Skip to main content
. 2020 Sep 24;27(11):1675–1687. doi: 10.1093/jamia/ocaa104

Table 1.

Summary information

Review Categorya n (%)b Chart Review Summary
T2DM (n = 84 821; cases = 9301, controls = 23 776)
 Cases with probability <50% 675 (0.80) 7 cases: none had T2DM diagnosis codes, but they did have a history of diabetes in clinical notes with medication and elevated blood glucose measurements.
3 could not be determined: no mention of T2DM in clinical notes but they had medication and elevated blood glucose measurements.
 Controls with probability ≥50% 24 (0.03) 1 case: clinical notes indicate well-controlled T2DM.
9 non-cases: no evidence that suggests T2DM, but they had other features that inflated the phenotype score, like blood glucose measurements or hypoglycemia.
 Unclassified with probability <50% 39 611 (46.69) 10 non-cases: no evidence that suggests T2DM.
 Unclassified with probability ≥50% 12 133 (14.31) 6 cases: history of T2DM in clinical notes.
4 non-cases: 3 have no evidence that suggests T2DM; 1 has prediabetes.
Dementia (n = 84 821; cases = 2989, controls = 77 688)
 Cases with probability <50% 1339 (1.58) 4 cases: history of dementia in clinical notes with no dementia diagnosis codes but they were identified as cases by the presence of an associated medication.
6 non-cases or could not be determined: no evidence that suggests dementia; identified as cases by the presence of an associated medication, where the medication was prescribed for reasons unrelated to dementia
 Controls with probability ≥50% 0 (0.00) 0 cases
 Unclassified with probability <50% 2894 (3.41) 10 non-cases: 7 diagnosed with delirium, but not dementia.
3 with a persistent mental disorder, but not dementia
 Unclassified with probability ≥50% 1250 (1.47) 5 cases: history of dementia in clinical notes but excluded from a case group for not having >5 diagnosis codes.
5 non-cases: persistent mental disorder, but not dementia
Hypothyroidism (n = 84 821; cases = 5170, controls = 27 926)
 Cases with probability <50% 16 (0.02) 10 cases: history of hypothyroidism in clinical notes, but specific diagnosis code were not captured or assigned a lower score by PheMap
 Controls with probability ≥50% 133 (0.16) 5 non-cases: goiter, but no hypothyroidism
5 non-cases: coma (assigned a high score by PheMap), but unrelated to hypothyroidism
 Unclassified with probability <50% 45 499 (55.64) 10 non-cases: no evidence that suggests hypothyroidism
 Unclassified with probability ≥50% 11 492 (13.55) 3 cases: history of hypothyroidism in clinical notes
3 cases: postsurgical or secondary due to recent contrast exposure
4 non-cases: goiter or coma (see “Controls with probability ≥50%”)

Note: Data are of subjects with inconsistent PheMap phenotype probabilities, eMERGE-defined case-control status, and results of chart reviews of randomly selected 10 samples per group. eMERGE: Electronic Medical Records and Genomics; T2DM: type 2 diabetes mellitus; XPRESS: eXtraction of Phenotypes from Records using Silver Standards.

a

Case, control, and unclassified refer to a patient’s eMERGE-defined case-control status.

b

Reporting as percentage of total patients (N = 84 821)