Overall positive predictive value (PPV) for acute myocardial infarction (AMI) identification, according to database
| Source | Coding system | Number of cases sampled | Number of cases retrieved | Number of cases confirmed (%) | Number of cases considered non-assessable | PPV, best case scenario* (95% CI) | PPV, worst case scenario† (95% CI) |
|---|---|---|---|---|---|---|---|
| General practitioner (GP)/specialist diagnoses (IPCI, the Netherlands) | ICPC | 200 | 200 | 93 (46.5) | 76 | 75.0 (67.4 to 82.6) | 46.5 (37.7 to 55.3) |
| Free text | 200 | 200 | 26 (13.0) | 68 | 19.7 (12.9 to 26.5) | 13.0 (7.3 to 18.7) | |
| GP/specialist diagnoses (HSD, Italy) | ICD9-CM | 187 | 187 | 112 (56.0) | 71 | 96.6 (93.2 to 99.9) | 59.9 (51.0 to 68.8) |
| free text | 13 | 13 | 3 (1.5) | 8 | 60 (17.1 to 100) | 23.1 (0 to 60.0) | |
| Primary hospital discharge diagnoses (Aarhus, Denmark) | ICD-10 | 200 | 148 | 148 | 0 | 100 (100 to 100) | 74.0 (66.9 to 81.1) |
*Best-case scenario: non-assessable and non-retrievable cases are not included in the numerator or denominator when the PPV is calculated.
†Worst-case scenario: both non-assessable and non-retrievable cases are included in the denominator when the PPV is calculated. For Aarhus, the number of cases that would have been retrieved per code was estimated based on the percentage of distribution of codes in the retrieved cases within the random sample.
ICPC, International Classification of Primary Care; IPCI, Integrated Primary Care Information.