The Journal of Nervous and Mental Disease (C) 2000 Lippincott Williams & Wilkins, Inc. Volume 188(3), March 2000, pp 135-140 A Critical Time Intervention with Mentally III Homeless Men: Impact On Psychiatric Symptoms [Articles] HERMAN, DANIEL D.S.W.1; OPLER, LEWIS M.D., Ph.D.2; FELIX, ALAN M.D.3; VALENCIA, ELIE J.D., M.A.4; WYATT, RICHARD JED M.D.5; SUSSER, EZRA M.D., Dr.PH.6 1 New York State Psychiatric Institute and Department of Psychiatry, Columbia University College of Physicians and Surgeons, 100 Haven Avenue, #31F, New York, New York 10032. Send reprint requests to Dr. Herman. 2 Albert Einstein College of Medicine and New York State Office of Mental Health. 3 Columbia Presbyterian Medical Center and Department of Psychiatry, Columbia University College of Physicians and Surgeons. 4 Joseph L. Mailman School of Public Health, Columbia University. 5 Neuropsychiatry Branch, National Institute of Mental Health and Department of Psychiatry, Columbia University College of Physicians and Surgeons. 6 New York State Psychiatric Institute and Joseph L. Mailman School of Public Health, Columbia University. This work was supported by grant MH-48041 (E. Valencia, PI) from the National Institute of Mental Health and NIMH Research Scientist Development Award MH-01204 and a NARSAD Young Investigator Award to Dr. Herman. ---------------------------------------------- Outline Abstract Methods Subjects Intervention Measures Data Analysis Results Discussion Limitations Conclusions References Graphics Table 1 Table 2 Abstract We describe the impact of a psychosocial intervention, critical time intervention (CTI), on the cardinal symptom dimensions of schizophrenia, namely negative, positive, and general psychopathology. Ninety-six men with schizophrenia and other psychotic disorders who were discharged from a homeless shelter were randomly assigned to receive either CTI or usual services only. CTI is a time-limited intervention designed to enhance continuity of care during the transition from institution to community. Symptom severity at baseline and at 6 months was assessed using the Positive and Negative Syndrome Scale. Using data on 76 subjects for whom we have complete symptom data, we assessed the impact of CTI on change in symptoms. The results suggest that CTI was associated with a statistically significant decrease in negative symptoms at the 6-month follow-up, reflecting modest clinical improvement. There was no significant effect on positive or general psychopathology symptoms. ---------------------------------------------- In an earlier publication, we reported the results of a randomized clinical trial demonstrating the effectiveness of critical time intervention (CTI) in preventing recurrent homelessness among homeless men with schizophrenia and other psychotic disorders (Susser et al., 1997). Here we describe the impact of this psychosocial intervention on the cardinal symptom dimensions of schizophrenia, namely negative, positive, and general psychopathology. Typically, traditional antipsychotic medications produce significant improvement in hallucinations, delusions, and disordered thinking-so-called positive symptoms-whereas they are relatively ineffective against the apathy, blunted affect, and social withdrawal that characterize the negative symptom domain in schizophrenia. Recent studies have suggested that the newer antipsychotics, in contrast to standard or conventional antipsychotics, might treat negative and general psychopathology symptoms in addition to positive symptoms (Beasley et al., 1997; Chouinard et al., 1993; Marder and Meibach 1994). Although few studies have focused specifically on the impact of psychosocial interventions on symptom clusters in schizophrenia, a recent meta-analysis of the effectiveness of psychosocial treatments in schizophrenia (Mojtabai et al., 1998) suggests that such interventions, in conjunction with somatic treatment, may have a greater effect on negative symptoms than on positive or general psychopathology symptoms. Discontinuity between institutional and community treatment is a critical problem (Dorwart and Hoover 1994; Olfson et al., 1998) that has contributed to a range of negative outcomes among persons with severe mental illness after institutional care. CTI is a time-limited intervention that overlaps the period before and after discharge from the institution to the community. A key aspect of CTI is that the post-discharge phase of the intervention is delivered by staff who have established relationships with patients during their institutional stay. CTI shares with long-term assertive community treatment models (Lehman et al., 1997) a focus on stabilizing patients in the community through "in vivo" development of independent living skills and by building effective support networks in the community. The emphasis, however, is on maintaining continuity of care during a critical period of transition, whereas primary responsibility is gradually passed to the existing supports in the community. In our randomized clinical trial of CTI among severely mentally ill men being discharged from a shelter to community living, we found the intervention to be associated with a significant, lasting reduction in post-discharge homelessness. Although CTI was designed specifically to prevent homelessness, we speculated that it might also have a favorable impact on important clinical outcomes. In the analyses below, we examine whether, in the context of treatment with maintenance antipsychotic pharmacotherapy, this psychosocial intervention is associated with differential effects on positive, negative, and general psychopathology symptoms, respectively, over a 6-month follow-up period. Methods Subjects The methods of the CTI study have been described in detail in our previous report (Susser et al., 1997). Briefly, the subjects were patients discharged from an on-site psychiatry program in a New York City men's shelter to community-based housing. All patients of the program had severe mental illness, usually schizophrenia or other psychotic disorders such as bipolar disorder. The vast majority were African-Americans and Latinos. The on-site program, described in detail elsewhere (Valencia et al., 1996; Valencia et al., 1997), treated 60 to 80 men at any given time, providing outreach and rehabilitation services and preparing them for housing in the community. Patients who completed the on-site program had access to a broad spectrum of supportive housing in New York City, ranging from intensively supervised community residences to single-room-occupancy hotels with on-site social services. Nonetheless, a significant minority of men could not or would not be accommodated in any of these settings and were discharged to family, friends, or miscellaneous other arrangements. Over a 2-year period (1991 to 1993), all 102 men who were being discharged to housing in the New York City region were invited to participate in the clinical trial. Ninety-six (94%) chose to participate and signed an informed consent form. After their housing placement had been selected, these men were randomly assigned to receive either CTI or usual services only (USO). The men in the CTI group received 9 months of CTI plus usual services and then usual services only for the following 9 months. The men in the USO group received usual services only for the entire 18 months of the study. Complete follow-up data on homelessness-the study's primary outcome-was obtained on the entire sample. To maximize follow-up on the primary outcome in the context of a limited budget, symptom ratings and other measures of clinical outcome were not collected on subjects enrolled during the latter stages of the study. We report here on the 76 subjects (38 in CTI and 38 in USO) for whom we have complete symptom data at baseline and at 6-month follow-up. When we compared these 76 subjects with the remaining 20 subjects on demographic, diagnostic, and homelessness variables, we found no significant differences between groups. Intervention CTI has two components. The first is to strengthen the individual's long-term ties to services, family, and friends. These supports are potentially available in varying measures, but mentally ill individuals and those upon whom they depend often need assistance to work with one another. The second component of CTI is providing emotional and practical support during the critical time of transition. During a stay in an institution such as a large shelter, an individual may develop strong ties and become used to having many of his social and physical needs met on-site. The return to a possibly isolated and destitute life in the community, however welcome, can be emotionally painful. For each man randomly assigned to CTI, the clinical team devised a plan for the transfer of care to other formal and informal supports. For this purpose, the team identified one or two specific areas of potential discontinuity that were related to the risk of homelessness for that individual, and in which intervention was likely to be effective in preventing homelessness, for instance, medication adherence or money management. Each man was then assigned to a "CTI worker" to implement the plan. This worker did not need to have a professional degree but did need to have experience working with this population and enough "street smarts" to work in vivo with these men in the community. A psychiatrist or other mental health professional supervised the CTI worker. To implement the first component of CTI, the strengthening of long-term ties, this CTI worker provided support for both the man and those who could assist him in the areas of treatment that were selected. This generally entailed visiting the family home or community residence, being present at appointments, and giving advice in crises. The CTI worker often functioned as a mediator, helping to ensure that the man and his supports communicated until they had negotiated workable relationships between themselves. When this approach failed, the CTI worker could intervene directly. Occasionally, the man was transferred to another community service. He could also be relocated to another residence where his particular form of disturbed behavior was more likely to be tolerated. To implement the second component of CTI, support during a critical period, during the first 2 weeks after discharge the CTI worker spent time with the patient in the community observing his physical and social surroundings and daily habits. Subsequent support was individually tailored; at one extreme some patients needed only a few follow-up visits, whereas other patients needed frequent visits and practical assistance for several months. In the USO group, the men were referred to mental health and rehabilitation programs that were generally of high quality. Following the usual model of discharge from an institution, the staff of the onsite shelter psychiatry program were available to these agencies for consultation on request but did not actively seek a role in the patient's care after discharge. The men were also referred as needed to community agencies for substance abuse, general health, income support, education, legal advocacy, and other services. The number and the range of potential service agencies in New York City was greater than in most urban areas. The connections among the various agencies in regard to the care of an individual, however, were generally weak and unsystematized. The CTI and USO interventions were applied in a standardized fashion, as specified in a manual (available from the first author). To ensure adherence to the specified protocols, the investigators met weekly with the clinical team. The clinical team presented cases in CTI and USO at these meetings, and in each case presentation they specified the relation of the treatment plan to the respective CTI or USO protocol. In this way, any departures from the intended experimental and control treatments were detected rapidly. Measures Psychiatric Symptoms. Symptom severity was assessed using the Positive and Negative Syndrome Scale (PANSS; Kay et al., 1992), a 30-item, 7-point severity scale that has been widely employed in both clinical and research settings, including a prior study of homelessness among chronically mentally ill adults (Opler et al., 1994). The PANSS generates scale scores in three domains-positive symptoms, negative symptoms, and general psychopathology. The positive scale is computed by summing ratings on delusions, conceptual disorganization, hallucinatory behavior, excitement, grandiosity, suspiciousness/persecution, and hostility. The negative scale is computed by summing ratings on blunted affect, emotional withdrawal, poor rapport, passive-apathetic social withdrawal, difficulty in abstract thinking, lack of spontaneity and flow of conversation, and stereotyped thinking. The general psychopathology scale consists of the sum of the following 16 items not included in the positive and negative scales: somatic concern, anxiety, guilt feelings, tension, mannerisms and posturing, depression, motor retardation, uncooperativeness, unusual thought content, disorientation, poor attention, lack of judgment and insight, disturbance of volition, poor impulse control, preoccupation, and active social avoidance. For each scale, lower scores indicate less psychopathology. Symptom ratings were performed at baseline and 6 months by trained raters blind to group assignment. Reliability for these measures was good (alpha for baseline positive, negative, and general symptoms were .80, .86, and .89, respectively). Data Analysis The primary purpose of the data analyses was to examine differences between the experimental and control groups on the outcomes of interest. First, we compared the experimental and control group means and standard deviations for each symptom scale at baseline and 6 months. To formally assess the effect of the intervention on psychiatric symptoms, we then performed three multivariate regression analyses in which the dependent variable was one of the three symptom scales measured at 6 months, the independent variable was group membership, and the respective baseline symptom factor score was entered as a covariate. We conducted further analyses stratifying by age group and ethnic group. The results of these analyses were unchanged. Results Table 1 provides the demographic and diagnostic characteristics of the sample. The sample was primarily African-American, and most men were over age 35 (median age = 36). The majority of the sample had been homeless for over a year during their lifetime. Schizophrenia and its related disorders (including schizoaffective and schizophreniform) were the most frequent diagnoses. ---------------------------------------------- TABLE 1 Descriptive Characteristics of the Sample (N = 76) ---------------------------------------------- There were 76 cases (38 CTI and 38 USO) for whom there were complete symptom data at base-line and at 6 months. Table 2 compares the CTI and the USO groups on mean symptom scale scores at these two time points (all scores and mean differences are unadjusted). On the negative symptoms scale (baseline range = 8 to 36), the mean change in the CTI group was -2.6 compared with a mean change of +1.0 in the USO group. On the positive symptoms scale (baseline range = 7 to 40), the mean change in the CTI group was -2.2 compared with a mean change of -2.7 in the USO group. On the general psychopathology scale (baseline range = 19 to 67), the mean change in the CTI group was -6.5 compared with a mean change of -3.1 in the USO group. In the regression analyses controlling for baseline score for the 6-month outcomes, there was a significant group effect for negative symptoms only (F = 5.7, p = .02). ---------------------------------------------- TABLE 2 PANSS Symptom Scores in USO and CTI Groups at Baseline and 6 Months ---------------------------------------------- When we confined the analysis to 57 subjects diagnosed with schizophrenia and related disorders (schizophrenia, schizoaffective, and schizophreni-form disorder), the pattern of results was quite similar. On the negative symptoms scale, the mean change in the CTI group was -2.9 compared with a mean change of +.5 in the USO group. On the positive symptoms scale, the mean change in the CTI group was -3.1 compared with a mean change of -3.0 in the USO group. On the general psychopathology scale, the mean change in the CTI group was -8.3 compared with a mean change of -4.4 in the USO group. In the regression analyses controlling for baseline score for the 6-month outcomes, there was a marginally significant group effect (F = 3.4, p = .07) for negative symptoms only (this larger p value reflects the reduced statistical power associated with the smaller sample size employed in these analyses). Discussion In this study, CTI was associated with a statistically significant decrease in negative symptoms at the 6-month follow-up, reflecting modest clinical improvement. This is an important finding as it is one of the few studies of which we are aware that has focused specifically on the effect of a psychosocial intervention on negative symptoms in psychotic disorders. It is also consistent with the conclusion of a recent meta-analysis of controlled outcome studies in this area (Mojtabai et al., 1998). There was no significant effect on positive or general psychopathology symptoms. We consider below several possible explanations for the observed reduction in negative symptoms. Negative symptoms are frequently cited as among the most disabling aspects of psychotic disorders and among the most resistant to treatment. Because standard antipsychotics have been shown to have little if any impact on negative symptoms (and none of the subjects in this study were receiving either the atypical antipsychotic clozapine or the novel antipsychotics risperidone, olanzpine, or quetiapine), it is unlikely that the observed reduction in negative symptoms in the experimental group can be explained by a different pattern of medication use in that group. Further, the lack of group differences on positive symptoms, which are responsive to antipsychotic medications, also suggests that it is not differences in medication use that account for the observed differences on negative symptom scores. Nevertheless, we compared the proportion of subjects in the experimental and control groups who reported having taken psychoactive medications over the course of the follow-up period. At 6 months, the proportion of experimental and control group members having taken antipsychotic medication in the prior month were virtually identical, providing further evidence that the treatment effect was not a function of differential rates of medication use. Although we consider it unlikely, because we did not collect data on medication dosing we cannot rule out the possibility that the USO group received higher doses of standard neuroleptics and/or lower doses of antiparkinsonian medications leading to higher levels of secondary negative symptoms (i.e., resulting from neuroleptic medications). As we have reported elsewhere (Susser et al., 1997) the treatment group experienced significantly fewer homeless nights over the follow-up period than did the control group. Therefore, we explored whether the observed effect on negative symptoms was mediated by the experience of homelessness. To investigate this possibility, we performed a hierarchical regression analysis predicting negative symptom score at 6 months in which we first entered baseline negative symptom score and treatment condition (step 1) followed by a variable indicating the number of homeless nights experienced during the 6-month follow-up period (step 2). We found no increase in explanatory power associated with the addition of the homelessness variable (F of change = .20, p = .65), indicating that reduced homelessness did not mediate the effect of the treatment on negative symptom score at 6 months. The evidence suggesting that neither medication nor homelessness accounted for the observed reduction in negative symptoms supports the idea that the effect can be attributed to the intervention. However, as in many experimental studies that test a composite "package" of services such as CTI, it can be difficult to identify the specific mechanisms through which the treatment produced the observed effect. In this case, the primary focus of CTI was to prevent recurrent homelessness after discharge from the shelter. How might the intervention have directly contributed to the apparent improvement in negative symptoms? Several studies have suggested that the degree of prefrontal cortical dysfunction in schizophrenia is related both to impairments in carrying out executive tasks and to severity of negative symptoms (Andreasen et al., 1992; Opler et al., 1991; Wolkin et al., 1992). One intriguing possibility is that by encouraging patients to focus on planning and organizing their transition from institution to community living, CTI provides cognitive remediation, helping to reactivate prefrontal cortical functions as they are summoned to carry out executive tasks that were underutilized during lengthy periods of passivity and regression. In future studies, we hope to investigate the prospect that this psychosocial intervention may improve prefrontally mediated cognitive functions as well as negative symptoms. It is also possible to offer a psychodynamic mechanism that may have contributed to the reduction in negative symptoms. For homeless mentally ill men who commonly have experienced multiple emotional losses during their lives, the withdrawal, apathy, and affective blunting that comprise negative symptoms may, in part, be a reaction to these important psychological events. CTI focuses explicitly on issues of relationship and social support. As described above, one of its key elements is that the strong ties that develop between patients and treatment staff are not abruptly severed at the point of discharge but rather are used to promote successful community reintegration through a successful transition process. Perhaps this experience plays a role in ameliorating the hopelessness and despair that has characterized these men's lives and, in turn, leads to an observed lessening in negative symptoms. In all likelihood, no single mechanism can account for the reduction in negative symptoms we observed in the experimental group. Rather, we suspect that this improvement may be attributable to the synergistic effect of cognitive, psychopharmacological, and psychodynamic mechanisms operating simultaneously. We hope to design future studies of the CTI model in such a way as to shed further light on this question. As noted above, there was no significant effect of CTI on positive or general psychopathology symptoms. The most recent quantitative review of the research on the role of psychosocial treatments in schizophrenia suggested that positive and general symptoms tend to be less responsive to psychosocial interventions than are negative symptoms (Mojtabai et al., 1998). Although the reasons for this are not well understood, it may be that negative symptoms, particularly among persons who have been ill for a number of years, are partly related to factors such as social isolation and demoralization, which might be more responsive to psychosocial interventions. Limitations There were a number of limitations to the study. First, the number of total participants was modest. Although 96 men were randomized into the two experimental conditions, symptom data at baseline and 6-month follow-up was available for only 76 men. Although there appear to be no systematic differences between these 76 men and the 20 for whom data are incomplete, it is possible that this may have affected the results to some unknown degree. Second, although the change in negative symptoms in the full group was statistically significant at the .05 level, the magnitude of the effect was modest. Third, the degree to which the observed effect on negative symptoms was a lasting one cannot be adequately assessed since our outcome data were collected at 6 months. However, in a subgroup of 47 subjects whose symptoms were measured at 18 months, the trend was toward continued lower levels of negative symptoms in the experimental group when compared with the controls (mean negative symptoms score in CTI group = 15.3, SD = 4.8; mean negative symptoms score in USO group = 17.6, SD = 6.5, p = .16). Finally, although we specified a priori that we would evaluate the effect of our intervention on psychiatric symptoms, the intriguing finding presented here did not conform with our original hypothesis, which predicted improvement in negative, positive, and general psychopathology symptoms. Conclusions In an earlier paper, we reported that CTI was effective in its primary intended outcome-reducing recurrent homelessness among severely mentally ill men after discharge from a shelter. The present report suggests that a time-limited intervention focused on the transitional period may also have had beneficial effects on an important clinical outcome, namely negative symptoms, that has stubbornly resisted the effect of treatment in many previous efforts. Because negative symptoms are considered to play an important role with respect to the quality of life of persons with severe mental disorders, this finding warrants further study. Although CTI was originally designed and tested with homeless men leaving a shelter, we believe that it may also be fruitfully applied in other settings including psychiatric hospitals and correctional facilities. References Andreasen N, Rezai K, Alliger R, Swayze V, Flaum M, Kirchner P, Cohen G, O'Leary D (1992) Hypofrontality in neuroleptic-naive patients and in patients with chronic schizophrenia: Assessment with xenon 133 single-photon emission computed tomography and the Tower of London. Arch Gen Psychiatry 49:943-958. Bibliographic Links Beasley CJ, Hamilton S, Crawford A, Delva M, Tollefson G, Tran P, Blin O, Beuzen J (1997) Olanzapine versus haloperidol: Acute phase results of the international double-blind olanzapine trial. Eur Neuropsychopharmacol 7:125-137. Bibliographic Links Carpenter W, Heinrichs D, Wagman A (1988) Deficit and nondeficit forms of schizophrenia: the concept. Am J Psychiatry 145:578-583. Bibliographic Links Chouinard G, Jones B, Remington G, Bloom D, Addington D, MacEwan W, Labelle A, Beauclair L, Arnott W (1993) A Canadian multicenter placebo-controlled study of fixed doses of risperidone and haloperidol in the treatment of chronic schizophrenic patients. J Clin Psychopharmacol 13:25-40. Bibliographic Links Dorwart RA, Hoover CW (1994) A national study of transitional hospital services in mental health. Am J Public Health 84:1229-1234. Full Text Bibliographic Links Kay S, Opler L, Fiszbein A (1992) Positive and negative syndrome scale (PANSS) manual. Toronto: Multi-Health Systems, Inc. Lehman AF, Dixon LB, Kernan E, DeForge BR, Postrado LT (1997) A randomized trial of assertive community treatment for homeless persons with severe mental illness. Arch Gen Psychiatry 54:1038-1043. Marder S, Meibach R (1994) Risperidone in the treatment of schizophrenia. Am J Psychiatry 141:825-835. Mojtabai R, Nicholson R, Carpenter B (1998) Role of psychosocial treatments in management of schizophrenia: A meta-analytic review of controlled outcome studies. Schizophr Bull 24:571-588. Olfson M, Mechanic D, Boyer CA, Hansell S (1998) Linking inpatients with schizophrenia to outpatient care. Psychiatr Services 49:911-917. Opler L, Ramirez P, Rosenkilde C, Fiszbein A (1991) Neurocognitive features of chronic schizophrenic inpatients. J Nerv Ment Dis 179:638-640. Bibliographic Links Opler LA, Caton CL, Shrout P, Dominguez B, Kass F (1994) Symptom profiles and homelessness in schizophrenia. J Nerv Ment Dis 182:174-177. Bibliographic Links Susser E, Valencia E, Conover S, Felix A, Tsai W, Wyatt R (1997) Preventing recurrent homelessness among mentally ill men: A "critical time" intervention after discharge from a shelter. Am J Public Health 87:256-262. Ovid Full Text Full Text Bibliographic Links Valencia E, Susser E, McQuiston H (1996) Critical time points in the clinical care of homeless mentally ill individuals. In J Vaccaro, G Clark (Eds), Practicing psychiatry in the community: A manual (pp 259-276). Washington, DC: American Psychiatric Press. Valencia E, Susser E, Torres J, Felix A, Conover S (1997) Critical time intervention for homeless mentally ill individuals in transition from shelter to community living. In W Breakey, J Thompson (Eds), Mentally ill and homeless: special programs for special needs (pp 75-94). Newark, NJ: Gordon and Breach Science Publishers. Wolkin A, Sanfilipo M, Wolf A, Angrist B, Brodie J, Rotrosen J (1992) Negative symptoms and hypofrontality in chronic schizophrenia. Arch Gen Psychiatry 49:959-963. Bibliographic Links ---------------------------------------------- Accession Number: 00005053-200003000-00002