Table 2.
Probability | |||
---|---|---|---|
Recover Progress1 | Lapse History2 | Lapse | Non−Lapse |
(−35 to −28) | Stay sober (0 to 0.5) | 33% | 67% |
(−28 to −21) | Stay sober (0 to 0.5) | 20% | 80% |
(−21 to −14) | Stay sober (0 to 0.5) | 11% | 89% |
(−14 to −7) | Stay sober (0 to 0.5) | 7% | 93% |
(−7 to 0) | Stay sober (0 to 0.5) | 7% | 93% |
(0 to 7) | Stay sober (0 to 0.5) | 4% | 96% |
(7 to 14) | Stay sober (0 to 0.5) | 4% | 96% |
(14 to 21) | Stay sober (0 to 0.5) | 2% | 98% |
(21 to 28) | Stay sober (0 to 0.5) | 2% | 98% |
(28 to 35) | Stay sober (0 to 0.5) | 2% | 98% |
(−35 to −28) | Lapse once (0.5 to 2) | 25% | 75% |
(−28 to −21) | Lapse once (0.5 to 2) | 50% | 50% |
(−21 to −14) | Lapse once (0.5 to 2) | 33% | 67% |
(−14 to −7) | Lapse once (0.5 to 2) | 41% | 59% |
(−7 to 0) | Lapse once (0.5 to 2) | 38% | 62% |
(0 to 7) | Lapse once (0.5 to 2) | 27% | 73% |
(7 to 14) | Lapse once (0.5 to 2) | 18% | 82% |
(14 to 21) | Lapse once (0.5 to 2) | 13% | 87% |
(21 to 28) | Lapse once (0.5 to 2) | 8% | 92% |
(28 to 35) | Lapse once (0.5 to 2) | 17% | 83% |
(−35 to −28) | Lapse more than once (>=2) | 80% | 20% |
(−28 to −21) | Lapse more than once (>=2) | 63% | 38% |
(−21 to −14) | Lapse more than once (>=2) | 80% | 20% |
(−14 to −7) | Lapse more than once (>=2) | 54% | 46% |
(−7 to 0) | Lapse more than once (>=2) | 67% | 33% |
(0 to 7) | Lapse more than once (>=2) | 59% | 41% |
(7 to 14) | Lapse more than once (>=2) | 60% | 40% |
(14 to 21) | Lapse more than once (>=2) | 50% | 50% |
(21 to 28) | Lapse more than once (>=2) | 33% | 67% |
(28 to 35) | Lapse more than once (>=2) | 50% | 50% |
Note:
Recovery progress (from −35 to +35) was divided into 10 equal discrete categories.
Lapse history were divided into three categories: values 0 to less than 0.5 were in the “sober” category; values 0.5 to less than 2 were in “lapse once” category; and values 2 and more were in “lapse more than once” category.
Probabilities in these two categories, (1) “Recovery Progress 14 to 21 and Lapse more than once” and (2) “Recovery Progress 28 to 35 and Lapse more than once”, were based on the initial model specification (see section 2.7.1. Model configuration for details) because the training dataset does not contain cases in these two categories.