Schizophrenia Bulletin Advance Access originally published online on February 16, 2005
Schizophrenia Bulletin 2005 31(3):751-758; doi:10.1093/schbul/sbi016
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Expressed Emotion: Relevance to Rehospitalization in Schizophrenia Over 7 Years
Senior Psychologist, Anxiety Disorders Unit, Geha Mental Health Center, Rabin Medical Center, Beilinson Campus, Petah Tikva, Israel; and Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel
former Head of Geha Mental Health Center, Rabin Medical Center; and Clinical Professor of Psychiatry at the Sackler Faculty of Medicine, Tel Aviv University
Professor of Psychiatry and Head of the Department of Psychiatry, University of Cambridge, Cambridge, U.K
Head of the Research Department, Geha Mental Health Center, Rabin Medical Center; Head of the Laboratory of Biological Psychiatry, Felsenstein Medical Research Center, Rabin Medical Center, Beilinson Campus; and Professor of Psychiatry at Sackler Faculty of Medicine, Tel Aviv University
Director of the Outpatient Department, Geha Mental Health Center, Rabin Medical Center; and Senior Lecturer of Psychiatry, Sackler Faculty of Medicine, Tel Aviv University
 |
Abstract |
|---|
 Top Abstract MethodsResultsDiscussionConclusions and Clinical...References |
|---|
Expressed emotion (EE) is an established factor in short-term relapse in schizophrenia. However, data on its long-term predictive ability are scarce. We extended our short-term investigation over 7 years' followup. The study population consisted of 108 patients, 93 with schizophrenia and 15 with schizoaffective disorder. EE of the key relatives was rated with the Five Minute Speech Sample (FMSS). Patient households were categorized by EE and its two components: criticism (CR) and emotional overinvolvement (EOI). High CR was associated with earlier first and second readmissions (Breslow p = 0.002 and 0.04, respectively). High CR was associated with a higher rate of readmissions (p = 0.01) and a longer hospital stay (p = 0.02) compared with low CR. Both compliance with pharmacotherapy and the interaction of high-CR x poor compliance were additional contributors to time to first readmission. This study is the first to demonstrate the prolonged predictive validity of EE. Our results support the value of CR as a prognostic indicator of the course of schizophrenia. The FMSS appears to have predictive power in respect to psychiatric hospitalization. Therapies aimed at lowering high EE seem warranted as a long-term preventive approach.
Keywords: Expressed emotion (EE) / long-term course / family relatives / hospitalizations
Expressed emotion (EE) is a significant and robust predictor in schizophrenia (Butzlaff and Hooley 1998). Prospective studies of schizophrenia have shown that patients living in families characterized by high levels of EE were significantly more likely to relapse than those residing in low-EE households (Kavanagh 1992; Bebbington and Kuipers 1994; Butzlaff and Hooley 1998). The crucial components of the EE construct were identified as criticism (CR) and emotional overinvolvement (EOI) (Leff and Vaughn 1985). Although the majority of studies treated EE as a unitary construct, some attempted to determine the differential predictive power of its two components (Barrelet et al. 1990; Stirling et al. 1993; Chambless and Stekete 1999; King and Dixon 1999). High EE was usually associated with high levels of CR (Vaughn and Leff 1976; MacMillan et al. 1986; Marom et al. 2002), and CR made a greater contribution to relapse (MacMillan et al. 1986; Barrelet et al. 1990; Lopez et al. 1999; Marom et al. 2002).
Most of the EE studies examined patients for followup periods of an average of 9 to 12 months and up to 24 months at most (Kavanagh 1992), with only a very few longitudinal studies of over 2 years' duration (McCreadie et al. 1993; Tarrier et al. 1994; Huguelet et al. 1995; Schulze Monking et al. 1997). In the study of Tarrier et al. (1994), 24 patients who had not relapsed at 2 years after discharge from index admission were followed up. High-EE subjects had more relapses than patients from low-EE households at both 5- and 8-year followup after discharge from index admission, but this difference did not reach significance, probably because of the small sample size. The 5-year results of McCreadie et al. (1993), based on half of their initial sample of 60 patients, support the notion of persistent association between higher relapse rate and high EE. Huguelet et al. (1995) conducted a 5-year followup study of 44 first admission patients with schizophrenia and noted a greater tendency toward more frequent hospitalizations in patients from high-EE than from low-EE families. Finally, Schulze Monking et al. (1997) found in an 8-year followup study that patients with schizophrenia from high-EE households had been hospitalized more often and for longer periods than patients from low-EE households, although this held true in only the subgroup of 28 chronic patients who had been ill for more than 4.5 years at baseline. However, these authors assessed baseline EE when patients had had an outpatient status for at least 8 weeks after discharge from index hospitalization, whereas the other studies examined EE at admission. Therefore, the results are hardly comparable.
The aim of the present study was to determine whether EE can also predict long-term outcome. This is an extension of our previous 9-month study (Marom et al. 2002). We hypothesized that high familial EE, compared with low familial EE, would be associated with worse indexes of hospitalization during a 7-year followup. Specifically, we intended to examine the effects of EE during a 6.25-year period beginning after the end of the first 9-month followup (Butzlaff and Hooley 1998).
Compliance with medication was evaluated to clarify the inconclusive findings to date with respect to its confounding effect on the predictive power of EE (Moline et al. 1985; Nuechterlein et al. 1986; Vaughan et al. 1992; Sellwood et al. 2003).
|
Methods |
|---|
|
TopAbstractMethodsResultsDiscussionConclusions and Clinical...References |
|---|
Subjects
The initial study group consisted of 114 consecutive patients newly admitted to Geha Mental Health Center between March 1993 and September 1994. All were hospitalized in an acute locked ward. Inclusion criteria were as follows: (1) diagnosis of schizophrenia or schizoaffective disorder according to DSM–III–R (APA 1987) criteria at admission; (2) psychotic state defined as a score of 4 or more on at least one of the following Brief Psychiatric Rating Scale (BPRS) (Overall and Gorham 1962) positive items: conceptual disorganization (#4), suspiciousness (#11), hallucinatory behavior (#12), or unusual thought content (#15); (3) age 20 to 65 years; (4) residence with a first degree relative for at least 1 year prior to index admission or residence alone but spending at least 35 hours per week with a close relative; (5) main language Hebrew; (6) absence of significant chronic physical disease; and (7) no history of or current substance abuse.
Six patients were later excluded from the analysis, three because they were transferred during index hospitalization to a long-term care hospital and three because they were still hospitalized at the data collection. Thus, the final sample included 108 patients. During the 7 years of followup, 4 of these patients (2 from low-EE and 2 from high-EE households) died, and they were included in the analyses under "intent to treat."
A total of 151 key relatives were interviewed to assess familial EE: for 60.2 percent of the patients, a single relative was interviewed (n = 65), and for the remainder, two relatives were interviewed (n = 86). Relatives were predominantly parents (66.7%), with a minority of spouses. For the patients for whom two relatives were interviewed, households were classified as high-EE when either one relative or both satisfied the criteria for individual high-EE status. For patients with only a single relative, the EE status of the interviewee was taken as the EE status of the household. These criteria are in accordance with the accumulated EE literature (Kavanagh 1992). No significant difference was found between number of relatives interviewed in high- and low-EE groups (
2 = 0.79, df = 1, p = 0.37).
The study was approved by the Geha Mental Health Center Review Board and Ethics Committee, and written informed consent was obtained after the procedure had been fully explained to the subjects.
Procedure
A prospective, longitudinal design was used. Within the first 2 days of admission, the patients underwent a psychiatric interview according to the guidelines of the Structured Clinical Interview for DSM–III–R (Spitzer et al. 1989) by two senior psychiatrists to establish the diagnosis, also using data from medical records. A consensus between the two senior psychiatrists was required for diagnosis, and it was reconfirmed close to discharge. The BPRS was administered by the treating therapist within the first 2 days of admission, at discharge, and 6 months after discharge. Within the first 4 days of admission, the relatives underwent an individual structured interview, conducted by the treating therapist, to collect background and medical data. The relatives completed the Five Minute Speech Sample (FMSS) (Magana et al. 1986), which was administered and rated by a qualified EE researcher (S.M.), to determine familial EE.
Patients were followed for 7 years after discharge. According to national hospitalization regulations, readmissions, when necessary, were mostly to Geha Mental Health Center, where readmission was determined by senior psychiatrists who were blinded to patients' household EE status. Psychiatric hospitalizations in Israel are free of charge, and there was only occasional case management.
Measures
Assessment of EE.
The administration, coding, and rating of the EE index were conducted according to the Manual for Coding Expressed Emotions From the Five Minute Speech Sample (Magana 1990). The interrater reliability of the FMSS was established by two independent raters (S.M. and another qualified rater) on a sample of the first 20 relatives, and kappa values (Cohen 1960) were 
= 0.86 for global EE, = 0.86 for CR, and = 0.76 for EOI.
Outcome measures. Outcome was measured by three variables: (1) time from discharge from index hospitalization to first and second readmissions to a psychiatric hospital; (2) total number of psychiatric admissions after index admission; and (3) total length of stay at psychiatric hospitals after discharge from index hospitalization. Both the demographic and the outcome clinical data were obtained from structured interviews conducted at admission and the patient psychiatric hospitalization file at the Central Registry of the Israel Ministry of Health, which lists the number and total duration of all psychiatric hospitalizations in the country.
Compliance with medication during the 2 years prior to index admission, determined by anamnesis and patient hospital file, was rated by the treating therapist as high (patient usually followed the therapist's instructions) or low (patient had not been taking the prescribed medication at least half of the time or usually did not take the medication). All patients except one had a previous history of prescribed neuroleptic medication.
Statistical Analysis. Low-EE and high-EE groups were compared by chi-square test; unpaired, 2-tailed t test; univariate analysis of variance; and Mann-Whitney test as appropriate, followed by Bonferroni test. Time to first and second readmissions according to EE was examined by survival analyses (Kaplan-Meier) (Lee 1992). The contribution of the variables that affected time to first readmission was examined by Cox regression analysis.
|
Results |
|---|
|
TopAbstractMethodsResultsDiscussionConclusions and Clinical...References |
|---|
The demographic data of the whole sample have been reported previously (Marom et al. 2002). The illness data of the whole sample and of the low-EE and high-EE groups are shown in table 1. Categorization as low EE or high EE proved to be unrelated to demographic variables, to clinical characteristics, and to compliance with medication. Half the patients were classified as compliant with medication and half as noncompliant. Ninety-three patients (86.1%) had a diagnosis of schizophrenia, and 15 (13.9%) had a diagnosis of schizoaffective disorder. Patients from low-EE households were younger than patients from high-EE households (32.9 [standard deviation (SD) = 9.4] vs. 38.1 [SD = 11.7]; t = 2.55, df = 106, p < 0.05). In addition to the Bonferroni correction, we conducted analysis of covariance with patient age as the covariate, which did not affect the analyses of parametric variables. Moreover, Cox regression analysis of readmission rate yielded nonsignificant results with respect to patients' age. Thus, differences between patients' ages did not seem to significantly affect the impact of familial EE on the outcome.
|
Classification by familial EE yielded two groups: low EE, 56 households (51.9%); and high EE, 52 households (48.1%). Classification by the two EE components yielded four subgroups: low CR-low EOI, 56 households (51.9%); high CR-low EOI, 31 households (28.7%); low CR-high EOI, 13 households (12.0%); high CR-high EOI, 8 households (7.4%). Slightly more than one-third of the households were classified as high CR, and about one-fifth as high EOI.
During the 9 months following discharge, 36 of 108 patients (33.3%) were readmitted for the first time; additionally, during the 6.25 years up to the end of the 7-year followup, 33 (30.6%) more patients were readmitted for the first time. Thus, during the 7 years following discharge, 69 of 108 patients (63.9%) were readmitted at least once: 34 of 52 patients (65.4%) from high-EE households and 35 of 56 (62.5%) from low-EE households. Comparison of the patients by global EE status with respect to time to first readmission yielded a nonsignificant difference between the survival curves by log-rank test (p = 0.23) and a trend for association by Breslow test (p = 0.06). Analysis by the EE components showed that 29 of 39 (74.4%) patients from high-CR households were readmitted, compared with 40 of 69 (57.9%) from low-CR households. On survival analysis by the CR component only (figure 1), the low-CR households had a significantly higher proportion of patients who were not readmitted than the high-CR households (log rank p = 0.008, Breslow p = 0.002). There was no significant difference in the survival curves based on EOI (log rank p = 0.34, Breslow p = 0.49). We also conducted a survival analysis of patients whose duration of illness at index admission was less than 5 years. This group constituted of 28 patients, 22 from low-CR households and 6 from high-CR households. On survival analysis by CR, the low-CR households had a significantly higher proportion of patients who were not readmitted than did the high-CR households (log rank p = 0.016, Breslow p = 0.007). Thus, our findings held similarly for patients with both short and chronic duration of illness.
|
A second survival analysis was conducted to examine whether second readmissions could be predicted by level of EE. Both low-EE and low-CR households had a significantly higher proportion of patients who were not readmitted for the second time than the high-EE and high-CR households (log rank and Breslow p = 0.04 and 0.04, respectively) (figure 2). Patients from high-CR households were readmitted more often compared with patients from low-CR households (25/29 [86.2%] and 26/40 [65%], respectively). There was no significant difference in the survival curves based on EOI by either log rank or Breslow test (p = 0.9 for both).
|
Among patients readmitted for the first time (n = 69), mean and median duration to first readmission were 21 and 9 months, respectively. Among patients with second readmissions (n = 51), mean and median duration from first to second readmission were 18 and 6.4 months, respectively.
Cox regression analysis was conducted to determine the contribution of the various variables, including EE status, to readmission over time. The analysis was done for CR, which had proved to be the significant EE component associated with time to first and second readmissions. CR, compliance with medication, and the interaction CR x compliance each predicted outcome (table 2).
|
On examination of each of the four combinations of level of CR and compliance status by survival analysis (figure 3), a significant difference was observed among the survival curves (log rank p = 0.005, Breslow p = 0.001). The interaction EE x compliance was similarly significant in predicting readmission. Survival analysis of each of the combinations of level of EE and compliance status yielded significant difference among the survival curves (log rank p = 0.05, Breslow p = 0.02).
|
For the 69 patients who were readmitted at least once, we conducted a second survival analysis to examine whether time to second readmission could be predicted by EE and compliance status. The four combinations of level of CR and compliance status did not differ significantly (figure 4). However, the survival curves were best predicted by the CR status and not by the compliance status. At the second readmission, patients of low-CR families, irrespective of compliance status, are less often readmitted, whereas those with high-CR status are more often readmitted.
|
Patients from high-CR households had a significantly higher number of psychiatric readmissions and longer cumulative length of stay at psychiatric hospitals than patients from low-CR households during the 7 years after the index hospitalization (U = 959, p = 0.01; U = 975, p = 0.02, respectively) (table 3). We also analyzed these data among readmitted patients only (table 3). In this subgroup, patients from high-CR households had a significantly higher number of psychiatric readmissions than patients from low-CR households during the 7 years after the index admission (U = 414, p = 0.04). Readmitted patients from high-CR households had a longer cumulative length of stay at psychiatric hospitals than readmitted patients from low-CR households during those years, but the difference was not significant (U = 440, p = 0.12).
|
|
Discussion |
|---|
|
TopAbstractMethodsResultsDiscussionConclusions and Clinical...References |
|---|
A major new finding of the present study was that index EE status is associated significantly with time to both first and second readmissions during the 7 years after discharge from the index hospitalization. Additionally, high CR at the time of index admission was prospectively related to higher number and longer length of hospitalizations during the 7 years. Taken together, the three results indicate that the well-established association between EE and relapse (Butzlaff and Hooley 1998) is not restricted to the first 9 months following discharge. Moreover, the significantly higher readmission rates among high-CR household patients held also for patients with a nonchronic illness at index admission. These results are novel in the sense that they extend the conclusions of many previous studies of shorter duration. Thus, EE appears to have harmful effects and to influence illness outcome for extended periods. There are only scarce data in the literature for comparison to our results because of the paucity of other long-followup EE studies. The current results support our prospective hypothesis that psychotic patients from high-EE households have a worse long-term prognosis than patients from low-EE households.
Further examination of the predictive validity of EE and its components with respect to time to readmission showed that high CR alone, as well as low compliance with medication, and the interaction CR x compliance with medication, were strongly related to outcome. Moreover, the survival curves of the second readmission, taking into consideration the compliance status (figure 4), although not significant, show that patients with low-CR status were less often readmitted, irrespective of their compliance level. In contrast, patients from the two high-CR subgroups were more often readmitted. Thus, CR appears to be as important to long-term outcome as compliance, namely adherence, with medications. These results provide additional evidence that familial EE is currently one of the powerful contributors to relapse over the short term, 9 months (Butzlaff and Hooley 1998), as well as over 7 years.
The present findings clearly support previous reports (Hogarty 1993) that good compliance with medication is an important preventive factor in relapse but alone is often not sufficient to counteract the burden of a stressful family. Our findings indicate that patients from high-CR households are at greater risk of relapse for the long term, even if they take their medication regularly. Our study stresses the long-term influence of the interaction CR x compliance with medication, literally the social and pharmacotherapeutic factors, in the daily life of schizophrenia patients.
Regarding readmission rate, our previous results (Marom et al. 2002) support its utility as a reasonable measure of relapse. In that short-term study with the same population, readmission was positively associated with deterioration in the mental status of patients, as reflected by significantly higher BPRS score at a 6-month followup. Similar results were obtained in other countries, such as Italy (Bertrando et al. 1992) and Australia (Parker et al. 1989), although for much shorter followup periods.
We based our prediction of outcome on data obtained at the arbitrary index admission. However, during 7 years of followup, there could be changes in EE levels of relatives, in familial living arrangements, and in patients' compliance with medication. Nevertheless, index EE level maintained its predictive validity regardless of probable changes in these factors among some patients and their relatives. This may suggest that the index high CR alone has at least a partially detrimental effect.
The predictive value of the initial measure of EE over a 7-year period does not necessarily resolve the debate on the direction of cause and effect. Rosenfarb et al. (2000) reported the oversusceptibility of schizophrenia patients with working memory deficits to develop psychotic thinking in reaction to interpersonal CR. Patients with specific cognitive deficits may need more hospitalizations when they live in deleterious highly critical families. We cannot rule out the possibility that poor prepsychosis characteristics and poor functioning alone may result in poorer outcome and elicitation of high EE in family members. Our results are best interpreted according to the model of reciprocal interaction between an increase in CR and psychotic symptomatology (Kavanagh 1992) or between interpersonal CR and neurocognitive vulnerability (Rosenfarb et al. 2000).
Our study highlights the value of EE in the prediction of the long-term hospitalization indexes of schizophrenia patients. The findings also suggest, in line with earlier ones (although these were related to short outcomes [Barrelet et al. 1990; Vaughan et al. 1992]), the importance of CR. Our results indicate that treating EE as a unitary index conceals the specific contribution of CR. The study suggests that CR may warrant more detailed consideration in EE research and that the separate roles of CR and EOI as targets of family interventions should be considered.
Clearly, there are limitations to defining outcome merely by time to readmission or by number and length of hospitalizations. Readmission is only a global indicator of outcome and not necessarily identical to emergence of a new major psychotic episode. However, the fact that hospitalizations in Israel are free and that there is a single central registry that allows reliable checking for inpatient admissions across the whole country may suggest that readmissions are good proxies for relapse. Moreover, neither case management nor private hospitalization is widely available in Israel. In addition, optimally, readmitted and non-readmitted patients should have been clinically examined. Thus, our findings on the predictive value of EE are confined to parameters of hospitalization. Additionally, our binary definition of compliance with medication might be too crude and insufficiently sensitive. Another limitation relates to the fact that we do not have information on the stability of EE levels, household arrangements, and compliance with medication during the followup period.
The FMSS (Magana et al. 1986) was used to assess EE. Although the Camberwell Family Interview (Vaughn and Leff 1976) may have yielded a higher rate of high EE and high EOI (Magana et al. 1986), our probable underdetection did not affect the predictive power of the EE indexes, although it might be relevant to local cultural features (e.g., relative frequency of the EE components). Research demonstrating the predictive validity of the FMSS is still scarce (e.g., Otsuka et al. 1994; Uehara et al. 1997). Our present study indicates its predictive validity over a long followup period.
The sample size in our study was significantly larger than in the majority of previous studies of EE. In addition, duration of illness was longer and mean number of prior admissions was larger. Thus, our study supports the prognostic value of high EE in the presumably difficult-to-treat chronic population of schizophrenia patients for long followup periods.
|
Conclusions and Clinical Implications |
|---|
|
TopAbstractMethodsResultsDiscussionConclusions and Clinical...References |
|---|
Our finding that the association between EE and outcome persists over a considerable period of time extends the validity of EE to the long-term clinical course of patients with schizophrenia and schizoaffective disorder. Moreover, it demonstrates the value of EE as a predictor of the clinical course of schizophrenia. The FMSS may serve as a convenient EE screening instrument to detect familial high CR and as a predictor of psychiatric rehospitalization. Further examination should focus on the differential role of the EE components in order to gain optimal outcome in familial psychoeducational interventions.
|
Footnotes |
|---|
Send reprint requests to Dr. S. Marom, Geha Mental Health Center, P.O. Box 102, Petah Tikva 49100, Israel; e-mail: smarom{at}clalit.org.il .
1 Significant difference between the groups (log rank p = 0.008, Breslow p = 0.002). 
1 Significant difference between the groups (log rank and Breslow p = 0.04 for both).
1 Significant difference between the groups (log rank p = 0.005, Breslow p = 0.001).
1 Nonsignificant difference between the groups (log rank p = 0.21, Breslow p = 0.23).
|
Acknowledgments |
|---|
The authors thank Ilana Gelernter, M.A., and Yoav Benjamini, Ph.D., from the Statistical Laboratory, Sackler School of Mathematics, Tel Aviv University, for the statistical consultation.
|
References |
|---|
|
TopAbstractMethodsResultsDiscussionConclusions and Clinical...References |
|---|
-
American Psychiatric Association. DSM–III–R: Diagnostic and Statistical Manual of Mental Disorders 3rd ed., Washington, DC: APA 1987 revised.
Barrelet, L, Ferrero, F, Szigethy, L, Giddey, C, Pellizer, G. Expressed emotion and first admission schizophrenia: Nine-month follow-up in a French cultural environment. British Journal of Psychiatry 1990; 156:357–362.
Bebbington, P and Kuipers, L. The predictive utility of expressed emotion in schizophrenia: An aggregate analysis. Psychological Medicine 1994; 24:707–717.[Web of Science][Medline]
Bertrando, P, Beltz, J, Bressi, C, Clerici, M, Farma, T, Invernizzi, G, Cazzullo, CL. Expressed emotion in schizophrenia in Italy: A study of an urban population. British Journal of Psychiatry 1992; 161:223–229.
Butzlaff, AM and Hooley, JM. Expressed emotion and psychiatric relapse. Archives of General Psychiatry 1998; 55:547–552.
Chambless, DL and Stekete, G. Expressed emotion and behavior therapy outcome: A prospective study with obsessive compulsive and agoraphobic patients. Journal of Consulting and Clinical Psychology 1999; 67:658–665.[CrossRef][Web of Science][Medline]
Cohen, J. A coefficient of agreement for nominal scales. Educational and Psychological Measurement 1960; 20:37–46.[CrossRef][Web of Science]
Hogarty, GE. Prevention of relapse in chronic schizophrenia patients. Journal of Clinical Psychiatry 1993; 54:18–23.
Huguelet, P, Favre, S, Binyet, S, Gonzales, C, Zabala, I. The use of expressed emotion index as a predictor of outcome in first admitted schizophrenic patients in French speaking area of Switzerland. Acta Psychiatrica Scandinavica 1995; 92:447–452.[Web of Science][Medline]
Kavanagh, DJ. Recent development in expressed emotion and schizophrenia. British Journal of Psychiatry 1992; 160:601–620.
King, S and Dixon, MJ. Expressed emotion and relapse in young schizophrenia outpatients. Schizophrenia Bulletin 1999; 25:377–386.
Lee, ET. Statistical Methods for Survival Data Analysis New York, NY: John Wiley 1992.
Leff, J and Vaughn, C. Expressed Emotion in Families: Its Significance for Mental Illness New York, NY: Guilford Press 1985.
Lopez, SR, Nelson, KA, Snyder, KS, Mintz, J. Attributions and affective reactions of family members and course of schizophrenia. Journal of Abnormal Psychology 1999; 108:307–314.[CrossRef][Web of Science][Medline]
MacMillan, JF, Gold, A, Crow, TJ, Johnson, AL, Johnstone, EC. The Northwick Park study of first episodes of schizophrenia: IV. Expressed emotion and relapse. British Journal of Psychiatry 1986; 148:133–143.
Magana, AA. Manual for Coding Expressed Emotions From the Five Minute Speech Sample Los Angeles, CA: UCLA Family Project 1990.
Magana, AB, Goldstein, MJ, Karno, M, Milkowitz, DJ, Jenkins, J, Falloon, IRH. A brief method for assessing expressed emotion in relatives of psychiatric patients. Psychiatry Research 1986; 17:203–212.[CrossRef][Web of Science][Medline]
Marom, S, Munitz, H, Jones, PB, Weizman, A, Hermesh, H. Familial expressed emotion: Outcome and course of Israeli patients with schizophrenia. Schizophrenia Bulletin 2002; 28:731–743.
McCreadie, RG, Robertson, LJ, Hall, D, Berry, II. The Nithdale schizophrenia surveys: XI. Relatives' expressed emotion. Stability over five years and its relation to relapse. British Journal of Psychiatry 1993; 162:393–397.
Moline, RA, Singh, S, Morris, A, Meltzer, HY. Family expressed emotion and relapse in schizophrenia in 24 urban American patients. American Journal of Psychiatry 1985; 142:1078–1081.
Nuechterlein, KH, Snyder, KS, Dawson, ME, Rappe, S, Gitlin, M, Fogelson, D. Expressed emotion, fixed dose fluphenazine decanoate maintenance and relapse in recent-onset schizophrenia. Psychopharmacology Bulletin 1986; 22:633–639.[Web of Science][Medline]
Otsuka, T, Nakane, Y, Ohta, Y. Symptoms and social adjustment of schizophrenic patients as evaluated by family members. Acta Psychiatrica Scandinavica 1994; 89:111–116.[Web of Science][Medline]
Overall, JE and Gorham, DR. The Brief Psychiatric Rating Scale. Psychological Reports 1962; 10:799–812.[Web of Science]
Parker, G, Hayward, L, Johnston, P. Factorial validity of the EE scales. Psychological Medicine 1989; 19:435–446.[Web of Science][Medline]
Rosenfarb, IS, Nuechterlein, KH, Goldstein, MJ, Subotnik, KL. Neurocognitive vulnerability, interpersonal criticism, and the emergence of unusual thinking by schizophrenic patients during family transactions. Archives of General Psychiatry 2000; 57:1174–1179.
Schulze Monking, H, Hornung, WP, Stricker, K, Buchkremer, G. Expressed emotion in an 8 year follow-up. European Psychiatry 1997; 12:105–110.[CrossRef]
Sellwood, W, Tarrier, N, Quinn, J, Barrowclough, C. The family and compliance in schizophrenia: The influence of clinical variables, relatives' knowledge and expressed emotion. Psychological Medicine 2003; 33:91–96.[CrossRef][Web of Science][Medline]
Spitzer, RL, Williams, JBW, Gibbon, N. Structured Clinical Interview for DSM–III–R New York, NY: Biometrics Research Division, New York State Psychiatric Institute 1989.
Stirling, J, Tantam, D, Thomas, P, Newby, D, Montague, L, Ring, N, Rowe, S. Expressed emotion and schizophrenia: The ontogeny of EE during an 18-month follow-up. Psychological Medicine 1993; 23:771–778.[Web of Science][Medline]
Tarrier, N, Barrowclough, C, Porceddu, K, Fitzpatrick, E. The Salford Family Intervention Project: Relapse rates of schizophrenia at five and eight years. British Journal of Psychiatry 1994; 165:829–832.
Uehara, T, Yokoyama, T, Goto, M, Kohmura, N, Nakano, Y, Toyooka, K, Ihda, S. Expressed emotion from the Five-Minute Speech Sample and relapse of outpatients with schizophrenia. Acta Psychiatrica Scandinavica 1997; 95:454–456.[Web of Science][Medline]
Vaughan, K, Doyle, M, McConaghi, N, Blaszcynski, A, Fox, A, Tarrier, N. The relationship between relatives' expressed emotion and schizophrenic relapse: An Australian replication. Social Psychiatry and Psychiatric Epidemiology 1992; 27:10–15.[CrossRef][Web of Science][Medline]
Vaughn, CE and Leff, JP. The influence of family and social factors on the course of psychiatric illness. British Journal of Psychiatry 1976; 129:125–137.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  W. A. Vega, M. Karno, M. Alegria, J. Alvidrez, G. Bernal, M. Escamilla, J. Escobar, P. Guarnaccia, J. Jenkins, A. Kopelowicz, et al. Research Issues for Improving Treatment of U.S. Hispanics With Persistent Mental Disorders Psychiatr Serv, March 1, 2007; 58(3): 385 - 394. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. I. Velligan, Y.-W. F. Lam, D. C. Glahn, J. A. Barrett, N. J. Maples, L. Ereshefsky, and A. L. Miller Defining and Assessing Adherence to Oral Antipsychotics: A Review of the Literature Schizophr Bull, October 1, 2006; 32(4): 724 - 742. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||