Nunes, A., Ardau, R., Berghöfer, A., Bocchetta, A., Chillotti, C., Deiana, V., Garnham, J., Grof, E., Hajek, T., Manchia, M., Müller-Oerlinghausen, B., Pinna, M., Pisanu, C., O’Donovan, C., Severino, G., Slaney, C., Suwalska, A., Zvolsky, P., Cervantes, P., ... Alda, M. (2020). Prediction of lithium response using clinical data. Acta Psychiatrica Scandinavica, 141(2), 131-141. https://doi.org/10.1111/acps.13122
Prediction of lithium response using clinical data. /
Nunes, A.; Ardau, R.; Berghöfer, A. et al.
In:
Acta Psychiatrica Scandinavica, Vol. 141, No. 2, 01.02.2020, p. 131-141.
Research output: Contribution to journal › Article › peer-review
Nunes, A, Ardau, R, Berghöfer, A, Bocchetta, A, Chillotti, C, Deiana, V, Garnham, J, Grof, E, Hajek, T, Manchia, M, Müller-Oerlinghausen, B, Pinna, M, Pisanu, C, O’Donovan, C, Severino, G, Slaney, C, Suwalska, A, Zvolsky, P, Cervantes, P, del Zompo, M, Grof, P, Rybakowski, J, Tondo, L, Trappenberg, T & Alda, M 2020, 'Prediction of lithium response using clinical data', Acta Psychiatrica Scandinavica, vol. 141, no. 2, pp. 131-141. https://doi.org/10.1111/acps.13122
Nunes A, Ardau R, Berghöfer A, Bocchetta A, Chillotti C, Deiana V et al. Prediction of lithium response using clinical data. Acta Psychiatrica Scandinavica. 2020 Feb 1;141(2):131-141. doi: 10.1111/acps.13122
Nunes, A. ; Ardau, R. ; Berghöfer, A. et al. / Prediction of lithium response using clinical data. In: Acta Psychiatrica Scandinavica. 2020 ; Vol. 141, No. 2. pp. 131-141.
@article{9ea3583324b841a1a506e731ec9c3f53,
title = "Prediction of lithium response using clinical data",
abstract = "Objective: Promptly establishing maintenance therapy could reduce morbidity and mortality in patients with bipolar disorder. Using a machine learning approach, we sought to evaluate whether lithium responsiveness (LR) is predictable using clinical markers. Method: Our data are the largest existing sample of direct interview-based clinical data from lithium-treated patients (n = 1266, 34.7% responders), collected across seven sites, internationally. We trained a random forest model to classify LR—as defined by the previously validated Alda scale—against 180 clinical predictors. Results: Under appropriate cross-validation procedures, LR was predictable in the pooled sample with an area under the receiver operating characteristic curve of 0.80 (95% CI 0.78–0.82) and a Cohen kappa of 0.46 (0.4–0.51). The model demonstrated a particularly low false-positive rate (specificity 0.91 [0.88–0.92]). Features related to clinical course and the absence of rapid cycling appeared consistently informative. Conclusion: Clinical data can inform out-of-sample LR prediction to a potentially clinically relevant degree. Despite the relevance of clinical course and the absence of rapid cycling, there was substantial between-site heterogeneity with respect to feature importance. Future work must focus on improving classification of true positives, better characterizing between- and within-site heterogeneity, and further testing such models on new external datasets.",
author = "A. Nunes and R. Ardau and A. Bergh{\"o}fer and A. Bocchetta and C. Chillotti and V. Deiana and J. Garnham and E. Grof and T. Hajek and M. Manchia and B. M{\"u}ller-Oerlinghausen and M. Pinna and C. Pisanu and C. O{\textquoteright}Donovan and G. Severino and C. Slaney and A. Suwalska and P. Zvolsky and P. Cervantes and {del Zompo}, M. and P. Grof and J. Rybakowski and L. Tondo and T. Trappenberg and M. Alda",
note = "Funding Information: Genome Canada (MA, AN), Dalhousie Department of Psychiatry Research Fund (MA, AN), Canadian Institutes of Health Research #64410 (MA), Nova Scotia Health Research Foundation Scotia Scholars Graduate Scholarship (AN), Killam Postgraduate Scholarship (AN). Publisher Copyright: {\textcopyright} 2019 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd",
year = "2020",
month = feb,
day = "1",
doi = "10.1111/acps.13122",
language = "English",
volume = "141",
pages = "131--141",
journal = "Acta Psychiatrica Scandinavica",
issn = "0001-690X",
publisher = "Wiley-Blackwell",
number = "2",
}
TY - JOUR
T1 - Prediction of lithium response using clinical data
AU - Nunes, A.
AU - Ardau, R.
AU - Berghöfer, A.
AU - Bocchetta, A.
AU - Chillotti, C.
AU - Deiana, V.
AU - Garnham, J.
AU - Grof, E.
AU - Hajek, T.
AU - Manchia, M.
AU - Müller-Oerlinghausen, B.
AU - Pinna, M.
AU - Pisanu, C.
AU - O’Donovan, C.
AU - Severino, G.
AU - Slaney, C.
AU - Suwalska, A.
AU - Zvolsky, P.
AU - Cervantes, P.
AU - del Zompo, M.
AU - Grof, P.
AU - Rybakowski, J.
AU - Tondo, L.
AU - Trappenberg, T.
AU - Alda, M.
N1 - Funding Information:
Genome Canada (MA, AN), Dalhousie Department of Psychiatry Research Fund (MA, AN), Canadian Institutes of Health Research #64410 (MA), Nova Scotia Health Research Foundation Scotia Scholars Graduate Scholarship (AN), Killam Postgraduate Scholarship (AN).
Publisher Copyright:
© 2019 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
PY - 2020/2/1
Y1 - 2020/2/1
N2 - Objective: Promptly establishing maintenance therapy could reduce morbidity and mortality in patients with bipolar disorder. Using a machine learning approach, we sought to evaluate whether lithium responsiveness (LR) is predictable using clinical markers. Method: Our data are the largest existing sample of direct interview-based clinical data from lithium-treated patients (n = 1266, 34.7% responders), collected across seven sites, internationally. We trained a random forest model to classify LR—as defined by the previously validated Alda scale—against 180 clinical predictors. Results: Under appropriate cross-validation procedures, LR was predictable in the pooled sample with an area under the receiver operating characteristic curve of 0.80 (95% CI 0.78–0.82) and a Cohen kappa of 0.46 (0.4–0.51). The model demonstrated a particularly low false-positive rate (specificity 0.91 [0.88–0.92]). Features related to clinical course and the absence of rapid cycling appeared consistently informative. Conclusion: Clinical data can inform out-of-sample LR prediction to a potentially clinically relevant degree. Despite the relevance of clinical course and the absence of rapid cycling, there was substantial between-site heterogeneity with respect to feature importance. Future work must focus on improving classification of true positives, better characterizing between- and within-site heterogeneity, and further testing such models on new external datasets.
AB - Objective: Promptly establishing maintenance therapy could reduce morbidity and mortality in patients with bipolar disorder. Using a machine learning approach, we sought to evaluate whether lithium responsiveness (LR) is predictable using clinical markers. Method: Our data are the largest existing sample of direct interview-based clinical data from lithium-treated patients (n = 1266, 34.7% responders), collected across seven sites, internationally. We trained a random forest model to classify LR—as defined by the previously validated Alda scale—against 180 clinical predictors. Results: Under appropriate cross-validation procedures, LR was predictable in the pooled sample with an area under the receiver operating characteristic curve of 0.80 (95% CI 0.78–0.82) and a Cohen kappa of 0.46 (0.4–0.51). The model demonstrated a particularly low false-positive rate (specificity 0.91 [0.88–0.92]). Features related to clinical course and the absence of rapid cycling appeared consistently informative. Conclusion: Clinical data can inform out-of-sample LR prediction to a potentially clinically relevant degree. Despite the relevance of clinical course and the absence of rapid cycling, there was substantial between-site heterogeneity with respect to feature importance. Future work must focus on improving classification of true positives, better characterizing between- and within-site heterogeneity, and further testing such models on new external datasets.
UR - http://www.scopus.com/inward/record.url?scp=85075483782&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85075483782&partnerID=8YFLogxK
U2 - 10.1111/acps.13122
DO - 10.1111/acps.13122
M3 - Article
C2 - 31667829
AN - SCOPUS:85075483782
SN - 0001-690X
VL - 141
SP - 131
EP - 141
JO - Acta Psychiatrica Scandinavica
JF - Acta Psychiatrica Scandinavica
IS - 2
ER -