EndoTODAY
| EndoATLAS
| OPD
Parasite
| Eso
| Sto
| Cancer
| ESD
Boxim
| DEX
| Sono
| Schedule
Home
| Recent
| Blog
| Links
EndoTODAY ³»½Ã°æ ±³½Ç
[ºÐ¼®Åë°èÀÇ ºÐ·ù]
ÀÇÇп¡¼ Åë°èºÐ¼®À̶õ (1) °¡¼³ÀÇ °ËÁ¤, (2) ¸ð¼öÀÇ ÃßÁ¤, (3) ¿¹ÈÄÀÎÀÚÀÇ ÆľÇÀ¸·Î ¿ä¾ÇÇÒ ¼ö ÀÖ´Ù. Æò±ÕÄ¡ÀÇ ºñ±³´Â µÎ °ª°£¿¡ Â÷ÀÌ°¡ ¾ø´Ù´Â ±Í¹«°¡¼³À» ¼³Á¤ÇÏ°í À̸¦ ºÐ¼®ÇÔÀ¸·Î½á ½ÇÁ¦·Î Åë°èÀûÀ¸·Î ÀǹÌÀÖ´Â Â÷ÀÌ°¡ ÀÖ´Â Áö¸¦ ¾Ë¾Æ³»´Â °ÍÀ̹ǷΠ°¡¼³ÀÇ °ËÁ¤¿¡ ÇØ´çÇÑ´Ù. »ýÁ¸·ü°è»êÀº ¸ð¾ÆÁø µ¥ÀÌŸ·ÎºÎÅÍ ¾ËÁö ¸øÇÏ´Â °ªÀ» Åë°èÀû ¹æ¹ýÀ¸·Î ÃßÃøÇÏ´Â °ÍÀ̹ǷΠ¸ð¼öÀÇ ÃßÁ¤¿¡ ÇØ´çÇÑ´Ù. Cox proportional hazards modelÀº ±³¶õº¯¼ö¸¦ ÅëÁ¦ÇÔÀ¸·Î½á ¾î¶°ÇÑ º¯¼ö°¡ ȯÀÚÀÇ »ýÁ¸·ü¿¡ Â÷À̸¦ °¡Á®¿À´ÂÁö ¾Ë¾Æº¸´Â °úÁ¤À̹ǷΠ¿¹ÈÄÀÎÀÚÀÇ Æľǿ¡ ÇØ´çÇÑ´Ù. ÀÌó·³ °¢°¢ÀÇ Åë°è¹æ¹ýÀÌ À§ ¼¼°¡ÁöÀÇ °úÁ¦ Áß ¾î¶² °ÍÀ» ¼öÇàÇϱâ À§ÇÑ ¹æ¹ýÀÎÁö¸¦ ¾Æ´Â °ÍÀÌ ÀüüÀÇ °³³äÀ» ÆľÇÇÏ´Â µ¥ µµ¿òÀÌ µÈ´Ù.
ºÐ¼®Åë°èÀÇ ¹æ¹ýµéÀº º¯¼öÀÇ Á¾·ù¿¡ µû¶ó º¹ÀâÇÏ°Ô ºÐ·ù°¡ µÇÁö¸¸ ±âº»ÀûÀ¸·Î Á¾¼Óº¯¼öÀÇ Á¾·ù, Áï Á¾¼Óº¯¼ö°¡ ¿¬¼Óº¯¼öÀÎÁö ºñ¿¬¼Óº¯¼öÀÎÁö¿¡ µû¶ó¼ Å©°Ô ³ª´©¾îÁø´Ù. Á¾¼Óº¯¼ö°¡ ¿¬¼Óº¯¼öÀÎ °æ¿ìÀÇ ´ëÇ¥ÀûÀÎ ¹æ¹ýÀº [t-test, ȸ±Í, »ó°ü]ÀÌ°í Á¾¼Óº¯¼ö°¡ ºñ¿¬¼Óº¯¼öÀÎ °æ¿ìÀÇ ´ëÇ¥ÀûÀÎ ¹æ¹ýÀº [chi-square, logistic model]ÀÌ´Ù.
¸ð¼öÀû ¹æ¹ý¿¡´Â ¸í¸íÇÑ »ç¶÷ÀÇ À̸§À» »ý·«Çϱ⵵ ÇÑ´Ù.
- Æò±ÕÄ¡ÀÇ ºÐ¼® (of continuous variable)
- »ó°üºÐ¼® (correlation of continuous variable): ¼±ÈÄ°ü°è°¡ ºÒ¸íÈ®ÇÑ º¯¼öÀÇ interdependencyÀÇ Á¤·®È
- ´Ü¼ø»ó°üºÐ¼®(bivariate correlation analysis)
- ¸ð¼öÀû°ËÁ¤¹ý : »ó°ü°è¼ö (Pearson's correlation coefficient)
- ºñ¸ð¼öÀû°ËÁ¤¹ý : Spearman »ó°ü°è¼ö
- Æí(partial)»ó°üºÐ¼® : ±³¶õº¯¼ö¸¦ ±³Á¤ÇÏ¿© ÁÖ´Â ¹æ¹ý
- ȸ±ÍºÐ¼® (regression of continuous variable): ¼±ÈÄ°ü°è ȤÀº Àΰú°ü°è°¡ ÀÖ´Â µ¶¸³º¯¼ö¿Í Á¾¼Óº¯¼öÀÇ dependencyÀÇ Á¤·®È
- ´Ü¼øȸ±ÍºÐ¼®(simple linear regression)
- ´ÙÁßȸ±ÍºÐ¼®(multiple regression analysis)
- ºñ¿¬¼Ó ¹üÁÖÇüº¯¼öÀÇ ºÐ¼® (±³Â÷ºÐ¼®)
- two-by-two table (2 X 2)
- Pearson's chi-square test : ¸ð¼öÀû¹æ¹ý
- chi-square test with Yate's correction
- Fisher's exact test : ºñ¸ð¼öÀû ¹æ¹ý, »ó±â µÎ ¹æ¹ýÀÇ °á°ú°¡ ÀÏÄ¡ÇÏÁö ¾ÊÀ» ¶§, cell³»ÀÇ ¼ýÀÚ°¡ 5º¸´Ù ÀÛÀ»¶§
- unadjusted likelihood ratio (¿ìµµºñ) test
- two-by-k table (2 X k) : dose-response¸¦ °üÂûÇϵµ·Ï ³ë·ÂÇؾß
- score test for trend : SPSS¿¡ ³»ÀåµÇÁö ¾ÊÀ½; linear by linear associationÀ» ´ë½Å »ç¿ë
- likelihood ratio test(LR test) for trend(¿ìµµºñ°æÇâºÐ¼®¹ý) : ·ÎÁþ¸ðµ¨±â¹ýÀÇ ÀÏÁ¾
- R(row)-by-C(column) table
- µ¶¸³¼º¿¡ ´ëÇÑ °ËÁ¤(µÎ º¯¼ö ¸ðµÎ nominal scaleÀÎ °æ¿ì) : chi-square test
- °æÇâºÐ¼® (ordinal scale º¯¼ö¿¡¼) : linear by linear associationÀ¸·Î ±Ù»çÀû Á¢±Ù
- ÀÏÄ¡µµ(degree of agreement) ÆÇÁ¤ : ÄÚÇîÀÇ kappa °ª
- confounding variableÀÇ ¿µÇâÀ» º¸Á¤ÇÏ´Â ¹æ¹ý
- stratified table(ÃþȺм®¹ý) : ¼Ò¼öÀÇ ±³¶õº¯¼ö¸¦ ÅëÁ¦
- multivariate linear logistic regression analysis(¼±Çü ·ÎÁþ¸ðµ¨À» ÀÌ¿ëÇÑ ´Ùº¯·® ȸ±ÍºÐ¼®¹ý) : ´Ù¼öÀÇ ±³¶õº¯¼ö¸¦ ÅëÁ¦ÇÏ°í °¢ º¯¼öÀÇ odds¸¦ ±¸ÇÒ ¼ö ÀÖ´Â °·ÂÇÑ ¹æ¹ý
- »ýÁ¸ºÐ¼®(survival analysis)
- ´©Àû»ýÁ¸À²ÀÇ »êÃâ : ¸ð¼ö¸¦ ÃßÁ¤ÇÏ´Â ´Ü°è
- »ý¸íÇ¥(actuarial)¹ý : 50 case ÀÌ»ó, ¿ø·¡ »ý¸íº¸Çèȸ»ç¿¡¼ »ç¿ëÇÏ´ø ¹æ¹ý
- Kaplan-Meier¹ý (product-limit¹ý) : 50 case ÀÌÇÏ
- µÎ »ýÁ¸°î¼±ÀÇ ºñ±³ : °¡¼³À» °ËÁ¤ÇÏ´Â ´Ü°è
- Mantel-Haenszel¹ý : »ý¸íÇ¥¹ýÀ¸·Î »êÃâÇÑ »ýÁ¸°î¼± ºñ±³
- log-rank¹ý : Kaplan Meier¹ýÀ¸·Î »êÃâÇÑ »ýÁ¸°î¼± ºñ±³, »ýÁ¸±â°£ÀÌ ±æ ¶§
- Gehan's generalized Wilcoxon¹ý : »ýÁ¸±â°£ÀÌ ÂªÀ» ¶§
- likelihood ratio test
- ±³¶õº¯¼öÀÇ ¿µÇâÀ» º¸Á¤ : ¿¹ÈÄÀÎÀÚ¸¦ È®ÀÎÇÏ´Â ´Ü°è
- Cox proportional hazards model
© ÀÏ¿ø³»½Ã°æ±³½Ç ¹Ù¸¥³»½Ã°æ¿¬±¸¼Ò ÀÌÁØÇà. EndoTODAY Endoscopy Learning Center. Lee Jun Haeng.