Table 5.
Linear mixed model | Logistic regression | LNPT | ||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
Clinical cut-off | Spike cut-off | |||||||||||
Variable | Transform | Slope | P | 95% CI | OR | P | 95% CI | OR | P | 95% CI | P | Interpretation of resultsb |
DG_SYM | 2.9527 | <0.0001 | 2.62 to 3.29 | <0.0001 | Significant, positive | |||||||
DGT_SP_bck | 0.2876 | <0.0001 | 0.21 to 0.36 | 0.0066 | Significant, positive | |||||||
DGT_SP_frw | 0.0666 | 0.0577 | −0.01 to 0.14 | 0.1966 | Not significant, positive | |||||||
TMT_A | log(x) | −0.0634 | <0.0001 | −0.07 to −0.05 | <0.0001 | Significant, positive | ||||||
TMT_B | log(x) | −0.0464 | <0.0001 | −0.06 to −0.04 | <0.0001 | Significant, positive | ||||||
VLMT_corr | sqrt(80 − x) | −0.3473 | <0.0001 | −0.40 to −0.30 | <0.0001 | Significant, positive | ||||||
VLMT_lss_d | log(8 + x) | −0.0144 | 0.0282 | −0.03 to −0.002 | 0.0847 | Significant, positive | ||||||
VLMT_lss_t | log(5 + x) | −0.0226 | 0.0473 | −0.05 to −0.00 | 0.0970 | Significant, positive | ||||||
VLMT_rec | 1.82c | 0.0918 | 0.93 to 3.77 | 1.40d | 0.0036 | 1.12 to 1.75 | 0.0796 | Borderline significant, positive | ||||
BDI | 1.01e | 0.9367 | 0.72 to 1.43 | 1.04 | 0.6303 | 0.90 to 1.19 | 0.8432 | Not significant, assumed stable | ||||
MSS | 2.86e | 0.0043 | 1.47 to 6.25 | 1.36 | <0.0001 | 1.18 to 1.57 | 0.0029 | Conflicting with YMRS, negative | ||||
GAF | sqrt(100 − x) | −0.0142 | 0.5811 | −0.06 to 0.04 | 1.24d | 0.0120 | 1.05 to 1.46 | 1.11 | 0.2053 | 0.95 to 1.29 | 0.9271 | Conflicting, assumed stable |
0.76c | 0.0027 | 0.63 to 0.91 | ||||||||||
GAF_cat | −0.0227 | 0.2480 | −0.06 to 0.01 | 0.9521 | ||||||||
IDS | 0.61e | 0.1081 | 0.32 to 1.10 | 0.89 | 0.1028 | 0.76 to 1.02 | 0.2854 | Not significant, assumed stable | ||||
PANSS | 1.07 | 0.3625 | 0.92 to 1.26 | 0.7953 | Not significant, assumed stable | |||||||
YMRS | 0.80 | <0.0001 | 0.70 to 0.93 | 0.2853 | Conflicting with MSS, positive |
LNPT, longitudinal non-parametric test; DG_SYM, Digit Symbol Test; DGT_SP, verbal Digit Span test; bck, backwards; frw, forwards; TMT, Trail Making Test, Parts A and B; VLMT, German version of the Rey Auditory Verbal Learning Test; corr, correctly recalled; lss_d, loss of correct words after distraction; lss_t, loss of correct words after time span; rec, recognition of words; BDI, Beck Depression Inventory II; MSS, Self-Report Manic Inventory [in German]; GAF, Global Assessment of Functioning; cat, categorical; IDS, 30-item Inventory of Depressive Symptomatology; MSS, Self-Report Manic Inventory [in German]; PANSS, Positive and Negative Syndrome Scale; YMRS, Young Mania Rating Scale.
The time effect was estimated using a linear mixed model (LMM) or logistic regression including age, gender and a random intercept, and checked for robustness using an LNPT. Estimated effects are reported as slopes for the LMM and OR for the logistic regression.
Summary of the results regarding the stability assumption. Positive indicates improved performance over time for the participants or fewer symptoms.
Cut-off for good function (could be interpreted as few symptoms).
Cut-off for perfect function (could be interpreted as no symptoms).
Clinical cut-off in logistic regression. Note that for VLMT_rec and GAF/GAF_cat one or two of the cut-offs fall together and are not listed twice.