\u53f3\u30af\u30ea\u30c3\u30af\u3057\u3066\u3001Properties\u304b\u3089\u30d5\u30a9\u30f3\u30c8\u3092\u5909\u3048\u305f\u308a\u3057\u3066\u307f\u307e\u3057\u305f\u3002
probt\u3068RR\u306edensity\uff08\u78ba\u7387\u5bc6\u5ea6\u5206\u5e03\uff09\u3082\u8868\u793a\u3057\u3066\u307f\u307e\u3057\u305f\u3002<\/p>

<\/div><\/div><\/div>\n\n\n\n
\"\"
density<\/figcaption><\/figure>\n\n\n\n
<\/div>

\u30ea\u30b9\u30af\u6bd4\u306f\u305d\u306e\u307e\u307e\u3060\u3068\u3001\u6b63\u898f\u5206\u5e03\u3067\u306f\u306a\u3044\u3067\u3059\u306d\u3002\u5bfe\u6570\u5909\u63db\u3059\u308b\u3068\u5de6\u53f3\u5bfe\u79f0\u306e\u5206\u5e03\u306b\u306a\u308b\u306f\u305a\u3067\u3059\u3002

\u3055\u3066\u3001\u4eca\u56de\u306fArm-based model\u3092\u7528\u3044\u305fNetwork Meta-analysis\u3092OpenBUGS\u3067\u884c\u3044\u307e\u3057\u305f\u3002\u540c\u3058\u89e3\u6790\u3092R\u306e\u30d1\u30c3\u30b1\u30fc\u30b8\u3067\u3042\u308bpcnetmeta<\/a>\u3092\u4f7f\u3063\u3066\u5b9f\u884c\u3059\u308b\u3053\u3068\u304c\u3067\u304d\u307e\u3059\u3002\u540c\u3058\u7d50\u679c\u304c\u5f97\u3089\u308c\u308b\u306f\u305a\u3067\u3059\u3002nma.ab.bin()\u3068\u3044\u3046\u95a2\u6570\u3092\u4f7f\u3046\u3093\u3060\u3051\u308c\u3069\u3001model=”het_cor”\u306b\u3057\u305f\u5834\u5408\u3001\u4eca\u56de\u306e\u30e2\u30c7\u30eb\u3068\u540c\u3058\u306b\u306a\u308a\u307e\u3059\u3002pcnetmeta\u3067\u306f\u3001\u4e0b\u306e\u3088\u3046\u306a\u3001Network Graph\u3082\u4f5c\u6210\u3067\u304d\u307e\u3059\u3002
<\/p>

<\/div><\/div><\/div>\n\n\n\n
\"\"<\/figure>\n\n\n\n
<\/div>

pcnetmeta\u3060\u3051\u3067\u306a\u304f\u3001R\u306e\u30d1\u30c3\u30b1\u30fc\u30b8\u306e\u30d7\u30ed\u30b0\u30e9\u30e0\u3084\u30c7\u30fc\u30bf\u30d5\u30a1\u30a4\u30eb\u306fGitHub<\/a>\u3067\u898b\u3089\u308c\u308b\u3093\u3067\u3059\u306d\u3002\u3053\u308c\u3063\u3066\u3059\u3054\u3044\u3067\u3059\u306d\u3002<\/p>

<\/div><\/div><\/div>\n\n\n\n
<\/div>

\u4eca\u56de\u306f\u3001OpenBUGS\u306e\u52d5\u304b\u3057\u65b9\u306b\u3064\u3044\u3066\u306e\u8aac\u660e\u3060\u3051\u3067\u3001Network Meta-analysis\u306e\u7d71\u8a08\u5b66\u7684\u30e2\u30c7\u30eb\u306e\u8aac\u660e\u306f\u3057\u306a\u304b\u3063\u305f\u3051\u308c\u3069\u3001\u672c\u5f53\u306f\u305d\u3063\u3061\u304c\u91cd\u8981\u3067\u3059\u306d\u3002Arm-based model\u306f\u4eca\u307e\u3067\u306e\u3001\u901a\u5e38\u306e\u30da\u30a2\u6bd4\u8f03\u3060\u3051\u306e\u30e1\u30bf\u30a2\u30ca\u30ea\u30b7\u30b9\u306b\u5bfe\u3057\u3066\u3082\u9069\u7528\u3067\u304d\u3066\u3001\u3057\u304b\u3082\u5358\u4e00\u7fa4\u306e\u3064\u307e\u308aSingle-arm\u306e\u30b3\u30db\u30fc\u30c8\u7814\u7a76\u3084\u75be\u60a3\u30ec\u30b8\u30b9\u30c8\u30ea\u30fc\u306e\u30c7\u30fc\u30bf\u306a\u3069\u3082\u6d3b\u7528\u3067\u304d\u308b\u53ef\u80fd\u6027\u304c\u3042\u308b\u3002Zhang J, et al: Bayesian hierarchical methods for meta-analysis combining randomized-controlled and single-arm studies. Stat Methods Med Res 2019;28:1293-1310. PMID: 29433407<\/a>\u306e\u8ad6\u6587\u3067\u308f\u304b\u308b\u3088\u3002\u305d\u308c\u3068\u3001\u9593\u63a5\u6bd4\u8f03\u306e\u969b\u306b\u3001\u6b20\u640d\u5024\u306e\u88dc\u5b8c\u3068\u3044\u3046\u8003\u3048\u3092\u9069\u7528\u3059\u308b\u3068\u3044\u3046Hong P\u3089\u306e\u65b9\u6cd5\u3082\u753b\u671f\u7684\u3067\u3059\u306d\u3002Hong H, et al: A Bayesian missing data framework for generalized multiple outcome mixed treatment comparisons. Res Synth Methods 2016;7:6-22. PMID: 26536149<\/a><\/p>

<\/div><\/div><\/div>\n\n\n\n
<\/div>

\u3053\u308c\u306f\u3059\u3054\u3044\u9032\u6b69\u3067\u3059\u306d\u3002\u672c\u5f53\u306b\u3002\u307e\u3060\u307e\u3060\u8b70\u8ad6\u304c\u7d9a\u304d\u305d\u3046\u306a\u6c17\u914d\u3067\u3059\u306d\u3002\u554f\u984c\u306f\u3001\u76f4\u63a5\u6bd4\u8f03\u3068\u9593\u63a5\u6bd4\u8f03\u306e\u30a4\u30f3\u30b3\u30d2\u30ec\u30f3\u30b9\u3060\u3051\u3067\u306f\u306a\u3044\u3067\u3059\u306d\u3002<\/p>

<\/div><\/div><\/div>\n\n\n\n
<\/div>

\u6700\u5f8c\u306bArm-based model\u3068Contrast-based model\u3092\u5f0f\u3067\u8868\u3057\u305f\u3082\u306e\u3092\u898b\u305b\u308b\u306d\u3002<\/p>

<\/div><\/div><\/div>\n\n\n\n
\"\"
Arm-based model<\/figcaption><\/figure>\n\n\n\n
\"\"
Contrast-based model<\/figcaption><\/figure>\n","protected":false},"excerpt":{"rendered":"

model { for(i in 1:tS) { p[i]<-phi(mu[t[i]]+ vi[s[i], t[i]]) # model r[i]~dbin(p[i], totaln[i]) # binomial … “Network Meta-analysis\u3092OpenBUGS\u3067” \u306e<\/span>\u7d9a\u304d\u3092\u8aad\u3080<\/a><\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[32,2],"tags":[],"class_list":["post-741","post","type-post","status-publish","format-standard","hentry","category-statistics","category-sr"],"_links":{"self":[{"href":"https:\/\/info.zanet.biz\/index.php?rest_route=\/wp\/v2\/posts\/741","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/info.zanet.biz\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/info.zanet.biz\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/info.zanet.biz\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/info.zanet.biz\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=741"}],"version-history":[{"count":15,"href":"https:\/\/info.zanet.biz\/index.php?rest_route=\/wp\/v2\/posts\/741\/revisions"}],"predecessor-version":[{"id":771,"href":"https:\/\/info.zanet.biz\/index.php?rest_route=\/wp\/v2\/posts\/741\/revisions\/771"}],"wp:attachment":[{"href":"https:\/\/info.zanet.biz\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=741"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/info.zanet.biz\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=741"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/info.zanet.biz\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=741"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}