{[ promptMessage ]}

Bookmark it

{[ promptMessage ]}

matched data and agreement

# matched data and agreement - MATCHED PAIR DATA AND...

This preview shows pages 1–2. Sign up to view the full content.

MATCHED PAIR DATA AND AGREEMENT Sometimes the best way to control confounding is to match on the confounding factor - Age distribution unexposed is young - Age distribution of exposed is older - Among 0-5 years, we will have enough unexposed and 5-15 have exposed, So we don’t have same age to compare - So, can’t make comparable comparisons - Can be matched for case, or other exposures Example matched case control study Case control study of acute prolapsed lumbar intervertebral disk. Kelsey et al. (1975) For each case, a matched control of the same age and gender was obtained from the same hospital or radiology service 217 matched pairs One exposure of interest: whether subject was a driver Each pari is a separate stratum - Analysis adjusted for matched factor - Cases and control are patched pair - Here, m=no of pair - Each pair creates its own 2*2 table - So, H 0 : All OR is same in all table=1 - H A : OR in those 2*2 table ≠1 Summary - There are four possible table for each pair Stratum specific Odds Ratio In this case Type of combination No of tables Both case and control exposed 144 Case exposed, control not 41 Control exposed, case not 19 Both case and control unexposed 13 Total pair 217 M-H gives common OR by aggregating the data from all tables Another summary Counts are pairs, not individuals M-H OR is ratio of number of discordant pairs: 41/19 = 2.16 Without matching - 185 = 144+41 63 = 144+19 - 32 = 13+19 54 = 13+41 Stratum- specific MH-Test Contributions χ 2 MH

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
This is the end of the preview. Sign up to access the rest of the document.

{[ snackBarMessage ]}