{"id":4854,"date":"2026-09-25T05:41:38","date_gmt":"2026-09-24T20:41:38","guid":{"rendered":"https:\/\/best-biostatistics.com\/toukei-er\/?p=4854"},"modified":"2026-09-25T05:41:39","modified_gmt":"2026-09-24T20:41:39","slug":"why-beforeafter-comparisons-mislead-evaluating-interventions-with-interrupted-time-series-analysis","status":"publish","type":"post","link":"https:\/\/best-biostatistics.com\/toukei-er\/entry\/why-beforeafter-comparisons-mislead-evaluating-interventions-with-interrupted-time-series-analysis\/","title":{"rendered":"\u4ecb\u5165\u306e\u524d\u5f8c\u6bd4\u8f03\u306f\u5371\u306a\u3044\uff5c\u6642\u7cfb\u5217\u30c7\u30fc\u30bf\u306e\u52b9\u679c\u691c\u8a3c\u306f\u5206\u5272\u6642\u7cfb\u5217\u89e3\u6790\uff08ITS\uff09\u3067"},"content":{"rendered":"\n<blockquote class=\"wp-block-quote is-layout-flow wp-block-quote-is-layout-flow\">\n<p class=\"wp-block-paragraph\"><strong>\u5f8c\u8f29<\/strong>\uff1a\u300c\u5148\u8f29\u3001\u5e02\u306e\u6d88\u9632\u5c40\u304b\u3089\u6708\u3054\u3068\u306e\u6551\u6025\u642c\u9001\u4ef6\u6570\u306e\u30c7\u30fc\u30bf\u3092\u3082\u3089\u3044\u307e\u3057\u305f\u30022018\u5e741\u6708\u304b\u30892025\u5e7412\u6708\u307e\u3067\u306e96\u304b\u6708\u5206\u3067\u3059\u300d<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>\u5148\u8f29<\/strong>\uff1a\u300c\u3044\u3044\u30c7\u30fc\u30bf\u3060\u306d\u3002\u4f55\u3092\u8abf\u3079\u305f\u3044\u3093\u3060\u3044\uff1f\u300d<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>\u5f8c\u8f29<\/strong>\uff1a\u300c\u3053\u306e\u5e02\u3067\u306f2022\u5e741\u6708\u306b <strong>\u6551\u6025\u5b89\u5fc3\u30bb\u30f3\u30bf\u30fc\u4e8b\u696d\uff08#7119\uff09<\/strong> \u304c\u59cb\u307e\u308a\u307e\u3057\u305f\u3002\u6551\u6025\u8eca\u3092\u547c\u3076\u304b\u8ff7\u3063\u305f\u3068\u304d\u306b\u96fb\u8a71\u3067\u76f8\u8ac7\u3067\u304d\u308b\u7a93\u53e3\u3067\u3059\u3002\u305d\u308c\u3067\u642c\u9001\u304c\u6e1b\u3063\u305f\u304b\u3092\u8abf\u3079\u305f\u3044\u3093\u3067\u3059\u3002\u3067\u3082\u6642\u7cfb\u5217\u306e\u89e3\u6790\u65b9\u6cd5\u3092\u8abf\u3079\u308b\u3068\u3001Mann-Kendall\u691c\u5b9a\u3001ADF\u691c\u5b9a\u3001KPSS\u691c\u5b9a\u3001ARIMA\u3001ITS\u3001CITS\u2026\u2026\u3068\u6b21\u3005\u306b\u51fa\u3066\u304d\u3066\u3001\u3069\u308c\u3092\u4f7f\u3048\u3070\u3044\u3044\u306e\u304b\u308f\u304b\u3089\u306a\u304f\u306a\u3063\u3066\u3057\u307e\u3063\u3066\u300d<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>\u5148\u8f29<\/strong>\uff1a\u300c\u624b\u6cd5\u306e\u540d\u524d\u304b\u3089\u9078\u307c\u3046\u3068\u3059\u308b\u3068\u3001\u307f\u3093\u306a\u305d\u3053\u3067\u8ff7\u3046\u3093\u3060\u3002\u9078\u3076\u9806\u756a\u304c\u9006\u306a\u3093\u3060\u3088\u3002<strong>\u5148\u306b\u300e\u4f55\u3092\u77e5\u308a\u305f\u3044\u304b\u300f\u3092\u6c7a\u3081\u308b<\/strong>\u3002\u305d\u3046\u3059\u308c\u3070\u3001\u4f7f\u3046\u624b\u6cd5\u306f\u307b\u307c\u81ea\u52d5\u7684\u306b\u6c7a\u307e\u308b\u300d<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>\u5f8c\u8f29<\/strong>\uff1a\u300c\u4f55\u3092\u77e5\u308a\u305f\u3044\u304b\u2026\u2026\u3002#7119\u304c\u59cb\u307e\u3063\u305f <strong>\u305d\u306e\u6642\u70b9\u3092\u5883\u306b<\/strong>\u3001\u642c\u9001\u4ef6\u6570\u304c\u5909\u308f\u3063\u305f\u304b\u3069\u3046\u304b\u3067\u3059\u300d<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>\u5148\u8f29<\/strong>\uff1a\u300c\u305d\u308c\u306a\u3089\u7b54\u3048\u306f <strong>\u5206\u5272\u6642\u7cfb\u5217\u89e3\u6790\uff08Interrupted Time Series\u3001ITS\uff09<\/strong> \u3060\u3002\u305f\u3060\u3001\u307b\u304b\u306e\u624b\u6cd5\u3068\u3069\u3046\u9055\u3046\u306e\u304b\u3082\u3042\u308f\u305b\u3066\u6574\u7406\u3057\u3066\u304a\u3053\u3046\u3002\u6b21\u306b\u5225\u306e\u30c7\u30fc\u30bf\u3092\u6e21\u3055\u308c\u305f\u3068\u304d\u306b\u3001\u3082\u3046\u8ff7\u308f\u306a\u304f\u3066\u6e08\u3080\u304b\u3089\u306d\u300d<\/p>\n<\/blockquote>\n\n\n\n<p class=\"wp-block-paragraph\">\u6642\u7cfb\u5217\u30c7\u30fc\u30bf\u306e\u89e3\u6790\u65b9\u6cd5\u306b\u60a9\u3080\u521d\u5b66\u8005\u306f\u591a\u3044\u3002\u6559\u79d1\u66f8\u3084Web\u306e\u8a18\u4e8b\u306f\u624b\u6cd5\u3054\u3068\u306b\u66f8\u304b\u308c\u3066\u3044\u308b\u306e\u306b\u3001\u73fe\u5834\u306e\u30c7\u30fc\u30bf\u306f\u300c\u3069\u306e\u624b\u6cd5\u3092\u4f7f\u3048\u3070\u3044\u3044\u304b\u300d\u3068\u3044\u3046\u554f\u3044\u304b\u3089\u59cb\u307e\u308b\u304b\u3089\u3067\u3042\u308b\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u672c\u8a18\u4e8b\u306f2\u90e8\u69cb\u6210\u3067\u3042\u308b\u3002<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>\u524d\u534a<\/strong>\uff1a\u6642\u7cfb\u5217\u89e3\u6790\u306e\u624b\u6cd5\u3092\u300c\u554f\u3044\u300d\u3067\u6574\u7406\u3057\u305f <strong>\u5730\u56f3<\/strong> \u3092\u793a\u3059\u3002\u7d71\u8a08ER\u3067\u89e3\u8aac\u3057\u3066\u304d\u305fMann-Kendall\u691c\u5b9a\u3001ADF\u691c\u5b9a\u3068KPSS\u691c\u5b9a\u3001CITS\uff08\u6bd4\u8f03\u5206\u5272\u6642\u7cfb\u5217\uff09\u304c\u3001\u5730\u56f3\u306e\u3069\u3053\u306b\u4f4d\u7f6e\u3059\u308b\u306e\u304b\u304c\u308f\u304b\u308b<\/li>\n\n\n\n<li><strong>\u5f8c\u534a<\/strong>\uff1a\u5730\u56f3\u306e\u4e2d\u304b\u3089 <strong>ITS<\/strong> \u3092\u53d6\u308a\u4e0a\u3052\u3001\u521d\u5b66\u8005\u304c\u30cf\u30de\u308b5\u3064\u306e\u7f60\u3068\u3001R\u3067\u306e\u5b9f\u88c5\u3092Step by Step\u3067\u78ba\u8a8d\u3059\u308b<\/li>\n<\/ul>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h2 class=\"wp-block-heading\">1. \u6642\u7cfb\u5217\u89e3\u6790\u306f\u300c\u554f\u3044\u300d\u3067\u9078\u3076\uff1a\u624b\u6cd5\u306e\u5730\u56f3<\/h2>\n\n\n\n<h3 class=\"wp-block-heading\">\u884c\u304d\u5148\u3092\u6c7a\u3081\u3066\u304b\u3089\u3001\u4e57\u308b\u96fb\u8eca\u3092\u9078\u3076<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">\u5927\u304d\u306a\u99c5\u306e\u8def\u7dda\u56f3\u306e\u524d\u3067\u3001\u884c\u304d\u5148\u3092\u6c7a\u3081\u305a\u306b\u300c\u3069\u306e\u96fb\u8eca\u306b\u4e57\u308d\u3046\u304b\u300d\u3068\u60a9\u3093\u3067\u3082\u7b54\u3048\u306f\u51fa\u306a\u3044\u3002\u307e\u305a\u884c\u304d\u5148\u3092\u6c7a\u3081\u308b\u3002\u305d\u3046\u3059\u308c\u3070\u3001\u4e57\u308b\u3079\u304d\u8def\u7dda\u306f\u81ea\u7136\u306b\u6c7a\u307e\u308b\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u6642\u7cfb\u5217\u89e3\u6790\u3082\u540c\u3058\u3067\u3042\u308b\u3002\u624b\u6cd5\u306e\u4e00\u89a7\u3092\u773a\u3081\u3066\u60a9\u3080\u3088\u308a\u5148\u306b\u3001<strong>\u81ea\u5206\u306e\u30c7\u30fc\u30bf\u306b\u6295\u3052\u304b\u3051\u305f\u3044\u554f\u3044<\/strong> \u3092\u306f\u3063\u304d\u308a\u3055\u305b\u308b\u3002\u4ee3\u8868\u7684\u306a\u554f\u3044\u3068\u624b\u6cd5\u306e\u5bfe\u5fdc\u306f\u3001\u6b21\u306e\u3068\u304a\u308a\u3067\u3042\u308b\u3002<\/p>\n\n\n\n<figure class=\"wp-block-table\"><table class=\"has-fixed-layout\"><thead><tr><th>\u77e5\u308a\u305f\u3044\u3053\u3068\uff08\u554f\u3044\uff09<\/th><th>\u624b\u6cd5<\/th><th>\u89e3\u8aac\u8a18\u4e8b<\/th><\/tr><\/thead><tbody><tr><td>\u4e00\u8cab\u3057\u3066\u5897\u3048\u3066\uff08\u6e1b\u3063\u3066\uff09\u3044\u308b\u304b\u3001\u3069\u308c\u304f\u3089\u3044\u306e\u30da\u30fc\u30b9\u304b<\/td><td>Mann-Kendall\u691c\u5b9a\uff0bSen&#8217;s slope<\/td><td><a href=\"https:\/\/best-biostatistics.com\/toukei-er\/entry\/how-to-test-whether-a-time-series-is-increasing-in-r-mann-kendall-test-and-sens-slope\/\">Mann-Kendall\u691c\u5b9a\u3068Sen&#8217;s slope<\/a><\/td><\/tr><tr><td>\u56de\u5e30\u3084ARIMA\u306e\u524d\u63d0\u3068\u306a\u308b\u300c\u5b9a\u5e38\u6027\u300d\u3092\u6e80\u305f\u3059\u304b<\/td><td>ADF\u691c\u5b9a\u00d7KPSS\u691c\u5b9a<\/td><td><a href=\"https:\/\/best-biostatistics.com\/toukei-er\/entry\/how-to-assess-stationarity-of-time-series-in-r-combining-the-adf-and-kpss-tests\/\">ADF\u691c\u5b9a\u3068KPSS\u691c\u5b9a\u306e\u7d44\u307f\u5408\u308f\u305b\u65b9<\/a><\/td><\/tr><tr><td><strong>\u3042\u308b\u6642\u70b9\u306e\u51fa\u6765\u4e8b\uff08\u653f\u7b56\u30fb\u4ecb\u5165\uff09\u3067\u3001\u6c34\u6e96\u3084\u50be\u304d\u304c\u5909\u308f\u3063\u305f\u304b<\/strong><\/td><td><strong>ITS\uff08\u5206\u5272\u6642\u7cfb\u5217\u89e3\u6790\uff09<\/strong><\/td><td><strong>\u672c\u8a18\u4e8b<\/strong><\/td><\/tr><tr><td>\u305d\u306e\u5909\u5316\u306f\u3001\u540c\u3058\u6642\u671f\u306e\u5225\u306e\u51fa\u6765\u4e8b\u306e\u305b\u3044\u3067\u306f\u306a\u3044\u3068\u8a00\u3048\u308b\u304b<\/td><td>CITS\uff08\u6bd4\u8f03\u5206\u5272\u6642\u7cfb\u5217\u89e3\u6790\uff09<\/td><td><a href=\"https:\/\/best-biostatistics.com\/toukei-er\/entry\/what-is-comparative-interrupted-time-series\/\">Comparative Interrupted Time Series \u3068\u306f<\/a><\/td><\/tr><tr><td>\u5c06\u6765\u306e\u5024\u3092\u4e88\u6e2c\u3057\u305f\u3044\u304b<\/td><td>ARIMA\u306a\u3069<\/td><td><a href=\"https:\/\/best-biostatistics.com\/toukei-er\/entry\/basics-of-time-series-analysis\/\">R\u3067\u6642\u7cfb\u5217\u30c7\u30fc\u30bf\u5206\u6790\u3092\u884c\u3046\u65b9\u6cd5<\/a><\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n<p class=\"wp-block-paragraph\">\u554f\u3044\u306e\u7acb\u3066\u65b9\u3067\u8ff7\u3063\u305f\u3068\u304d\u306f\u3001\u6b21\u306e\u9806\u306b\u81ea\u5206\u306b\u554f\u3044\u304b\u3051\u3066\u307f\u308b\u3068\u3088\u3044\u3002<\/p>\n\n\n\n<pre class=\"wp-block-code\"><code>\u3010\u51fa\u767a\u70b9\u3011\u307e\u305a\u30b0\u30e9\u30d5\u3068\u30b3\u30ec\u30ed\u30b0\u30e9\u30e0\u3067\u30c7\u30fc\u30bf\u306e\u7d20\u9854\u3092\u898b\u308b\n    \u2502\n    \u251c\u2500 Q1. \u7279\u5b9a\u306e\u6642\u70b9\u306b\u300c\u51fa\u6765\u4e8b\u300d\uff08\u653f\u7b56\u30fb\u4ecb\u5165\u30fb\u4e8b\u6545\u306a\u3069\uff09\u304c\u3042\u3063\u305f\u304b\uff1f\n    \u2502     \u2502\n    \u2502     \u251c\u2500 Yes \u2192 Q2. \u540c\u3058\u6642\u671f\u306b\u8d77\u304d\u305f\u5225\u306e\u51fa\u6765\u4e8b\u3068\u5207\u308a\u5206\u3051\u305f\u3044\u304b\uff1f\n    \u2502     \u2502          \u2502    \u307e\u305f\u3001\u4ecb\u5165\u3092\u53d7\u3051\u3066\u3044\u306a\u3044\u6bd4\u8f03\u5bfe\u8c61\u306e\u7cfb\u5217\u306f\u3042\u308b\u304b\uff1f\n    \u2502     \u2502          \u251c\u2500 \u5bfe\u7167\u7cfb\u5217\u3042\u308a \u2192 CITS\n    \u2502     \u2502          \u2514\u2500 \u5bfe\u7167\u7cfb\u5217\u306a\u3057 \u2192 ITS\uff08\u672c\u8a18\u4e8b\uff09\n    \u2502     \u2502\n    \u2502     \u2514\u2500 No \u2192 Q3. \u77e5\u308a\u305f\u3044\u306e\u306f\u300c\u904e\u53bb\u306e\u50be\u5411\u300d\u304b\u300c\u5c06\u6765\u306e\u4e88\u6e2c\u300d\u304b\uff1f\n    \u2502                \u251c\u2500 \u904e\u53bb\u306e\u50be\u5411 \u2192 Mann-Kendall\u691c\u5b9a\uff0bSen's slope\n    \u2502                \u2514\u2500 \u5c06\u6765\u306e\u4e88\u6e2c \u2192 ARIMA\u306a\u3069\n    \u2502\n    \u2514\u2500 \u3069\u306e\u9053\u3092\u9078\u3093\u3067\u3082\uff1a\u7cfb\u5217\u306e\u6027\u8cea\u3092\u77e5\u308b\u305f\u3081\u3001ADF\u00d7KPSS\u3067\u5b9a\u5e38\u6027\u3092\u78ba\u8a8d\u3057\u3066\u304a\u304f\u3068\u5b89\u5fc3\n       \uff08ITS\u3067\u306f\u300c\u4ecb\u5165\u524d\u306e\u671f\u9593\u3060\u3051\u300d\u3067\u78ba\u8a8d\u3059\u308b \u2192 \u7b2c3\u7ae0\u306e\u7f60\u2464\uff09<\/code><\/pre>\n\n\n\n<p class=\"wp-block-paragraph\">\u30dd\u30a4\u30f3\u30c8\u306f\u3001<strong>\u3053\u308c\u3089\u306e\u624b\u6cd5\u306f\u4e8c\u8005\u629e\u4e00\u3067\u306f\u306a\u3044<\/strong> \u3068\u3044\u3046\u3053\u3068\u3067\u3042\u308b\u3002ADF\u00d7KPSS\u306f\u3001\u3069\u306e\u9053\u3092\u9078\u3093\u3067\u3082\u5f79\u306b\u7acb\u3064\u300c\u5065\u5eb7\u8a3a\u65ad\u300d\u306e\u3088\u3046\u306a\u3082\u306e\u3060\u3002\u5fc5\u9808\u306e\u624b\u9806\u3067\u306f\u306a\u3044\u304c\u3001\u7cfb\u5217\u306e\u6027\u8cea\uff08\u4e00\u5b9a\u306e\u5742\u3092\u4e0a\u3063\u3066\u3044\u308b\u306e\u304b\u3001\u6f02\u6d41\u3057\u3066\u3044\u308b\u306e\u304b\uff09\u3092\u77e5\u3063\u3066\u304a\u304f\u3068\u3001\u30e2\u30c7\u30eb\u306e\u9078\u3073\u65b9\u3084\u7d50\u679c\u306e\u89e3\u91c8\u3067\u8ff7\u3044\u306b\u304f\u304f\u306a\u308b\u3002ITS\u3084Mann-Kendall\u691c\u5b9a\u3092\u4f7f\u3046\u3068\u304d\u3082\u3001\u81ea\u5df1\u76f8\u95a2\u3068\u5b63\u7bc0\u6027\u306e\u78ba\u8a8d\u306f\u6b20\u304b\u305b\u306a\u3044\u3002\u672c\u8a18\u4e8b\u306e\u5f8c\u534a\u306eR\u306e\u5b9f\u88c5\u3067\u3082\u3001MK\u8a18\u4e8b\u3084ADF-KPSS\u8a18\u4e8b\u3067\u4f7f\u3063\u305f\u9053\u5177\u304c\u7e70\u308a\u8fd4\u3057\u767b\u5834\u3059\u308b\u3002<\/p>\n\n\n\n<blockquote class=\"wp-block-quote is-layout-flow wp-block-quote-is-layout-flow\">\n<p class=\"wp-block-paragraph\"><strong>\u203b\u30b3\u30e9\u30e0\uff1a\u306a\u305c\u666e\u901a\u306et\u691c\u5b9a\u3084\u56de\u5e30\u3067\u306f\u30c0\u30e1\u306a\u306e\u304b<\/strong><br>t\u691c\u5b9a\u3084\u901a\u5e38\u306e\u56de\u5e30\u306f\u3001\u300c\u30c7\u30fc\u30bf\u4e00\u3064\u3072\u3068\u3064\u304c\u4e92\u3044\u306b\u72ec\u7acb\u3067\u3042\u308b\u300d\u3053\u3068\u3092\u524d\u63d0\u306b\u3057\u3066\u3044\u308b\u3002\u3068\u3053\u308d\u304c\u6642\u7cfb\u5217\u30c7\u30fc\u30bf\u306f\u3001\u305f\u3044\u3066\u3044\u300c\u5148\u6708\u306e\u72b6\u614b\u3092\u5f15\u304d\u305a\u308b\u300d\u3002MK\u8a18\u4e8b\u3067\u306f\u3001\u3053\u308c\u3092 <strong>\u4ef2\u826f\u3057\u30b0\u30eb\u30fc\u30d7\u304c\u6df7\u3056\u3063\u305f\u591a\u6570\u6c7a<\/strong> \u306b\u55a9\u3048\u305f\u300230\u4eba\u306e\u30af\u30e9\u30b9\u3067\u3082\u3001\u3044\u3064\u3082\u540c\u3058\u610f\u898b\u3092\u8a00\u30465\u4eba\u7d44\u304c6\u3064\u3042\u308c\u3070\u3001\u5b9f\u8cea\u306f6\u7968\u5206\u306e\u610f\u898b\u3057\u304b\u306a\u3044\u3002\u305d\u308c\u309230\u7968\u3068\u3057\u3066\u6570\u3048\u308b\u3068\u3001\u5b9f\u614b\u4ee5\u4e0a\u306b\u81ea\u4fe1\u6e80\u3005\u306a\u7d50\u8ad6\u304c\u51fa\u3066\u3057\u307e\u3046\u3002<br>\u6642\u7cfb\u5217\u30c7\u30fc\u30bf\u306b\u666e\u901a\u306e\u691c\u5b9a\u3092\u305d\u306e\u307e\u307e\u4f7f\u3046\u3068\u3001\u540c\u3058\u3053\u3068\u304c\u8d77\u304d\u308b\u3002\u6a19\u6e96\u8aa4\u5dee\u304c\u5c0f\u3055\u304f\u898b\u7a4d\u3082\u3089\u308c\u3001P\u5024\u304c\u5c0f\u3055\u304f\u51fa\u3059\u304e\u308b\u306e\u3067\u3042\u308b\u3002\u6642\u7cfb\u5217\u5c02\u7528\u306e\u624b\u6cd5\u304c\u5fc5\u8981\u306a\u306e\u306f\u3001\u3053\u306e\u305f\u3081\u3067\u3042\u308b\u3002<\/p>\n<\/blockquote>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h2 class=\"wp-block-heading\">2. ITS\u3068\u306f\uff1a\u300c\u4ecb\u5165\u304c\u306a\u304b\u3063\u305f\u4e16\u754c\u300d\u306e\u5ef6\u9577\u7dda\u3068\u6bd4\u3079\u308b<\/h2>\n\n\n\n<h3 class=\"wp-block-heading\">\u30c0\u30a4\u30a8\u30c3\u30c8\u306e\u52b9\u679c\u3092\u6b63\u3057\u304f\u6e2c\u308b\u306b\u306f<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">\u30c0\u30a4\u30a8\u30c3\u30c8\u3092\u59cb\u3081\u305f\u4eba\u304c\u3001\u300c\u958b\u59cb\u524d\u306e1\u5e74\u9593\u306e\u5e73\u5747\u4f53\u91cd\u300d\u3068\u300c\u958b\u59cb\u5f8c\u306e1\u5e74\u9593\u306e\u5e73\u5747\u4f53\u91cd\u300d\u3092\u6bd4\u3079\u3066\u3001\u300c5kg\u6e1b\u3063\u305f\u3001\u30c0\u30a4\u30a8\u30c3\u30c8\u6210\u529f\uff01\u300d\u3068\u559c\u3093\u3067\u3044\u308b\u3068\u3057\u3088\u3046\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u3057\u304b\u3057\u3001\u3082\u3057\u3053\u306e\u4eba\u304c <strong>\u30c0\u30a4\u30a8\u30c3\u30c8\u3092\u59cb\u3081\u308b\u524d\u304b\u3089\u6bce\u67080.3kg\u305a\u3064\u6e1b\u3063\u3066\u3044\u305f<\/strong> \u3068\u3057\u305f\u3089\u3069\u3046\u3060\u308d\u3046\u3002\u4f55\u3082\u3057\u306a\u304f\u3066\u30821\u5e74\u30673.6kg\u6e1b\u3063\u3066\u3044\u305f\u306f\u305a\u306a\u306e\u3067\u3001\u30c0\u30a4\u30a8\u30c3\u30c8\u306e\u672c\u5f53\u306e\u52b9\u679c\u306f5kg\u3088\u308a\u305a\u3063\u3068\u5c0f\u3055\u3044\u3002\u9006\u306b\u3001\u958b\u59cb\u524d\u306f\u6bce\u6708\u5897\u3048\u3066\u3044\u305f\u4eba\u306a\u3089\u3001\u300c\u524d\u5f8c\u3067\u5909\u308f\u3089\u306a\u3044\u300d\u3053\u3068\u304c\u7acb\u6d3e\u306a\u52b9\u679c\u304b\u3082\u3057\u308c\u306a\u3044\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u6b63\u3057\u3044\u6bd4\u3079\u65b9\u306f\u3001<strong>\u300c\u30c0\u30a4\u30a8\u30c3\u30c8\u3092\u59cb\u3081\u306a\u304b\u3063\u305f\u3089\u3001\u4f53\u91cd\u306f\u305d\u308c\u307e\u3067\u306e\u30da\u30fc\u30b9\u3067\u3053\u3046\u63a8\u79fb\u3057\u305f\u306f\u305a\u3060\u300d\u3068\u3044\u3046\u5ef6\u9577\u7dda\u3092\u5f15\u304d\u3001\u5b9f\u969b\u306e\u4f53\u91cd\u3068\u6bd4\u3079\u308b\u3053\u3068<\/strong> \u3067\u3042\u308b\u3002\u3053\u306e\u5ef6\u9577\u7dda\u3092 <strong>\u53cd\u4e8b\u5b9f\uff08counterfactual\uff09<\/strong> \u3068\u547c\u3076\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u3053\u308c\u304cITS\u306e\u8003\u3048\u65b9\u306e\u3059\u3079\u3066\u3067\u3042\u308b\u3002\u4ecb\u5165\u524d\u306e\u30c7\u30fc\u30bf\u304b\u3089\u300c\u4ecb\u5165\u304c\u306a\u304b\u3063\u305f\u4e16\u754c\u300d\u306e\u5ef6\u9577\u7dda\u3092\u63cf\u304d\u3001\u4ecb\u5165\u5f8c\u306e\u5b9f\u969b\u306e\u30c7\u30fc\u30bf\u3068\u306e\u305a\u308c\u3092\u4ecb\u5165\u52b9\u679c\u3068\u307f\u306a\u3059\u3002<\/p>\n\n\n\n<blockquote class=\"wp-block-quote is-layout-flow wp-block-quote-is-layout-flow\">\n<p class=\"wp-block-paragraph\"><strong>\u203b\u30b3\u30e9\u30e0\uff1a\u53cd\u4e8b\u5b9f\uff08counterfactual\uff09\u3068\u306f<\/strong><br>\u300c\u3082\u3057\u301c\u304c\u306a\u304b\u3063\u305f\u3089\u3001\u3069\u3046\u306a\u3063\u3066\u3044\u305f\u304b\u300d\u3068\u3044\u3046\u3001\u5b9f\u969b\u306b\u306f\u89b3\u6e2c\u3067\u304d\u306a\u3044\u4eee\u60f3\u306e\u4e16\u754c\u306e\u3053\u3068\u3067\u3042\u308b\u3002\u56e0\u679c\u63a8\u8ad6\u3067\u306f\u3001\u4ecb\u5165\u306e\u52b9\u679c\u3092\u300c\u5b9f\u969b\u306e\u4e16\u754c\u300d\u3068\u300c\u53cd\u4e8b\u5b9f\u306e\u4e16\u754c\u300d\u306e\u5dee\u3068\u3057\u3066\u5b9a\u7fa9\u3059\u308b\u3002RCT\u3067\u306f\u3001\u30e9\u30f3\u30c0\u30e0\u306b\u5272\u308a\u4ed8\u3051\u305f\u5bfe\u7167\u7fa4\u304c\u53cd\u4e8b\u5b9f\u306e\u4ee3\u308f\u308a\u3092\u52d9\u3081\u308b\u3002ITS\u3067\u306f\u3001<strong>\u4ecb\u5165\u524d\u306e\u30c8\u30ec\u30f3\u30c9\u306e\u5ef6\u9577\u7dda<\/strong> \u304c\u305d\u306e\u4ee3\u308f\u308a\u3092\u52d9\u3081\u308b\u3002\u3060\u304b\u3089ITS\u3067\u306f\u3001\u300c\u4ecb\u5165\u524d\u306e\u30c8\u30ec\u30f3\u30c9\u304c\u3001\u4ecb\u5165\u304c\u306a\u3051\u308c\u3070\u305d\u306e\u307e\u307e\u7d9a\u3044\u305f\u306f\u305a\u3060\u300d\u3068\u3044\u3046\u4eee\u5b9a\u304c\u7d50\u8ad6\u306e\u571f\u53f0\u306b\u306a\u308b\u3002<\/p>\n<\/blockquote>\n\n\n\n<h3 class=\"wp-block-heading\">\u30bb\u30b0\u30e1\u30f3\u30c8\u56de\u5e30\uff1a\u300c\u6bb5\u5dee\u300d\u3068\u300c\u50be\u304d\u306e\u5909\u5316\u300d\u3092\u5206\u3051\u3066\u6e2c\u308b<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">ITS\u306e\u6a19\u6e96\u7684\u306a\u89e3\u6790\u65b9\u6cd5\u306f\u3001<strong>\u30bb\u30b0\u30e1\u30f3\u30c8\u56de\u5e30\uff08segmented regression\uff09<\/strong> \u3067\u3042\u308b\uff08Wagner et al., 2002\uff09\u3002\u6570\u5f0f\u306f1\u672c\u3060\u3051\u793a\u3059\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">$$<br>Y_t = \\beta_0 + \\beta_1 \\cdot \\text{time}_t + \\beta_2 \\cdot \\text{level}_t + \\beta_3 \\cdot \\text{time\\_after}_t + \\varepsilon_t<br>$$<\/p>\n\n\n\n<figure class=\"wp-block-table\"><table class=\"has-fixed-layout\"><thead><tr><th>\u5909\u6570<\/th><th>\u610f\u5473<\/th><\/tr><\/thead><tbody><tr><td>$\\text{time}_t$<\/td><td>\u89b3\u5bdf\u958b\u59cb\u304b\u3089\u306e\u7d4c\u904e\u6642\u9593\uff081, 2, 3, \u2026\uff09<\/td><\/tr><tr><td>$\\text{level}_t$<\/td><td>\u4ecb\u5165\u524d\u306f0\u3001\u4ecb\u5165\u5f8c\u306f1<\/td><\/tr><tr><td>$\\text{time\\_after}_t$<\/td><td>\u4ecb\u5165\u5f8c\u306e\u7d4c\u904e\u6642\u9593\uff08\u4ecb\u5165\u524d\u306f0\u3001\u4ecb\u5165\u3057\u305f\u6708\u30920\u3068\u3057\u3066 0, 1, 2, \u2026\uff09<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n<figure class=\"wp-block-table\"><table class=\"has-fixed-layout\"><thead><tr><th>\u4fc2\u6570<\/th><th>\u610f\u5473<\/th><th>\u30c0\u30a4\u30a8\u30c3\u30c8\u306e\u4f8b\u3067\u306f<\/th><\/tr><\/thead><tbody><tr><td>$\\beta_1$<\/td><td>\u4ecb\u5165\u524d\u306e\u50be\u304d\uff08\u3082\u3068\u3082\u3068\u306e\u30c8\u30ec\u30f3\u30c9\uff09<\/td><td>\u958b\u59cb\u524d\u306e\u6bce\u6708\u306e\u4f53\u91cd\u306e\u5909\u5316<\/td><\/tr><tr><td>$\\beta_2$<\/td><td><strong>\u6c34\u6e96\u306e\u5909\u5316<\/strong>\uff1a\u4ecb\u5165\u76f4\u5f8c\u306e\u6bb5\u5dee<\/td><td>\u59cb\u3081\u305f\u76f4\u5f8c\u306b\u30b9\u30c8\u30f3\u3068\u843d\u3061\u305f\u5206<\/td><\/tr><tr><td>$\\beta_3$<\/td><td><strong>\u50be\u304d\u306e\u5909\u5316<\/strong>\uff1a\u4ecb\u5165\u524d\u3068\u6bd4\u3079\u3066\u3001\u50be\u304d\u304c\u3069\u308c\u3060\u3051\u5909\u308f\u3063\u305f\u304b<\/td><td>\u6bce\u6708\u306e\u6e1b\u308a\u65b9\u304c\u52a0\u901f\u3057\u305f\u5206<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n<p class=\"wp-block-paragraph\">\u52b9\u679c\u3092 <strong>\u300c\u6bb5\u5dee\u300d\u3068\u300c\u50be\u304d\u306e\u5909\u5316\u300d\u306e2\u3064\u306b\u5206\u3051\u3066<\/strong> \u6e2c\u308c\u308b\u306e\u304c\u3001ITS\u306e\u5927\u304d\u306a\u5f37\u307f\u3067\u3042\u308b\u3002#7119\u306e\u3088\u3046\u306a\u4e8b\u696d\u306a\u3089\u3001\u59cb\u307e\u3063\u305f\u76f4\u5f8c\u306b\u642c\u9001\u304c\u6e1b\u308b\uff08\u6bb5\u5dee\uff09\u3060\u3051\u3067\u306a\u304f\u3001\u5e02\u6c11\u306b\u5468\u77e5\u3055\u308c\u308b\u306b\u3064\u308c\u3066\u6e1b\u308a\u65b9\u304c\u5927\u304d\u304f\u306a\u308b\uff08\u50be\u304d\u306e\u5909\u5316\uff09\u3053\u3068\u3082\u8003\u3048\u3089\u308c\u308b\u3002<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<div id=\"biost-986890293\" class=\"biost- biost-entity-placement\"><p style=\"text-align: center;\"><span style=\"font-size: 20px;\"><strong><a href=\"https:\/\/best-biostatistics.com\/kmhl\">\uff1e\uff1e\u3082\u3046\u7d71\u8a08\u3067\u60a9\u3080\u306e\u306f\u7d42\u308f\u308a\u306b\u3057\u307e\u305b\u3093\u304b\uff1f\u00a0<\/a><\/strong><\/span><\/p>\r\n<a href=\"https:\/\/best-biostatistics.com\/kmhl\"><img class=\"aligncenter wp-image-2794 size-full\" src=\"https:\/\/best-biostatistics.com\/wp\/wp-content\/uploads\/2023\/11\/bn_r_03.png\" alt=\"\" width=\"500\" height=\"327\" \/><\/a>\r\n<p style=\"text-align: center;\"><span style=\"color: #ff0000; font-size: 20px;\"><strong><span class=\"marker2\">\u21911\u4e07\u4eba\u4ee5\u4e0a\u306e\u533b\u7642\u5f93\u4e8b\u8005\u304c\u8cfc\u8aad\u4e2d<\/span><\/strong><\/span><\/p><\/div><h2 class=\"wp-block-heading\">3. \u521d\u5b66\u8005\u304c\u30cf\u30de\u308bITS\u306e5\u3064\u306e\u7f60<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">ITS\u306f\u5f37\u529b\u306a\u624b\u6cd5\u3060\u304c\u3001\u843d\u3068\u3057\u7a74\u3082\u591a\u3044\u3002\u3053\u3053\u3067\u306f\u521d\u5b66\u8005\u304c\u7279\u306b\u30cf\u30de\u308a\u3084\u3059\u30445\u3064\u3092\u6319\u3052\u308b\u3002\u3044\u305a\u308c\u3082\u7d71\u8a08ER\u306e\u65e2\u5b58\u8a18\u4e8b\u3068\u3064\u306a\u304c\u3063\u3066\u3044\u308b\u3002<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">\u7f60\u2460\uff1a\u4ecb\u5165\u524d\u5f8c\u306e\u5e73\u5747\u3092\u5358\u7d14\u306b\u6bd4\u3079\u308b<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">\u30c0\u30a4\u30a8\u30c3\u30c8\u306e\u4f8b\u3067\u898b\u305f\u3068\u304a\u308a\u3067\u3042\u308b\u3002\u3082\u3068\u3082\u3068\u5897\u52a0\u50be\u5411\u306b\u3042\u308b\u7cfb\u5217\u3067\u306f\u3001\u4ecb\u5165\u306b\u52b9\u679c\u304c\u3042\u3063\u3066\u3082\u3001\u4ecb\u5165\u5f8c\u306e\u5e73\u5747\u306e\u65b9\u304c\u5927\u304d\u304f\u306a\u308b\u3053\u3068\u304c\u3042\u308b\u3002\u5f8c\u534a\u306eStep 3\u3067\u306f\u3001\u3053\u306e\u7f60\u304c\u5b9f\u969b\u306b\u8d77\u304d\u308b\u69d8\u5b50\u3092R\u3067\u78ba\u304b\u3081\u308b\u3002<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">\u7f60\u2461\uff1a\u81ea\u5df1\u76f8\u95a2\u3092\u7121\u8996\u3059\u308b<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">\u30b3\u30e9\u30e0\u3067\u8ff0\u3079\u305f\u300c\u4ef2\u826f\u3057\u30b0\u30eb\u30fc\u30d7\u306e\u591a\u6570\u6c7a\u300d\u306e\u554f\u984c\u3067\u3042\u308b\u3002\u81ea\u5df1\u76f8\u95a2\u3092\u7121\u8996\u3057\u305f\u901a\u5e38\u306e\u6700\u5c0f\u4e8c\u4e57\u6cd5\uff08OLS\uff09\u3067\u306f\u3001\u6a19\u6e96\u8aa4\u5dee\u304c\u5c0f\u3055\u304f\u306a\u308a\u3059\u304e\u3001\u4fe1\u983c\u533a\u9593\u304c\u72ed\u304f\u306a\u308a\u3059\u304e\u308b\u3002Turner et al.\uff082021\uff09\u306e\u30b7\u30df\u30e5\u30ec\u30fc\u30b7\u30e7\u30f3\u7814\u7a76\u3067\u306f\u3001\u81ea\u5df1\u76f8\u95a2\u3092\u8003\u616e\u3059\u308b\u65b9\u6cd5\u3092\u4f7f\u3063\u3066\u3082\u81ea\u5df1\u76f8\u95a2\u306e\u5927\u304d\u3055\u306f\u904e\u5c0f\u8a55\u4fa1\u3055\u308c\u304c\u3061\u3067\u3001\u305d\u306e\u7d50\u679c\u3001\u4fe1\u983c\u533a\u9593\u306e\u88ab\u8986\u7387\u304c\u540d\u76ee\u306e95%\u3092\u4e0b\u56de\u308b\u3053\u3068\u304c\u793a\u3055\u308c\u3066\u3044\u308b\u3002\u540c\u7814\u7a76\u3067\u306f\u3001<strong>\u5236\u9650\u4ed8\u304d\u6700\u5c24\u6cd5\uff08REML\uff09<\/strong> \u304c\u81ea\u5df1\u76f8\u95a2\u3092\u6700\u3082\u504f\u308a\u306a\u304f\u63a8\u5b9a\u3057\u300112\u70b9\u4ee5\u4e0a\u306e\u7cfb\u5217\u3067\u306fREML\u304c\u63a8\u5968\u3055\u308c\u3066\u3044\u308b\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u2192 MK\u8a18\u4e8b\u306e <a href=\"https:\/\/best-biostatistics.com\/toukei-er\/entry\/how-to-test-whether-a-time-series-is-increasing-in-r-mann-kendall-test-and-sens-slope\/#index_id8\">\u7f60\u2460\uff1a\u81ea\u5df1\u76f8\u95a2<\/a> \u3068\u540c\u3058\u69cb\u56f3\u3067\u3042\u308b\u3002<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">\u7f60\u2462\uff1a\u5b63\u7bc0\u6027\u3092\u8abf\u6574\u3057\u306a\u3044<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">\u6551\u6025\u642c\u9001\u306f\u51ac\uff08\u611f\u67d3\u75c7\u30fb\u5fc3\u8840\u7ba1\u75be\u60a3\u30fb\u8ee2\u5012\uff09\u306b\u591a\u304f\u3001\u5730\u57df\u306b\u3088\u3063\u3066\u306f\u590f\uff08\u71b1\u4e2d\u75c7\uff09\u306b\u3082\u5897\u3048\u308b\uff08\u5f8c\u534a\u306e\u7591\u4f3c\u30c7\u30fc\u30bf\u3067\u306f\u3001\u5358\u7d14\u5316\u306e\u305f\u3081\u51ac\u306e\u30d4\u30fc\u30af\u3060\u3051\u3092\u5165\u308c\u3066\u3044\u308b\uff09\u3002\u4ecb\u5165\u306e\u76f4\u524d\u304c\u51ac\u306e\u30d4\u30fc\u30af\u3067\u3001\u76f4\u5f8c\u304c\u6625\u306e\u8c37\u306a\u3089\u3001\u5b63\u7bc0\u306e\u6ce2\u3060\u3051\u3067\u300c\u4ecb\u5165\u5f8c\u306b\u6e1b\u3063\u305f\u300d\u3088\u3046\u306b\u898b\u3048\u3066\u3057\u307e\u3046\u3002\u6708\u3092\u30ab\u30c6\u30b4\u30ea\u5909\u6570\u3068\u3057\u3066\u5165\u308c\u308b\uff08\u6708\u30c0\u30df\u30fc\uff09\u304b\u3001\u30b5\u30a4\u30f3\u30fb\u30b3\u30b5\u30a4\u30f3\u306e\u8abf\u548c\u9805\u3092\u5165\u308c\u3066\u8abf\u6574\u3059\u308b\uff08Lopez Bernal et al., 2017\uff09\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u2192 MK\u8a18\u4e8b\u306e <a href=\"https:\/\/best-biostatistics.com\/toukei-er\/entry\/how-to-test-whether-a-time-series-is-increasing-in-r-mann-kendall-test-and-sens-slope\/#index_id9\">\u7f60\u2461\uff1a\u5b63\u7bc0\u6027<\/a> \u3067\u7d39\u4ecb\u3057\u305f\u300c7\u6708\u30681\u6708\u3067\u80cc\u6bd4\u3079\u3092\u3057\u306a\u3044\u300d\u3068\u540c\u3058\u767a\u60f3\u3067\u3042\u308b\u3002<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">\u7f60\u2463\uff1a\u540c\u3058\u6642\u671f\u306b\u8d77\u304d\u305f\u5225\u306e\u51fa\u6765\u4e8b\u3092\u898b\u843d\u3068\u3059<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">ITS\u306e\u6700\u5927\u306e\u5f31\u70b9\u304c\u3053\u308c\u3067\u3042\u308b\u3002ITS\u306f1\u3064\u306e\u96c6\u56e3\u306e\u4e2d\u3067\u524d\u5f8c\u3092\u6bd4\u3079\u308b\u306e\u3067\u3001<strong>\u4ecb\u5165\u3068\u307b\u307c\u540c\u6642\u306b\u8d77\u304d\u305f\u5225\u306e\u51fa\u6765\u4e8b\uff08\u5171\u4ecb\u5165\uff09<\/strong> \u306e\u5f71\u97ff\u3092\u533a\u5225\u3067\u304d\u306a\u3044\uff08Lopez Bernal et al., 2018\uff09\u3002\u4f8b\u3048\u3070\u3001#7119\u3068\u540c\u3058\u6642\u671f\u306b\u611f\u67d3\u75c7\u306e\u6d41\u884c\u304c\u53ce\u307e\u3063\u305f\u308a\u3001\u8fd1\u96a3\u306e\u75c5\u9662\u304c\u6551\u6025\u306e\u53d7\u3051\u5165\u308c\u4f53\u5236\u3092\u5909\u3048\u305f\u308a\u3059\u308c\u3070\u3001\u305d\u306e\u5f71\u97ff\u3082\u300c#7119\u306e\u52b9\u679c\u300d\u306b\u6df7\u3056\u3063\u3066\u3057\u307e\u3046\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u5bfe\u51e6\u6cd5\u306f\u3001<strong>\u4ecb\u5165\u3092\u53d7\u3051\u3066\u3044\u306a\u3044\u6bd4\u8f03\u5bfe\u8c61\u306e\u7cfb\u5217\uff08\u5bfe\u7167\u7cfb\u5217\uff09\u3092\u52a0\u3048\u308b<\/strong> \u3053\u3068\u3067\u3042\u308b\u3002\u4f8b\u3048\u3070\u3001#7119\u3092\u5c0e\u5165\u3057\u3066\u3044\u306a\u3044\u8fd1\u96a3\u306e\u5e02\u306e\u642c\u9001\u4ef6\u6570\u3092\u4e26\u3079\u3066\u6bd4\u3079\u308b\u3002\u3053\u308c\u304c <strong>CITS\uff08\u6bd4\u8f03\u5206\u5272\u6642\u7cfb\u5217\u89e3\u6790\uff09<\/strong> \u3067\u3042\u308b\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u2192 <a href=\"https:\/\/best-biostatistics.com\/toukei-er\/entry\/what-is-comparative-interrupted-time-series\/\">CITS\u306e\u8a18\u4e8b<\/a> \u3067\u3001\u5bfe\u7167\u7fa4\u306e\u9078\u3073\u65b9\u3068\u30e2\u30c7\u30eb\u306e\u4f7f\u3044\u5206\u3051\u3092\u89e3\u8aac\u3057\u3066\u3044\u308b\u3002<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">\u7f60\u2464\uff1a\u4ecb\u5165\u3092\u307e\u305f\u3044\u3067\u5b9a\u5e38\u6027\u3092\u691c\u5b9a\u3059\u308b\uff0f\u4ecb\u5165\u524d\u306e\u6642\u70b9\u304c\u5c11\u306a\u3059\u304e\u308b<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">\u56de\u5e30\u306e\u524d\u63d0\u3092\u78ba\u304b\u3081\u308b\u305f\u3081\u306bADF\u691c\u5b9a\u3084KPSS\u691c\u5b9a\u3092\u4f7f\u3046\u3068\u304d\u3001<strong>\u4ecb\u5165\u3092\u307e\u305f\u3044\u3060\u7cfb\u5217\u5168\u4f53<\/strong> \u3067\u691c\u5b9a\u3057\u3066\u306f\u3044\u3051\u306a\u3044\u3002\u4ecb\u5165\u306b\u3088\u308b\u6bb5\u5dee\u306f\u3001ADF\u691c\u5b9a\u304b\u3089\u898b\u308b\u3068\u300c\u30b7\u30e7\u30c3\u30af\u304c\u6d88\u3048\u305a\u306b\u6b8b\u3063\u305f\u300d\u3088\u3046\u306b\u898b\u3048\u3001\u5358\u4f4d\u6839\u3068\u53d6\u308a\u9055\u3048\u3089\u308c\u3084\u3059\u3044\u304b\u3089\u3067\u3042\u308b\uff08Perron, 1989\uff09\u3002<strong>\u4ecb\u5165\u524d\u306e\u671f\u9593\u3060\u3051\u3067<\/strong> \u691c\u5b9a\u3059\u308b\u306e\u304c\u5b89\u5168\u3067\u3042\u308b\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u307e\u305f\u3001\u4ecb\u5165\u524d\u306e\u6642\u70b9\u304c\u5c11\u306a\u3059\u304e\u308b\u3068\u3001\u300c\u4ecb\u5165\u524d\u306e\u30c8\u30ec\u30f3\u30c9\u300d\u81ea\u4f53\u304c\u4e0d\u5b89\u5b9a\u306b\u306a\u308a\u3001\u53cd\u4e8b\u5b9f\u306e\u5ef6\u9577\u7dda\u304c\u4fe1\u983c\u3067\u304d\u306a\u304f\u306a\u308b\u3002\u6708\u6b21\u30c7\u30fc\u30bf\u306a\u3089\u3001\u5b63\u7bc0\u6027\u3092\u63a8\u5b9a\u3059\u308b\u305f\u3081\u306b\u3082\u3001\u4ecb\u5165\u524d\u306b\u5c11\u306a\u304f\u3068\u3082\u6570\u5e74\u5206\u306e\u30c7\u30fc\u30bf\u304c\u3042\u308b\u3053\u3068\u304c\u671b\u307e\u3057\u3044\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u2192 <a href=\"https:\/\/best-biostatistics.com\/toukei-er\/entry\/how-to-assess-stationarity-of-time-series-in-r-combining-the-adf-and-kpss-tests\/#index_id15\">ADF-KPSS\u8a18\u4e8b<\/a> \u306e\u30b3\u30e9\u30e0\u300c\u69cb\u9020\u5909\u5316\u304c\u3042\u308b\u3068\u3001\u5224\u5b9a\u3092\u8aa4\u308a\u3084\u3059\u3044\u300d\u3067\u3001\u30ca\u30a4\u30eb\u5ddd\u306e\u6d41\u91cf\u30c7\u30fc\u30bf\u3092\u4f7f\u3063\u3066\u3053\u306e\u7f60\u3092\u5b9f\u6f14\u3057\u3066\u3044\u308b\u3002<\/p>\n\n\n\n<blockquote class=\"wp-block-quote is-layout-flow wp-block-quote-is-layout-flow\">\n<p class=\"wp-block-paragraph\"><strong>\u203b\u30b3\u30e9\u30e0\uff1a\u4ef6\u6570\u30c7\u30fc\u30bf\u306a\u3089\u30dd\u30a2\u30bd\u30f3\u56de\u5e30\u3082<\/strong><br>\u6551\u6025\u642c\u9001\u300c\u4ef6\u6570\u300d\u306e\u3088\u3046\u306a\u30ab\u30a6\u30f3\u30c8\u30c7\u30fc\u30bf\u3067\u306f\u3001\u30dd\u30a2\u30bd\u30f3\u56de\u5e30\uff08\u904e\u5206\u6563\u304c\u3042\u308c\u3070\u6e96\u30dd\u30a2\u30bd\u30f3\u56de\u5e30\u3084\u8ca0\u306e\u4e8c\u9805\u56de\u5e30\uff09\u3092\u4f7f\u3046\u306e\u304c\u672c\u6765\u306e\u5f62\u3067\u3042\u308b\u3002Lopez Bernal et al.\uff082017\uff09\u306e\u30c1\u30e5\u30fc\u30c8\u30ea\u30a2\u30eb\u3067\u3082\u3001\u6e96\u30dd\u30a2\u30bd\u30f3\u56de\u5e30\uff08quasi-Poisson\uff09\u3092\u4f7f\u3063\u3066\u3044\u308b\u3002\u4eba\u53e3\u3092\u88dc\u6b63\u3059\u308b\u5834\u5408\u306f\u3001\u4eba\u53e3\u306e\u5bfe\u6570\u3092\u30aa\u30d5\u30bb\u30c3\u30c8\u3068\u3057\u3066\u5165\u308c\u308c\u3070\u300c\u4eba\u53e3\u3042\u305f\u308a\u306e\u7387\u300d\u3092\u30e2\u30c7\u30eb\u5316\u3067\u304d\u308b\u3002\u3053\u306e\u5834\u5408\u3001\u52b9\u679c\u306f\u300c\u4ef6\u6570\u306e\u5dee\u300d\u3067\u306f\u306a\u304f <strong>\u300c\u7387\u306e\u6bd4\uff08\u4f55%\u5909\u5316\u3057\u305f\u304b\uff09\u300d<\/strong> \u3068\u3057\u3066\u8868\u3055\u308c\u308b\u3002<br>\u672c\u8a18\u4e8b\u3067\u306f\u3001\u8003\u3048\u65b9\u3092\u3064\u304b\u3080\u3053\u3068\u3092\u512a\u5148\u3057\u3066\u7dda\u5f62\u30e2\u30c7\u30eb\uff08\u6b63\u898f\u5206\u5e03\uff09\u3067\u793a\u3059\u30021\u304b\u6708\u306b1,000\u4ef6\u524d\u5f8c\u3068\u4ef6\u6570\u304c\u5341\u5206\u306b\u591a\u3044\u5834\u5408\u3001\u7dda\u5f62\u30e2\u30c7\u30eb\u3067\u3082\u8fd1\u3044\u7d50\u679c\u306b\u306a\u308b\u3053\u3068\u304c\u591a\u3044\u3002\u305f\u3060\u3057\u3001\u4ef6\u6570\u304c\u5c11\u306a\u3044\u7cfb\u5217\uff08\u4f8b\uff1a\u6708\u306b\u6570\u4ef6\u301c\u6570\u5341\u4ef6\uff09\u3067\u306f\u3001\u30dd\u30a2\u30bd\u30f3\u7cfb\u306e\u30e2\u30c7\u30eb\u3092\u4f7f\u3046\u3079\u304d\u3067\u3042\u308b\u3002<\/p>\n<\/blockquote>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h2 class=\"wp-block-heading\">4. \u6b63\u653b\u6cd5\uff1a\u6587\u732e\u3067\u78ba\u8a8d\u3059\u308b\u6a19\u6e96\u7684\u306a\u624b\u9806<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">ITS\u306e\u65b9\u6cd5\u8ad6\u306f\u3001\u533b\u7642\u653f\u7b56\u30fb\u516c\u8846\u885b\u751f\u306e\u5206\u91ce\u3067\u6574\u5099\u3055\u308c\u3066\u304d\u305f\u3002\u521d\u5b66\u8005\u304c\u307e\u305a\u62bc\u3055\u3048\u3066\u304a\u304d\u305f\u3044\u6587\u732e\u3092\u6319\u3052\u308b\u3002<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>Wagner et al.\uff082002\uff09<\/strong>\uff1a\u533b\u85ac\u54c1\u306e\u4f7f\u7528\u306b\u95a2\u3059\u308b\u653f\u7b56\u30fb\u6559\u80b2\u4ecb\u5165\u306e\u8a55\u4fa1\u306b\u3001\u30bb\u30b0\u30e1\u30f3\u30c8\u56de\u5e30\u3092\u4f53\u7cfb\u7684\u306b\u9069\u7528\u3057\u305f\u53e4\u5178\u7684\u306a\u8ad6\u6587\u3067\u3042\u308b\u3002ITS\u3092\u300c\u6e96\u5b9f\u9a13\u30c7\u30b6\u30a4\u30f3\u306e\u4e2d\u3067\u6700\u3082\u5f37\u529b\u306a\u30a2\u30d7\u30ed\u30fc\u30c1\u300d\u3068\u4f4d\u7f6e\u3065\u3051\u3066\u3044\u308b\u3002<\/li>\n\n\n\n<li><strong>Lopez Bernal et al.\uff082017\uff09<\/strong>\uff1a\u516c\u8846\u885b\u751f\u4ecb\u5165\u306eITS\u306e\u30c1\u30e5\u30fc\u30c8\u30ea\u30a2\u30eb\u3067\u3042\u308b\u3002\u30a4\u30bf\u30ea\u30a2\u306e\u5168\u56fd\u7981\u7159\u6cd5\uff082005\u5e741\u6708\u65bd\u884c\uff09\u306b\u3064\u3044\u3066\u3001\u30b7\u30c1\u30ea\u30a2\u5dde\u306e\u6025\u6027\u51a0\u30a4\u30d9\u30f3\u30c8\u306e\u30c7\u30fc\u30bf\u3092\u4f8b\u306b\u3001\u5b63\u7bc0\u6027\u306e\u8abf\u6574\u3001\u81ea\u5df1\u76f8\u95a2\u306e\u78ba\u8a8d\u3001\u52b9\u679c\u306e\u5f62\uff08\u6bb5\u5dee\u304b\u50be\u304d\u304b\uff09\u306e\u4e8b\u524d\u8a2d\u5b9a\u307e\u3067\u3092\u89e3\u8aac\u3057\u3066\u3044\u308b\u3002\u30c7\u30fc\u30bf\u3068R\u306e\u30b3\u30fc\u30c9\u3082\u516c\u958b\u3055\u308c\u3066\u3044\u308b\u3002\u306a\u304a\u30012020\u5e74\u306b\u8a02\u6b63\uff08Corrigendum\uff09\u304c\u51fa\u3055\u308c\u3001\u30e2\u30c7\u30eb\u306e\u5b9a\u7fa9\u3068R\u306e\u30b3\u30fc\u30c9\u304c\u4fee\u6b63\u3055\u308c\u3066\u3044\u308b\u3002GitHub\u3067\u516c\u958b\u3055\u308c\u3066\u3044\u308b\u306e\u306f\u8a02\u6b63\u5f8c\u306e\u30b3\u30fc\u30c9\u3067\u3042\u308b\u3002<\/li>\n\n\n\n<li><strong>Kontopantelis et al.\uff082015, BMJ\uff09<\/strong>\uff1a\u30e9\u30f3\u30c0\u30e0\u5316\u304c\u3067\u304d\u306a\u3044\u3068\u304d\u306e\u6e96\u5b9f\u9a13\u30c7\u30b6\u30a4\u30f3\u3068\u3057\u3066\u306eITS\u306b\u3064\u3044\u3066\u3001\u5404\u30e2\u30c7\u30ea\u30f3\u30b0\u65b9\u6cd5\u306e\u9577\u6240\u30fb\u77ed\u6240\u30fb\u524d\u63d0\u3092\u5b9f\u4f8b\u3068\u3068\u3082\u306b\u89e3\u8aac\u3057\u3066\u3044\u308b\u3002<\/li>\n\n\n\n<li><strong>Turner et al.\uff082021\uff09<\/strong>\uff1a\u81ea\u5df1\u76f8\u95a2\u306e\u6271\u3044\u65b9\uff08OLS\u3001Newey-West\u3001Prais-Winsten\u3001REML\u3001ARIMA\u306a\u3069\uff09\u3092\u30b7\u30df\u30e5\u30ec\u30fc\u30b7\u30e7\u30f3\u3067\u6bd4\u8f03\u3057\u305f\u7814\u7a76\u3067\u3042\u308b\u3002<\/li>\n\n\n\n<li><strong>Schaffer et al.\uff082021\uff09<\/strong>\uff1a\u30bb\u30b0\u30e1\u30f3\u30c8\u56de\u5e30\u306e\u4ee3\u308f\u308a\u306b <strong>ARIMA\u30e2\u30c7\u30eb<\/strong> \u3067ITS\u3092\u884c\u3046\u65b9\u6cd5\u306e\u30ac\u30a4\u30c9\u3067\u3042\u308b\u3002\u81ea\u5df1\u76f8\u95a2\u3084\u5b63\u7bc0\u6027\u3092ARIMA\u306e\u67a0\u7d44\u307f\u3067\u67d4\u8edf\u306b\u8868\u73fe\u3067\u304d\u3001<a href=\"https:\/\/best-biostatistics.com\/toukei-er\/entry\/basics-of-time-series-analysis\/\">\u6642\u7cfb\u5217\u306e\u57fa\u790e\u8a18\u4e8b<\/a>\u3067\u6271\u3063\u305f <code>auto.arima()<\/code> \u3068\u3082\u5730\u7d9a\u304d\u3067\u3042\u308b\u3002\u672c\u8a18\u4e8b\u3067\u306f\u3001\u89e3\u91c8\u306e\u3057\u3084\u3059\u3055\u3092\u512a\u5148\u3057\u3066\u30bb\u30b0\u30e1\u30f3\u30c8\u56de\u5e30\u3092\u4f7f\u3046\u3002<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">\u81e8\u5e8a\u30fb\u516c\u8846\u885b\u751f\u3067\u306e\u5fdc\u7528\u4f8b\u3068\u3057\u3066\u306f\u3001<strong>Barone-Adesi et al.\uff082011\uff09<\/strong> \u304c\u3042\u308b\u3002\u30c1\u30e5\u30fc\u30c8\u30ea\u30a2\u30eb\u306e\u4f8b\u3068\u540c\u3058\u7981\u7159\u6cd5\u306b\u3064\u3044\u3066\u3001\u30a4\u30bf\u30ea\u30a2\u5168\u56fd20\u5dde\u306e\u30c7\u30fc\u30bf\u3092\u4f7f\u3044\u3001\u65bd\u884c\u524d\u5f8c\u306e\u6025\u6027\u51a0\u30a4\u30d9\u30f3\u30c8\u306b\u3088\u308b\u5165\u9662\u7387\u306e\u5909\u5316\u3092ITS\u3067\u8a55\u4fa1\u3057\u3066\u3044\u308b\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u3053\u308c\u3089\u306e\u6587\u732e\u3092\u8e0f\u307e\u3048\u305f\u6a19\u6e96\u7684\u306a\u624b\u9806\u306f\u3001\u6b21\u306e\u3068\u304a\u308a\u3067\u3042\u308b\u3002<\/p>\n\n\n\n<ol class=\"wp-block-list\">\n<li><strong>\u4ecb\u5165\u6642\u70b9\u3068\u3001\u60f3\u5b9a\u3059\u308b\u52b9\u679c\u306e\u5f62\uff08\u6bb5\u5dee\u30fb\u50be\u304d\u306e\u5909\u5316\u30fb\u79fb\u884c\u671f\u9593\uff09\u3092\u3001\u30c7\u30fc\u30bf\u3092\u898b\u308b\u524d\u306b\u6c7a\u3081\u308b<\/strong><\/li>\n\n\n\n<li><strong>\u30b0\u30e9\u30d5\u3092\u63cf\u304f<\/strong>\uff1a\u30c8\u30ec\u30f3\u30c9\u3001\u5b63\u7bc0\u6027\u3001\u5916\u308c\u5024\u3001\u4ecb\u5165\u524d\u5f8c\u306e\u5909\u5316\u3092\u76ee\u3067\u78ba\u8a8d\u3059\u308b<\/li>\n\n\n\n<li><strong>\u4ecb\u5165\u524d\u306e\u671f\u9593\u3067\u5b9a\u5e38\u6027\u3092\u78ba\u8a8d\u3059\u308b<\/strong>\uff08\u7f60\u2464\uff09<\/li>\n\n\n\n<li><strong>\u30bb\u30b0\u30e1\u30f3\u30c8\u56de\u5e30\u3092\u5f53\u3066\u306f\u3081\u308b<\/strong>\uff1a\u5b63\u7bc0\u6027\u3092\u8abf\u6574\u3059\u308b\uff08\u7f60\u2462\uff09<\/li>\n\n\n\n<li><strong>\u6b8b\u5dee\u306e\u81ea\u5df1\u76f8\u95a2\u3092\u78ba\u8a8d\u3057\u3001\u8003\u616e\u3057\u305f\u30e2\u30c7\u30eb\u3067\u63a8\u5b9a\u3059\u308b<\/strong>\uff08\u7f60\u2461\uff09<\/li>\n\n\n\n<li><strong>\u53cd\u4e8b\u5b9f\u3068\u6bd4\u3079\u3066\u3001\u52b9\u679c\u3092\u5177\u4f53\u7684\u306a\u6570\u5b57\u3067\u8a9e\u308b<\/strong>\uff1a\u300c\u25ef\u304b\u6708\u5f8c\u306b\u6708\u3042\u305f\u308a\u25ef\u4ef6\uff08\u25ef%\uff09\u5c11\u306a\u3044\u300d<\/li>\n\n\n\n<li><strong>\u5171\u4ecb\u5165\u306e\u53ef\u80fd\u6027\u3092\u691c\u8a0e\u3059\u308b<\/strong>\uff1a\u5fc5\u8981\u306a\u3089\u5bfe\u7167\u7cfb\u5217\u3092\u52a0\u3048\u308b\uff08\u7f60\u2463\u3001CITS\uff09<\/li>\n<\/ol>\n\n\n\n<p class=\"wp-block-paragraph\">\u624b\u98061\u304c\u6700\u521d\u306b\u3042\u308b\u3053\u3068\u306b\u6ce8\u610f\u3057\u3066\u307b\u3057\u3044\u3002\u30c7\u30fc\u30bf\u3092\u898b\u3066\u304b\u3089\u300c\u6bb5\u5dee\u3060\u3051\u306e\u30e2\u30c7\u30eb\u300d\u300c\u50be\u304d\u3082\u5165\u308c\u305f\u30e2\u30c7\u30eb\u300d\u306a\u3069\u3092\u53d6\u3063\u66ff\u3048\u5f15\u3063\u66ff\u3048\u3057\u3066\u3001\u4e00\u756a\u90fd\u5408\u306e\u3088\u3044\u7d50\u679c\u3092\u5831\u544a\u3059\u308b\u306e\u306f\u3001MK\u8a18\u4e8b\u3067\u3082\u89e6\u308c\u305fP\u5024\u30cf\u30c3\u30ad\u30f3\u30b0\u3067\u3042\u308b\u3002<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h2 class=\"wp-block-heading\">5. R\u3067\u5b9f\u88c5\u3057\u3066\u307f\u3088\u3046\uff08Step by Step\uff09<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">\u3053\u3053\u304b\u3089\u306f\u3001\u5192\u982d\u306e\u5f8c\u8f29\u306e\u30c7\u30fc\u30bf\u3092\u6a21\u3057\u305f\u7591\u4f3c\u30c7\u30fc\u30bf\u3067\u3001\u3053\u3053\u307e\u3067\u306e\u5185\u5bb9\u3092\u78ba\u8a8d\u3057\u3066\u3044\u304f\u3002\u4f7f\u3046\u30d1\u30c3\u30b1\u30fc\u30b8\u306f\u6b21\u306e4\u3064\u3067\u3042\u308b\u3002<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><code>nlme<\/code>\uff1a<code>gls()<\/code> \u3067\u3001\u81ea\u5df1\u76f8\u95a2\uff08AR(1)\uff09\u3092\u8003\u616e\u3057\u305f\u56de\u5e30\u3092\u884c\u3046<\/li>\n\n\n\n<li><code>lmtest<\/code>\uff1a<code>dwtest()<\/code> \u3067Durbin-Watson\u691c\u5b9a\u3001<code>coeftest()<\/code> \u3067\u6a19\u6e96\u8aa4\u5dee\u3092\u5dee\u3057\u66ff\u3048\u305f\u4fc2\u6570\u306e\u691c\u5b9a\u3092\u884c\u3046<\/li>\n\n\n\n<li><code>sandwich<\/code>\uff1a<code>NeweyWest()<\/code> \u3067\u3001\u81ea\u5df1\u76f8\u95a2\u306b\u9811\u5065\u306a\u6a19\u6e96\u8aa4\u5dee\u3092\u8a08\u7b97\u3059\u308b<\/li>\n\n\n\n<li><code>tseries<\/code>\uff1aADF-KPSS\u8a18\u4e8b\u3068\u540c\u3058\u304f\u3001<code>adf.test()<\/code> \u3068 <code>kpss.test()<\/code> \u3092\u4f7f\u3046<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">\u4ee5\u4e0b\u306e\u7d50\u679c\u306f\u3001R 4.6.1 \u3067\u5b9f\u969b\u306b\u5b9f\u884c\u3057\u3066\u78ba\u8a8d\u3057\u305f\u3082\u306e\u3067\u3042\u308b\u3002\u4e71\u6570\u3092\u4f7f\u3046\u305f\u3081\u3001<code>set.seed()<\/code> \u306e\u5024\u3084R\u306e\u30d0\u30fc\u30b8\u30e7\u30f3\u304c\u9055\u3046\u3068\u6570\u5024\u306f\u5909\u308f\u308b\u3002<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">Step 0\uff1a\u30d1\u30c3\u30b1\u30fc\u30b8\u306e\u6e96\u5099<\/h3>\n\n\n\n<pre class=\"wp-block-code\"><code># install.packages(c(\"nlme\", \"lmtest\", \"sandwich\", \"tseries\"))\nlibrary(nlme)\nlibrary(lmtest)\nlibrary(sandwich)\nlibrary(tseries)<\/code><\/pre>\n\n\n\n<h3 class=\"wp-block-heading\">Step 1\uff1a\u7591\u4f3c\u30c7\u30fc\u30bf\u3092\u4f5c\u308b<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">\u3042\u308b\u5e02\u306e\u6708\u6b21\u306e\u6551\u6025\u642c\u9001\u4ef6\u6570\uff082018\u5e741\u6708\u301c2025\u5e7412\u6708\u300196\u304b\u6708\uff09\u3092\u3001\u6b21\u306e\u8981\u7d20\u3092\u8db3\u3057\u5408\u308f\u305b\u3066\u4f5c\u308b\u3002<strong>\u30c7\u30fc\u30bf\u306f\u67b6\u7a7a\u306e\u5e02\u306e\u7591\u4f3c\u30c7\u30fc\u30bf\u3067\u3042\u308a\u3001\u5b9f\u5728\u306e#7119\u4e8b\u696d\u306e\u52b9\u679c\u3092\u793a\u3059\u3082\u306e\u3067\u306f\u306a\u3044\u3002<\/strong><\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>\u4ecb\u5165\u524d\u306e\u30c8\u30ec\u30f3\u30c9<\/strong>\uff1a1\u304b\u6708\u3042\u305f\u308a3\u4ef6\u305a\u3064\u5897\u3048\u308b\uff08\u9ad8\u9f62\u5316\u306b\u3088\u308b\u5897\u52a0\u3092\u60f3\u5b9a\uff09<\/li>\n\n\n\n<li><strong>\u5b63\u7bc0\u6027<\/strong>\uff1a1\u6708\u3092\u30d4\u30fc\u30af\u306b\u3001\u51ac\u306f\uff0b60\u4ef6\u3001\u590f\u306f\u221260\u4ef6\u307b\u3069\u4e0a\u4e0b\u3059\u308b\u6ce2<\/li>\n\n\n\n<li><strong>\u81ea\u5df1\u76f8\u95a2\u306e\u3042\u308b\u8aa4\u5dee<\/strong>\uff1a\u524d\u6708\u306e\u8aa4\u5dee\u306e5\u5272\u3092\u5f15\u304d\u7d99\u3050\u30ce\u30a4\u30ba\uff08AR(1)\u3001\u4fc2\u65700.5\uff09<\/li>\n\n\n\n<li><strong>#7119\u306e\u52b9\u679c\uff082022\u5e741\u6708\u304b\u3089\uff09<\/strong>\uff1a\u6c34\u6e96\u304c60\u4ef6\u4e0b\u304c\u308a\uff08\u6bb5\u5dee\uff09\u3001\u50be\u304d\u304c1\u304b\u6708\u3042\u305f\u308a2\u4ef6\u5c0f\u3055\u304f\u306a\u308b\uff08\u5897\u52a0\u30da\u30fc\u30b9\u304c\u67083\u4ef6\u304b\u3089\u67081\u4ef6\u306b\u920d\u308b\uff09<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">MK\u8a18\u4e8b\u3084ADF-KPSS\u8a18\u4e8b\u3068\u540c\u3058\u304f\u3001\u7f60\u2461\uff08\u81ea\u5df1\u76f8\u95a2\uff09\u3068\u7f60\u2462\uff08\u5b63\u7bc0\u6027\uff09\u304c\u518d\u73fe\u3055\u308c\u308b\u3088\u3046\u306b\u3001\u3042\u3048\u3066\u4ed5\u8fbc\u3093\u3067\u3044\u308b\u3002<\/p>\n\n\n\n<pre class=\"wp-block-code\"><code>set.seed(7119)\n\nn     &lt;- 96\ntime  &lt;- 1:n                              # \u6642\u9593\uff081\u301c96\u304b\u6708\u76ee\uff09\nmonth &lt;- rep(1:12, times = 8)             # \u6708\uff081\u301c12\uff09\nyear  &lt;- rep(2018:2025, each = 12)        # \u5e74\n\nlevel      &lt;- as.numeric(time &gt;= 49)      # \u4ecb\u5165\uff082022\u5e741\u6708\uff1d49\u304b\u6708\u76ee\u4ee5\u964d\u304c1\uff09\ntime_after &lt;- pmax(0, time - 49)          # \u4ecb\u5165\u5f8c\u306e\u7d4c\u904e\u6708\u6570\uff082022\u5e741\u6708\u304c0\uff09\n\nseason &lt;- 60 * cos(2 * pi * (month - 1) \/ 12)                            # 1\u6708\u30d4\u30fc\u30af\u306e\u5b63\u7bc0\u306e\u6ce2\nnoise  &lt;- as.numeric(arima.sim(model = list(ar = 0.5), n = n, sd = 25))  # \u81ea\u5df1\u76f8\u95a2\u306e\u3042\u308b\u8aa4\u5dee\n\ncalls &lt;- round(1000 + 3 * time - 60 * level - 2 * time_after + season + noise)\n\nd &lt;- data.frame(year, month, time, level, time_after, calls)\ncalls_ts &lt;- ts(calls, start = c(2018, 1), frequency = 12)<\/code><\/pre>\n\n\n\n<p class=\"wp-block-paragraph\">ITS\u306e\u89e3\u6790\u3067\u4e00\u756a\u5927\u4e8b\u306a\u306e\u306f\u3001<code>level<\/code> \u3068 <code>time_after<\/code> \u306e2\u5217\u3092\u6b63\u3057\u304f\u4f5c\u308b\u3053\u3068\u3067\u3042\u308b\u3002\u4ecb\u5165\u306e\u524d\u5f8c\u3067\u3001\u30c7\u30fc\u30bf\u306f\u6b21\u306e\u3088\u3046\u306b\u306a\u308b\u3002<\/p>\n\n\n\n<pre class=\"wp-block-code\"><code> <code>  year month time level time_after calls\n48 2021    12   48     0          0  1215\n49 2022     1   49     1          0  1182\n50 2022     2   50     1          1  1174<\/code><\/code><\/pre>\n\n\n\n<h3 class=\"wp-block-heading\">Step 2\uff1a\u4ecb\u5165\u7dda\u3092\u5165\u308c\u305f\u6642\u7cfb\u5217\u30b0\u30e9\u30d5<\/h3>\n\n\n\n<pre class=\"wp-block-code\"><code>plot(calls_ts, type = \"o\", pch = 16, cex = 0.6,\n     xlab = \"Year\", ylab = \"Ambulance transports (per month)\",\n     main = \"Monthly ambulance transports\")\nabline(v = 2022, col = \"red\", lty = 2)<\/code><\/pre>\n\n\n\n<figure class=\"wp-block-image size-full\"><img decoding=\"async\" width=\"800\" height=\"500\" src=\"https:\/\/best-biostatistics.com\/toukei-er\/wp-content\/uploads\/2026\/09\/step2-2.png\" alt=\"\" class=\"wp-image-4859\" srcset=\"https:\/\/best-biostatistics.com\/toukei-er\/wp-content\/uploads\/2026\/09\/step2-2.png 800w, https:\/\/best-biostatistics.com\/toukei-er\/wp-content\/uploads\/2026\/09\/step2-2-300x188.png 300w, https:\/\/best-biostatistics.com\/toukei-er\/wp-content\/uploads\/2026\/09\/step2-2-768x480.png 768w\" sizes=\"(max-width: 800px) 100vw, 800px\" \/><\/figure>\n\n\n\n<blockquote class=\"wp-block-quote is-layout-flow wp-block-quote-is-layout-flow\">\n<p class=\"wp-block-paragraph\">\u30b0\u30e9\u30d5\u306e\u30bf\u30a4\u30c8\u30eb\u3084\u8ef8\u30e9\u30d9\u30eb\u306f\u82f1\u8a9e\u306b\u3057\u3066\u3044\u308b\u3002Mac\u306a\u3069\u3067\u306f\u3001\u30d5\u30a9\u30f3\u30c8\u3092\u6307\u5b9a\u305b\u305a\u306b\u65e5\u672c\u8a9e\u3092\u4f7f\u3046\u3068\u6587\u5b57\u304c\u300c\u25a1\u300d\uff08\u3044\u308f\u3086\u308b\u8c46\u8150\uff09\u306b\u5316\u3051\u308b\u3053\u3068\u304c\u3042\u308b\u305f\u3081\u3060\u3002<\/p>\n<\/blockquote>\n\n\n\n<p class=\"wp-block-paragraph\">\u6bce\u5e74\u51ac\u306b\u5c71\u3001\u590f\u306b\u8c37\u304c\u6765\u308b\u5b63\u7bc0\u306e\u6ce2\u304c\u306f\u3063\u304d\u308a\u898b\u3048\u308b\u3002\u4ecb\u5165\u524d\uff08\u8d64\u3044\u70b9\u7dda\u306e\u5de6\u5074\uff09\u306f\u3001\u305d\u306e\u6ce2\u5168\u4f53\u304c\u53f3\u80a9\u4e0a\u304c\u308a\u306b\u5897\u3048\u3066\u3044\u308b\u3002\u4ecb\u5165\u5f8c\u306f\u30012022\u5e74\u306e\u51ac\u306e\u5c71\u304c2021\u5e74\u3088\u308a\u4f4e\u304f\u306a\u308a\u3001\u305d\u306e\u5f8c\u3082\u5897\u3048\u65b9\u304c\u3086\u308b\u3084\u304b\u306b\u306a\u3063\u3066\u3044\u308b\u3088\u3046\u306b\u898b\u3048\u308b\u3002\u305f\u3060\u3057\u3001\u76ee\u8996\u3060\u3051\u3067\u306f\u3001\u5b63\u7bc0\u306e\u6ce2\u306b\u57cb\u3082\u308c\u305f\u5909\u5316\u306e\u5927\u304d\u3055\u306f\u308f\u304b\u3089\u306a\u3044\u3002<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">Step 3\uff1a\u3010\u30c0\u30e1\u306a\u4f8b\u3011\u4ecb\u5165\u524d\u5f8c\u306e\u5e73\u5747\u3092\u6bd4\u3079\u308b<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">\u307e\u305a\u3001\u7f60\u2460\u3092\u308f\u3056\u3068\u8e0f\u3093\u3067\u307f\u308b\u3002\u4ecb\u5165\u524d48\u304b\u6708\u3068\u4ecb\u5165\u5f8c48\u304b\u6708\u306e\u5e73\u5747\u3092\u3001t\u691c\u5b9a\u3067\u6bd4\u3079\u308b\u3002<\/p>\n\n\n\n<pre class=\"wp-block-code\"><code>t.test(calls ~ level, data = d)<\/code><\/pre>\n\n\n\n<pre class=\"wp-block-code\"><code>\tWelch Two Sample t-test\n\ndata:  calls by level\nt = -2.5969, df = 92.523, p-value = 0.01094\n...\nsample estimates:\nmean in group 0 mean in group 1 \n       1073.958        1104.625<\/code><\/pre>\n\n\n\n<p class=\"wp-block-paragraph\">\u4ecb\u5165\u524d\u306e\u5e73\u5747\u306f\u7d041,074\u4ef6\u3001\u4ecb\u5165\u5f8c\u306e\u5e73\u5747\u306f\u7d041,105\u4ef6\u3067\u3042\u3063\u305f\u3002<strong>#7119\u306e\u5c0e\u5165\u5f8c\u306e\u65b9\u304c\u3001\u6708\u3042\u305f\u308a\u7d0431\u4ef6\u591a\u3044<\/strong> \u3068\u3044\u3046\u7d50\u679c\u3067\u3042\u308b\uff08P = 0.011\uff09\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u300c#7119\u3092\u59cb\u3081\u305f\u3089\u3001\u304b\u3048\u3063\u3066\u6551\u6025\u642c\u9001\u304c\u5897\u3048\u305f\u300d\u3068\u5831\u544a\u3057\u3066\u3057\u307e\u3044\u305d\u3046\u306b\u306a\u308b\u3002\u3057\u304b\u3057\u7591\u4f3c\u30c7\u30fc\u30bf\u306b\u306f\u3001#7119\u306b\u3088\u3063\u3066\u642c\u9001\u304c <strong>\u6e1b\u308b<\/strong> \u52b9\u679c\u3092\u5165\u308c\u3066\u3042\u308b\u3002\u3053\u306e\u9006\u8ee2\u304c\u8d77\u304d\u305f\u306e\u306f\u3001\u3082\u3068\u3082\u3068\u306e\u5897\u52a0\u50be\u5411\uff08\u67083\u4ef6\u305a\u3064\u5897\u3048\u308b\uff09\u3092\u7121\u8996\u3057\u305f\u304b\u3089\u3067\u3042\u308b\u3002\u30c0\u30a4\u30a8\u30c3\u30c8\u306e\u4f8b\u3068\u3061\u3087\u3046\u3069\u9006\u5411\u304d\u306e\u5931\u6557\u3060\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u52a0\u3048\u3066\u3001\u3053\u306et\u691c\u5b9a\u306f\u81ea\u5df1\u76f8\u95a2\u3082\u5b63\u7bc0\u6027\u3082\u7121\u8996\u3057\u3066\u3044\u308b\u306e\u3067\u3001P\u5024\u305d\u306e\u3082\u306e\u3082\u4fe1\u7528\u3067\u304d\u306a\u3044\u3002<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">Step 4\uff1a\u4ecb\u5165\u524d\u306e\u671f\u9593\u3060\u3051\u3067\u5b9a\u5e38\u6027\u3092\u78ba\u8a8d\u3059\u308b\uff08ADF\u00d7KPSS\uff09<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">\u7f60\u2464\u306b\u5f93\u3044\u3001<strong>\u4ecb\u5165\u524d\u306e48\u304b\u6708\u3060\u3051<\/strong> \u3092\u53d6\u308a\u51fa\u3057\u3066\u3001<a href=\"https:\/\/best-biostatistics.com\/toukei-er\/entry\/how-to-assess-stationarity-of-time-series-in-r-combining-the-adf-and-kpss-tests\/\">ADF-KPSS\u8a18\u4e8b<\/a>\u3068\u540c\u3058\u624b\u9806\u3067\u5b9a\u5e38\u6027\u3092\u78ba\u8a8d\u3059\u308b\u3002\u4ecb\u5165\u524d\u306f\u53f3\u80a9\u4e0a\u304c\u308a\u306a\u306e\u3067\u3001\u3069\u3061\u3089\u3082\u30c8\u30ec\u30f3\u30c9\u9805\u3092\u5165\u308c\u305f\u578b\u3067\u691c\u5b9a\u3059\u308b\uff08<code>adf.test()<\/code> \u306f\u5e38\u306b\u30c8\u30ec\u30f3\u30c9\u9805\u5165\u308a\u3001<code>kpss.test()<\/code> \u306f <code>null = \"Trend\"<\/code> \u3092\u6307\u5b9a\uff09\u3002<\/p>\n\n\n\n<pre class=\"wp-block-code\"><code>pre &lt;- subset(d, level == 0)\nadf.test(pre$calls)\nkpss.test(pre$calls, null = \"Trend\")<\/code><\/pre>\n\n\n\n<pre class=\"wp-block-code\"><code>Dickey-Fuller = -4.8327, Lag order = 3, p-value = 0.01\nKPSS Trend = 0.034537, Truncation lag parameter = 3, p-value = 0.1<\/code><\/pre>\n\n\n\n<p class=\"wp-block-paragraph\">ADF\u691c\u5b9a\u306f\u5358\u4f4d\u6839\u3092\u68c4\u5374\u3057\uff08P \u2264 0.01\u3002<code>tseries<\/code> \u306f0.01\u672a\u6e80\u306eP\u5024\u30920.01\u3068\u8868\u793a\u3059\u308b\uff09\u3001KPSS\u691c\u5b9a\u306f\u30c8\u30ec\u30f3\u30c9\u5b9a\u5e38\u3092\u68c4\u5374\u3057\u306a\u304b\u3063\u305f\uff08P \u2265 0.10\uff09\u3002ADF-KPSS\u8a18\u4e8b\u306e\u5224\u5b9a\u8868\u306b\u5f53\u3066\u306f\u3081\u308b\u3068\u3001<strong>2\u4eba\u306e\u88c1\u5224\u5b98\u306e\u610f\u898b\u304c\u300c\u30c8\u30ec\u30f3\u30c9\u5b9a\u5e38\u300d\u3067\u4e00\u81f4<\/strong> \u3059\u308b\u3002\u4ecb\u5165\u524d\u306e\u7cfb\u5217\u3092\u300c\u76f4\u7dda\u30c8\u30ec\u30f3\u30c9\uff0b\u5b9a\u5e38\u306a\u63fa\u3089\u304e\u300d\u3068\u307f\u306a\u3057\u3066\u30bb\u30b0\u30e1\u30f3\u30c8\u56de\u5e30\u3092\u884c\u3046\u3053\u3068\u306b\u3001\u5927\u304d\u306a\u554f\u984c\u306f\u306a\u3055\u305d\u3046\u3067\u3042\u308b\u3002<\/p>\n\n\n\n<blockquote class=\"wp-block-quote is-layout-flow wp-block-quote-is-layout-flow\">\n<p class=\"wp-block-paragraph\"><strong>\u6ce8\u610f<\/strong>\uff1a48\u70b9\u3068\u3044\u3046\u7cfb\u5217\u306f\u3001\u5358\u4f4d\u6839\u691c\u5b9a\u306b\u306f\u77ed\u3044\u90e8\u985e\u3067\u3042\u308b\u3002\u5b9f\u969b\u3001\u6708\u30c0\u30df\u30fc\u3067\u5b63\u7bc0\u6027\u3092\u9664\u3044\u305f\u6b8b\u5dee\u3067\u691c\u5b9a\u3057\u76f4\u3059\u3068\u3001ADF\u691c\u5b9a\u306f\u5358\u4f4d\u6839\u3092\u68c4\u5374\u3067\u304d\u305a\uff08P = 0.41\uff09\u3001KPSS\u691c\u5b9a\u3082\u68c4\u5374\u3057\u306a\u3044\u300c\u5224\u65ad\u4fdd\u7559\u300d\u306b\u306a\u3063\u305f\u3002\u77ed\u3044\u7cfb\u5217\u3067\u306f\u3001\u691c\u5b9a\u306e\u7d50\u679c\u3060\u3051\u3067\u767d\u9ed2\u3092\u3064\u3051\u305a\u3001\u30b0\u30e9\u30d5\u3084\u30b3\u30ec\u30ed\u30b0\u30e9\u30e0\u3001\u5206\u91ce\u306e\u77e5\u8b58\u3068\u5408\u308f\u305b\u3066\u5224\u65ad\u3057\u3066\u307b\u3057\u3044\u3002<\/p>\n<\/blockquote>\n\n\n\n<h3 class=\"wp-block-heading\">Step 5\uff1a\u30bb\u30b0\u30e1\u30f3\u30c8\u56de\u5e30\uff08OLS\uff09<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">\u3044\u3088\u3044\u3088ITS\u306e\u672c\u4f53\u3067\u3042\u308b\u3002\u7b2c2\u7ae0\u306e\u5f0f\u306b\u3001\u5b63\u7bc0\u6027\u3092\u8abf\u6574\u3059\u308b\u305f\u3081\u306e\u6708\u30c0\u30df\u30fc\uff08<code>factor(month)<\/code>\uff09\u3092\u52a0\u3048\u3066\u3001\u307e\u305a\u306f\u901a\u5e38\u306e\u6700\u5c0f\u4e8c\u4e57\u6cd5\uff08OLS\uff09\u3067\u5f53\u3066\u306f\u3081\u308b\u3002<\/p>\n\n\n\n<pre class=\"wp-block-code\"><code>m_ols &lt;- lm(calls ~ time + level + time_after + factor(month), data = d)\nround(coef(summary(m_ols))&#91;2:4, ], 4)<\/code><\/pre>\n\n\n\n<pre class=\"wp-block-code\"><code><code>           Estimate Std. Error t value Pr(&gt;|t|)\ntime         2.6339     0.2733  9.6361        0\nlevel      -56.5672    10.7990 -5.2382        0\ntime_after  -1.6677     0.3803 -4.3852        0<\/code><\/code><\/pre>\n\n\n\n<p class=\"wp-block-paragraph\">\uff08<code>Pr(&gt;|t|)<\/code> \u306e0\u306f\u3001\u5c0f\u6570\u7b2c4\u4f4d\u3067\u4e38\u3081\u305f\u7d50\u679c\u3067\u3042\u308a\u3001\u3044\u305a\u308c\u3082 P &lt; 0.001 \u3068\u3044\u3046\u610f\u5473\u3067\u3042\u308b\uff09<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><code>time<\/code>\uff08\u03b21\uff1d\u4ecb\u5165\u524d\u306e\u50be\u304d\uff09\uff1a\u6708\u3042\u305f\u308a\u7d04 <strong>+2.6\u4ef6<\/strong><\/li>\n\n\n\n<li><code>level<\/code>\uff08\u03b22\uff1d\u6c34\u6e96\u306e\u5909\u5316\uff09\uff1a\u4ecb\u5165\u76f4\u5f8c\u306b\u7d04 <strong>\u221257\u4ef6<\/strong> \u306e\u6bb5\u5dee<\/li>\n\n\n\n<li><code>time_after<\/code>\uff08\u03b23\uff1d\u50be\u304d\u306e\u5909\u5316\uff09\uff1a\u50be\u304d\u304c\u6708\u3042\u305f\u308a\u7d04 <strong>\u22121.7\u4ef6<\/strong> \u5c0f\u3055\u304f\u306a\u3063\u305f<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">Step 3\u3068\u306f\u9055\u3044\u3001\u300c#7119\u306e\u5c0e\u5165\u5f8c\u306b\u642c\u9001\u306f\u6e1b\u3063\u305f\u300d\u3068\u3044\u3046\u3001\u8a2d\u5b9a\u3069\u304a\u308a\u306e\u65b9\u5411\u306e\u7d50\u679c\u306b\u306a\u3063\u305f\u3002\u3082\u3068\u3082\u3068\u306e\u5897\u52a0\u50be\u5411\u3092\u300c\u4ecb\u5165\u524d\u306e\u50be\u304d\u300d\u3068\u3057\u3066\u5206\u3051\u3066\u63a8\u5b9a\u3057\u305f\u304b\u3089\u3067\u3042\u308b\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u305f\u3060\u3057\u3001\u3053\u306e\u6a19\u6e96\u8aa4\u5dee\u306f\u3001\u307e\u3060\u7f60\u2461\uff08\u81ea\u5df1\u76f8\u95a2\uff09\u3092\u7121\u8996\u3057\u305fOLS\u306e\u3082\u306e\u3067\u3042\u308b\u3002<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">Step 6\uff1a\u6b8b\u5dee\u306e\u81ea\u5df1\u76f8\u95a2\u3092\u78ba\u8a8d\u3059\u308b<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">OLS\u306e\u6b8b\u5dee\u306b\u81ea\u5df1\u76f8\u95a2\u304c\u6b8b\u3063\u3066\u3044\u306a\u3044\u304b\u3092\u3001\u30b3\u30ec\u30ed\u30b0\u30e9\u30e0\u3067\u78ba\u8a8d\u3059\u308b\u3002<\/p>\n\n\n\n<pre class=\"wp-block-code\"><code>acf(resid(m_ols), lag.max = 24,\n    main = \"Correlogram of residuals (segmented regression, OLS)\")\nround(acf(resid(m_ols), lag.max = 3, plot = FALSE)$acf&#91;2:4], 2)<\/code><\/pre>\n\n\n\n<figure class=\"wp-block-image size-full\"><img decoding=\"async\" width=\"800\" height=\"500\" src=\"https:\/\/best-biostatistics.com\/toukei-er\/wp-content\/uploads\/2026\/09\/step6.png\" alt=\"\" class=\"wp-image-4863\" srcset=\"https:\/\/best-biostatistics.com\/toukei-er\/wp-content\/uploads\/2026\/09\/step6.png 800w, https:\/\/best-biostatistics.com\/toukei-er\/wp-content\/uploads\/2026\/09\/step6-300x188.png 300w, https:\/\/best-biostatistics.com\/toukei-er\/wp-content\/uploads\/2026\/09\/step6-768x480.png 768w\" sizes=\"(max-width: 800px) 100vw, 800px\" \/><\/figure>\n\n\n\n<pre class=\"wp-block-code\"><code>&#91;1]  0.32  0.06 -0.04<\/code><\/pre>\n\n\n\n<p class=\"wp-block-paragraph\">\u6708\u30c0\u30df\u30fc\u3092\u5165\u308c\u305f\u306e\u3067\u5b63\u7bc0\u306e\u6ce2\u306f\u6d88\u3048\u3066\u3044\u308b\u304c\u3001<strong>\u30e9\u30b01\u306b\u7d040.32\u306e\u81ea\u5df1\u76f8\u95a2<\/strong> \u304c\u6b8b\u308a\u3001\u9752\u3044\u70b9\u7dda\u306e\u7bc4\u56f2\u3092\u306f\u307f\u51fa\u3057\u3066\u3044\u308b\u3002\u30e9\u30b02\u4ee5\u964d\u306f\u307b\u307c0\u3067\u3042\u308b\u3002\u300c\u524d\u6708\u306e\u30af\u30bb\u30921\u304b\u6708\u5206\u3060\u3051\u5f15\u304d\u305a\u308b\u300d\u3068\u3044\u3046\u3001AR(1)\u306b\u8fd1\u3044\u30d1\u30bf\u30fc\u30f3\u3067\u3042\u308b\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Durbin-Watson\u691c\u5b9a\u3067\u3082\u78ba\u8a8d\u3057\u3066\u304a\u304f\u3002<\/p>\n\n\n\n<pre class=\"wp-block-code\"><code>dwtest(m_ols)<\/code><\/pre>\n\n\n\n<pre class=\"wp-block-code\"><code>DW = 1.3165, p-value = 0.000246\nalternative hypothesis: true autocorrelation is greater than 0<\/code><\/pre>\n\n\n\n<p class=\"wp-block-paragraph\">DW\u7d71\u8a08\u91cf\u306f\u3001\u81ea\u5df1\u76f8\u95a2\u304c\u306a\u3051\u308c\u30702\u306b\u8fd1\u304f\u3001\u6b63\u306e\u81ea\u5df1\u76f8\u95a2\u304c\u3042\u308b\u30682\u3088\u308a\u5c0f\u3055\u304f\u306a\u308b\u3002\u3053\u3053\u3067\u306f1.32\u3067\u3001\u6b63\u306e\u81ea\u5df1\u76f8\u95a2\u304c\u3042\u308b\u3068\u5224\u5b9a\u3055\u308c\u305f\uff08P &lt; 0.001\uff09\u3002<\/p>\n\n\n\n<blockquote class=\"wp-block-quote is-layout-flow wp-block-quote-is-layout-flow\">\n<p class=\"wp-block-paragraph\"><strong>\u6ce8\u610f<\/strong>\uff1aTurner et al.\uff082021\uff09\u306f\u3001Durbin-Watson\u691c\u5b9a\u306f <strong>\u7cfb\u5217\u304c\u9577\u304f\u3001\u81ea\u5df1\u76f8\u95a2\u304c\u5927\u304d\u3044\u5834\u5408\u3092\u9664\u3044\u3066\u3001\u81ea\u5df1\u76f8\u95a2\u306e\u691c\u51fa\u529b\u304c\u4f4e\u3044<\/strong> \u3053\u3068\u3092\u793a\u3057\u3001\u3053\u306e\u691c\u5b9a\u306b\u983c\u308b\u3079\u304d\u3067\u306f\u306a\u3044\u3068\u3057\u3066\u3044\u308b\u3002\u300cDW\u691c\u5b9a\u3067\u6709\u610f\u3067\u306a\u304b\u3063\u305f\u304b\u3089\u3001\u81ea\u5df1\u76f8\u95a2\u306f\u306a\u3044\u300d\u3068\u5224\u65ad\u3059\u308b\u306e\u306f\u5371\u967a\u3067\u3042\u308b\u3002\u30b3\u30ec\u30ed\u30b0\u30e9\u30e0\u3068\u3042\u308f\u305b\u3066\u898b\u3066\u3001\u8ff7\u3063\u305f\u3089\u81ea\u5df1\u76f8\u95a2\u3092\u8003\u616e\u3057\u305f\u30e2\u30c7\u30eb\u3092\u4f7f\u3046\u65b9\u304c\u5b89\u5168\u3067\u3042\u308b\u3002<\/p>\n<\/blockquote>\n\n\n\n<h3 class=\"wp-block-heading\">Step 7\uff1a\u81ea\u5df1\u76f8\u95a2\u3092\u8003\u616e\u3057\u305f\u30bb\u30b0\u30e1\u30f3\u30c8\u56de\u5e30\uff08GLS\uff0bAR(1)\u3001REML\uff09<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\"><code>nlme<\/code> \u30d1\u30c3\u30b1\u30fc\u30b8\u306e <code>gls()<\/code> \u3067\u3001\u8aa4\u5dee\u306bAR(1)\u306e\u81ea\u5df1\u76f8\u95a2\u3092\u4eee\u5b9a\u3057\u305f\u30bb\u30b0\u30e1\u30f3\u30c8\u56de\u5e30\u3092\u884c\u3046\u3002<code>correlation = corAR1(form = ~ time)<\/code> \u304c\u300c\u524d\u306e\u6708\u306e\u8aa4\u5dee\u3092\u5f15\u304d\u305a\u308b\u300d\u3068\u3044\u3046\u6307\u5b9a\u3067\u3042\u308a\u3001\u63a8\u5b9a\u6cd5\u306b\u306f Turner et al.\uff082021\uff09\u3067\u63a8\u5968\u3055\u308c\u3066\u3044\u308bREML\u3092\u4f7f\u3046\u3002<\/p>\n\n\n\n<pre class=\"wp-block-code\"><code>m_gls &lt;- gls(calls ~ time + level + time_after + factor(month), data = d,\n             correlation = corAR1(form = ~ time), method = \"REML\")\nround(summary(m_gls)$tTable&#91;2:4, ], 4)\nround(intervals(m_gls, which = \"coef\")$coef&#91;2:4, ], 2)\ncoef(m_gls$modelStruct$corStruct, unconstrained = FALSE)   # AR(1)\u4fc2\u6570<\/code><\/pre>\n\n\n\n<pre class=\"wp-block-code\"><code> <code>             Value Std.Error t-value p-value\ntime         2.6825    0.3997  6.7117  0.0000\nlevel      -56.3415   15.3333 -3.6744  0.0004\ntime_after  -1.7703    0.5695 -3.1083  0.0026\n\n            lower   est.  upper\ntime         1.89   2.68   3.48\nlevel      -86.85 -56.34 -25.83\ntime_after  -2.90  -1.77  -0.64\n\n      Phi \n0.3814738<\/code> <\/code><\/pre>\n\n\n\n<p class=\"wp-block-paragraph\">OLS\uff08Step 5\uff09\u3068\u6bd4\u3079\u3066\u307f\u3088\u3046\u3002<\/p>\n\n\n\n<figure class=\"wp-block-table\"><table class=\"has-fixed-layout\"><thead><tr><th><\/th><th>OLS\uff1a\u63a8\u5b9a\u5024\uff08\u6a19\u6e96\u8aa4\u5dee\uff09<\/th><th>GLS\uff0bAR(1)\uff1a\u63a8\u5b9a\u5024\uff08\u6a19\u6e96\u8aa4\u5dee\uff09<\/th><th>\u7591\u4f3c\u30c7\u30fc\u30bf\u306e\u771f\u306e\u5024<\/th><\/tr><\/thead><tbody><tr><td>\u4ecb\u5165\u524d\u306e\u50be\u304d \u03b21<\/td><td>2.63\uff080.27\uff09<\/td><td>2.68\uff080.40\uff09<\/td><td>3<\/td><\/tr><tr><td>\u6c34\u6e96\u306e\u5909\u5316 \u03b22<\/td><td>\u221256.6\uff0810.8\uff09<\/td><td>\u221256.3\uff0815.3\uff09<\/td><td>\u221260<\/td><\/tr><tr><td>\u50be\u304d\u306e\u5909\u5316 \u03b23<\/td><td>\u22121.67\uff080.38\uff09<\/td><td>\u22121.77\uff080.57\uff09<\/td><td>\u22122<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>\u63a8\u5b9a\u5024\u306f\u307b\u3068\u3093\u3069\u5909\u308f\u3089\u306a\u3044\u304c\u3001\u6a19\u6e96\u8aa4\u5dee\u306f1.4\u301c1.5\u500d\u306b\u5e83\u304c\u3063\u305f<\/strong>\u3002OLS\u306f\u3001\u300c\u4ef2\u826f\u3057\u30b0\u30eb\u30fc\u30d7\u300d\u3092\u72ec\u7acb\u306a\u7968\u3068\u3057\u3066\u6570\u3048\u3066\u3044\u305f\u3076\u3093\u3001\u63a8\u5b9a\u306e\u78ba\u304b\u3055\u3092\u904e\u5927\u306b\u898b\u7a4d\u3082\u3063\u3066\u3044\u305f\u306e\u3067\u3042\u308b\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">GLS\u306e\u7d50\u679c\u3001\u6c34\u6e96\u306e\u5909\u5316\u306f \u221256.3\u4ef6\uff0895%\u4fe1\u983c\u533a\u9593 \u221286.9\u301c\u221225.8\uff09\u3001\u50be\u304d\u306e\u5909\u5316\u306f\u6708\u3042\u305f\u308a \u22121.77\u4ef6\uff0895%\u4fe1\u983c\u533a\u9593 \u22122.90\u301c\u22120.64\uff09\u3068\u63a8\u5b9a\u3055\u308c\u305f\u3002\u3069\u3061\u3089\u306e\u533a\u9593\u30820\u3092\u542b\u307e\u306a\u3044\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u3053\u3053\u3067\u3001\u03b23\u306e\u8aad\u307f\u65b9\u306b\u6ce8\u610f\u3057\u3066\u307b\u3057\u3044\u3002\u03b23\u306f <strong>\u50be\u304d\u306e\u300c\u5909\u5316\u300d<\/strong> \u3067\u3042\u308a\u3001\u4ecb\u5165\u5f8c\u306e\u50be\u304d\u305d\u306e\u3082\u306e\u3067\u306f\u306a\u3044\u3002\u4ecb\u5165\u5f8c\u306e\u50be\u304d\u306f \u03b21 + \u03b23 \u3067\u6c42\u3081\u308b\u3002\u70b9\u63a8\u5b9a\u5024\u306f\u8db3\u3057\u7b97\u3067\u51fa\u305b\u308b\u304c\u300195%\u4fe1\u983c\u533a\u9593\u306b\u306f2\u3064\u306e\u4fc2\u6570\u306e\u5171\u5206\u6563\u3082\u5fc5\u8981\u306a\u306e\u3067\u3001<code>vcov()<\/code> \u3092\u4f7f\u3063\u3066\u8a08\u7b97\u3059\u308b\u3002<\/p>\n\n\n\n<pre class=\"wp-block-code\"><code># \u4ecb\u5165\u5f8c\u306e\u50be\u304d\uff08\u03b21 + \u03b23\uff09\u306895%\u4fe1\u983c\u533a\u9593\nb_gls  &lt;- coef(m_gls)\nw_post &lt;- setNames(rep(0, length(b_gls)), names(b_gls))\nw_post&#91;c(\"time\", \"time_after\")] &lt;- 1\npost_slope &lt;- sum(w_post * b_gls)\npost_se    &lt;- sqrt(drop(t(w_post) %*% vcov(m_gls) %*% w_post))\nround(c(slope = post_slope,\n        lower = post_slope - 1.96 * post_se,\n        upper = post_slope + 1.96 * post_se), 2)<\/code><\/pre>\n\n\n\n<pre class=\"wp-block-code\"><code>slope lower upper \n 0.91  0.13  1.70<\/code><\/pre>\n\n\n\n<p class=\"wp-block-paragraph\">\u4ecb\u5165\u5f8c\u306e\u50be\u304d\u306f \u03b21 + \u03b23 \u2248 2.68 \u2212 1.77 = <strong>\u6708\u3042\u305f\u308a\u7d04+0.9\u4ef6<\/strong>\uff0895%\u4fe1\u983c\u533a\u9593 0.13\u301c1.70\uff09\u3067\u3042\u308b\u3002\u3064\u307e\u308a\u3001#7119\u306e\u5c0e\u5165\u5f8c\u3082\u642c\u9001\u4ef6\u6570\u306f <strong>\u5897\u3048\u7d9a\u3051\u3066\u3044\u308b\u304c\u3001\u5897\u3048\u65b9\u304c\u67083\u4ef6\u5f31\u304b\u3089\u67081\u4ef6\u5f31\u306b\u920d\u3063\u305f<\/strong> \u3068\u3044\u3046\u306e\u304c\u6b63\u3057\u3044\u89e3\u91c8\u3067\u3042\u308b\u3002\u300c\u5c0e\u5165\u5f8c\u306f\u6bce\u67081.8\u4ef6\u305a\u3064\u6e1b\u3063\u3066\u3044\u308b\u300d\u3068\u8aad\u3080\u306e\u306f\u8aa4\u308a\u3060\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u63a8\u5b9a\u3055\u308c\u305fAR(1)\u4fc2\u6570\uff08<code>Phi<\/code>\uff09\u306f0.38\u3067\u3001\u7591\u4f3c\u30c7\u30fc\u30bf\u306b\u4ed5\u8fbc\u3093\u3060\u771f\u306e\u50240.5\u3088\u308a\u5c0f\u3055\u304b\u3063\u305f\u3002\u4e71\u65701\u56de\u5206\u306e\u7d50\u679c\u306a\u306e\u3067\u65ad\u5b9a\u306f\u3067\u304d\u306a\u3044\u304c\u3001Turner et al.\uff082021\uff09\u304c\u793a\u3057\u305f\u300c\u81ea\u5df1\u76f8\u95a2\u306f\u3069\u306e\u65b9\u6cd5\u3067\u3082\u904e\u5c0f\u8a55\u4fa1\u3055\u308c\u304c\u3061\u300d\u3068\u3044\u3046\u5831\u544a\u3068\u6574\u5408\u7684\u3067\u3042\u308b\u3002<strong>GLS\u306e\u4fe1\u983c\u533a\u9593\u3067\u3055\u3048\u3001\u5b9f\u969b\u306b\u306f\u3084\u3084\u72ed\u3059\u304e\u308b\u53ef\u80fd\u6027\u304c\u3042\u308b<\/strong> \u3053\u3068\u3092\u982d\u306b\u5165\u308c\u3066\u304a\u304d\u305f\u3044\u3002<\/p>\n\n\n\n<blockquote class=\"wp-block-quote is-layout-flow wp-block-quote-is-layout-flow\">\n<p class=\"wp-block-paragraph\"><strong>\u53c2\u8003\uff1aNewey-West\u6a19\u6e96\u8aa4\u5dee<\/strong><br>OLS\u306e\u63a8\u5b9a\u5024\u306f\u305d\u306e\u307e\u307e\u306b\u3001\u6a19\u6e96\u8aa4\u5dee\u3060\u3051\u3092\u81ea\u5df1\u76f8\u95a2\u306b\u9811\u5065\u306a\u3082\u306e\u306b\u5dee\u3057\u66ff\u3048\u308b\u65b9\u6cd5\u3082\u3042\u308b\u3002\u7d4c\u6e08\u5b66\u306a\u3069\u3067\u3088\u304f\u4f7f\u308f\u308c\u308b <strong>Newey-West\u6a19\u6e96\u8aa4\u5dee<\/strong> \u3067\u3042\u308b\u3002<\/p>\n\n\n\n<pre class=\"wp-block-code\"><code>round(coeftest(m_ols, vcov = NeweyWest(m_ols, lag = 3, prewhite = FALSE))&#91;2:4, ], 4)\n#            Estimate Std. Error t value Pr(&gt;|t|)\n# time         2.6339     0.2867  9.1878    0e+00\n# level      -56.5672     9.7766 -5.7860    0e+00\n# time_after  -1.6677     0.4533 -3.6794    4e-04<\/code><\/pre>\n\n\n\n<p class=\"wp-block-paragraph\">\u3053\u306e\u30c7\u30fc\u30bf\u3067\u306f\u3001\u6c34\u6e96\u306e\u5909\u5316\u306e\u6a19\u6e96\u8aa4\u5dee\u304cOLS\u3088\u308a\u3082\u3080\u3057\u308d\u5c0f\u3055\u304f\u306a\u3063\u305f\uff0810.8 \u2192 9.8\uff09\u3002Newey-West\u6a19\u6e96\u8aa4\u5dee\u306f\u5927\u6a19\u672c\u3092\u524d\u63d0\u3068\u3057\u305f\u65b9\u6cd5\u3067\u3042\u308a\u300196\u70b9\u7a0b\u5ea6\u306e\u7cfb\u5217\u3067\u306f\u81ea\u5df1\u76f8\u95a2\u3092\u5341\u5206\u306b\u53cd\u6620\u3067\u304d\u306a\u3044\u3053\u3068\u304c\u3042\u308b\u3002Turner et al.\uff082021\uff09\u3067\u3082\u300112\u70b9\u4ee5\u4e0a\u306e\u7cfb\u5217\u3067\u306fREML\u304c\u63a8\u5968\u3055\u308c\u3066\u3044\u308b\u3002<\/p>\n<\/blockquote>\n\n\n\n<h3 class=\"wp-block-heading\">Step 8\uff1a\u53cd\u4e8b\u5b9f\u3068\u6bd4\u3079\u3066\u3001\u52b9\u679c\u3092\u5177\u4f53\u7684\u306a\u6570\u5b57\u3067\u8a9e\u308b<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">\u6700\u5f8c\u306b\u3001ITS\u306e\u6838\u5fc3\u3067\u3042\u308b <strong>\u53cd\u4e8b\u5b9f\u306e\u7dda<\/strong> \u3092\u63cf\u304f\u3002<code>level<\/code> \u3068 <code>time_after<\/code> \u30920\u306b\u3057\u305f\u30c7\u30fc\u30bf\u3067\u4e88\u6e2c\u3059\u308c\u3070\u3001\u300c#7119\u304c\u306a\u304b\u3063\u305f\u5834\u5408\u306e\u5ef6\u9577\u7dda\u300d\u304c\u5f97\u3089\u308c\u308b\u3002<\/p>\n\n\n\n<pre class=\"wp-block-code\"><code>cf &lt;- d\ncf$level      &lt;- 0\ncf$time_after &lt;- 0\nd$fit &lt;- predict(m_gls)                  # \u30e2\u30c7\u30eb\u306e\u5f53\u3066\u306f\u3081\u5024\nd$cf  &lt;- predict(m_gls, newdata = cf)    # #7119\u304c\u306a\u304b\u3063\u305f\u5834\u5408\u306e\u4e88\u6e2c\uff08\u53cd\u4e8b\u5b9f\uff09\n\nplot(d$time, d$calls, pch = 16, cex = 0.6, col = \"grey40\", xaxt = \"n\",\n     xlab = \"Year\", ylab = \"Ambulance transports (per month)\",\n     main = \"Observed, fitted, and counterfactual\",\n     ylim = c(900, max(c(d$calls, d$cf))))\naxis(1, at = seq(1, 96, by = 12), labels = 2018:2025)\nlines(d$time, d$fit, col = \"blue\", lwd = 2)\nlines(d$time&#91;d$level == 1], d$cf&#91;d$level == 1], col = \"red\", lwd = 2, lty = 2)\nabline(v = 48.5, lty = 3)\nlegend(\"bottomright\",\n       c(\"Observed\", \"Fitted (GLS)\", \"Counterfactual (no #7119)\"),\n       col = c(\"grey40\", \"blue\", \"red\"), pch = c(16, NA, NA),\n       lty = c(NA, 1, 2), lwd = c(NA, 2, 2), bty = \"n\", cex = 0.85)<\/code><\/pre>\n\n\n\n<figure class=\"wp-block-image size-full\"><img decoding=\"async\" width=\"800\" height=\"500\" src=\"https:\/\/best-biostatistics.com\/toukei-er\/wp-content\/uploads\/2026\/09\/step8.png\" alt=\"\" class=\"wp-image-4866\" srcset=\"https:\/\/best-biostatistics.com\/toukei-er\/wp-content\/uploads\/2026\/09\/step8.png 800w, https:\/\/best-biostatistics.com\/toukei-er\/wp-content\/uploads\/2026\/09\/step8-300x188.png 300w, https:\/\/best-biostatistics.com\/toukei-er\/wp-content\/uploads\/2026\/09\/step8-768x480.png 768w\" sizes=\"(max-width: 800px) 100vw, 800px\" \/><\/figure>\n\n\n\n<p class=\"wp-block-paragraph\">\u9752\u3044\u5b9f\u7dda\u304c\u5b9f\u969b\u306e\u30c7\u30fc\u30bf\u306b\u5f53\u3066\u306f\u3081\u305f\u30e2\u30c7\u30eb\u3001\u8d64\u3044\u70b9\u7dda\u304c\u300c#7119\u304c\u306a\u304b\u3063\u305f\u4e16\u754c\u300d\u306e\u5ef6\u9577\u7dda\u3067\u3042\u308b\u30022\u672c\u306e\u7dda\u306e <strong>\u9699\u9593<\/strong> \u304c\u3001#7119\u306e\u52b9\u679c\u306e\u63a8\u5b9a\u5024\u3067\u3042\u308b\u3002\u4ecb\u5165\u76f4\u5f8c\u306e\u6bb5\u5dee\u304c\u3001\u6642\u9593\u3068\u3068\u3082\u306b\u958b\u3044\u3066\u3044\u304f\u69d8\u5b50\u304c\u308f\u304b\u308b\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u9699\u9593\u306e\u5927\u304d\u3055\u3092\u6570\u5b57\u3067\u51fa\u3059\u306b\u306f\u3001\u4ecb\u5165\u304b\u3089 k \u304b\u6708\u5f8c\u306e\u52b9\u679c \u03b22 + \u03b23 \u00d7 k \u3068\u3001\u305d\u306e95%\u4fe1\u983c\u533a\u9593\u3092\u8a08\u7b97\u3059\u308b\u3002\u53cd\u4e8b\u5b9f\u306b\u5bfe\u3059\u308b\u76f8\u5bfe\u5909\u5316\u3082\u3042\u308f\u305b\u3066\u793a\u3059\u3002<\/p>\n\n\n\n<pre class=\"wp-block-code\"><code>effect_at &lt;- function(model, data, k) {\n  b &lt;- coef(model)\n  V &lt;- vcov(model)\n  w &lt;- setNames(rep(0, length(b)), names(b))\n  w&#91;\"level\"] &lt;- 1\n  w&#91;\"time_after\"] &lt;- k\n  est &lt;- sum(w * b)                          # \u03b22 + \u03b23 \u00d7 k\n  se  &lt;- sqrt(drop(t(w) %*% V %*% w))        # \u305d\u306e\u6a19\u6e96\u8aa4\u5dee\n  row &lt;- which(data$level == 1 &amp; data$time_after == k)\n  data.frame(month    = sprintf(\"%d-%02d\", data$year&#91;row], data$month&#91;row]),\n             effect   = round(est, 1),\n             lower    = round(est - 1.96 * se, 1),\n             upper    = round(est + 1.96 * se, 1),\n             relative = paste0(round(100 * est \/ data$cf&#91;row], 1), \"%\"))\n}\ndo.call(rbind, lapply(c(0, 12, 23), function(k) effect_at(m_gls, d, k)))<\/code><\/pre>\n\n\n\n<pre class=\"wp-block-code\"><code> <code>   month effect  lower upper relative\n1 2022-01  -56.3  -86.4 -26.3    -4.7%\n2 2023-01  -77.6 -110.7 -44.5    -6.2%\n3 2023-12  -97.1 -136.9 -57.2    -7.7%<\/code><\/code><\/pre>\n\n\n\n<p class=\"wp-block-paragraph\">\uff08\u4fe1\u983c\u533a\u9593\u306f\u6b63\u898f\u8fd1\u4f3c\uff08\u00b11.96\u00d7\u6a19\u6e96\u8aa4\u5dee\uff09\u3067\u8a08\u7b97\u3057\u3066\u3044\u308b\u305f\u3081\u3001Step 7\u306e <code>intervals()<\/code>\uff08t\u5206\u5e03\u306b\u57fa\u3065\u304f\uff09\u3068\u306f\u308f\u305a\u304b\u306b\u7570\u306a\u308b\uff09<\/p>\n\n\n\n<blockquote class=\"wp-block-quote is-layout-flow wp-block-quote-is-layout-flow\">\n<p class=\"wp-block-paragraph\"><strong>\u6ce8\u610f<\/strong>\uff1a\u76f8\u5bfe\u5909\u5316\uff08%\uff09\u306e\u5206\u6bcd\u306f\u3001\u305d\u306e\u6708\u306e\u53cd\u4e8b\u5b9f\u306e\u5024\u3067\u3042\u308b\u3002\u53cd\u4e8b\u5b9f\u306b\u306f\u5b63\u7bc0\u306e\u6ce2\u304c\u542b\u307e\u308c\u308b\u306e\u3067\u3001\u540c\u3058\u4ef6\u6570\u306e\u5dee\u3067\u3082\u3001\u5206\u6bcd\u304c\u5927\u304d\u3044\u51ac\u306e\u6708\u3067\u306f%\u304c\u5c0f\u3055\u304f\u3001\u5206\u6bcd\u304c\u5c0f\u3055\u3044\u590f\u306e\u6708\u3067\u306f%\u304c\u5927\u304d\u304f\u306a\u308b\u3002%\u3092\u5831\u544a\u3059\u308b\u3068\u304d\u306f\u3001<strong>\u3069\u306e\u6708\u306e\u5024\u304b<\/strong> \u3092\u5fc5\u305a\u6dfb\u3048\u3066\u307b\u3057\u3044\u3002<\/p>\n<\/blockquote>\n\n\n\n<p class=\"wp-block-paragraph\">\u3053\u308c\u3067\u3001\u7d50\u679c\u3092\u73fe\u5834\u306e\u8a00\u8449\u306b\u7ffb\u8a33\u3067\u304d\u308b\u3002<\/p>\n\n\n\n<blockquote class=\"wp-block-quote is-layout-flow wp-block-quote-is-layout-flow\">\n<p class=\"wp-block-paragraph\">#7119\u306e\u5c0e\u5165\u76f4\u5f8c\uff082022\u5e741\u6708\uff09\u3001\u6551\u6025\u642c\u9001\u4ef6\u6570\u306f\u53cd\u4e8b\u5b9f\u3068\u6bd4\u3079\u3066\u6708\u3042\u305f\u308a\u7d0456\u4ef6\uff08\u7d044.7%\uff09\u5c11\u306a\u304b\u3063\u305f\u3002\u305d\u306e\u5dee\u306f\u6642\u9593\u3068\u3068\u3082\u306b\u5e83\u304c\u308a\u3001\u5c0e\u5165\u304b\u308924\u304b\u6708\u76ee\uff082023\u5e7412\u6708\uff09\u306b\u306f\u6708\u3042\u305f\u308a\u7d0497\u4ef6\uff0895%\u4fe1\u983c\u533a\u9593 57\u301c137\u4ef6\u3001\u7d047.7%\uff09\u5c11\u306a\u3044\u3068\u63a8\u5b9a\u3055\u308c\u305f\u3002<\/p>\n<\/blockquote>\n\n\n\n<p class=\"wp-block-paragraph\">P\u5024\u3060\u3051\u3092\u5831\u544a\u3059\u308b\u306e\u3067\u306f\u306a\u304f\u3001<strong>\u300c\u4f55\u4ef6\uff08\u4f55%\uff09\u6e1b\u3063\u305f\u304b\u300d\u3092\u4fe1\u983c\u533a\u9593\u3068\u3068\u3082\u306b\u793a\u3059<\/strong> \u3053\u3068\u304c\u5927\u5207\u3067\u3042\u308b\u3002P\u5024\u306f\u300c\u5076\u7136\u3068\u306f\u8003\u3048\u306b\u304f\u3044\u304b\u300d\u3092\u5224\u65ad\u3059\u308b\u6750\u6599\u306b\u3059\u304e\u305a\u3001\u52b9\u679c\u306e\u5927\u304d\u3055\u3092\u8868\u3059\u3082\u306e\u3067\u306f\u306a\u3044\u3002<\/p>\n\n\n\n<blockquote class=\"wp-block-quote is-layout-flow wp-block-quote-is-layout-flow\">\n<p class=\"wp-block-paragraph\"><strong>\u6ce8\u610f<\/strong>\uff1a\u53cd\u4e8b\u5b9f\u306e\u5ef6\u9577\u7dda\u306f\u3001\u300c\u4ecb\u5165\u524d\u306e\u30c8\u30ec\u30f3\u30c9\u304c\u3001\u4ecb\u5165\u304c\u306a\u3051\u308c\u3070\u305d\u306e\u307e\u307e\u7d9a\u3044\u305f\u300d\u3068\u3044\u3046\u4eee\u5b9a\u306b\u7acb\u3063\u3066\u3044\u308b\u3002\u4ecb\u5165\u304b\u3089\u6642\u9593\u304c\u305f\u3064\u307b\u3069\u3001\u3053\u306e\u4eee\u5b9a\u306f\u4e0d\u78ba\u304b\u306b\u306a\u308b\u3002\u672c\u8a18\u4e8b\u3067\u52b9\u679c\u3092\u793a\u3057\u305f\u306e\u309224\u304b\u6708\u76ee\u307e\u3067\u306b\u3068\u3069\u3081\u305f\u306e\u306f\u305d\u306e\u305f\u3081\u3067\u3042\u308b\u3002\u9060\u3044\u5c06\u6765\u307e\u3067\u5ef6\u9577\u7dda\u3092\u4f38\u3070\u3057\u3066\u300c5\u5e74\u5f8c\u306b\u306f\u25ef\u4ef6\u6e1b\u3063\u305f\u300d\u3068\u8a9e\u308b\u306e\u306f\u3001\u614e\u91cd\u3067\u3042\u308b\u3079\u304d\u3060\u3002<\/p>\n<\/blockquote>\n\n\n\n<blockquote class=\"wp-block-quote is-layout-flow wp-block-quote-is-layout-flow\">\n<p class=\"wp-block-paragraph\"><strong>\u203b\u30b3\u30e9\u30e0\uff1aITS\u306e\u7d50\u679c\u3092\u300c\u56e0\u679c\u300d\u3068\u547c\u3093\u3067\u3088\u3044\u304b<\/strong><br>\u4eca\u56de\u306e\u7591\u4f3c\u30c7\u30fc\u30bf\u306b\u306f #7119 \u4ee5\u5916\u306e\u51fa\u6765\u4e8b\u3092\u5165\u308c\u3066\u3044\u306a\u3044\u306e\u3067\u3001\u63a8\u5b9a\u5024\u306f\u8a2d\u5b9a\u3057\u305f\u52b9\u679c\u306b\u8fd1\u304f\u306a\u3063\u305f\u3002\u3057\u304b\u3057\u5b9f\u969b\u306e\u30c7\u30fc\u30bf\u3067\u306f\u3001\u7f60\u2463\u306e\u5171\u4ecb\u5165\u3092\u5426\u5b9a\u3067\u304d\u306a\u3044\u9650\u308a\u3001\u300c#7119\u306b\u3088\u3063\u3066\u6e1b\u3063\u305f\u300d\u3068\u65ad\u5b9a\u3059\u308b\u306e\u306f\u8a00\u3044\u904e\u304e\u3067\u3042\u308b\u3002\u8ad6\u6587\u3067\u306f\u300c#7119\u306e\u5c0e\u5165 <strong>\u3068\u95a2\u9023\u3057\u3066<\/strong> \u6e1b\u5c11\u3057\u305f\uff08was associated with\uff09\u300d\u3068\u66f8\u304d\u3001\u5171\u4ecb\u5165\u306e\u53ef\u80fd\u6027\u3092\u9650\u754c\uff08Limitations\uff09\u3067\u8b70\u8ad6\u3059\u308b\u306e\u304c\u8aa0\u5b9f\u3067\u3042\u308b\u3002\u5bfe\u7167\u7cfb\u5217\u3092\u7528\u610f\u3067\u304d\u308b\u306a\u3089\u3001<a href=\"https:\/\/best-biostatistics.com\/toukei-er\/entry\/what-is-comparative-interrupted-time-series\/\">CITS<\/a>\u3078\u9032\u3080\u3053\u3068\u3067\u3001\u3053\u306e\u5f31\u70b9\u3092\u88dc\u5f37\u3067\u304d\u308b\u3002<\/p>\n<\/blockquote>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h2 class=\"wp-block-heading\">6. \u8ad6\u6587\u3067\u305d\u306e\u307e\u307e\u4f7f\u3048\u308b\u82f1\u6587\u30c6\u30f3\u30d7\u30ec\u30fc\u30c8<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">\u4ee5\u4e0b\u306e\u6570\u5024\u306f\u3001\u672c\u8a18\u4e8b\u306e\u7591\u4f3c\u30c7\u30fc\u30bf\u3092\u5b9f\u969b\u306b\u89e3\u6790\u3057\u305f\u7d50\u679c\u3067\u3042\u308b\u3002\u81ea\u5206\u306e\u30c7\u30fc\u30bf\u306b\u4f7f\u3046\u3068\u304d\u306f\u3001\uff3b \uff3d\u5185\u3092\u5fc5\u305a\u81ea\u5206\u306e\u89e3\u6790\u7d50\u679c\u306b\u7f6e\u304d\u63db\u3048\u3066\u307b\u3057\u3044\u3002<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">Methods<\/h3>\n\n\n\n<blockquote class=\"wp-block-quote is-layout-flow wp-block-quote-is-layout-flow\">\n<p class=\"wp-block-paragraph\">We conducted an interrupted time series analysis to evaluate changes in [monthly ambulance transports] associated with the introduction of [the emergency telephone consultation service (#7119)] in [January 2022]. Data from [January 2018] to [December 2025] ([48] months before and [48] months after the introduction) were analysed using segmented regression (Wagner et al., 2002), which included terms for the pre-intervention trend, the change in level immediately after the intervention, and the change in slope after the intervention. The expected form of the effect (a change in both level and slope) was specified a priori. Seasonality was adjusted for by including calendar month as a categorical variable. Stationarity of the pre-intervention series was assessed using the augmented Dickey\u2013Fuller and Kwiatkowski\u2013Phillips\u2013Schmidt\u2013Shin tests. Autocorrelation was assessed using the autocorrelation function of the residuals, and the final model was fitted by generalized least squares with a first-order autoregressive error structure, estimated by restricted maximum likelihood (Turner et al., 2021). Absolute and relative effects were estimated by comparing the fitted values with the counterfactual values predicted from the pre-intervention trend. All analyses were performed using R version [4.6.1] with the nlme package.<\/p>\n<\/blockquote>\n\n\n\n<p class=\"wp-block-paragraph\">\uff08\u548c\u8a33\u306e\u8981\u70b9\uff1aITS\u30c7\u30b6\u30a4\u30f3\u3067\u3042\u308b\u3053\u3068\u3001\u671f\u9593\u3068\u4ecb\u5165\u6642\u70b9\u3001\u30bb\u30b0\u30e1\u30f3\u30c8\u56de\u5e30\u306e3\u3064\u306e\u9805\u3001\u52b9\u679c\u306e\u5f62\u3092\u4e8b\u524d\u306b\u6c7a\u3081\u305f\u3053\u3068\u3001\u6708\u30c0\u30df\u30fc\u306b\u3088\u308b\u5b63\u7bc0\u6027\u306e\u8abf\u6574\u3001\u4ecb\u5165\u524d\u306e\u5b9a\u5e38\u6027\u306e\u78ba\u8a8d\u3001\u6b8b\u5dee\u306e\u81ea\u5df1\u76f8\u95a2\u306e\u78ba\u8a8d\u3068GLS\uff0bAR(1)\uff08REML\uff09\u3001\u53cd\u4e8b\u5b9f\u3068\u306e\u6bd4\u8f03\u3067\u52b9\u679c\u3092\u8868\u3057\u305f\u3053\u3068\uff09<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">Results<\/h3>\n\n\n\n<blockquote class=\"wp-block-quote is-layout-flow wp-block-quote-is-layout-flow\">\n<p class=\"wp-block-paragraph\">Before the introduction of [#7119], [monthly ambulance transports] increased by [2.7] transports per month (95% CI, [1.9] to [3.5]). The residuals of the segmented regression showed positive first-order autocorrelation (lag-1 autocorrelation, [0.32]). After the introduction, there was an immediate decrease in level of [56.3] transports per month (95% CI, [25.8] to [86.9]) and a decrease in slope of [1.77] transports per month (95% CI, [0.64] to [2.90]), resulting in a post-intervention trend of [0.91] transports per month (95% CI, [0.13] to [1.70]). At [24] months after the introduction, the number of transports was estimated to be [97.1] per month (95% CI, [57.2] to [136.9]) lower than expected in the absence of the intervention, corresponding to a relative reduction of [7.7]%.<\/p>\n<\/blockquote>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>\u67fb\u8aad\u8005\u306b\u300c\u306a\u305c\u524d\u5f8c\u6bd4\u8f03\u3067\u306f\u306a\u304fITS\u306a\u306e\u304b\u300d\u3068\u805e\u304b\u308c\u305f\u3068\u304d\u306e\u4e00\u6587<\/strong><\/p>\n\n\n\n<blockquote class=\"wp-block-quote is-layout-flow wp-block-quote-is-layout-flow\">\n<p class=\"wp-block-paragraph\">A simple before\u2013after comparison would have been biased by the underlying increasing trend in ambulance transports; the interrupted time series design accounts for this pre-existing trend by using it to estimate the counterfactual.<\/p>\n<\/blockquote>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>\u9650\u754c\uff08Limitations\uff09\u306e\u8a18\u8ff0\u4f8b<\/strong><\/p>\n\n\n\n<blockquote class=\"wp-block-quote is-layout-flow wp-block-quote-is-layout-flow\">\n<p class=\"wp-block-paragraph\">As with all single-series interrupted time series studies, we cannot exclude the possibility that other events occurring around the time of the intervention contributed to the observed changes.<\/p>\n<\/blockquote>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h2 class=\"wp-block-heading\">\u307e\u3068\u3081<\/h2>\n\n\n\n<ul class=\"wp-block-list\">\n<li>\u6642\u7cfb\u5217\u89e3\u6790\u3067\u8ff7\u3063\u305f\u3089\u3001<strong>\u624b\u6cd5\u306e\u540d\u524d\u304b\u3089\u3067\u306f\u306a\u304f\u300c\u4f55\u3092\u77e5\u308a\u305f\u3044\u304b\u300d\u304b\u3089\u9078\u3076<\/strong>\u3002\n<ul class=\"wp-block-list\">\n<li>\u4e00\u8cab\u3057\u3066\u5897\u3048\u3066\u3044\u308b\u304b \u2192 <a href=\"https:\/\/best-biostatistics.com\/toukei-er\/entry\/how-to-test-whether-a-time-series-is-increasing-in-r-mann-kendall-test-and-sens-slope\/\">Mann-Kendall\u691c\u5b9a\uff0bSen&#8217;s slope<\/a><\/li>\n\n\n\n<li>\u7cfb\u5217\u306e\u6027\u8cea\u3092\u77e5\u308b\uff08\u3069\u306e\u9053\u3067\u3082\u5f79\u7acb\u3064\u5065\u5eb7\u8a3a\u65ad\uff09 \u2192 <a href=\"https:\/\/best-biostatistics.com\/toukei-er\/entry\/how-to-assess-stationarity-of-time-series-in-r-combining-the-adf-and-kpss-tests\/\">ADF\u691c\u5b9a\u00d7KPSS\u691c\u5b9a<\/a><\/li>\n\n\n\n<li>\u3042\u308b\u6642\u70b9\u306e\u51fa\u6765\u4e8b\u3067\u5909\u308f\u3063\u305f\u304b \u2192 <strong>ITS\uff08\u672c\u8a18\u4e8b\uff09<\/strong><\/li>\n\n\n\n<li>\u540c\u6642\u671f\u306e\u5225\u306e\u51fa\u6765\u4e8b\u3068\u5207\u308a\u5206\u3051\u305f\u3044 \u2192 <a href=\"https:\/\/best-biostatistics.com\/toukei-er\/entry\/what-is-comparative-interrupted-time-series\/\">CITS<\/a><\/li>\n\n\n\n<li>\u5c06\u6765\u3092\u4e88\u6e2c\u3057\u305f\u3044 \u2192 <a href=\"https:\/\/best-biostatistics.com\/toukei-er\/entry\/basics-of-time-series-analysis\/\">ARIMA\u306a\u3069<\/a><\/li>\n<\/ul>\n<\/li>\n\n\n\n<li>ITS\u306f\u3001\u4ecb\u5165\u524d\u306e\u30c8\u30ec\u30f3\u30c9\u306e\u5ef6\u9577\u7dda\uff08<strong>\u53cd\u4e8b\u5b9f<\/strong>\uff09\u3068\u5b9f\u969b\u306e\u30c7\u30fc\u30bf\u3092\u6bd4\u3079\u308b\u624b\u6cd5\u3067\u3042\u308b\u3002\u30bb\u30b0\u30e1\u30f3\u30c8\u56de\u5e30\u3067\u3001\u52b9\u679c\u3092 <strong>\u6c34\u6e96\u306e\u5909\u5316\uff08\u6bb5\u5dee\uff09<\/strong> \u3068 <strong>\u50be\u304d\u306e\u5909\u5316<\/strong> \u306b\u5206\u3051\u3066\u63a8\u5b9a\u3059\u308b\u3002\u03b23\u306f\u50be\u304d\u306e\u300c\u5909\u5316\u300d\u3067\u3042\u308a\u3001\u4ecb\u5165\u5f8c\u306e\u50be\u304d\u306f \u03b21 + \u03b23 \u3067\u8aad\u3080\u3002<\/li>\n\n\n\n<li>\u4ecb\u5165\u524d\u5f8c\u306e\u5e73\u5747\u3092\u5358\u7d14\u306b\u6bd4\u3079\u308b\u3068\u3001\u3082\u3068\u3082\u3068\u306e\u30c8\u30ec\u30f3\u30c9\u306e\u305b\u3044\u3067 <strong>\u52b9\u679c\u306e\u5411\u304d\u3055\u3048\u9006\u8ee2\u3059\u308b<\/strong> \u3053\u3068\u304c\u3042\u308b\uff08Step 3\uff09\u3002<\/li>\n\n\n\n<li>\u5b63\u7bc0\u6027\u306f\u6708\u30c0\u30df\u30fc\u306a\u3069\u3067\u8abf\u6574\u3059\u308b\u3002\u81ea\u5df1\u76f8\u95a2\u306f\u30b3\u30ec\u30ed\u30b0\u30e9\u30e0\u3067\u78ba\u8a8d\u3057\u3001<strong>GLS\uff0bAR(1)\uff08REML\uff09<\/strong> \u306a\u3069\u3067\u8003\u616e\u3059\u308b\u3002Durbin-Watson\u691c\u5b9a\u3060\u3051\u306b\u983c\u3089\u306a\u3044\u3002<\/li>\n\n\n\n<li>\u5b9a\u5e38\u6027\u306f <strong>\u4ecb\u5165\u524d\u306e\u671f\u9593\u3060\u3051\u3067<\/strong> \u78ba\u8a8d\u3059\u308b\u3002<\/li>\n\n\n\n<li>\u7d50\u679c\u306f <strong>\u300c\u4f55\u4ef6\uff08\u4f55%\uff09\u6e1b\u3063\u305f\u304b\u300d\u3092\u4fe1\u983c\u533a\u9593\u3068\u3068\u3082\u306b<\/strong> \u8a9e\u308b\u3002\u5171\u4ecb\u5165\u3092\u5426\u5b9a\u3067\u304d\u306a\u3051\u308c\u3070\u3001CITS\u3067\u88dc\u5f37\u3059\u308b\u3002<\/li>\n<\/ul>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h2 class=\"wp-block-heading\">\u4e00\u62ec\u30b3\u30d4\u30da\u7528R\u30b9\u30af\u30ea\u30d7\u30c8<\/h2>\n\n\n\n<pre class=\"wp-block-code\"><code># ============================================================\n# \u5206\u5272\u6642\u7cfb\u5217\uff08ITS\uff09\u89e3\u6790\uff1a#7119\u5c0e\u5165\u524d\u5f8c\u306e\u6551\u6025\u642c\u9001\u4ef6\u6570\uff08\u7591\u4f3c\u30c7\u30fc\u30bf\uff09\n# ============================================================\n\n# --- \u30d1\u30c3\u30b1\u30fc\u30b8 ---\n# install.packages(c(\"nlme\", \"lmtest\", \"sandwich\", \"tseries\"))\nlibrary(nlme)\nlibrary(lmtest)\nlibrary(sandwich)\nlibrary(tseries)\n\n# --- Step 1\uff1a\u7591\u4f3c\u30c7\u30fc\u30bf\u306e\u4f5c\u6210\uff082018\u5e741\u6708\u301c2025\u5e7412\u6708\u300196\u304b\u6708\uff09---\nset.seed(7119)\n\nn     &lt;- 96\ntime  &lt;- 1:n                              # \u6642\u9593\uff081\u301c96\u304b\u6708\u76ee\uff09\nmonth &lt;- rep(1:12, times = 8)             # \u6708\uff081\u301c12\uff09\nyear  &lt;- rep(2018:2025, each = 12)        # \u5e74\n\nlevel      &lt;- as.numeric(time &gt;= 49)      # \u4ecb\u5165\uff082022\u5e741\u6708\uff1d49\u304b\u6708\u76ee\u4ee5\u964d\u304c1\uff09\ntime_after &lt;- pmax(0, time - 49)          # \u4ecb\u5165\u5f8c\u306e\u7d4c\u904e\u6708\u6570\uff082022\u5e741\u6708\u304c0\uff09\n\nseason &lt;- 60 * cos(2 * pi * (month - 1) \/ 12)                            # 1\u6708\u30d4\u30fc\u30af\u306e\u5b63\u7bc0\u306e\u6ce2\nnoise  &lt;- as.numeric(arima.sim(model = list(ar = 0.5), n = n, sd = 25))  # \u81ea\u5df1\u76f8\u95a2\u306e\u3042\u308b\u8aa4\u5dee\n\ncalls &lt;- round(1000 + 3 * time - 60 * level - 2 * time_after + season + noise)\n\nd &lt;- data.frame(year, month, time, level, time_after, calls)\ncalls_ts &lt;- ts(calls, start = c(2018, 1), frequency = 12)\n\n# --- Step 2\uff1a\u4ecb\u5165\u7dda\u3092\u5165\u308c\u305f\u6642\u7cfb\u5217\u30b0\u30e9\u30d5 ---\nplot(calls_ts, type = \"o\", pch = 16, cex = 0.6,\n     xlab = \"Year\", ylab = \"Ambulance transports (per month)\",\n     main = \"Monthly ambulance transports\")\nabline(v = 2022, col = \"red\", lty = 2)\n\n# --- Step 3\uff1a\u3010\u30c0\u30e1\u306a\u4f8b\u3011\u4ecb\u5165\u524d\u5f8c\u306e\u5e73\u5747\u3092\u6bd4\u3079\u308b ---\nt.test(calls ~ level, data = d)\n\n# --- Step 4\uff1a\u4ecb\u5165\u524d\u306e\u671f\u9593\u3060\u3051\u3067\u5b9a\u5e38\u6027\u3092\u78ba\u8a8d\uff08ADF\u00d7KPSS\uff09---\npre &lt;- subset(d, level == 0)\nadf.test(pre$calls)\nkpss.test(pre$calls, null = \"Trend\")\n\n# --- Step 5\uff1a\u30bb\u30b0\u30e1\u30f3\u30c8\u56de\u5e30\uff08OLS\u3001\u6708\u30c0\u30df\u30fc\u3067\u5b63\u7bc0\u6027\u3092\u8abf\u6574\uff09---\nm_ols &lt;- lm(calls ~ time + level + time_after + factor(month), data = d)\nround(coef(summary(m_ols))&#91;2:4, ], 4)\n\n# --- Step 6\uff1a\u6b8b\u5dee\u306e\u81ea\u5df1\u76f8\u95a2\u3092\u78ba\u8a8d ---\nacf(resid(m_ols), lag.max = 24,\n    main = \"Correlogram of residuals (segmented regression, OLS)\")\nround(acf(resid(m_ols), lag.max = 3, plot = FALSE)$acf&#91;2:4], 2)\ndwtest(m_ols)\n\n# --- Step 7\uff1a\u81ea\u5df1\u76f8\u95a2\u3092\u8003\u616e\u3057\u305f\u30bb\u30b0\u30e1\u30f3\u30c8\u56de\u5e30\uff08GLS\u3001AR(1)\u3001REML\uff09---\nm_gls &lt;- gls(calls ~ time + level + time_after + factor(month), data = d,\n             correlation = corAR1(form = ~ time), method = \"REML\")\nround(summary(m_gls)$tTable&#91;2:4, ], 4)\nround(intervals(m_gls, which = \"coef\")$coef&#91;2:4, ], 2)\ncoef(m_gls$modelStruct$corStruct, unconstrained = FALSE)   # AR(1)\u4fc2\u6570\n\n# \u4ecb\u5165\u5f8c\u306e\u50be\u304d\uff08\u03b21 + \u03b23\uff09\u306895%\u4fe1\u983c\u533a\u9593\nb_gls  &lt;- coef(m_gls)\nw_post &lt;- setNames(rep(0, length(b_gls)), names(b_gls))\nw_post&#91;c(\"time\", \"time_after\")] &lt;- 1\npost_slope &lt;- sum(w_post * b_gls)\npost_se    &lt;- sqrt(drop(t(w_post) %*% vcov(m_gls) %*% w_post))\nround(c(slope = post_slope,\n        lower = post_slope - 1.96 * post_se,\n        upper = post_slope + 1.96 * post_se), 2)\n\n# \u53c2\u8003\uff1aNewey-West\u6a19\u6e96\u8aa4\u5dee\nround(coeftest(m_ols, vcov = NeweyWest(m_ols, lag = 3, prewhite = FALSE))&#91;2:4, ], 4)\n\n# --- Step 8\uff1a\u53cd\u4e8b\u5b9f\u3068\u6bd4\u3079\u3066\u52b9\u679c\u3092\u8a9e\u308b ---\ncf &lt;- d\ncf$level      &lt;- 0\ncf$time_after &lt;- 0\nd$fit &lt;- predict(m_gls)                  # \u30e2\u30c7\u30eb\u306e\u5f53\u3066\u306f\u3081\u5024\nd$cf  &lt;- predict(m_gls, newdata = cf)    # #7119\u304c\u306a\u304b\u3063\u305f\u5834\u5408\u306e\u4e88\u6e2c\uff08\u53cd\u4e8b\u5b9f\uff09\n\nplot(d$time, d$calls, pch = 16, cex = 0.6, col = \"grey40\", xaxt = \"n\",\n     xlab = \"Year\", ylab = \"Ambulance transports (per month)\",\n     main = \"Observed, fitted, and counterfactual\",\n     ylim = c(900, max(c(d$calls, d$cf))))\naxis(1, at = seq(1, 96, by = 12), labels = 2018:2025)\nlines(d$time, d$fit, col = \"blue\", lwd = 2)\nlines(d$time&#91;d$level == 1], d$cf&#91;d$level == 1], col = \"red\", lwd = 2, lty = 2)\nabline(v = 48.5, lty = 3)\nlegend(\"bottomright\",\n       c(\"Observed\", \"Fitted (GLS)\", \"Counterfactual (no #7119)\"),\n       col = c(\"grey40\", \"blue\", \"red\"), pch = c(16, NA, NA),\n       lty = c(NA, 1, 2), lwd = c(NA, 2, 2), bty = \"n\", cex = 0.85)\n\n# \u4ecb\u5165\u304b\u3089k\u304b\u6708\u5f8c\u306e\u52b9\u679c\uff08\u03b22 + \u03b23 \u00d7 k\uff09\u306895%\u4fe1\u983c\u533a\u9593\u3001\u53cd\u4e8b\u5b9f\u306b\u5bfe\u3059\u308b\u76f8\u5bfe\u5909\u5316\neffect_at &lt;- function(model, data, k) {\n  b &lt;- coef(model)\n  V &lt;- vcov(model)\n  w &lt;- setNames(rep(0, length(b)), names(b))\n  w&#91;\"level\"] &lt;- 1\n  w&#91;\"time_after\"] &lt;- k\n  est &lt;- sum(w * b)\n  se  &lt;- sqrt(drop(t(w) %*% V %*% w))\n  row &lt;- which(data$level == 1 &amp; data$time_after == k)\n  data.frame(month    = sprintf(\"%d-%02d\", data$year&#91;row], data$month&#91;row]),\n             effect   = round(est, 1),\n             lower    = round(est - 1.96 * se, 1),\n             upper    = round(est + 1.96 * se, 1),\n             relative = paste0(round(100 * est \/ data$cf&#91;row], 1), \"%\"))\n}\ndo.call(rbind, lapply(c(0, 12, 23), function(k) effect_at(m_gls, d, k)))<\/code><\/pre>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h2 class=\"wp-block-heading\">\u95a2\u9023\u8a18\u4e8b<\/h2>\n\n\n\n<ul class=\"wp-block-list\">\n<li><a href=\"https:\/\/best-biostatistics.com\/toukei-er\/entry\/how-to-test-whether-a-time-series-is-increasing-in-r-mann-kendall-test-and-sens-slope\/\">\u6642\u7cfb\u5217\u30c7\u30fc\u30bf\u306e\u300c\u5897\u3048\u3066\u3044\u308b\u300d\u3092R\u3067\u5224\u5b9a\u3059\u308b\u65b9\u6cd5\uff5cMann-Kendall\u691c\u5b9a\u3068Sen&#8217;s slope<\/a><\/li>\n\n\n\n<li><a href=\"https:\/\/best-biostatistics.com\/toukei-er\/entry\/how-to-assess-stationarity-of-time-series-in-r-combining-the-adf-and-kpss-tests\/\">\u6642\u7cfb\u5217\u30c7\u30fc\u30bf\u306e\u5b9a\u5e38\u6027\u3092R\u3067\u5224\u5b9a\u3059\u308b\u65b9\u6cd5\uff5cADF\u691c\u5b9a\u3068KPSS\u691c\u5b9a\u306e\u7d44\u307f\u5408\u308f\u305b\u65b9<\/a><\/li>\n\n\n\n<li><a href=\"https:\/\/best-biostatistics.com\/toukei-er\/entry\/what-is-comparative-interrupted-time-series\/\">Comparative Interrupted Time Series \u3068\u306f\u4f55\u304b \u3069\u3093\u306a\u6642\u306b\u4f7f\u3046\u306e\u304c\u826f\u3044\u304b \u308f\u304b\u308a\u3084\u3059\u304f\u89e3\u8aac<\/a><\/li>\n\n\n\n<li><a href=\"https:\/\/best-biostatistics.com\/toukei-er\/entry\/basics-of-time-series-analysis\/\">R\u3067\u6642\u7cfb\u5217\u30c7\u30fc\u30bf\u5206\u6790\u3092\u884c\u3046\u65b9\u6cd5<\/a><\/li>\n<\/ul>\n\n\n\n<h2 class=\"wp-block-heading\">\u53c2\u8003\u6587\u732e<\/h2>\n\n\n\n<ol class=\"wp-block-list\">\n<li>Wagner AK, Soumerai SB, Zhang F, Ross-Degnan D. Segmented regression analysis of interrupted time series studies in medication use research. <em>Journal of Clinical Pharmacy and Therapeutics<\/em>. 2002;27(4):299\u2013309. <a href=\"https:\/\/doi.org\/10.1046\/j.1365-2710.2002.00430.x\">https:\/\/doi.org\/10.1046\/j.1365-2710.2002.00430.x<\/a><\/li>\n\n\n\n<li>Lopez Bernal J, Cummins S, Gasparrini A. Interrupted time series regression for the evaluation of public health interventions: a tutorial. <em>International Journal of Epidemiology<\/em>. 2017;46(1):348\u2013355. <a href=\"https:\/\/doi.org\/10.1093\/ije\/dyw098\">https:\/\/doi.org\/10.1093\/ije\/dyw098<\/a> \uff08Corrigendum: <em>International Journal of Epidemiology<\/em>. 2020;49(4):1414. <a href=\"https:\/\/doi.org\/10.1093\/ije\/dyaa118\">https:\/\/doi.org\/10.1093\/ije\/dyaa118<\/a> \uff0f\u30c7\u30fc\u30bf\u3068R\u30b3\u30fc\u30c9\uff1a<a href=\"https:\/\/github.com\/gasparrini\/2017_lopezbernal_IJE_codedata\">https:\/\/github.com\/gasparrini\/2017_lopezbernal_IJE_codedata<\/a> \uff09<\/li>\n\n\n\n<li>Lopez Bernal J, Cummins S, Gasparrini A. The use of controls in interrupted time series studies of public health interventions. <em>International Journal of Epidemiology<\/em>. 2018;47(6):2082\u20132093. <a href=\"https:\/\/doi.org\/10.1093\/ije\/dyy135\">https:\/\/doi.org\/10.1093\/ije\/dyy135<\/a><\/li>\n\n\n\n<li>Kontopantelis E, Doran T, Springate DA, Buchan I, Reeves D. Regression based quasi-experimental approach when randomisation is not an option: interrupted time series analysis. <em>BMJ<\/em>. 2015;350:h2750. <a href=\"https:\/\/doi.org\/10.1136\/bmj.h2750\">https:\/\/doi.org\/10.1136\/bmj.h2750<\/a><\/li>\n\n\n\n<li>Turner SL, Forbes AB, Karahalios A, Taljaard M, McKenzie JE. Evaluation of statistical methods used in the analysis of interrupted time series studies: a simulation study. <em>BMC Medical Research Methodology<\/em>. 2021;21:181. <a href=\"https:\/\/doi.org\/10.1186\/s12874-021-01364-0\">https:\/\/doi.org\/10.1186\/s12874-021-01364-0<\/a><\/li>\n\n\n\n<li>Barone-Adesi F, Gasparrini A, Vizzini L, Merletti F, Richiardi L. Effects of Italian smoking regulation on rates of hospital admission for acute coronary events: a country-wide study. <em>PLoS One<\/em>. 2011;6(3):e17419. <a href=\"https:\/\/doi.org\/10.1371\/journal.pone.0017419\">https:\/\/doi.org\/10.1371\/journal.pone.0017419<\/a><\/li>\n\n\n\n<li>Schaffer AL, Dobbins TA, Pearson SA. Interrupted time series analysis using autoregressive integrated moving average (ARIMA) models: a guide for evaluating large-scale health interventions. <em>BMC Medical Research Methodology<\/em>. 2021;21:58. <a href=\"https:\/\/doi.org\/10.1186\/s12874-021-01235-8\">https:\/\/doi.org\/10.1186\/s12874-021-01235-8<\/a><\/li>\n\n\n\n<li>Perron P. The great crash, the oil price shock, and the unit root hypothesis. <em>Econometrica<\/em>. 1989;57(6):1361\u20131401. <a href=\"https:\/\/www.jstor.org\/stable\/1913712\">https:\/\/www.jstor.org\/stable\/1913712<\/a><\/li>\n<\/ol>\n\n\n\n<h2 class=\"wp-block-heading\">\u304a\u3059\u3059\u3081\u66f8\u7c4d<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">\u6642\u7cfb\u5217\u89e3\u6790\u306e\u571f\u53f0\u3092\u56fa\u3081\u305f\u3044\u4eba\u5411\u3051\u306b\u3001R\u3067\u5b66\u3079\u308b\u5165\u9580\u66f8\u3068\u3001\u7121\u6599\u3067\u8aad\u3081\u308b\u5b9f\u8df5\u7684\u306a\u6559\u79d1\u66f8\u3092\u7d39\u4ecb\u3059\u308b\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>1. \u99ac\u5834\u771f\u54c9\u300e\u6642\u7cfb\u5217\u5206\u6790\u3068\u72b6\u614b\u7a7a\u9593\u30e2\u30c7\u30eb\u306e\u57fa\u790e\uff1aR\u3068Stan\u3067\u5b66\u3076\u7406\u8ad6\u3068\u5b9f\u88c5\u300f\uff08\u30d7\u30ec\u30a2\u30c7\u30b9\u51fa\u7248\uff09<\/strong><br>\u81ea\u5df1\u76f8\u95a2\u3001\u5b9a\u5e38\u6027\u3001ARIMA\u3068\u3044\u3063\u305f\u6642\u7cfb\u5217\u5206\u6790\u306e\u57fa\u672c\u3092\u3001R\u306e\u30b3\u30fc\u30c9\u3068\u3068\u3082\u306b\u4e00\u304b\u3089\u5b66\u3079\u308b\u3002\u672c\u8a18\u4e8b\u3067\u4f7f\u3063\u305f\u300c\u30b3\u30ec\u30ed\u30b0\u30e9\u30e0\u300d\u300cAR(1)\u300d\u3092\u3001\u3082\u3046\u4e00\u6b69\u6df1\u304f\u7406\u89e3\u3057\u305f\u3044\u4eba\u306e\u6700\u521d\u306e1\u518a\u306b\u5411\u3044\u3066\u3044\u308b\u3002<br>&#x1f449; <a href=\"https:\/\/amzn.to\/4yY0tXr\">https:\/\/www.amazon.co.jp\/dp\/4903814874<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>2. Hyndman RJ, Athanasopoulos G. <em>Forecasting: Principles and Practice<\/em>, 3rd ed.\uff08\u7121\u6599\u306eWeb\u7248\uff09<\/strong><br>\u5b63\u7bc0\u6027\u306e\u6271\u3044\u3001\u56de\u5e30\u30e2\u30c7\u30eb\u306e\u6b8b\u5dee\u306e\u81ea\u5df1\u76f8\u95a2\u3001ARIMA\u3092\u3001\u5b9f\u30c7\u30fc\u30bf\u3067\u5b9f\u8df5\u7684\u306b\u5b66\u3079\u308b\u6559\u79d1\u66f8\u3067\u3042\u308b\u3002Web\u4e0a\u3067\u5168\u6587\u3092\u7121\u6599\u3067\u8aad\u3081\u308b\u3002ARIMA\u3092\u4f7f\u3063\u305fITS\uff08Schaffer et al., 2021\uff09\u306b\u9032\u307f\u305f\u3044\u4eba\u306e\u8db3\u304c\u304b\u308a\u306b\u3082\u306a\u308b\u3002<br>&#x1f449; <a href=\"https:\/\/otexts.com\/fpp3\/\">https:\/\/otexts.com\/fpp3\/<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><\/p>\n","protected":false},"excerpt":{"rendered":"<p>\u5f8c\u8f29\uff1a\u300c\u5148\u8f29\u3001\u5e02\u306e\u6d88\u9632\u5c40\u304b\u3089\u6708\u3054\u3068\u306e\u6551\u6025\u642c\u9001\u4ef6\u6570\u306e\u30c7\u30fc\u30bf\u3092\u3082\u3089\u3044\u307e\u3057\u305f\u30022018\u5e741\u6708\u304b\u30892025\u5e7412\u6708\u307e\u3067\u306e96\u304b\u6708\u5206\u3067\u3059\u300d \u5148\u8f29\uff1a\u300c\u3044\u3044\u30c7\u30fc\u30bf\u3060\u306d\u3002\u4f55\u3092\u8abf\u3079\u305f\u3044\u3093\u3060\u3044\uff1f\u300d \u5f8c\u8f29\uff1a\u300c\u3053\u306e\u5e02\u3067\u306f2022\u5e741\u6708\u306b \u6551\u6025\u5b89 [&hellip;]<\/p>\n","protected":false},"author":2,"featured_media":4870,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"swell_btn_cv_data":"","_jetpack_newsletter_access":"","_jetpack_dont_email_post_to_subs":false,"_jetpack_newsletter_tier_id":0,"_jetpack_memberships_contains_paywalled_content":false,"_jetpack_feature_clip_id":0,"_jetpack_memberships_contains_paid_content":false,"footnotes":"","jetpack_publicize_message":"","jetpack_publicize_feature_enabled":true,"jetpack_social_post_already_shared":true,"jetpack_social_options":{"image_generator_settings":{"template":"highway","default_image_id":0,"font":"","enabled":false},"version":2},"jetpack_post_was_ever_published":false},"categories":[54],"tags":[],"class_list":["post-4854","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-54"],"jetpack_publicize_connections":[],"jetpack_sharing_enabled":true,"jetpack_featured_media_url":"https:\/\/best-biostatistics.com\/toukei-er\/wp-content\/uploads\/2026\/09\/ITS-scaled.jpeg","_links":{"self":[{"href":"https:\/\/best-biostatistics.com\/toukei-er\/wp-json\/wp\/v2\/posts\/4854","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/best-biostatistics.com\/toukei-er\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/best-biostatistics.com\/toukei-er\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/best-biostatistics.com\/toukei-er\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/best-biostatistics.com\/toukei-er\/wp-json\/wp\/v2\/comments?post=4854"}],"version-history":[{"count":13,"href":"https:\/\/best-biostatistics.com\/toukei-er\/wp-json\/wp\/v2\/posts\/4854\/revisions"}],"predecessor-version":[{"id":4872,"href":"https:\/\/best-biostatistics.com\/toukei-er\/wp-json\/wp\/v2\/posts\/4854\/revisions\/4872"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/best-biostatistics.com\/toukei-er\/wp-json\/wp\/v2\/media\/4870"}],"wp:attachment":[{"href":"https:\/\/best-biostatistics.com\/toukei-er\/wp-json\/wp\/v2\/media?parent=4854"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/best-biostatistics.com\/toukei-er\/wp-json\/wp\/v2\/categories?post=4854"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/best-biostatistics.com\/toukei-er\/wp-json\/wp\/v2\/tags?post=4854"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}