Table 2.
Validation of the clinical staging models.
| Objective vs Validation/Feasibility | Sample | Stages | Conclusions | ||
|---|---|---|---|---|---|
| McGorry et al. (2010) | Berendensen et al. (2018) | To examine the construct validity of the staging model by measuring differences in severity of clinical profiles and therapeutic improvement between clinical stages. | n = 258 |
Stage 2 = 48 Stage 3b = 100 Stage 3c = 81 Stage 4 = 29 |
Only stages 3c and 4 showed adequate construct validity [significant differences were found for negative symptoms (F = 4.56, p < 0.010), number of psychotic episodes (F = 13.65, p < 0.010), and premorbid functioning (F = 7.33, p < 0.001) according to stages]. |
| Berendensen et al. (2019) |
To determine the inter-rater reliability of the clinical staging. To investigate whether a short course can improve reliability. |
n = 114 (no training) |
Stage 2 = 22 Stage 3a = 1 Stage 3b = 39 Stage 3c = 41 Stage 4 = 11 |
The inter-rater reliability in clinical staging was better after training (ICC = 0.57 vs ICC = 0.75). |
|
|
n = 100 (with training) |
Stage 2 = 22 Stage 3a = 1 Stage 3b = 50 Stage 3c = 22 Stage 4 = 5 |
||||
| Godin et al. [18] |
To classify patients according to the model. To use clinical, cognitive, and treatment variables to explore validity. To explore the stability of the model. |
n = 770 |
Stage 2a = 89 Stage 2b = 272 Stage 3a = 241 Stage 3b = 112 Stage 4 = 56 |
Follow-up at one year showed good stability (62% of the sample remained stable). | |
| Berendensen et al. (2021) | To examine differences in severity for dimensional symptoms of psychosis between stages. | n = 291 |
Stage 2 = 62 Stage 3a = 9 Stage 3b = 127 Stage 2b = 75 Stage 4 = 18 |
Significant differences in the severity of symptoms only were found in stages 3c and 4 [hallucinations (H = 14.34, p = 0.006), negative symptoms (H = 19.67, p = 0.001), and cognitive deficits (H = 26.29, p < 0.001)]. | |
| Hickie et al. [17] | Hickie et al. [17] | To demonstrate the inter-rater reliability of the model. | n = 209 |
Stage 1a = 21 Stage 1b = 112 Stage 2 = 53 |
The inter-rater reliability was acceptable (K = 0.72, p < 0.001). |
| Romanowska et al. [25] | To assess neurocognition in a sample of patients in the first stages of schizophrenia. | n = 243 |
Stage 0 = 41 Stage 1a = 52 Stage 1b = 108 Controls = 42 |
Patients in stage 1b presented significantly poorer cognitive performance (MATRICS Overall Composite F = 5.70, p < 0.001). | |
| Addington et al. [26, 27] |
To identify sample that met different stages of risk for the development of a serious mental illness (SMI) based on a published clinical staging model. To determine whether participants allocated to the different stages were a good fit to the model. |
n = 243 |
Stage 0 = 41 Stage 1a = 52 Stage 1b = 108 Controls = 42 |
Patients in stage 1b had significantly more severe symptoms than participants in lower stages [functioning (F = 77.10, p < 0.002), depressive symptoms (F = 30.10, p < 0.002), and prodromal psychotic symptoms (F = 37.30, p < 0.002)]. | |
| Addington et al. [28] | To describe changes in participants over 12 months to understand the course of illness progression in its earliest stages. | n = 243 |
Stage 0 = 41 Stage 1a = 53 Stage 1b = 107 Controls = 42 |
Follow-up at one year showed stability (only 7–9% of the participants changed stage in the follow-up). |
ICC Intraclass Correlation Coefficient, K Kappa Statistics, MATRICS The Measurement and Treatment Research to Improve Cognition in Schizophrenia.