èªãã ã J Clin Epidemiol. 1995 Dec;48(12):1503-10. ãã©ã¡ã¼ã¿æ°ã ã®ãã¼ã¿ãæ¨æ¬æ° å観測ãã¦éå帰ãªããªãããã®å帰åæãè¡ãã®ã ãããæ¨æ¬æ°ã¯ããã¤ãã£ãããããã§ãã?ãã¨èããããã¨ãå¤ã ãããããããã¨ãããã©ã¡ã¼ã¿ã®10å(10)ãã£ããããããããªãã£ãããã(é©å½ ã¨çãããã¨ãå¤ãã®ã ãããã©ã¡ã¼ã¿æ°ã«å¯¾ããæ¨æ¬æ°ã®å²å(EPV, event per variables)ã10 ãªãããããããªãã£ããã(é©å½ ã¨è¨ã£ã¦ããè«æã è«æä¸ã§ã¯673人ã®æ£è ã«å¯¾ãã¦7ã¤ã®ãã©ã¡ã¼ã¿ã測å®ããã¦ãã¦ã252件ã®æ»äº¡ã«ã¤ãã¦Cox å帰ãè¡ããããã¼ã¿ã»ããããã¹ã¦ä½¿ãã°252/7 = 36 EPV ã®ã¨ãããã2, 5, 10, 15, 20 ã¨EPV ã®å²åãå¤ãã¦ãã¼ã¿ã»ããããªãµã³ããªã³ã°ãã¦Cox å帰ãç¹°ãè¿ãããã©ã¡ã¼
{{#tags}}- {{label}}
{{/tags}}