Table 1. Predictor Comparisons between Samples Used in Model Development and Internal/External Validation.
| Predictor | Sample | |||
|---|---|---|---|---|
| Development | External Validation | |||
| CAMEO EIS | Birmingham EIS | Pooled Development Sample | SLaM EIS Validation Sample | |
| Final Included Sample size, N. | 539 | 246 | 785 | 1,110 |
| Male Sex, N. (%) | 328 (60.9%) | 146 (59.3%) | 474 (60.4%) | 692 (62.3%) |
| Age in Years, mean (SD) min, max |
30.23 (12.00) 14, 65 |
23.86 (4.87) 15, 37 |
28.24 (10.73) 14, 65 |
28.82 (9.94) 17.5, 64 |
| White/unrecorded Ethnicity, N. (%) | 449 (83.3%) | 70 (28.4%) | 519 (66.1%) | 378 (34.0%) |
| Black/African-Caribbean Ethnicity, N. (%) | 21 (3.9%) | 57 (23.2%) | 78 (9.9%) | 507 (45.7%) |
| Asian Ethnicity, N. (%) | 69 (12.8%) | 119 (48.4%) | 188 (23.9%) | 225 (20.3%) |
| Triglycerides, mmol/L, mean (SD) | 1.42 (1.07) | 1.55 (1.30) | 1.46 (1.15) | 1.25 (0.96) |
| Alkaline phosphatase (ALP), U/L, mean (SD) | 78.58 (25.55) | 82.67 (25.78) | 79.86 (25.68) | 75.03 (22.01) |
| Lymphocyte count, billion/L, mean (SD) | 1.91 (0.69) | 2.22 (0.79) | 2.01 (0.74) | 1.98 (0.64) |
| Smoking, N. (%) | 201 (37.3%) | 124 (50.4%) | 325 (41.4%) | 468 (42.2%) |
| Body mass index (BMI), kg/m2, mean (SD) | 24.68 (6.65) | 25.74 (5.78) | 25.01 (6.41) | 24.07 (5.58) |
| Random plasma glucose (mmol/L), mean (SD) | 5.25 (1.73) | 4.87 (1.28) | 5.13 (1.61) | 5.11 (1.80) |
| Neutrophil count, billion/L, mean (SD) | 4.60 (2.00) | 4.14 (2.06) | 4.46 (2.03) | 4.01 (1.98) |
| Follow-up time, years, mean (SD) min, max |
4.45 (1.57) 0.58*, 8.50 |
3.55 (0.58) 2.67, 4.58 |
4.17 (1.40) | 4.41 (1.76) 0.75*, 8.75 |
| TRS at Follow-up, N. (%) | 35 (6.5%) | 23 (9.3%) | 58 (7.4%) | 102 (9.2%) |
SLaM, South London and Maudsley NHS Foundation Trust; EIS, Early Intervention Service; CAMEO, Cambridgeshire and Peterborough Assessing, Managing and Enhancing Outcomes EIS (Cambridgeshire and Peterborough Foundation NHS Trust); SD, standard deviation; TRS, treatment-resistant schizophrenia.
*
only TRS cases have been included if they had follow-up < 2 years