The predictive effect of insight on adverse clinical outcomes in bipolar I disorder: A two-year prospective study

Transcription

1 Journal of Affective Disorders 108 (2008) Research report The predictive effect of insight on adverse clinical outcomes in bipolar I disorder: A two-year prospective study Cheng-Fang Yen a,b, Cheng-Sheng Chen a,b,c,d,, Ju-Yu Yen b,c,d, Chih-Hung Ko b,d a Department of Psychiatry, Faculty of Medicine, College of Medicine, Kaohsiung Medical University, Kaohsiung, Taiwan b Department of Psychiatry, Kaohsiung Medical University Hospital, Kaohsiung, Taiwan c Department of Psychiatry, Kaohsiung Municipal Hsiao-Kang Hospital, Kaohsiung, Taiwan d Graduate Institute of Medicine, College of Medicine, Kaohsiung Medical University, Kaohsiung, Taiwan Received 27 August 2007; received in revised form 7 October 2007; accepted 8 October 2007 Available online 13 November 2007 Abstract Research has revealed that a lack of insight is associated with poorer clinical outcomes in schizophrenia; however, the predictive value of insight on adverse clinical outcomes among bipolar patients is quite understudied. The aim of this prospective study was to examine the impact of insight on adverse clinical outcomes among the patients with bipolar I disorder over a 2-year period. Sixtyfive remitted bipolar I disorder patients received follow-up assessments at 3, 6, 9, 12, 18, and 24 months to detect the adverse clinical outcomes defined by the incidence of bipolar-related psychiatric hospitalization, emergency room visits, violent or suicidal behavior. The Schedule of Assessment of Insight was used to provide a baseline insight score. Cox regression analysis was used to examine the predictive value of insight on the adverse clinical outcomes. Impaired insight into treatment and a greater number of previous hospitalizations significantly increased the risk of adverse clinical outcomes with bipolar disorder in the 2-year period. However, insight into recognition of the illness and re-labeling of psychotic phenomena did not have any significant effect on adverse clinical outcomes. Bipolar patients' insight into treatment is an independent predictor of adverse clinical outcomes. Improving insight into treatment might be a promising target for a better outcome Elsevier B.V. All rights reserved. Keywords: Bipolar disorder; Hospitalization; Insight; Outcome; Suicide; Violence 1. Introduction Bipolar disorder is characterized as a chronic mental disorder usually involving marked affective and behavioral dysfunctions (Angst and Sellaro, 2000). Most of bipolar patients have a chronic course with frequent Corresponding author. Department of Psychiatry, Kaohsiung Medical University, No. 100, Tzyou 1st Rd, Kaohsiung 807, Taiwan. Tel.: x6822; fax: address: sheng@cc.kmu.edu.tw (C.-S. Chen). relapses, typically characterized by exacerbation of affective symptoms and re-hospitalizations (Angst and Sellaro, 2000; Gitlin et al., 1995; Tohen et al., 1990), even though modern pharmacotherapies have been introduced (Gitlin et al., 1995). Suicide (Rihmer and Kiss, 2002) and violence (Feldmann, 2001) are also the main clinical issues in the treatment of patients with bipolar disorder. Therefore, the prevention of relapses of affective episodes and occurrence of suicidal and violent behaviors is the major goal in the treatment of bipolar disorders /$ - see front matter 2007 Elsevier B.V. All rights reserved. doi: /j.jad

2 122 C.-F. Yen et al. / Journal of Affective Disorders 108 (2008) Insight of individuals with psychiatric disorders generally indicates an awareness of the psychiatric symptoms experienced, the presence of psychiatric disorders, and awareness of the achieved effect of treatment (Amador et al., 1994). Research has revealed that a lack of insight is associated with poorer clinical outcomes (Yen et al., 2002b), impaired psychosocial functioning (Amador et al., 1994), compromised treatment compliance (Smith et al., 1999), and increased risk of violent (Buckley et al., 2004) and suicidal behaviors (Amador et al., 1996) for patients with severe mental disorders. However, the study of insight has focused primarily on schizophrenia, and insight among patients with bipolar disorder is quite understudied and poorly understood empirically and conceptually (Ghaemi and Rosenquist, 2004). The present study is the first to prospectively investigate the predictive value of insight for the occurrence of indicators for adverse clinical outcomes, including the incidence of bipolar-related psychiatric hospitalization, emergency room visits, suicidal and violent behaviors among a group of patients with bipolar I disorder during a 2-year period of follow-up using a Cox regression analysis. 2. Methods 2.1. Participants The current investigation is based on data from the Psychosis Insight Study in Taiwan, a prospective research program studying the association of insight and outcomes in 216 patients with schizophrenia or bipolar disorder enrolled from a university hospital psychiatric outpatient unit. A total of 71 patients with bipolar I disorder were assessed. Of these subjects, 65 patients agreed to enter this 2-year prospective study. No difference in sex, age, education, marriage, and duration of illness was found between the patients who agree and did not agree to enter the follow-up. The inclusion criteria were a DSM-IV diagnosis of bipolar I disorder (APA, 1994). A current state of remission which was defined by a Young's Mania Rating Scale (YMRS) (Young et al., 1978) score of 6 and the 24-item Hamilton Rating Scale for Depression (HAM-D) (Hamilton, 1960) score of 6 (Altshuler et al., 2002). The exclusion criteria were bipolar patients with comorbid mental retardation (2 patients) evaluated according to the education level and daily function. Those who had the DSM-IV diagnosis of substance use disorder (5 patients) diagnosed according to the self-reported information were also excluded, no matter substance use was related to mood symptoms or not. Clinical and demographic variables of this subject group are described in detail elsewhere (Yen et al., 2004). The protocol was approved by the Institutional Review Board of Kaohsiung Medical University. All patients provided written informed consent Procedures Four research psychiatrists (Yen CF, Chen CS, Yen JY, Ko CH) made the DSM-IV diagnosis of bipolar I disorder based on the results of the clinical diagnostic interviews, the interviewers' observations of behavior, and reviews of the subjects' medical records. Research psychiatrists also performed a semi-structured interview using the Schedule of Assessment of Insight (SAI) (David, 1990) to assess a patient's level of insight. The SAI taps three separate dimensions of insight that have been considered as the core concepts of insight: acceptance of the need for treatment (SAI-treatment, 2 items), recognition of the illness (SAI-illness, 3 items) and re-labeling of psychotic phenomena (SAI-psychosis, 2 items) (Amador et al., 1994; David, 1990). The range of the score is 0 4 for SAI-treatment, 0 6 for SAI-illness, and 0 4 for SAI-psychosis. A higher score indicates better insight. We have developed the semistructured interviewing schedule to collect information needed to complete the SAI, and used the SAI to evaluate the insight of patients with schizophrenia and bipolar disorder in previous studies (Yen et al., 2002a,b). In this study insight was assessed in the subjects who were receiving treatment and were in remitted state. It is the insight level under the treated condition that was the Table 1 Insight and clinical and sociodemographic characteristics of the sample at initiation of the study (N=65) Mean (S.D.) Range n (%) Demographic variable Gender (male) 31 (47.7) Age (yr) 39.3 (11.9) Education (yr) 11.7 (3.4) Having stable employment 37 (56.9) Clinical variable Duration of illness (yr) 12.3 (9.8) Numbers of previous hospitalizations 5.3 (5.4) 0 26 Having psychotic features 33 (50.8) Receiving 2 mood stabilizers 35 (53.8) Medical adherence on the MABS 23.7 (6.1) 7 28 Insight subscale scores SAI-treatment 3.1 (1.3) 0 4 SAI-awareness of one's illness 3.9 (1.5) 0 6 SAI-psychotic experiences 3.0 (1.2) 0 4 MABS: Medication Adherence Behavior Scale.

3 C.-F. Yen et al. / Journal of Affective Disorders 108 (2008) Table 2 Incidences of adverse clinical outcomes occurred during the 2-year period of follow-up n (%) Events of incidences of adverse clinical outcome Psychiatric hospitalization only 9 (13.8) Violent behavior only 2 (3.1) Emergency visiting and then psychiatric hospitalization 5 (7.7) Violent behavior and then psychiatric hospitalization 3 (4.6) Suicidal behavior and then psychiatric hospitalization 2 (3.1) Violent behavior and emergency room visiting 1 (1.5) Violent behavior and then emergency room visiting and 13 (20.0) psychiatric hospitalization Suicidal behavior and then emergency room visiting and 1 (1.5) psychiatric hospitalization Violent and suicidal behaviors and then emergency room 4 (6.2) visiting and psychiatric hospitalization No any adverse outcome 25 (38.5) predictor. The inter-rater reliability of the SAI total scores among the four research psychiatrist in this study ranged from 0.82 to The median of SAI subscale scores was adopted to divide the cohort into a higherinsight group (Nmedian) and a lower-insight group ( median). We also conducted an interview to determine a subject's medical adherence based on the Medication Adherence Behavior Scale (MABS) (Yen et al., 2005). A higher MABS score indicates greater medication adherence. During the 2-year follow-up period, patients were evaluated at 3-month intervals in the first year and then at 6-month intervals in the second year. At each evaluation, patients were interviewed by research psychiatrists who were blinded to the baseline SAI scores. The patient was asked about any instance of bipolar-related psychiatric hospitalization, emergency unit visiting, and any suicidal or violent behaviors since the last evaluation, which were defined as the indicators of adverse clinical outcomes in this study. Additional information from caregivers or medical chart records was used to support the occurrence of such events. We conducted an interview to determine a subject's violent and suicidal behavior based on the Violence and Suicide Assessment Scale (VASA) (Feinstein and Plutchik, 1990). According to the VASA, subjects that had any self-injury behavior or had exhibited violent behaviors toward people or properties were considered to have suicidal or violent behavior. Adverse clinical outcomes were defined by any presence of bipolar-related psychiatric hospitalization, an emergency room visiting related to symptomatic exacerbation of bipolar disorder, deliberate self-injury, or violent behavior Statistical analysis We calculated the rates of occurrence of indicators for adverse clinical outcomes defined as above and time to the first occurrence. A requirement that the number of independent predictors should not exceed 10% of the total number of subjects in the Cox regression analysis has been suggested (Christensen, 1987). Thus, we selected the potential risk factors, including demographic and clinical characteristics and the SAI subscale scores, initially using the uncontrolled Cox regression Table 3 Uncontrolled Cox regression analyses for examining the prediction of level of insight and clinical and demographic factors for poor clinical outcomes B S.E. Wald's χ 2 p Hazard ratio 95% CI of the hazard ratio Demographic variable Age Gender Education Having stable employment Clinical variable Duration of illness b b Numbers of previous hospitalizations b Medical adherence on the MABS Receiving 2 mood stabilizers Having psychotic features Insight subscale SAI-treatment SAI-illness SAI-psychosis MABS: Medication Adherence Behavior Scale.

4 124 C.-F. Yen et al. / Journal of Affective Disorders 108 (2008) Table 4 Controlled Cox regression analyses to examine the prediction of level of insight into treatment, numbers of previous hospitalizations and receiving therapy with two or more mood stabilizers for poor clinical outcomes B S.E. Wald's χ 2 p Hazard ratio 95% CI of the hazard ratio SAI-treatment (low/high) Number of previous hospitalizations Receiving 2 mood stabilizers analyses in order to limit the number of predictors not to exceed six in the controlled Cox regression analysis with the likelihood-ratio (LR) test. The hazard ratios with 95% confidence intervals were calculated. An alpha of 0.05 was used for all statistical tests. 3. Results Clinical and demographic variables of the 65 subjects are presented in Table 1. All 65 subjects have received six follow-up interviews during the 2-year period of follow-up. Based on the definition of indicators for adverse clinical outcomes in this study, 40 events of incidences of indicators were detected during the 2-year period of follow-up (Table 2). Most participants (72.5%) experienced two or more indicators of adverse clinical outcome simultaneously. The most common event of adverse outcome was occurrence of violent behavior and then visiting the emergency room and psychiatric hospitalization (32.5%), followed by psychiatric hospitalization only (22.5%) and emergency room visiting and then psychiatric hospitalization (12.5%). The median time to the first incidence was days (mean, days; range, days) after entry. A total of 25 participants who did not experience the event of adverse clinical outcome were censored. Table 3 shows the results of the uncontrolled Cox regression analysis examining the prediction of demographic and clinical factors and SAI insight subscale score for the time to the first incidence of the indicator for adverse clinical outcome and their impacts on the hazard. Results revealed that SAI-treatment subscale (χ 2 =4.42, pb0.04), number of previous hospitalizations (χ 2 =8.27, pb0.004), and receiving therapy with two or more mood stabilizers (χ 2 =4.27, pb0.04) were associated with the risk of adverse clinical outcomes during the 2-year period. Neither the SAI-illness and SAI-psychosis subscales, gender, age, education, occupational function, duration of illness, medical adherence at the index interview, nor the presence of psychotic features were found to have increased the risk of adverse clinical outcomes. Thus, three predictors, the SAItreatment subscale, the number of previous hospitalizations, and having received therapy with two or more mood stabilizers, were included in the controlled Cox regression analysis. As shown in Table 4, the SAItreatment subscale and the number of previous hospitalizations significantly increased the risk of adverse clinical outcomes (χ 2 =14.84, df=3, p b0.01). Bipolar patients with lower SAI-treatment scores were more likely to be at risk of adverse clinical outcomes with a hazard ratio of 2.04 (95% CI: ) when Fig. 1. Cox regression analysis of time to occurrence of indicators for poor clinical outcomes in bipolar patients assigned to high or low insight into treatment.

5 C.-F. Yen et al. / Journal of Affective Disorders 108 (2008) compared to the group with higher SAI-treatment scores. The addition of a single previous hospitalization increased the chance of adverse clinical outcomes by approximately 8.0%. Receiving therapy with two or more mood stabilizers did not show any risk of adverse clinical outcomes in the multivariate Cox regression analysis. Fig. 1 illustrates the impact of a lower level of SAI insight into treatment at study entry on the time to the first incidence of the indicator for adverse clinical outcomes. The cumulative rate of incidences of adverse clinical outcomes in the group with a higher level of SAI insight into treatment was 55.2% (95%CI: 39.6% 70.8%) and in the group with a lower level of SAI insight into treatment was 72.3% (95%CI: 55.4% 89.2%). 4. Discussion Findings from this prospective study revealed that bipolar patients in a state of remission with lower levels of insight into treatment or with greater numbers of previous hospitalizations were likely to experience adverse clinical outcomes during a two years followup. The strength of this study was that we could determine the temporal relationship between insight and adverse clinical outcomes by using a prospective research design. Meanwhile, by using Cox regression analysis, it could be concluded that the level of insight into treatment has an impact beyond the confounding effect of clinical and demographic variables. To our knowledge, thus far the 2-year period of follow-up used in this study is the longest design among previous similar studies. The principal finding of this study was that a low level of insight into treatment had an impact on adverse clinical outcomes with bipolar I disorder. Patients' compliance with the treatment regime may mediate the association between insight into treatment and adverse clinical outcomes. A great number of studies have suggested the importance of compliance with the treatment regime on prevention of relapse and suicide in bipolar patients (Kessing et al., 2005; Schumann et al., 1999; Tondo et al., 1997). Also, a patient's illness concept is a powerful predictor of medical adherence (Kleindienst and Greil, 2004; Yen et al., 2005). Therefore, it was highly possible that bipolar patients with negative attitude toward treatment may have become non-adherent to the treatment regime shortly before the occurrence of indicators for adverse clinical outcomes. Patients' insight into treatment may underpin subsequent non-adherent behavior. Clinically, patients are often unwilling to report non-adherent behavior to their family and medical health workers. This allows fewer chances to offer effective interventions to improve a patient's compliance and furthermore to prevent a relapse. On the other hand, patients' insight into treatment is an underlying concept that endures for a long period (Yen et al., 2007), and it often happened before emergence of real non-adherent behavior. Similar findings of a schizophrenia study have been reported that acceptance of the need for medication predicted a good outcome (Rathod et al., 2005). In the treatment of bipolar disorder patients' insight into treatment is also a promising target for psychoeducation and psychotherapy (Kleindienst and Greil, 2004). Medical health providers can identify patients with lower levels of insight, even those with a status of good adherence to treatment, and offer effective interventions in advance. Although the scores on the three SAI insight dimensions were correlated with each other (Pearson's correlation = ), the dimensions of insight into recognition of the illness and re-labeling of psychotic phenomena had no predictive value in the uncontrolled Cox regression analyses. These results support that there are a variety of relationships between different dimensions of insight and adverse clinical outcomes with bipolar disorder. We suggest that insight should no longer be considered a single concept of a dichotomized symptom, as this limits its clinical value for predicting outcomes with bipolar disorder. Congruent with the results of previous studies (Kessing et al., 2005; Nolen et al., 2004), our secondary principal finding was that a greater number of previous hospitalizations for affective episodes is a predictor of adverse clinical outcomes. Patients with repeated hospitalizations are vulnerable to developing further mood disturbance attacks. Although a previous randomized controlled study indicated that therapy with combined mood stabilizers had a greater efficacy on prevention of relapse than monotherapy (Denicoff et al., 1997), this study found no difference in prediction of adverse clinical outcomes between the two treatment models. All patients recruited in this study had initially received monotherapy, and then another mood stabilizer or antipsychotic was added if the monotherapy did not work. Combined therapy not having been randomly assigned to these patients might partially explain the discrepancy between the studies. A previous prospective study on patients with schizophrenia had found that insight into treatment was associated with less hospitalization and better social adjustment, but insight into mental illness or psychotic experience was not (Yen et al., 2002b). One explanation for this positive association might be that such patients were more like to conform to prescribed drug schedules (Buchanan, 1992). Although patients with bipolar disorder

6 126 C.-F. Yen et al. / Journal of Affective Disorders 108 (2008) had different level of insight compared with those with schizophrenia (Pini et al., 2004; Yen et al., 2002a), the results of this study on the association between insight and clinical outcome in patients with bipolar I disorder were similar with those in patients with schizophrenia. This study further confirmed the clinical value of evaluating insight into treatment across different diagnoses. Some limitations of this study must be taken into account. First, sample size in this study is relatively small. Further studies are needed to re-examine the predictive value of insight for clinical outcomes in bipolar disorder in a large sample and in a sample with a short duration of illness. Second, some variables that have been found to be associated with outcomes in bipolar disorder, such as a family history of drug abuse and a history of child abuse were not controlled. Third, there may have been some patients who had affective symptoms but who were not admitted to a hospital, did not visit an emergency room, or who did not engage in suicidal or violent behavior. Lastly, we did not examine whether the kinds of drugs used to treat bipolar disorder have ever been switched or not in the two-year followup period, and the effect of drugs on the insight could not be determined in this study. However, our previous study has found that the level of insight of consistently stable patients with bipolar I disorder was quite steady, and frequent mood disturbance attacks may cause patients' insight to deteriorate (Yen et al., 2007). Role of funding source This study was supported by a grant (NSC H & NSC H ) from the National Science Council (NSC), Taiwan (ROC). The NSC had no further role in study design; in the collection, analysis and interpretation of data; in the writing of the report; and in the decision to submit the paper for publication. Conflict of interest No author of this manuscript has commercial or financial involvements that might pose a conflict of interest in connection with this manuscript. References Altshuler, L., Mintz, J., Leight, K., The Life Functioning Questionnaire (LFQ): a brief, gender-neutral scale assessing functional outcome. Psychiatry Res. 112, Amador, X.F., Flaum, M., Andreasen, N.C., Strauss, D.H., Yale, S.A., Clark, S.C., Gorman, J.M., Awareness of illness in schizophrenia and schizoaffective and mood disorders. Arch. Gen. Psychiatry 51, Amador, X.F., Friedman, J.H., Kasapis, C., Yale, S.A., Flaum, M., Gorman, J.M., Suicidal behavior in schizophrenia and its relationship to awareness of illness. Am. J. Psychiatry 153, Angst, J., Sellaro, R., Historical perspectives and natural history of bipolar disorder. Biol. Psychiatry 48, APA, In: Association, A.P. (Ed.), Diagnostic and Statistic Manual of Mental Disorders. American Psychiatric Association, Washington, D.C. Buchanan, A., A two-year prospective study of treatment compliance in patients with schizophrenia. Psychol. Med. 22, Buckley, P.F., Hrouda, D.R., Friedman, L., Noffsinger, S.G., Resnick, P.J., Camlin-Shingler, K., Insight and its relationship to violent behavior in patients with schizophrenia. Am. J. Psychiatry 161, Christensen, E., Multivariate survival analysis using Cox's regression model. Hepatology 7, David, A.S., Insight and psychosis. Br. J. Psychiatry 156, Denicoff, K.D., Smith-Jackson, E.E., Disney, E.R., Ali, S.O., Leverich, G.S., Post, R.M., Comparative prophylactic efficacy of lithium, carbamazepine, and the combination in bipolar disorder. J. Clin. Psychiatry 58, Feinstein, R., Plutchik, R., Violence and suicide risk assessment in the psychiatric emergency room. Compr. Psychiatry 31, Feldmann, T.B., Bipolar disorder and violence. Psychiatr. Q. 72, Ghaemi, S.N., Rosenquist, K.J., Is insight in mania statedependent?: A meta-analysis. J. Nerv. Ment. Dis. 192, Gitlin, M.J., Swendsen, J., Heller, T.L., Hammen, C., Relapse and impairment in bipolar disorder. Am. J. Psychiatry 152, Hamilton, M., A rating scale for depression. J. Neurol. Neurosurg. Psychiatry 23, Kessing, L.V., Sondergard, L., Kvist, K., Andersen, P.K., Suicide risk in patients treated with lithium. Arch. Gen. Psychiatry 62, Kleindienst, N., Greil, W., Are illness concepts a powerful predictor of adherence to prophylactic treatment in bipolar disorder? J. Clin. Psychiatry 65, Nolen, W.A., Luckenbaugh, D.A., Altshuler, L.L., Suppes, T., McElroy, S.L., Frye, M.A., Kupka, R.W., Keck Jr., P.E., Leverich, G.S., Post, R.M., Correlates of 1-year prospective outcome in bipolar disorder: results from the Stanley Foundation Bipolar Network. Am. J. Psychiatry 161, Pini, S., de Queiroz, V., Dell'Osso, L., Abelli, M., Mastrocinque, C., Saettoni, M., Catena, M., Cassano, G.B., Cross-sectional similarities and differences between schizophrenia, schizoaffective disorder and mania or mixed mania with mood-incongruent psychotic features. Eur. Psychiatry 19, Rathod, S., Kingdon, D., Smith, P., Turkington, D., Insight into schizophrenia: the effects of cognitive behavioural therapy on the components of insight and association with sociodemographics data on a previously published randomised controlled trial. Schizophr. Res. 74, Rihmer, Z., Kiss, K., Bipolar disorders and suicidal behaviour. Bipolar Disorders 4 (Suppl 1), Schumann, C., Lenz, G., Berghofer, A., Muller-Oerlinghausen, B., Non-adherence with long-term prophylaxis: a 6-year naturalistic follow-up study of affectively ill patients. Psychiatry Res. 89, Smith, T.E., Hull, J.W., Goodman, M., Hedayat-Harris, A., Willson, D.F., Israel, L.M., Munich, R.L., The relative influences of symptoms, insight, and neurocognition on social adjustment in schizophrenia and schizoaffective disorder. J. Nerv. Ment. Dis. 187,

7 C.-F. Yen et al. / Journal of Affective Disorders 108 (2008) Tohen, M., Waternaux, C.M., Tsuang, M.T., Outcome in Mania. A 4-year prospective follow-up of 75 patients utilizing survival analysis. Arch. Gen. Psychiatry 47, Tondo, L., Jamison, K.R., Baldessarini, R.J., Effect of lithium maintenance on suicidal behavior in major mood disorders. Ann. N.Y. Acad. Sci. 836, Yen, C.F., Chen, C.S., Yeh, M.L., Yen, J.Y., Ker, J.H., Yang, S.J., 2002a. Comparison of insight in patients with schizophrenia and bipolar disorder in remission. J. Nerv. Ment. Dis. 190, Yen, C.F., Yeh, M.L., Chen, C.S., Chung, H.H., 2002b. Predictive value of insight for suicide, violence, hospitalization, and social adjustment for outpatients with schizophrenia: a prospective study. Compr. Psychiatry 43, Yen, C.F., Chen, C.S., Yeh, M.L., Ker, J.H., Yang, S.J., Yen, J.Y., Correlates of insight among patients with bipolar I disorder in remission. J. Affect. Disord. 78, Yen, C.F., Chen, C.S., Ko, C.H., Yeh, M.L., Yang, S.J., Yen, J.Y., Huang, C.F., Wu, C.C., Relationships between insight and medication adherence in outpatients with schizophrenia and bipolar disorder: prospective study. Psychiatry Clin. Neurosci. 59, Yen, C.F., Chen, C.S., Ko, C.H., Yen, J.Y., Huang, C.F., Changes in insight among patients with bipolar I disorder: a twoyear prospective study. Bipolar Disorders 9, Young,R.C.,Biggs,J.T.,Ziegler,V.E.,Meyer,D.A.,1978.Aratingscale for mania: reliability, validity and sensitivity. Br. J. Psychiatry 133,