INTRODUCTION
Case management for severe mental illness (SMI), defined as “an approach to service delivery that ensures that people with mental illness who have complex multiple problems and disabilities receive all the services they need” [1], has been a primary component of community mental health services in Korea since the early 1990s. The 2016 revision of the Korean Mental Health Act further reinforced the focus on community-based mental health services and dehospitalization. Consequently, case management services for SMI have gained increased attention as core offerings of community mental health welfare centers across Korea.
Case management for SMI can be categorized by the level of direct service provision and the case manager-to-patient ratio [2]. Traditional models, such as broker, clinical, strength, and rehabilitation, typically maintain a staff-to-patient ratio of 1:20 to 1:50. As of 2023, case management services in Korea have aligned with these models, with a national average ratio of 1:25.2 [3].
However, due to intrinsic limitations in the effectiveness of low-intensity service models [4], intensive case management (ICM) models, such as assertive community treatment (ACT), were developed and are regarded as more effective and holistic. Existing literature broadly defines ICM as a package of care modeled after ACT, assertive outreach, or similar case management approaches, with a caseload of up to 20 people [5]. Compared to standard care (outpatient care without case management), ICM has demonstrated effectiveness in improving outcomes for individuals with SMI, including reduced hospital admissions, better retention in care, and improved social functioning, particularly in models with high fidelity to ACT [5-7]. However, the advantage of ICM over standard case management (non-ICM) is not consistently defined in prior literature [5].
In Korea, although the case manager-to-patient ratio of 1:25.2 represents a significant improvement from the 2000s, this figure can be misleading as case managers in community mental health welfare centers cannot focus solely on case management. These centers are involved in numerous community mental health services, including suicide prevention, disaster-related trauma programs, general mental health promotion, and various government-mandated programs. The authors frequently observed case manager-to-patient ratios exceeding 1:50, with instances where the ratio surpassed 1:100 or more being common.
Since 2017, ICM services have been developed and implemented to support patients with SMI and complex needs through regional mental health welfare centers in Seoul, Busan, and Gyeonggi.8-10 This approach aims to redirect case management efforts toward the most critical needs, enhancing the effectiveness of community care for patients with SMI.
However, there is a paucity of research on the effectiveness of ICM in Korea. One study evaluated the effectiveness of the Seoul-intensive case management (S-ICM) service using monitoring data from April 2019 to March 2020.11 The findings showed that patients during the ICM period had shorter hospital stays than those in the pre-ICM period. Additionally, subgroup analysis indicated that psychiatric hospitalizations were shorter in the post-ICM observation period compared to the pre-ICM period. However, this study had limitations as a retrospective pre-post comparison. Given these limitations, the current study was designed to confirm the effectiveness of S-ICM through an randomized controlled trial (RCT), comparing S-ICM to non-ICM.
METHODS
Study design
The study consists of two phases: a 3-month S-ICM trial detailed in this study and a 6-month post-trial observation to be published after the study’s completion. We report the trial results of the first phase to address the urgent need for evidence on the effectiveness of ICM in reforming case management services in Korea.
As the intervention (S-ICM) is a human service, study participants could not be blinded. Therefore, the study was designed as an observer-blind RCT to evaluate the efficacy of SICM compared to non-ICM. Study participants were recruited from community mental health welfare centers around Seoul. The inclusion criteria were: 1) primary diagnosis of schizophrenia, schizoaffective disorder, bipolar disorder, or recurrent depressive disorder; 2) age between 18 years and 65 years; and 3) living in the community. Individuals currently receiving SICM services or already flagged for placement under S-ICM services were excluded from the study. A 6-month washout period was required for those with prior S-ICM experience to be eligible for participation. Additionally, individuals with comorbid intellectual disabilities, pervasive developmental disorders, cognitive disorders, or organic mental disorders were excluded. Investigators verified the inclusion and exclusion criteria, obtained informed consent from eligible participants, and conducted baseline assessments. Individuals already receiving S-ICM or not meeting the 6-month washout period were excluded. Additionally, those already assigned to S-ICM by their referral centers were excluded due to ethical constraints preventing randomization.
Following the baseline assessment, participants were randomly assigned to either the S-ICM or the non-ICM control (treatment-as-usual) group. The group assignment was communicated to each community mental health welfare center, which provided the respective S-ICM or non-ICM services without the knowledge of the investigators. During the study, the S-ICM group received S-ICM services for 3 months, while the control group received non-ICM services for the same period. Participants in both groups were re-evaluated by investigators blinded to group assignment at the end of the 3-month trial.
The core differences between S-ICM and non-ICM in Korea are: 1) a low patient-to-staff ratio (fewer than five patients per case manager); 2) weekly supervisions and monthly reviews for case managers provided by each community mental health welfare center’s ICM committee; 3) a limited service duration where S-ICM reverts to non-ICM upon achieving the objectives set by individual service plans; and 4) at least weekly face-to-face, 40-minute case management sessions provided in the patient’s living environment. More details regarding SICM service’s profile can be found in prior publications and the service manual [8,11].
Participants
Eighty-six individuals from 11 community mental health welfare centers in Seoul were referred to the study. Of these, six did not meet the inclusion/exclusion criteria, resulting in 80 participants eligible for enrollment and randomization. Following randomization, 41 participants were assigned to the S-ICM group and 39 to the control group. However, two participants in the control group withdrew their consent before the baseline evaluation, reducing the control group size to 37. No participants dropped out during the trial (Figure 1).
Outcome
We examined demographic factors, diagnoses, reasons for enrollment, service-related profiles, major life events related to mental health, and psychiatric symptoms of participants at baseline and the 3-month outcome evaluations. Service-related profiles included case management duration and the occurrence of treatment-related events before and during the trial. Data on major life events related to mental and physical health, including psychiatric and non-psychiatric admissions, emergency service use, self-harm or suicide attempts, criminal offenses, and experiences of abuse or neglect, were collected with the assistance of case managers who provided services to participants during the 3-month trial. Overall, psychiatric symptom severity was assessed using the Brief Psychiatric Rating Scale (BPRS) [12], and mood symptoms were evaluated with the Montgomery-Åsberg Depression Rating Scale (MADRS) [13,14]. The Health of Nations Outcome Scale (HoNOS) measured psychiatric outcomes across four domains [15]. Quality of life, disability, and perceived social support were assessed using the short form of the WHO Quality of Life scale (WHOQOL-BREF) [16,17], the WHO Disability Assessment Schedule 2.0 (WHODAS 2.0) [18,19], and the Multidimensional Scale of Perceived Social Support (MSPSS) [20], respectively. Additionally, the Clinical Global Impression-Improvement (CGI-I) [21] scale was used to evaluate the global impression of improvement.
Data analysis
Demographic characteristics and diagnoses between the SICM and non-ICM groups were compared using χ2-tests. Pretrial and during-trial service-related characteristics were compared using independent t-tests. The risk of major treatmentrelated events during the 3-month trial was assessed, with adjusted odds ratios (aORs) for each event obtained through multivariate logistic regression analysis using the backward elimination method. To compare changes in BPRS, MADRS, HoNOS, WHOQOL-BREF, WHODAS, and MSPSS between the two groups, an analysis of covariance (ANCOVA) was conducted to account for baseline differences and control for potential confounding variables. In the final model, age groups and the license qualifications of case managers were included as significant risk factors based on odds ratio analysis for major life events.
The trial phase of this study was approved by the Institutional Review Board of the Seoul Medical Center (IRB No.2023-03-005), and the observation phase was approved by the Institutional Review Board of Seoul National University Hospital (H-2405-076-1536). Statistical analyses were conducted using IBM SPSS Statistics, version 29 (IBM Corp.).
RESULTS
Table 1 provides the demographic and clinical characteristics of participants. No significant differences between the SICM and control groups regarding demographic characteristics and diagnoses were observed. Among all participants, 56.6% were female and 43.4% were male. Approximately onethird (34.6%) of participants were aged 50 to 59, followed by 23.1% in the 40 to 49 years age range. Additionally, 61.5% of participants lived alone. Approximately two-thirds (62.8%) of participants were supported by medicaid, and the majority (88.5%) were unemployed, indicating a lower socio-economic status among the study population. Diagnostically, 62.8% of participants were diagnosed with psychotic disorders, including schizophrenia and schizoaffective disorders, while the remaining participants were diagnosed with mood disorders, such as bipolar disorder and recurrent major depressive disorder.
Table 2 summarizes the pre-trial treatment history and service-related characteristics, as well as during-trial service-related characteristics. For treatment history, the mean duration of psychiatric treatment was 14.6 years for the S-ICM group and 16.7 years for the control group. The mean duration of untreated psychosis was 36.9 months for the S-ICM group and 23.0 months for the control group. The mean lifetime duration of psychiatric hospitalization was 79.3 months for the S-ICM group, compared to 42.4 months for the control group. However, these characteristics showed no significant differences between the S-ICM and control groups. Regarding case management history, the mean duration under case management was 60.8 months for the S-ICM group and 186.6 months for the control group, indicating a significantly longer duration of case management for the control group. In the year preceding the trial, the mean frequency of case manager contact was identical for both groups, with a mean of 1.1 contacts. As expected, the participants in the S-ICM group engaged with their case managers more frequently during the trial. The experience levels of case managers in both the S-ICM and control groups were comparable. However, there was a notable difference in the Mental Health Professional licensing status, a government qualification for mental health specialists, including nurses, social workers, clinical psychologists, and occupational therapists in Korea, among the case managers assigned to participants. In the S-ICM group, only 12.2% of case managers were non-licensed, compared to 32.4% in the control group.
We examined the occurrence of treatment-related events, including psychiatric admissions, emergency room (ER) visits, self-harm or suicide attempts, criminal offenses, and experiences of abuse or neglect during the trial. Preventing such incidents is a core objective of S-ICM, making them primary targets for this intervention. Logistic regression analysis was conducted to compare the risk of these events over the 3-month trial period between the S-ICM and control groups (Table 3). In the unadjusted model, there was no significant difference in the risk of life events between the S-ICM and control groups. However, when adjusted for age group (in 10-year increments), case manager licensing status, history of psychiatric admissions in the past 5 years, diagnostic group (psychotic disorders vs. mood disorders), and health insurance status (health insurance vs. medicaid), the odds ratio for the risk of self-harm or suicide attempts was significantly lower in the S-ICM group (aOR=0.30, 95% confidence interval [CI]: 0.09-0.94). After adjustments, other events did not show significant differences between the S-ICM and control groups. We examined the occurrence of negative life events, including interpersonal problems, divorce or separation, taking a loan, financial issues, the death of family members, job loss, and reduced welfare support. These events were rare, with job loss being the most frequent (n=3). When aggregated, 61.0% of the SICM group and 75.7% of the control group reported at least one major negative life event, with an aOR of 0.27 (95% CI: 0.08-0.92).
The changes in psychiatric symptoms, disability, quality of life, and perceived social support during the 3-month trial were evaluated using ANCOVA adjusted for baseline scores and controlled for age groups and the license qualifications of case managers (Table 4). The S-ICM group exhibited significantly improved BPRS and MSPSS scores compared to the control group. The S-ICM group demonstrated an average change of -11.78 (SD=15.23) on the BPRS, while the control group showed an average change of -7.19 (SD=13.77). The MSPSS indicated a significant improvement for the SICM group, with an average change of 1.15 (SD=10.66) compared to the control group’s change of 0.41 (SD=7.72). Other measures, such as MADRS, HoNOS total and subscales, WHOQOL domains, and WHODAS, did not show significant differences in score changes between the two groups.
Overall improvement, assessed using the CGI-I, revealed that the S-ICM group had an improvement ratio of 70.7% (improvement defined as CGI-I scores of 1 to 3, encompassing very much improved to minimally improved), while the control group had an improvement ratio of 29.7%. The aOR for CGI-I improvement, calculated using multivariate logistic regression and adjusted for age groups and case manager license status, was 8.20 (95% CI: 2.66-25.32). As noted earlier, follow-up research on the observed effects from this trial is ongoing as of 2024, with findings to be published in a subsequent article.
DISCUSSION
In Korea, the effectiveness of ICM has not been well documented. One community mental health welfare center in Gyunggi province provided ACT for individuals with SMI starting in 2008, nearly a decade before the formal introduction of ICM services at the provincial level and reported favorable outcomes over a 1-year period in a small, non-equivalent controlled observational study [22]. Aside from that study and some incidental, unpublished internal reports from a few community mental health welfare centers, there has been no literature in peer-reviewed journals until an observational study reported the effectiveness of S-ICM following its introduction in Seoul [11]. Thus, to the best of the authors’ knowledge, this study represents one of the few, if not the only, RCTs on ICM services in Korea.
Several noteworthy outcomes emerged from this trial. The finding of a significantly lower risk of self-harm or suicide attempts demonstrates that S-ICM, even over a brief 3-month period, can be more effective than non-ICM in Korea. Additionally, the observed lower risk of negative life events may be considered an indirect effect of S-ICM. The impact of SICM on reducing self-harm and suicide attempts, as well as negative life events, may be partially attributed to the strengthened social networks developed through increased engagement with S-ICM case managers. Many individuals with serious mental illness often experience social isolation. Expanding their social networks could protect against adverse life events that might lead to self-harm or suicidal behavior. The significantly improved perceived social support, measured by the MSPSS during the trial, supports this hypothesis.
The results further indicate that the 3-month ICM service led to modest but significant improvements in psychiatric symptoms, compared to non-ICM. The symptomatic relief observed was unexpected, as it was not a primary objective of S-ICM. It is possible that the supportive, cognitive, and behavioral approaches employed during individual S-ICM sessions led to improved disease and medication management.
However, reductions in hospitalization, which are a primary goal for community services for SMI, were not observed in the S-ICM group in this study. Although a prior observational study using a pre-post design showed decreased hospitalization rates for the S-ICM group [11], the authors anticipated finding reduced hospitalizations or at least fewer ER visits among S-ICM participants, which did not materialize in this trial.
Several possible explanations exist for the non-effectiveness of hospitalization observed in this trial. Firstly, the overall frequency of hospitalization among participants was low, which may account for the lack of significant change in hospital days. It is possible that the 3-month intervention duration was insufficient to detect an effect on hospitalization. This limitation may be addressed in the second part of the study, where a 6-month post-intervention outcome will be observed. Additionally, the number of participants in this trial may have been insufficient in detecting changes in such infrequent hospitalization events. The aforementioned observational study of S-ICM in Korea included a much larger sample size of 759 participants. Lastly, ICMs, including S-ICM, are designed to support individuals with more severe and uncontrolled symptoms compared to traditional case management services. However, there is concern that the severity of symptoms among participants in this trial may have been relatively low. This is due to research ethics considerations and the clearance period for study participation, which led to the exclusion of individuals who were already receiving SICM due to high severity at community mental health centers. This exclusion likely resulted in a sample that did not include the most severely ill individuals, who would be more prone to hospitalization. Previous meta-analyses of S-ICM have concluded that its effect on overall hospital admissions is modest, with a significant reduction in hospital time primarily observed in patients with high baseline hospital use [5].
Additional limitations exist in this study. The nature of ICM services as human services made it impossible to conduct a double-blinded trial. Additionally, a significant portion of case managers responsible for delivering services to both the S-ICM and control groups were not licensed mental health professionals. This was unavoidable due to the recent shortage of licensed professionals in the Korean community mental health field. Moreover, the lower proportion of licensed mental health professionals among case managers in the control group raises concerns about potential selection bias. However, other demographic and clinical factors were randomized adequately, and, importantly, the average years of experience of case managers were comparable between the two groups. Non-licensed case managers are still healthcare professionals, such as registered nurses and social workers. Lastly, unlike ACT, the S-ICM lacks a structured mechanism for measuring fidelity to service, which raises potential concerns about quality variability among community mental health welfare centers in Seoul. However, the Seoul Metropolitan Mental Health Welfare Center, which develops and provides manuals and protocols for S-ICM, conducts continuous monitoring and feedback to these centers to ensure service quality [23]. Therefore, concerns about quality differences may be minimal.
Despite the brief 3-month intervention period for participants with relatively low initial severity, the overall outcomes for the intervention group were superior to those of the control group in terms of reducing suicidal and self-harm behaviors, major life event changes, significant improvements in psychopathological symptoms, and preventing social deterioration. The efficacy of the intervention would be more pronounced with a longer intervention period and a higher initial severity among participants.
Case management is a principal function of Mental Health Welfare Centers, serving as community hubs for mental health services in Korea. However, there is increasing recognition of the need for more intensive, integrated, and proactive services for individuals with SMI. This study provides foundational evidence for the potential expansion of ICM services nationwide. However, to implement ICM nation-wide, several strategies to overcome the chronic personnel shortages in community mental health centers are required. Raising awareness among policymakers and the public regarding prioritizing individuals with serious mental illnesses in community mental health services is important. This can drive an increase in budget and personnel for community mental health systems in Korea. A standardized ICM service manual that accounts for regional socio-demographic variations across Korea should be developed. This manual should incorporate adaptable service components and a supervision system that addresses the diverse needs of local populations while maintaining the fundamental principles of ICM. Robust quality assurance for ICM should be implemented. This involves both monitoring by regional or national administrations and regular multidisciplinary supervision and support within community mental health welfare centers.
Further research on ICM is also required, including studies on long-term efficacy and the development of standard protocols for nationwide implementation. These issues will be addressed in the second part of this study.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
Print
