The necessity of validity diagnostics when drawing causal inferences from observational data: lessons from a multi-database evaluation of the risk of non-infectious uveitis among patients exposed to Remicade®

Table 3 Confusion matrix contingency cell counts and misclassification errors for the primary non-infectious uveitis outcome definition across databases

Database	TP	TN	FP	FN	Sensitivity	Specificity	PPV	NPV
Amb EMR	496	1,412,594	241	465	0.516129	0.999829	0.672999	0.99967
Pharmetrics	2249	1,967,965	1244	6378	0.260577	0.999368	0.643573	0.996769
Optum^® EHR	1369	1,946,703	317	3271	0.295043	0.999838	0.811981	0.998323
Clinformatics^®	6007	1,658,007	1673	6725	0.471762	0.998992	0.782031	0.99596
CCAE	4044	1,912,446	949	6283	0.391595	0.999504	0.809934	0.996725

Key –Amb EMR: IQVIA Ambulatory Electronic Medical Records, Pharmetrics: IQVIA Adjudicated Health Plan Claims Data, Optum^® EHR: Optum^® De-Identified Electronic Health Record, Clinformatics^®: Optum^® De-Identified Clinformatics^® Data Mart Database, CCAE: Merative™ MarketScan^® Commercial Database, TP: true positives, TN: true negatives, FP: false positives, FN: false negatives, Sensitivity = TP/(TP + FN), Specificity = TN/(TN + FP), PPV = positive predictive value = TP/(TP + FP), NPV = negative predictive value = TN/(TN + FN), Primary outcome definition: [first occurrence of a NIU code with a second NIU code occurrence between 31 days and 365 days relative to first occurrence] OR [first occurrence of a NIU code during an ophthalmology visit]

ISSN: 1471-2288