Table 3 Handling of treatment in the analyses of prognostic model studies
From: Treatment use in prognostic model research: a systematic review of cardiovascular prognostic studies
| Approach taken to account for treatment use | Development studies n = 124 (%) | Incremental value studies n = 135 (%) | Validation studies n = 146 (%) |
|---|---|---|---|
| Treated patients excluded from the analysis | 20 (15) | 53 (39) | 38 (26) |
| Antihypertensive medication users | 4 (3) | 6 (4) | 6 (4) |
| Lipid-lowering medication users | 6 (5) | 10 (7) | 16 (11) |
| Other medication users | 1 (1) | 2 (1) | 1 (1) |
| Lifestyle interventions | 0 | 0 | 0 |
| Patients who received surgery | 14 (10) | 39 (29) | 22 (15) |
| Untreated patients-only sensitivity analysis | 9 (7) | 5 (4) | 4 (3) |
| Stratification by treatment use | 1 (1) | 0 | 0 |
| Treatment included in the outcome | 23 (19) | 58 (43) | 35 (24) |
| Treatment modelled as a predictor | 54 (44) | 48 (59)a | – |
| Antihypertensive medication use | 49 (40) | 44 (54)a | – |
| Lipid-lowering medication use | 12 (10) | 15 (11)a | – |
| Other medication use | 2 (2) | 5 (4)a | – |
| Lifestyle interventions | 1 (1) | 0a | – |
| Surgical interventions | 0 | 0a | – |
| Type of treatment information modelled | |||
| Modelled directly (not a composite predictorb) | 37 (30) | 44 (54)a | – |
| Baseline treatment | 41 (33) | 36 (44)a | – |
| Changes in treatment during follow-up | 0 | 0a | – |
| Treatment at the end of follow-up | 0 | 1 (1)a | – |
| Not clearly reported | 12 (10) | 11 (8) | – |
| Statistical interactions with treatment considered | 21 (17) | 7 (5)a | – |