Skip to main content
. 2016 Sep 11;24(e1):e121–e128. doi: 10.1093/jamia/ocw123

Table 1.

Components and algorithm for 7 EHR phenotypes for diabetes

Abbreviation ICD_250 CCW A1c Med DDC Sup-DM NW A1c_Med
Full name ICD-9 code 250.xx Expanded ICD-9 codes Abnormal A1c (≥ 6.5%) DM Med DDC SUPREME DM eMERGE NW Abnormal A1c + DM Meda
Based on ICD-9 billing codes Other DM codes HEDIS misses NYC A1c h ealth department registry PBM Population Health Type 2 DM Population Health Any DM Exclusively Type 2 DM A combination of abnormal A1c and PBM
Components
1a ICD-9 250.x0, 250.x2 IP or AMB x1 IP x1 or AMB x2 (–) (–) IP, AMB, or ED x 1 IP x1 or AMB x2 IP, AMB, or ED x1 (–)
1b ICD-9 250.x1, 250.x3 IP or AMB x1 IP x1 or AMB x2 (–) (–) (–) IP x1 or AMB x2 (–) (–)
1c ICD-9 249.xx (–) IP x1 or AMB x2 (–) (–) IP, AMB, or ED x1 (–) (–) (–)
1d ICD-9 other codes (–) IP x1 or AMB x2 (–) (–) IP, AMB, or ED x1 IP x1 or AMB x2 (–) (–)
2a Diabetes Med Group 1: insulin (–) (–) (–) 1 or more DM med on AMB med rec 1 + AMB med rec 1+ AMB med rec Excludes any patient on insulin or amylin 1 or more DM med on AMB med rec
2b Diabetes Med Group 2: insulin secretagogues and incretins (–) (–) (–) 1+ AMB med rec 1+ AMB med rec Must have type 2 code
2c Diabetes Med Group 3: thiazolidinediones and metformin (–) (–) (–) 1+ AMB med rec Excludes if this is the only criteria met
3 Abnormal glucose lab (–) (–) (–) 2 abnormal labs in past 365 d 2 abnormal labs in past 720 d (–)
3a A1c ≥ 6.5% once (–) (–) 1+ abnormal (–) 1+ abnormal
3b A1c ≥ 6.5 % twice (–) (–) 1+ abnormal (–) 1+ abnormal
3c Fasting glucose ≥ 126 × 2 (–) (–) (–) (–) (–)
3d Random glucose ≥ 200 × 2 (–) (–) (–) (–) (–)
3e Abnormal OGTT (–) (–) (–) (–) (–)
3f Two of above (–) (–) (–) (–) (–)
Simplified algorithm 1a or 1b 1a or 1b or 1c or 1d 3a 2a or 2b or 2c 1a or 1c or 1d or 2a or 2b or 2c or 3b or 3c or 3d or 3e or 3f 1a or 1b or 1d or 2a or 2b or 3a or 3b or 3c or 3d or 3e or 3f 1a but never 1b and never 2a unless controlled on oral agents 3a or 2
No. patients identified by phenotype 18 893 16 320 12 182 11 800 22 050 18 958 11 620 15 478
Extrapolated no. patients with type 2 DM knowing sensitivity/specificity that phenotype would identify 13 906 12 804 10 507 9481 14 414 13 422 10 073 12 480
Extrapolated no. patients with any DM knowing sensitivity/specificity that phenotype would identify 15 833 14 521 11 741 10 668 16 387 15 281 10 408 13 904
PPV type 2 DM 0.74 0.78 0.87 0.80 0.66 0.71 0.86 0.81
PPV any DM 0.84 0.89 0.97 0.91 0.75 0.81 0.89 0.90

aThe eighth phenotype, A1c_Med, was a combination of phenotypes 3 (A1c) and 4 (Med) and was devised and studied after the sampling strategy.

AMB, ambulatory; DDC, Durham Diabetes Coalition; DM, diabetes mellitus; ED, emergency department; eMERGE NW, Electronic Medical Records and Genomics Network, Northwestern University; HEDIS, Healthcare Effectiveness Data and Information Set; ICD-9, International Classification of Diseases, Ninth Revision; IP, inpatient; med, medication; NYC, New York City; PBM, pharmacy benefit manager; OGTT, oral glucose tolerance test; PPV, positive predictive value; rec, reconciliation; SUPREME DM, Surveillance, Prevention, and Management of Diabetes Mellitus Project; T2, type 2.