Predicting schizophrenia at risk of readmissions in the short- and long-term using decision tree model
Article Sidebar
Main Article Content
Abstract
This study aims to develop readmission prediction models using a decision tree technique in data mining for predicting patients with schizophrenia at risk of readmission for four different time periods after discharge: ≤ 6 months, 6-12 months, 1-2years, > 2years. Information on the socio-demographic and clinical characteristics data were collected from the registered medical files of patients. Of the 2,285 patients admitted to Prasrimahabhodi Psychiatric Hospital between January 2007 and December 2012, 778 (34.05%) were read-missions. Almost 30% of these patients were readmitted within 6 months of discharge. The non-compliance with medication patients who were diagnoses of F20.3, F20.5 and F20.8 tend to be readmitted within 6 month, while subtype diagnoses of F20.1, F20.2 and F20.4 tend to be readmitted between 6 months and 1 year. Fur-thermore, patients who were subtype diagnoses of F20.2, F20.3, F20.4, F20.5 and F20.8 tend to be readmitted after 2 years. Among the patients who had low compliance to medication with diagnoses of F20.0 and F20.1 if they are unmarried, widowed and divorced, they tend to be readmitted after 2 years. The experimental results also showed that schizophrenia readmission prediction model achieved the highest accuracy, true positive rate, and true negative rate of short-term readmission up to 93.38%, 94.07% and 92.68%, and long-term read-mission up to 97.40%, 98.05% and 96.44%, respectively. The implications of this study may help to increase our understanding of early intervention and will enable clinicians and practitioners in planning care.
Article Details
References
(1) WHO. Schizophrenia. 2012; Available from: http://www.who.int/mental_health/management/ schizophrenia/en/.
(2) Phanthuname P, Vos T, Whiteford H, Bertram M, Udomratn P. Schizophrenia in Thailand : prevalence and burden of disease Population Health Metrics. 2010;8(24): 1-8.
(3) Memarian H, Balasun SK. Comparison between multi-layer perceptron and radial basis function networks for sediment load estimation in a tropical watershed. Water Resource and Protection. 2012;4:870-6.
(4) Tam O, Lam S, Shum H, Lau C, Chan K, Yan W. Characteristics of patients readmitted to intensive care unit: a nested case-control study. Hong Kong Med Journal. 2014;20(3).
(5) Madi N, Zhao H, Li J. Hospital readmissions for patients with mental illness in Cannada. Healthcare Quarterly. 2007;10(2):30-2.
(6) Kessler R, Berglund P, Demler O, Jin R, Merikangas K, Walters E. Lifetime prevalence and age-of-onset distributions of DMS IV disorder in the national comorbidity survey replication. Archives of General Psychiatry. 2005;62:593-600.
(7) Bunevicius R, Liaugaudaite V, Peceliuniene J, Raskauskiene N, Bunevicius A, Mickuviene N. Factors affecting the presence of depression, anxiety disorders, and suicidal ideation in patients attending primary health care service in Lithuania. Prim Health Care. 2014;32(1):24-9.
(8) Claveria O, Torra S. Forecasting tourism demand to catalonia: neural networks vs. time series models. Economic Modelling. 2014;36(2014):220-8.
(9) Charlesworth B, Sacks J, Templer DI, Thackrey M. Negative emotion as predictor of relapse in persons with schizophrenia living in board and care homes. Community Mental Health. 1993;29(3):261-8.
(10) Bowersox NW, Saunders SM, Berger BD. Predictors of rehospitalization in high-utilizing patients in the va psychiatric medical system. The Psychiatric Quarterly. 2012;83(1):53-64.
(11) Emsley R, Medori R, Koen L, Oosthuizen P, Niehaus D, Rabinowitz J. Long-acting injectable risperidone in the treatment of subjects with recent-onset psychosis: a preliminary study. Clin Psychopharmacol. 2008;28:210-3.
(12) Laan W, van der Does Y, Sezgi B, Smeets HM, Stolker JJ, Wit NJ, et al. Low treatment adherence with antipsychotics is associated with relapse in psychotic disorders within six months after discharge. Pharmacopsychiatry. 2010;43(6):221-4.
(13) Nelson EA, Maruish ME, Axler JL. Effects of discharge planning and compliance with outpatient appointments on readmission rates. Psychiatr Services. 2000;51(7):885-9.
(14) Emsley R, Oosthuizen P, Koen L, Niehaus D, Martinez L. Comparison of treatment response in second-episode versus first-episode schizophrenia. Clin Psychopharmacol. 2013;33(1):80-3.
(15) Olfson M, Mechanic D, Boyer C, Hansell S, Walkup J, Weiden P. Assessing clinical predictors of early rehospitalization in schizophrenia. Nervous and Mental Disease. 1999;187: 721-9.
(16) Bimerew M, Sonn F, Kortenbout W. Substance abuse and the risk of readmission of people with schizophrenia at amanuel psychiatric hospital, ethiopia. Curationis. 2007;30(2):74-81.
(17) Paimer A, Montaflo J, Sese A. Designing and artificial neural metwork for forcasting tourism time-series. Tour Manag. 2006;27:781-90.
(18) Nelson E, Maruish M, Axler J. Effects of discharge planning and compliance with outpatient appointments on readmission rates. Psychiatr Services. 2000;51(7):885-9.
(19) Lien L. Are readmission rates influenced by how psychiatric services are organized? . Nord J Psychiatry. 2002;56(1):23-8.
(20) Barekatain M, Maracy MR, Hassannejad R, Hosseini R. Factors associated with readmission of patients at a university hospital psychiatric ward in Iran. Psychiatry. 2013;2013:1-5.
(21) Hendryx M, Russo J, Stegner B, Dyck G, Ries R, Roy-Byrne P. Predicting rehospitalization and outpatient services from administration and clinical databases. Behavioral Health Services & Research. 2003;30(3):342-51.
(22) Han J, Kamber M. Data mining: Concepts and techniques. 2nd. ed. San Francisco: Morgan Kaufmann, Elsevier Science; 2006.
(23) Bellaachia A, Guven E. Predicting breast cancer survivability using data mining techniques. George Washington University. 2006: 1-4.
(24) Yao Z, Liu P, Lei L, Yin J, editors. R-c4.5 decision tree model and its applications to health care dataset. International Conference on Services Systems and Services Management; 2005: 1099-1103.
(25) Quinlan JR. C4.5: Programs for machine learning. San Mateo, California: Morgan Kaufmann; 1993.
(26) Thongkam J, Sukmak V. Cervical cancer survivability prediction models using machine learning techniques. Journal of Convergence Information Technology. 2013;8(15):13-22.
(27) Hui C, Wong G, Tang J, Chang W, Chan S, Lee E, et al. Predicting 1-year risk for relapse in patients who have discontinued or continued quetiapine after remission from first-episode psychosis. Schizophr Research. 2013;150(1):297-302.
(28) Tan P-N, Steinbach M, Kumar V. Introduction to data mining. Boston: Pearson Addison Wesley; 2006.
(29) Witten IH, Frank E. Data mining: Practical machine learning tools and techniques. 2ed ed. San Francisco: Morgan Kaufmann; 2005.
(30) Leelanuntakit T, Udomratn P, Kerdpongbunchote C. The results of prelapse program in thailand : Comparison one year before and after. Journal of the Psychiatrist Association of Thailand. 1999;44(1):3-11.
(31) Beratis S, Gabriel J, Hoidas S. Gender differences in the frequency of schizophrenic subtypes in unselected hospitalized patients. Schizophr Research. 1997;28:239-44.
(32) Mortensen P, Eaton W. Predictors for readmission risk in schizophrenia. Psychological Medicine. 1994;24:223-32.
(33) Doering S, Müller E, Köpcke W, Pietzcker A, Gaebel W, Linden M, et al. Predictors of relapse and rehospitalization in schizophrenia and schizoaffective disorder. schizophrenia bulletin. 1998;24:87-98.
(34) American Psychiatric Association. Diagnostic and statistical manual of mental disorders. 5th ed. Washington, DC.2013.
(35) Olfson M, Mechanic D, Boyer C, Hansell S, Walkup J, Weiden P. Assessing clinical predictions of early hospitalizations in schizophrenia. Nervous and Mental Disease. 1999;187:721-9.
(36) Alvarez-Jimenez M, Priedec A, Hetrick S, Bendall S, Killackey E, Parker A, et al. Risk factors for relapse following treatment for first episode psychosis: A systematic review and meta-analysis of longitudinal studies. Schizophrenia Research. 2012;139(1-3):116-28.