Does case management reduce the rehospitalizationvrate? Rossler W, Loffler W, Fatkenheuer B, Riecher-Rossler A. Does case management reduce the rehospitalization rate? Acta Psychiatr Scand 1992: 86: 445-449. It was expected from deinstitutionalization that outpatient care could replace hospital care. But many empirical studies proved that the number of admissions to psychiatric hospitals rose when community-based care developed. This might be due to the lack of coordination and cooperation of extramural services. The concept of case management therefore originated in the last few years. In this study the effect of case management on the rehospitalization rate is examined by analyzing the data of 4 social-psychiatric services responsible for defined catchment areas. During the evaluation period of 2.5 years, 162 patients dismissed from psychiatric hospitals were referred to those services. For each of these index patients a matched control patient was identified, each identical in diagnosis, sex, age, living conditions and number of previous inpatient episodes. The results of a survival analysis show that there are no significant effects of case management on the rate of rehospitalization nor on the length of time in hospital in case of a rehospitalization.
It was expected from deinstitutionalization that outpatient care could replace hospital care. But many empirical studies suggested an opposite development. In the United States, for example, the number of admissions to psychiatric hospitals rose when community mental health centers came into existence (1). A comparable Danish study concluded that outpatient treatment cannot replace inpatient treatment (2). Similar results were found in Germany (3). Using case register data, we showed that the implementation of a community-based care system in the industrial city of Mannheim, Germany resulted in complex changes: between 1974 and 1980 there was a total increase of about 100% in service usage. The utilization of outpatient services increased by 162 %, while the utilization of inpatient services increased by 43 %, but the increase differed between diagnostic groups (4). Utilization data from case registers can show trends in utilization but valid statements on the effectiveness of single components of a psychiatric care system are not possible. Thus, several authors carried out meta-analyses on impact studies of extramural institutions or studies concerning aftercare measures for chronic patients. May & Simpson (9, for example, found that, on average, intensive aftercare for chronic schizophrenic patients can improve working capacity, promote reintegration in society and reduce schizophrenic symptoms. Test & Stein (6) take a comparably optimistic view; they found a
W. Rossler, W. Loffler, B. Fatkenheuer, A. Riecher-Rossler Central Institute of Mental Health, Mannheim, Germany
Key words: case management: deinstitutionalization
W. Rossler, Central Institute of Mental Health, J5, D-6800 Mannheim 1, Germany Accepted for publication August 14, 1992
positive influence of outpatient treatment on the rate of hospitalization in all studies reviewed except one. On the other hand, Braun et al. (7) did not find an influence of extramural services on the hospitalization rate. One reason for these inconsistent results is a methodological problem. So the term “chronic” is hardly exactly defined with respect to severity of the illness or disability. Also, the complexity of extramural institutions and aftercare measures makes it very difficult to compare different care systems. The fact that many studies did not find a decisive impact of extramural care on the hospitalization rate also might be due to a structural deficiency of most fragmented care systems, in particular the lack of coordination and cooperation of extramural services. Therefore, as a key coordinating mechanism, the concept of case management originated in the last few years. Case management focuses on all aspects of the physical and social environment (8). The core elements of case management are the assessment of patient needs, the development of comprehensive service plans for the patients and arrangement of service delivery (9- 11). In the United States, case management is accepted as an indispensible element in community support systems for mentally handicapped and chronically mentally ill (12). The concept of case management also became just lately part of the British government policy for community care (13). But despite an apparent consensus in the lit445
Rossler et al.
erature about aims and objectives of case management (14), several empirical studies did not show a definite impact of case management on the rehospitalization rate (1 5 , 16). Other studies, however, found a decrease in hospital utilization (17-19). By means of a case-control study, overcoming some methodological problems of the named studies, we wanted to examine the effect of case management on the rehospitalization rate by analyzing the utilization data of four social-psychiatric services responsible for defined catchment areas. These newly established outpatient services were part of a model program initiated by the German federal state of Baden-Wurttemberg from 1982 until 1987.” On average, each of these social-psychiatric services was staffed with 4 social workers. The designated aims and objectives of social-psychiatric services were predominantly case management for chronically mentally ill people, especially for those discharged from psychiatric hospitals. Material and methods
The 4 catchment areas included are part of the federal state of Baden-Wurttemberg, a state with mixed urban and rural; areas. About 9.2 million inhabitants live in this state: in the studied areas, about 850,000 inhabitants, or 9.2% of the total population. The demographic characteristics of the catchment areas (Table 1) are representative for this federal state. A bed ratio of 0.7 psychiatric hospital beds per 1000 inhabitants in Baden-Wurttemberg indicates a relatively well developed psychiatric care system (20). General practitioners and psychiatrists in office practice provide a dense net of psychiatric outpatient care. Additionally, social help is offered in a differentiated system of social and complementary services. This care system reflects the highly complex structure of health and welfare systems in industrialized countries. The evaluation is based on the utilization data from January 1, 1984 until June 30, 1986. During that time, 10,920 unduplicated patients were discharged from all 7 psychiatric hospitals taking responsibility for the named catchment areas; 162 of these patients came into the care of the socialpsychiatric services. For all patients - including the 162 patients referred to the social-psychiatric services - a sociodemographic data set was recorded. The patients were diagnosed according to ICD-9. a The federal state of Baden-Wiirttemberg has initiated a program of 41 outpatient complementary model services directed to the care of chronically mentally ill and disabled of different age groups. Patients with alcohol- and drug-related diseases were excluded.
446
Table 1, Demographic characteristics of the catchment areas Catchment area
Type of inpatient treatment
Number of inhabitants January 1, 1985
Sigmaringen Konstanz
Departments of psychiatry State mental hospital (Reichenaul State mental hospital (Nordschwarzwald-Klinik) Department of psychiatry State mental hospital (Winnenden) State mental hospital (Weinsberg)
86,243
94.6
186,288
282.2
238,707
491.1
338,191
627.4
Boblingen Ludwigsburg
Population density/km*
Source: Statistisches Landesamt Baden-Wurttemberg (29). Statistisches Landesamt Baden-Wurttemberg (30).
A comparison of the group of patients in the aftercare of the social-psychiatric services after inpatient treatment (index group) with a group of patients that received no extramural care by the socialpsychiatric services after discharge from psychiatric hospital (control group) allows an estimate of the impact of the social-psychiatric services on the rehospitalization rate. For the index group, we expected a lower rate of rehospitalization, a longer period in the community between two inpatient episodes and a shorter time in hospital if rehospitalized. Besides case management, additional factors influence the rate of rehospitalization; it is important to control for such factors. We therefore used the design of a case-control study for our analysis. Known risk factors for rehospitalization are mainly diagnosis, sex, age, living conditions after discharge from psychiatric hospital and the number of previous inpatient treatment episodes (21). We therefore identified retrospectively for each of the 162 index patients a matched control patient identical in diagnosis and the other risk factors named above who was not referred to a social-psychiatric service after discharge from psychiatric hospital. By means of a survival analysis (using SPSS-X), we compared the groups with respect to the survival time in the community after discharge. Then we compared the length of inpatient treatment for the 2 rehospitalized subgroups again with survival techniques. The survival techniques, originally developed for the analysis of life tables, can be used for all problems in which the dependent variable is the interval between an initial and a termination event. For example the response can be the survival time in the community between discharge from psychiatric hospital and readmission to inpatient treatment. Survival function is the estimated cumulative probability of patients surviving from the beginning to the ith time interval, if a person has reached this interval without response. Censored observations mean
Does case management reduce the rehospitalization rate? that, for some cases, the termination event does not occur within the investigation period. So the survival time is the interval investigated. Both censored and uncensored observations are used in calculating survival time. With a D-statistic (D is asymptotically distributed as chi-square - Lee-Desu statistics) calculating differences between two survival functions, one can test the null hypothesis that subgroups are samples of the same survival distribution. The larger the D statistic, the greater the probability of (signific antly) different survival distributions. Results
Table2 shows the characteristics of the two samples. Both comprise 162 patients, 162 patients referred to the social-psychiatric services and 162 matched controls (twins), each identical in sex, age, living conditions, number of previous inpatient episodes and diagnosis who were not cared for by the social-psychiatric services. The studied samples had an excess of woman and the age group 25-44. For about one third of the patients it was the first inpatient treatment episode; another third of the patients had 1-2 previous psychiatric admissions. About one fifth of the patients had been hospitalized more than Table 2. Patient characteristics - identical for index and control group Sex male female
45% 55%
Age
5 24 years 2 5 4 4 years 45-64 years 2 6 5 years
16% 50% 26% 8%
Living conditions alone living in family no fixed abode unknown
29% 70% 1% 1%
Table 3 Face-to-face activities in social-psychiatric services - percentage of patients with the respective number of activities Number of monthly activities Kind of activity
0
51
52
3-4
5-8
>8
Assessment, planning counselling
8.8
17.7
27.2
33.1
9.5
3.8
Training in independent living skills
46.8
23.2
14.5
11.6
2.5
1.4
Coordination of professional support
32.7
29.8
19.8
12.1
3.9
1.8
Consultation with relatives
44.0
29.6
14.7
9.1
2.0
0.6
Maintenance and expansion of social networks
68.4
17.2
6.3
5.8
1.2
1.1
Home visits
35.4
27.1
20.2
13.3
2.8
1.2
Others
59.4
21.2
10.0
5.8
2.0
1.6
5 times. The predominant diagnosis was schizophrenia; about two thirds suffered from psychotic disorders. The average duration of care by the socialpsychiatric services for the index group was 6.2 months after discharge from psychiatric hospital. Table 3 shows the face-to-face-activities for index patients by the social-psychiatric services. The care offered for those patients is intensive. Assessment, planning and counselling was the basis of case management for about 90% of the index patients. These activities were constantly repeated: for more than half of the index patients twice a month, for about one third 3-4 times a month. About two thirds of these patients were visited at home, 37.5;/, more than twice a month. In general, the studied socialpsychiatric services displayed the whole array of case management activities.
Number of previous inpatient episodes
0 1-2 inpatient treatment episodes 2-5 inpatient treatment episodes > 5 inpatient treatment episodes unknown
31% 32% 18% 18% -
control group (n= 162) 0.7
0.6
Diagnosis schizophrenic/paranoiddisorders
(295, 297, 298) affective psychoses (296) neuroses, personality disorders, non-psychotic disorders (3OC-302, 306, 307, 31 1, 316.0) organic disorders (290, 293, 294, 310) acuteiadjustment reactions (308, 309) others
59%
P
0.5
index group (n.162)
12% 0.2
16% 2% 4% 6%
0.1
0
10
20
30
40
50
60
70
80
90
100
110
120
130
weeks
Fig. I . Comparison of the “survival time” in the community after discharge from hospital between index and control group by survival analysis
447
Rossler et al. Fig. 1 shows the survival analysis for the two samples compared. On the abscissa the analyzed time interval is marked in weeks. The ordinate shows the probability (P) of remaining in the community after discharge from hospital. So, for example, the probability of living in the community 10 weeks after discharge from hospital is P 2 0 . 9 6 for the index group, while the probability for the control group is P 2 0.89. About 40 weeks after discharge from hospital the probability of living in the community is almost identical for both groups. After that time the risk of being rehospitalized increases steadily for the index group, and remains fairly stable for the control group. At the end of the examined time interval (130 weeks), the probability of living in the community is P 2 0 . 5 4 for the index group, for the control group P20.79. The survival functions of the compared samples were not significantly different. For the patients who were rehospitalized during the evaluation period, we examined in a further step the time spent in hospital during this second admission. But again, the survival functions of the group that had been under the care of the social-psychiatric services and the control group were not significantly different ( P I 0.11).
Discussion Case management has been introduced to improve service delivery, mainly for chronically mentally ill people. One outcome criterion of case management is the rate of rehospitalization or the length of time in hospital in case of rehospitalization. The results of this study show that, during an observation period of 2.5 years, case management had no significant effects on the risk of rehospitalization or on the length of time in hospital in case of rehospitalization. These results have to be viewed against the background of intensive case management activities by the studied social-psychiatric services. Inevitably, the case load of the social-psychiatric services is low. This is a well-known phenomenon in case management (22). Considering the costs of such services, one is inclined to devalue case management at first view. Integrating our results into the reviewed literature, we have to discuss carefully the advantages and disadvantages of our study design. During deinstitutionalization, one major objective of the psychiatric reforms was reducing the number of psychiatric beds. Psychiatric hospitals were not supposed to provide residential care any longer but to focus mainly on psychiatric treatment. In 1984 the average length of hospitalization in a state mental hospital in Baden-Wurttemberg was 59 days and 33 days in a psychiatric department (23). This is a relatively short period compared with the traditional length of hospitalization. It therefore seems proba-
448
ble that hospital care in the studied catchment areas was mainly provided to treat mental disorders. One can doubt whether a decisive further reduction of length of time spent in hospital can be reached in such areas (ceiling effect). Although we know of no study on case management with a comparable thoroughly chosen control group, we also have to consider some methodological aspects. When selecting the samples, we chose well known sociodemographic and disease-related characteristics as criteria for the risk of rehospitalization after discharge from hospital. Nevertheless, we cannot be sure about the true extent of disability or chronicity in the patients of both groups, as the criteria chosen might be too coarse. The fact that more patients under the care of the social-psychiatric services were rehospitalized could still be because the patients referred to a social-psychiatric service were more disabled and vulnerable and thus more endangered for rehospitalization. One name given to these problem patients is “system misfits” (24,25), expressing that our care system cannot meet the needs of these patients. Some of the characteristics of the examined samples indicate that the social-psychiatric services deal with problem patients, in particular the young male schizophrenic, living alone and with several relapses in his psychiatric history. With respect to this patient group, it is easily understandable that intensive professional support can increase rather than decrease the probability of a referral to a psychiatric hospital, especially as one core element of case management is the link with resources. So a higher rehospitalization rate might indicate good quality of care. Additionally, as the utilization of a socialpsychiatric service is voluntary, patients of the control group might have refused a referral to a socialpsychiatric service. And for such patients there is a high probability that they also refuse rehospitalization. Indeed, we have to deal more frequently with problem patients who do not utilize the psychiatric care system according to their needs (26). This does not mean that this group does not need care. But some patients who are hostile and uncooperative with mental health professionals disappear from our records. After all, we have to ask whether the question “Does case management reduce the rehospitalization rate?” is the right question. This question might often be asked for economic reasons. For chronically mentally ill people themselves and their quality of life, discharge to the community - even interrupted by repeated, but short rehospitalizations might be preferable to a long - maybe life-long - stay in the back ward of a psychiatric hospital. On the other hand, case management for chronically mentally ill people in the community is rather regularly
Does case management reduce the rehospitalization rate? associated with a marked increase of service utilization in the community (8, 16, 27) and thus with an increase in costs. There is therefore an urgent need to identify the most effective elements in case management. In another study (28) on the differential effectiveness of case management activities for chronically mentally ill people, case management focusing on community functioning proved to be most effective when used within a long-term but low-contact treatment. Regarding the above-mentioned case management activities, training in independent living skills was more effective than therapeutic counselling. In view of coordination activities, linkage with psychiatric treatment made case management most effective. Thus, successful case management is nothing else but good psychiatric and rehabilitative treatment. Acknowledgement The scientific evaluation conducted by the Central Institute for Mental Health in Mannheim was financed by the Ministry of Labour, Housing, Family and Social Order, Baden-Wurttemberg.
References 1. ELLISONDL, RIEKERP, MARXJH. Organisational adaption to community mental health. In: ROMANPM et al., ed. Sociological perspectives on community mental health. Philadelphia: Davis Company, 1974. 2. KASTRUP M, NAKANEY, DUPONTA, BILLEM. Psychiatric treatment in a delimited population - with particular reference to outpatients. Acta Psychiatr Scand 1976: 53: 3550. 3. BAUERM. Sektorisierte Psychiatrie im Rahmen einer Universitatsklinik - Anspruch, Wirklichkeit und praktische Erfahrungen. Stuttgart: Enke, 1977. 4. ROSSLERW, HAFNERH. Psychiatrische Versorgungsplanung. Neuropsychiatrie 1985: 1: 8-17. 5. MAY PRA, SIMPSONGM. Schizophrenia: evaluation of treatment methods. In: KAPLANHI et al., ed. Comprehensive textbook of psychiatry. Vol. 111. Baltimore: Williams & Wilkins, 1980. 6. TESTMA, STEINLI. Community treatment of the chronic patient: research overview. Schizophr Bull 1978: 4: 350-364. 7. BRAWNP, KOCHANSKI G, SHAPIROR et al. Deinstitutionalization of psychiatric patients, a critical review of outcome studies. Am J Psychiatry 1981: 138: 736-749. 8. KANTERJ. Clinical case management: definition, principles, components. Hosp Community Psychiatry 1989: 40: 362368. 9. National Institute of Mental Health. Towards a model plan for a comprehensive community-based mental health system. Administrative document. Washington, DC: US Department of Health and Human Services, 1987.
10. CLIFFORD P, CRAIGT. Case management systems for the long-term mentally ill. A proposed interagency initiative. London: NUPRD, 1988. 11. RENSHAW J. Care planning and case management. Br J Soc Work 1987: 18: 79-105. 12. INTAGLIATAJ. Improving the quality of care for the chronic mentally disabled: the role of case management. Schizophr Bull 1982: 8: 655-674. 13. Caring for people. London: HMSO, 1989. 14. HOLLOWAY F. Case management for the mentally ill: looking at the evidence. Int J SOCPsychiatry 1991: 37: 2-13. IS. FRANKLIN J, SOLOWITZB, MASONM. CLEMONSJR, MILLERGE. An evaluation of case management. Am J Public Health 1987: 77: 674-678. 16. BORLAND A, MCRAEJ, LYCANC. Outcomes of five years of continuous intensive case management. Hosp Community Psychiatry 1989: 40: 369-376. 17. STEINLI, TESTMA. Alternative to mental hospital treatment. Arch Gen Psychiatry 1980: 37: 32-37. 18. HOULTJ, REYNOLDS I, CHARBONNEAU-POURIS M. Psychiatric hospital versus community treatment: the results of a randomized trial. Aust NZ Psychiatry 1983: 17: 160-167. 19. MARKSI, CONNOLLY J, MUIJENM. The Maudsley Daily Living Program. Bull R Coll Psychiatr 1988: 12: 22-24. 20. FREEMANHL, FRYERS TH, HENDERSONJH. Mental health services in Europe: 10 years on. Copenhagen: World Health Organization Regional Office for Europe, 1985. 21. AN DER HEIDENW, KRUMMB, HAFNERH. Die Wirksamkeit ambulanter psychiatrischer Versorgung - Ein Modell zur Evaluation extramuraler Dienste. Berlin: Springer, 1989. 22. THORNICROFT G. Case managers for the mentally ill. SOC Psychiatry Psychiatr Epidemiol 1990: 25: 141-143. 23. ROSSLERw , HAFNERH, MARTINIH, AN DER HEIDENw , JUNGE, LOFFLERW. Landesprogramm zur Weiterentwicklung der auljerstationaren psychiatrischen Versorgung BadenWurttemberg - Analysen, Konzepte, Erfahrungen. Weinheim: Deutscher Studien Verlag, 1987. 24. BEESONPG. The bureaucratic context of mental health care. In: LITTRELWB et al. ed. Bureaucracy as a social problem. Greenwich, CT: JAI Press, 1983. 25. JOHANSEN KH. The impact of patients with chronic character pathology on a hospital inpatient unit. Hosp Community Psychiatry 1983: 34: 842-846. 26. BACHRACH LL, TALBOTT JA, MEYERSON AT. The chronic psychiatric patient as a “difficult” patient: a conceptual analAT, ed. Barriers to treating the chronic ysis. In: MEYERSON mentally ill. New directions in mental health services. 1987: 35-50. RD, WARDR. As27. BOUDGR, MILLERLD, KRUMWIED sertive case management in three CMHC’s: a controlled study. Hosp Community Psychiatry 1988: 39: 41 1-418. 28. ROSSLERW, FATKENHEUER B, LOFFLERB. Soziale Rehabilitation Schizophrener in Sozialpsychiatrischen Diensten. Stuttgart: Enke (in press). 29. Statistisches Landesamt Baden-Wurttemberg. Statistische Berichte. Voraussichtliche Entwicklung der Bevolkerung auf der Basis vom 1.1.1984. Stuttgart, 1985. 30. Statistisches Landesamt Baden-Wurttemberg. Gemeindestatistik 1985. Amtliches Gemeindeverzeichnis BadenWurttemberg 1985. Stuttgart, 1985.
449
It was expected from deinstitutionalization that outpatient care could replace hospital care. But many empirical studies suggested an opposite development. In the United States, for example, the number of admissions to psychiatric hospitals rose when community mental health centers came into existence (1). A comparable Danish study concluded that outpatient treatment cannot replace inpatient treatment (2). Similar results were found in Germany (3). Using case register data, we showed that the implementation of a community-based care system in the industrial city of Mannheim, Germany resulted in complex changes: between 1974 and 1980 there was a total increase of about 100% in service usage. The utilization of outpatient services increased by 162 %, while the utilization of inpatient services increased by 43 %, but the increase differed between diagnostic groups (4). Utilization data from case registers can show trends in utilization but valid statements on the effectiveness of single components of a psychiatric care system are not possible. Thus, several authors carried out meta-analyses on impact studies of extramural institutions or studies concerning aftercare measures for chronic patients. May & Simpson (9, for example, found that, on average, intensive aftercare for chronic schizophrenic patients can improve working capacity, promote reintegration in society and reduce schizophrenic symptoms. Test & Stein (6) take a comparably optimistic view; they found a
W. Rossler, W. Loffler, B. Fatkenheuer, A. Riecher-Rossler Central Institute of Mental Health, Mannheim, Germany
Key words: case management: deinstitutionalization
W. Rossler, Central Institute of Mental Health, J5, D-6800 Mannheim 1, Germany Accepted for publication August 14, 1992
positive influence of outpatient treatment on the rate of hospitalization in all studies reviewed except one. On the other hand, Braun et al. (7) did not find an influence of extramural services on the hospitalization rate. One reason for these inconsistent results is a methodological problem. So the term “chronic” is hardly exactly defined with respect to severity of the illness or disability. Also, the complexity of extramural institutions and aftercare measures makes it very difficult to compare different care systems. The fact that many studies did not find a decisive impact of extramural care on the hospitalization rate also might be due to a structural deficiency of most fragmented care systems, in particular the lack of coordination and cooperation of extramural services. Therefore, as a key coordinating mechanism, the concept of case management originated in the last few years. Case management focuses on all aspects of the physical and social environment (8). The core elements of case management are the assessment of patient needs, the development of comprehensive service plans for the patients and arrangement of service delivery (9- 11). In the United States, case management is accepted as an indispensible element in community support systems for mentally handicapped and chronically mentally ill (12). The concept of case management also became just lately part of the British government policy for community care (13). But despite an apparent consensus in the lit445
Rossler et al.
erature about aims and objectives of case management (14), several empirical studies did not show a definite impact of case management on the rehospitalization rate (1 5 , 16). Other studies, however, found a decrease in hospital utilization (17-19). By means of a case-control study, overcoming some methodological problems of the named studies, we wanted to examine the effect of case management on the rehospitalization rate by analyzing the utilization data of four social-psychiatric services responsible for defined catchment areas. These newly established outpatient services were part of a model program initiated by the German federal state of Baden-Wurttemberg from 1982 until 1987.” On average, each of these social-psychiatric services was staffed with 4 social workers. The designated aims and objectives of social-psychiatric services were predominantly case management for chronically mentally ill people, especially for those discharged from psychiatric hospitals. Material and methods
The 4 catchment areas included are part of the federal state of Baden-Wurttemberg, a state with mixed urban and rural; areas. About 9.2 million inhabitants live in this state: in the studied areas, about 850,000 inhabitants, or 9.2% of the total population. The demographic characteristics of the catchment areas (Table 1) are representative for this federal state. A bed ratio of 0.7 psychiatric hospital beds per 1000 inhabitants in Baden-Wurttemberg indicates a relatively well developed psychiatric care system (20). General practitioners and psychiatrists in office practice provide a dense net of psychiatric outpatient care. Additionally, social help is offered in a differentiated system of social and complementary services. This care system reflects the highly complex structure of health and welfare systems in industrialized countries. The evaluation is based on the utilization data from January 1, 1984 until June 30, 1986. During that time, 10,920 unduplicated patients were discharged from all 7 psychiatric hospitals taking responsibility for the named catchment areas; 162 of these patients came into the care of the socialpsychiatric services. For all patients - including the 162 patients referred to the social-psychiatric services - a sociodemographic data set was recorded. The patients were diagnosed according to ICD-9. a The federal state of Baden-Wiirttemberg has initiated a program of 41 outpatient complementary model services directed to the care of chronically mentally ill and disabled of different age groups. Patients with alcohol- and drug-related diseases were excluded.
446
Table 1, Demographic characteristics of the catchment areas Catchment area
Type of inpatient treatment
Number of inhabitants January 1, 1985
Sigmaringen Konstanz
Departments of psychiatry State mental hospital (Reichenaul State mental hospital (Nordschwarzwald-Klinik) Department of psychiatry State mental hospital (Winnenden) State mental hospital (Weinsberg)
86,243
94.6
186,288
282.2
238,707
491.1
338,191
627.4
Boblingen Ludwigsburg
Population density/km*
Source: Statistisches Landesamt Baden-Wurttemberg (29). Statistisches Landesamt Baden-Wurttemberg (30).
A comparison of the group of patients in the aftercare of the social-psychiatric services after inpatient treatment (index group) with a group of patients that received no extramural care by the socialpsychiatric services after discharge from psychiatric hospital (control group) allows an estimate of the impact of the social-psychiatric services on the rehospitalization rate. For the index group, we expected a lower rate of rehospitalization, a longer period in the community between two inpatient episodes and a shorter time in hospital if rehospitalized. Besides case management, additional factors influence the rate of rehospitalization; it is important to control for such factors. We therefore used the design of a case-control study for our analysis. Known risk factors for rehospitalization are mainly diagnosis, sex, age, living conditions after discharge from psychiatric hospital and the number of previous inpatient treatment episodes (21). We therefore identified retrospectively for each of the 162 index patients a matched control patient identical in diagnosis and the other risk factors named above who was not referred to a social-psychiatric service after discharge from psychiatric hospital. By means of a survival analysis (using SPSS-X), we compared the groups with respect to the survival time in the community after discharge. Then we compared the length of inpatient treatment for the 2 rehospitalized subgroups again with survival techniques. The survival techniques, originally developed for the analysis of life tables, can be used for all problems in which the dependent variable is the interval between an initial and a termination event. For example the response can be the survival time in the community between discharge from psychiatric hospital and readmission to inpatient treatment. Survival function is the estimated cumulative probability of patients surviving from the beginning to the ith time interval, if a person has reached this interval without response. Censored observations mean
Does case management reduce the rehospitalization rate? that, for some cases, the termination event does not occur within the investigation period. So the survival time is the interval investigated. Both censored and uncensored observations are used in calculating survival time. With a D-statistic (D is asymptotically distributed as chi-square - Lee-Desu statistics) calculating differences between two survival functions, one can test the null hypothesis that subgroups are samples of the same survival distribution. The larger the D statistic, the greater the probability of (signific antly) different survival distributions. Results
Table2 shows the characteristics of the two samples. Both comprise 162 patients, 162 patients referred to the social-psychiatric services and 162 matched controls (twins), each identical in sex, age, living conditions, number of previous inpatient episodes and diagnosis who were not cared for by the social-psychiatric services. The studied samples had an excess of woman and the age group 25-44. For about one third of the patients it was the first inpatient treatment episode; another third of the patients had 1-2 previous psychiatric admissions. About one fifth of the patients had been hospitalized more than Table 2. Patient characteristics - identical for index and control group Sex male female
45% 55%
Age
5 24 years 2 5 4 4 years 45-64 years 2 6 5 years
16% 50% 26% 8%
Living conditions alone living in family no fixed abode unknown
29% 70% 1% 1%
Table 3 Face-to-face activities in social-psychiatric services - percentage of patients with the respective number of activities Number of monthly activities Kind of activity
0
51
52
3-4
5-8
>8
Assessment, planning counselling
8.8
17.7
27.2
33.1
9.5
3.8
Training in independent living skills
46.8
23.2
14.5
11.6
2.5
1.4
Coordination of professional support
32.7
29.8
19.8
12.1
3.9
1.8
Consultation with relatives
44.0
29.6
14.7
9.1
2.0
0.6
Maintenance and expansion of social networks
68.4
17.2
6.3
5.8
1.2
1.1
Home visits
35.4
27.1
20.2
13.3
2.8
1.2
Others
59.4
21.2
10.0
5.8
2.0
1.6
5 times. The predominant diagnosis was schizophrenia; about two thirds suffered from psychotic disorders. The average duration of care by the socialpsychiatric services for the index group was 6.2 months after discharge from psychiatric hospital. Table 3 shows the face-to-face-activities for index patients by the social-psychiatric services. The care offered for those patients is intensive. Assessment, planning and counselling was the basis of case management for about 90% of the index patients. These activities were constantly repeated: for more than half of the index patients twice a month, for about one third 3-4 times a month. About two thirds of these patients were visited at home, 37.5;/, more than twice a month. In general, the studied socialpsychiatric services displayed the whole array of case management activities.
Number of previous inpatient episodes
0 1-2 inpatient treatment episodes 2-5 inpatient treatment episodes > 5 inpatient treatment episodes unknown
31% 32% 18% 18% -
control group (n= 162) 0.7
0.6
Diagnosis schizophrenic/paranoiddisorders
(295, 297, 298) affective psychoses (296) neuroses, personality disorders, non-psychotic disorders (3OC-302, 306, 307, 31 1, 316.0) organic disorders (290, 293, 294, 310) acuteiadjustment reactions (308, 309) others
59%
P
0.5
index group (n.162)
12% 0.2
16% 2% 4% 6%
0.1
0
10
20
30
40
50
60
70
80
90
100
110
120
130
weeks
Fig. I . Comparison of the “survival time” in the community after discharge from hospital between index and control group by survival analysis
447
Rossler et al. Fig. 1 shows the survival analysis for the two samples compared. On the abscissa the analyzed time interval is marked in weeks. The ordinate shows the probability (P) of remaining in the community after discharge from hospital. So, for example, the probability of living in the community 10 weeks after discharge from hospital is P 2 0 . 9 6 for the index group, while the probability for the control group is P 2 0.89. About 40 weeks after discharge from hospital the probability of living in the community is almost identical for both groups. After that time the risk of being rehospitalized increases steadily for the index group, and remains fairly stable for the control group. At the end of the examined time interval (130 weeks), the probability of living in the community is P 2 0 . 5 4 for the index group, for the control group P20.79. The survival functions of the compared samples were not significantly different. For the patients who were rehospitalized during the evaluation period, we examined in a further step the time spent in hospital during this second admission. But again, the survival functions of the group that had been under the care of the social-psychiatric services and the control group were not significantly different ( P I 0.11).
Discussion Case management has been introduced to improve service delivery, mainly for chronically mentally ill people. One outcome criterion of case management is the rate of rehospitalization or the length of time in hospital in case of rehospitalization. The results of this study show that, during an observation period of 2.5 years, case management had no significant effects on the risk of rehospitalization or on the length of time in hospital in case of rehospitalization. These results have to be viewed against the background of intensive case management activities by the studied social-psychiatric services. Inevitably, the case load of the social-psychiatric services is low. This is a well-known phenomenon in case management (22). Considering the costs of such services, one is inclined to devalue case management at first view. Integrating our results into the reviewed literature, we have to discuss carefully the advantages and disadvantages of our study design. During deinstitutionalization, one major objective of the psychiatric reforms was reducing the number of psychiatric beds. Psychiatric hospitals were not supposed to provide residential care any longer but to focus mainly on psychiatric treatment. In 1984 the average length of hospitalization in a state mental hospital in Baden-Wurttemberg was 59 days and 33 days in a psychiatric department (23). This is a relatively short period compared with the traditional length of hospitalization. It therefore seems proba-
448
ble that hospital care in the studied catchment areas was mainly provided to treat mental disorders. One can doubt whether a decisive further reduction of length of time spent in hospital can be reached in such areas (ceiling effect). Although we know of no study on case management with a comparable thoroughly chosen control group, we also have to consider some methodological aspects. When selecting the samples, we chose well known sociodemographic and disease-related characteristics as criteria for the risk of rehospitalization after discharge from hospital. Nevertheless, we cannot be sure about the true extent of disability or chronicity in the patients of both groups, as the criteria chosen might be too coarse. The fact that more patients under the care of the social-psychiatric services were rehospitalized could still be because the patients referred to a social-psychiatric service were more disabled and vulnerable and thus more endangered for rehospitalization. One name given to these problem patients is “system misfits” (24,25), expressing that our care system cannot meet the needs of these patients. Some of the characteristics of the examined samples indicate that the social-psychiatric services deal with problem patients, in particular the young male schizophrenic, living alone and with several relapses in his psychiatric history. With respect to this patient group, it is easily understandable that intensive professional support can increase rather than decrease the probability of a referral to a psychiatric hospital, especially as one core element of case management is the link with resources. So a higher rehospitalization rate might indicate good quality of care. Additionally, as the utilization of a socialpsychiatric service is voluntary, patients of the control group might have refused a referral to a socialpsychiatric service. And for such patients there is a high probability that they also refuse rehospitalization. Indeed, we have to deal more frequently with problem patients who do not utilize the psychiatric care system according to their needs (26). This does not mean that this group does not need care. But some patients who are hostile and uncooperative with mental health professionals disappear from our records. After all, we have to ask whether the question “Does case management reduce the rehospitalization rate?” is the right question. This question might often be asked for economic reasons. For chronically mentally ill people themselves and their quality of life, discharge to the community - even interrupted by repeated, but short rehospitalizations might be preferable to a long - maybe life-long - stay in the back ward of a psychiatric hospital. On the other hand, case management for chronically mentally ill people in the community is rather regularly
Does case management reduce the rehospitalization rate? associated with a marked increase of service utilization in the community (8, 16, 27) and thus with an increase in costs. There is therefore an urgent need to identify the most effective elements in case management. In another study (28) on the differential effectiveness of case management activities for chronically mentally ill people, case management focusing on community functioning proved to be most effective when used within a long-term but low-contact treatment. Regarding the above-mentioned case management activities, training in independent living skills was more effective than therapeutic counselling. In view of coordination activities, linkage with psychiatric treatment made case management most effective. Thus, successful case management is nothing else but good psychiatric and rehabilitative treatment. Acknowledgement The scientific evaluation conducted by the Central Institute for Mental Health in Mannheim was financed by the Ministry of Labour, Housing, Family and Social Order, Baden-Wurttemberg.
References 1. ELLISONDL, RIEKERP, MARXJH. Organisational adaption to community mental health. In: ROMANPM et al., ed. Sociological perspectives on community mental health. Philadelphia: Davis Company, 1974. 2. KASTRUP M, NAKANEY, DUPONTA, BILLEM. Psychiatric treatment in a delimited population - with particular reference to outpatients. Acta Psychiatr Scand 1976: 53: 3550. 3. BAUERM. Sektorisierte Psychiatrie im Rahmen einer Universitatsklinik - Anspruch, Wirklichkeit und praktische Erfahrungen. Stuttgart: Enke, 1977. 4. ROSSLERW, HAFNERH. Psychiatrische Versorgungsplanung. Neuropsychiatrie 1985: 1: 8-17. 5. MAY PRA, SIMPSONGM. Schizophrenia: evaluation of treatment methods. In: KAPLANHI et al., ed. Comprehensive textbook of psychiatry. Vol. 111. Baltimore: Williams & Wilkins, 1980. 6. TESTMA, STEINLI. Community treatment of the chronic patient: research overview. Schizophr Bull 1978: 4: 350-364. 7. BRAWNP, KOCHANSKI G, SHAPIROR et al. Deinstitutionalization of psychiatric patients, a critical review of outcome studies. Am J Psychiatry 1981: 138: 736-749. 8. KANTERJ. Clinical case management: definition, principles, components. Hosp Community Psychiatry 1989: 40: 362368. 9. National Institute of Mental Health. Towards a model plan for a comprehensive community-based mental health system. Administrative document. Washington, DC: US Department of Health and Human Services, 1987.
10. CLIFFORD P, CRAIGT. Case management systems for the long-term mentally ill. A proposed interagency initiative. London: NUPRD, 1988. 11. RENSHAW J. Care planning and case management. Br J Soc Work 1987: 18: 79-105. 12. INTAGLIATAJ. Improving the quality of care for the chronic mentally disabled: the role of case management. Schizophr Bull 1982: 8: 655-674. 13. Caring for people. London: HMSO, 1989. 14. HOLLOWAY F. Case management for the mentally ill: looking at the evidence. Int J SOCPsychiatry 1991: 37: 2-13. IS. FRANKLIN J, SOLOWITZB, MASONM. CLEMONSJR, MILLERGE. An evaluation of case management. Am J Public Health 1987: 77: 674-678. 16. BORLAND A, MCRAEJ, LYCANC. Outcomes of five years of continuous intensive case management. Hosp Community Psychiatry 1989: 40: 369-376. 17. STEINLI, TESTMA. Alternative to mental hospital treatment. Arch Gen Psychiatry 1980: 37: 32-37. 18. HOULTJ, REYNOLDS I, CHARBONNEAU-POURIS M. Psychiatric hospital versus community treatment: the results of a randomized trial. Aust NZ Psychiatry 1983: 17: 160-167. 19. MARKSI, CONNOLLY J, MUIJENM. The Maudsley Daily Living Program. Bull R Coll Psychiatr 1988: 12: 22-24. 20. FREEMANHL, FRYERS TH, HENDERSONJH. Mental health services in Europe: 10 years on. Copenhagen: World Health Organization Regional Office for Europe, 1985. 21. AN DER HEIDENW, KRUMMB, HAFNERH. Die Wirksamkeit ambulanter psychiatrischer Versorgung - Ein Modell zur Evaluation extramuraler Dienste. Berlin: Springer, 1989. 22. THORNICROFT G. Case managers for the mentally ill. SOC Psychiatry Psychiatr Epidemiol 1990: 25: 141-143. 23. ROSSLERw , HAFNERH, MARTINIH, AN DER HEIDENw , JUNGE, LOFFLERW. Landesprogramm zur Weiterentwicklung der auljerstationaren psychiatrischen Versorgung BadenWurttemberg - Analysen, Konzepte, Erfahrungen. Weinheim: Deutscher Studien Verlag, 1987. 24. BEESONPG. The bureaucratic context of mental health care. In: LITTRELWB et al. ed. Bureaucracy as a social problem. Greenwich, CT: JAI Press, 1983. 25. JOHANSEN KH. The impact of patients with chronic character pathology on a hospital inpatient unit. Hosp Community Psychiatry 1983: 34: 842-846. 26. BACHRACH LL, TALBOTT JA, MEYERSON AT. The chronic psychiatric patient as a “difficult” patient: a conceptual analAT, ed. Barriers to treating the chronic ysis. In: MEYERSON mentally ill. New directions in mental health services. 1987: 35-50. RD, WARDR. As27. BOUDGR, MILLERLD, KRUMWIED sertive case management in three CMHC’s: a controlled study. Hosp Community Psychiatry 1988: 39: 41 1-418. 28. ROSSLERW, FATKENHEUER B, LOFFLERB. Soziale Rehabilitation Schizophrener in Sozialpsychiatrischen Diensten. Stuttgart: Enke (in press). 29. Statistisches Landesamt Baden-Wurttemberg. Statistische Berichte. Voraussichtliche Entwicklung der Bevolkerung auf der Basis vom 1.1.1984. Stuttgart, 1985. 30. Statistisches Landesamt Baden-Wurttemberg. Gemeindestatistik 1985. Amtliches Gemeindeverzeichnis BadenWurttemberg 1985. Stuttgart, 1985.
449
