Skip to main content

Advertisement

Table 5 Model fit statistics for the PSSA method in the numeric example

From: Combining population-based administrative health records and electronic medical records for disease surveillance

Model Overall Males Females 18–44 years 45–64 years 65+ years
1 167,249 73,418 93,565 42,925 71,311 26,719
2 166,994 73,405 93,493 42,921 70,983 26,554
3 166,506 73,181 93,351 42,421 71,033 26,622
4 165,719 72,706 92,688 42,220 70,483 26,525
  1. PSSA Probabilistic-based sensitivity-specificity adjusted method, PSSA, model 1 covariates are sex, age group, region, income quintile, Charlson comorbidity score, chronic obstructive pulmonary disease (A, E), diabetes (A, E), depression (A, E), dementia (A, E), obesity (A, E), cerebrovascular disease (A), congestive heart failure (A), coronary heart disease (A), renal disease (A), substance abuse (A); PSSA, model 2 covariates sex, age group, region, income quintile, chronic obstructive pulmonary disease (E), diabetes (E), depression (E), dementia (E), obesity (E), cerebrovascular disease (A), congestive heart failure (A), coronary heart disease (A), renal disease (A), substance abuse (A); PSSA, model 3 covariates are sex, age group, region, income quintile, chronic obstructive pulmonary disease (E), diabetes (E), depression (E), dementia (E), obesity (E), coronary heart disease (A), renal disease (A), substance abuse (A); PSSA, model 4 covariates are sex, age group, chronic obstructive pulmonary disease (E), diabetes (E), obesity (E), coronary heart disease (A), congestive heart failure (A), substance abuse (A); A and E denote disease-specific markers that were identified from AHRs and EMRs, respectively; Values in bold-face font represent the best-fitting model