Clinical Study
Comparison of a Bayesian Network with a Logistic Regression Model to Forecast IgA Nephropathy
Table 2
Characteristics of patients in the learning versus validation sample.
| | Group | | Learning
() | Validation
() | P |
| | Age (years) | Mean ± SEM | | | NS | | | Age < 40 years | 28 (37%) | 25 (34%) | NS | | | 40 ≤ age < 60 years | 26 (35%) | 25 (34%) | | | | Age ≥ 60 years | 21 (28%) | 24 (32%) | | | Male gender | | 44 (59%) | 50 (68%) | NS | | History of hypertension | | 31 (41%) | 35 (47%) | NS | | Microhematuria | | 32 (43%) | 34 (46%) | NS | | Macrohematuria | | 12 (16%) | 13 (18%) | NS | | Family history of hematuria | | 3 (4%) | 1 (1%) | NS | | History of diabetes | | 5 (7%) | 12 (16%) | NS | | Mean GFR (MDRD mL/min/1.73 m²) | Mean ± SEM | | | NS | | Stage of renal failure | eGFR ≤ 29 mL/min/1.73 m² | 11 (15%) | 11 (15%) | NS | | | 30 ≤ eGFR ≤ 59 mL/min/1.73 m² | 25 (33%) | 21 (28%) | | | | 60 mL/min/1.73 m² ≤ eGFR | 39 (52%) | 42 (57%) | | | Serum Ig A | Increased | 14 (19%) | 13 (18%) | NS | | | Normal | 25 (33%) | 26 (35%) | | | | Not performed | 36 (48%) | 35 (47%) | | | Proteinuria (g/24 h) | Mean ± SEM | | | NS | | | Proteinuria < 0.3 g/24 h | 8 (10%) | 10 (14%) | NS | | | 0.3 g/24 h ≤ proteinuria < 1 g/24 h | 17 (23%) | 13 (18%) | NS | | | 1 g/24 h ≤ proteinuria < 3 g/24 h | 26 (35%) | 24 (32%) | | | | Proteinuria ≥ 3 g/24 h | 24 (32%) | 27 (36%) | | | Number of IgAN | | 26 (35%) | 18 (24%) | NS |
|
|
