Introduction
In the context of deinstitutionalization, it is widely accepted that most people with severe mental illness do not need to reside in the hospital if they receive support from community mental health services [
1,
2]. Case (care) management, a method to coordinate services for meeting their needs [
3], is a key component of assistance for people with mental illness to live in the community. There are several models of case management with varying levels of intensity. In broker case management, the case manager’s role is to connect people with mental illness to necessary services; but with a staff-to-client ratio of approximately 50, interventions by the case manager and outreach services are poor. Broker case management was shown to be ineffective, particularly for people with severe mental illness [
4]. A case management model, which includes a staff-to-client ratio of less than 20, services directly provided by the case manager, and outreach services is generally considered intensive case management (ICM). For instance, assertive community treatment is a well-known model of ICM [
5]. A Cochrane review showed that ICM services reduce hospitalization rates, increase retention in care, and improve social functioning compared to standard care [
6]. However, ICM services appear to be considerably effortful and costly to provide for all patients with mental disorders; mental health resources in the real world are clearly limited. Considering the efficiency of community mental health services [
7], it is imperative to assess the proper allocation of ICM services to people with severe mental illness who need them.
In Japan, the mental health system has been shifting from hospitalization-based to community-based care [
8]. In addition, some organizations have gradually begun to provide ICM services. However, psychiatric hospitals and mental health clinics in Japan accept all people who potentially have mental health problems regardless of the care level needed [
9]. Thus, it is important that psychiatrists and other professionals assess the need for ICM services to efficiently provide services.
People who need ICM services may be defined as those who have problems living in the community without ICM services. The current level of functioning is associated with the level of care [
10]; thus, it seems to be useful as a screen for ICM need. Moreover, since ICM is effective for reducing hospitalization rates and improving social functioning [
6], ICM is considered necessary for patients at risk of frequent readmissions and challenges in adapting to living in the community. Systematic reviews have pointed out that medication adherence and family involvement are factors associated with readmission [
11,
12]. Prisoners, inmates, and probationers frequently have mental illness [
13], and recidivism is high in people with and without mental illness [
14]. Previous reviews have pointed out that poverty might cause mental illness, and mental illness worsens socioeconomics status [
15,
16]. Therefore, we hypothesized that the need for ICM services is based not only on current functioning, but also on these predictors of readmission and future challenges, i.e., medication adherence, family involvement, risk of criminal activity, and financial problems.
Assessment of functioning and prediction of readmission and future challenges in people with mental illness rely on multidisciplinary assessment by psychiatrists, mental health nurses, social workers, and occupational therapists. Such assessments are made in different contexts. Multidisciplinary professionals such as nurses, social workers, and occupational therapists are involved in the care of people with mental illness under different conditions [
17]. For example, nurses in general have many opportunities to interact with and observe people in medical settings [
18]. Social workers help their patients receive social and community services [
19]; thus, they usually have information about family, financial, and social circumstances. Occupational therapists may support their patients in acquiring occupational and daily skills [
20]; they usually assess functioning and characteristics related to occupational experiences and daily activities through their clinical practice. In summary, the views of multidisciplinary professionals can shed light on multiple aspects of people with mental illness that might be helpful for identifying the need for ICM services.
Over the past three decades, several scales have been developed to assess levels of care and functioning [
21‐
25]. The Global Assessment of Functioning (GAF) [
21] and Personal and Social Performance Scale (PSP) [
22] are common tools measuring functioning in people with mental illness. Lyons et al. [
23] developed the Severity of Psychiatric Illness Scale to assess the need for psychiatric admission. These measures do not include all of the factors associated with the need for ICM services, since they were not specifically created for assessing this need. Furthermore, algorithms for determining levels of care were proposed in previous studies [
26,
27]. The Level of Care Utilization System for Psychiatric and Addiction Services (LOCUS) was developed by the American Association of Community Psychiatrists [
26] and used by some Japanese institutions [
28]. LOCUS consists of six dimensions of five detailed grading systems. It classifies people with mental illness into six levels, from outpatients via ICM to inpatients [
26]. Raters need training to precisely evaluate people with mental illness according to the five detailed grading systems. In addition, if clinical staff members use LOCUS, they may need much effort to gather the needed information. In other words, it may be hard for professionals in various disciplines to acquire the relevant information in routine clinical practice to use LOCUS. Taken together, for the Japanese mental health system, we considered it important to develop a scale that comprehensively includes factors associated with the need for ICM and is easy for professionals from various disciplines to rate.
In this context, we developed the Intensive Case Management Screening Sheet (ICMSS) to briefly assess the need for ICM services in people with mental illness. ICMSS consists of items relevant to current functioning and predictors for readmission and future challenges in living independently. In addition, we developed ICMSS for professionals from various disciplines to use easily in clinical practice. Using data from two institutions that voluntarily provided ICM services to people with mental illness [
29,
30], we aimed to examine the psychometric properties of ICMSS and its discriminative ability for users and non-users of ICM services.
Discussion
In this study, we developed the ICMSS to assess the need for ICM services among people with mental illness and examined its psychometric properties and discriminative ability. Exploratory factor analysis showed a one-factor structure and 14 items were selected by the factor loadings. Confirmatory factor analysis also supported the factor structure. ICMSS had a significantly larger AUC than GAF and PSP, indicating that it had better discriminative ability. Previous studies have identified some predictors such as medication adherence and family involvement [
11,
12]. GAF and PSP are associated with current functioning, whereas ICMSS was developed to assess both current functioning and predictors for readmission and future challenges in living independently. ICM services are necessary not only to support people with current problems, but also to prevent readmission and future challenges in living independently. Thus, we suggest that ICMSS is useful for assessing the need for ICM services in people with mental illness.
For ICMSS, sensitivity was high for both datasets when the cutoff was ≥ 1, but drastically lower with a cutoff of ≥ 2. Hence, we suggest that a cutoff of ≥ 1 is appropriate for screening whether ICM services are needed. However, the AUC was not very large, even in the stepwise logistic regression analysis. In this study, ICMSS was rated by a single mental health professional on a multidisciplinary treatment team; thus, the ICMSS score for each participant does not reflect the views of professionals in other disciplines. Flexible clinical judgment seems important for decision-making about whether to provide ICM services. Therefore, we thought that the need for ICM services was being determined by quantitative assessments (i.e., ICMSS) and clinical judgment.
We performed ROC analysis to evaluate the discriminative ability of ICMSS. For one dataset (
n = 715), we excluded unknown ICMSS responses. Unknown and missing responses were grouped together for ICMSS scoring for the other dataset (
n = 911). Regardless of the difference in scoring, the results from the two datasets showed similar discriminative ability. Grouping unknown and missing responses appeared to be useful for clinical ICMSS scoring. Moreover, in the Camberwell Assessment of Need, which is used broadly in clinical settings, a similar scoring procedure is applied for unknown responses [
35]. Although it is necessary to gather as much information about users as possible for ICMSS scoring, we may recommend grouping together unknown and missing responses for scoring in a clinical setting.
Algorithms for determining the level of care were proposed in previous studies [
26,
27]. Ideally, these algorithms can be used in the Japanese system. It is expected that community mental health services, including ICM services, are provided according to actual resource and community characteristics [
7]. In Japan, the mental health system is shifting from hospitalization-based to community-based care [
8]. However, ICM services are not fully supported by the Japanese system; the costs of ICM services are borne by the medical institution. Thus, the criteria for providing ICM services changed flexibly according to resource and staff assignments at each institution in Japan. Algorithms provide a precise classification of the level of care, but they are inflexible. ICMSS helps in the acquisition of information by professionals of various disciplines that we consider useful for flexible judgment in Japanese institutions.
In confirmatory factor analysis, the χ2 test was significant, suggesting poor model fit. Generally, the χ2 test is sensitive to sample size and tends to reject models with a large sample. Our sample was relatively large. However, other indices showed excellent fit. Thus, we believe that a one-factor model is appropriate for ICMSS. The age range of our participants was broad (20–92 years). In the stepwise logistic regression analysis, ICMSS remained significant in the model including age. However, in Japan, persons aged > 65 years can choose to use welfare system services for elderly people, which might have affected the present results. This study included patients from two psychiatric institutions. The generalizability of findings may be limited. In addition, participants were recruited from patients who received psychiatric outpatient services, psychiatric day care, or outreach services. We did not obtain data from current and recently discharged inpatients of psychiatric hospitals. The results might vary when focusing on inpatients.
In summary, we developed ICMSS for nurses, social workers, and occupational therapists to use easily in clinical practice. ROC analysis showed that ICMSS has better discriminative ability between users and non-users of ICM services than GAF and PSP. Thus, we suggest that ICMSS is a useful screening tool that reflects the views of professionals of various disciplines. However, the AUC of ICMSS was not large. The need for ICM services might be associated with quantitative assessment (i.e., ICMSS) and clinical judgment. In Japan, ICM services are provided only by a limited number of advanced organizations. We expect that ICMSS will be used as a measure that reflects the views of professionals from various disciplines in several Japanese institutions, which may contribute to the further spread of ICM services in Japan.
Authors’ contributions
KS: made substantial contribution to the analysis and writing of the manuscript. SY CF: involved in designing the study, data collecting, analyzing the data, and drafting the manuscript. YK, KN, TA, NH: involved in data collection and drafting and revising the manuscript. All authors read and approved the final manuscript.