Community Ment Health J DOI 10.1007/s10597-013-9652-0
ORIGINAL PAPER
Can Assertive Community Treatment Remedy Patients Dropping Out of Treatment Due to Fragmented Services? Marjan Drukker • Wijnand Laan • Fred Dreef Ger Driessen • Hugo Smeets • Jim Van Os
•
Received: 11 October 2012 / Accepted: 14 October 2013 Ó Springer Science+Business Media New York 2013
Abstract Previously, many patients with severe mental illness had difficulties to engage with fragmented mental health services, thus not receiving care. In a Dutch city, Assertive Community Treatment (ACT) was introduced to cater specifically for this group of patients. In a pre–post comparison, changes in mental health care consumption were examined. All mental health care contacts, ACT and non-ACT, of patients in the newly started ACT-teams were extracted from the regional Psychiatric Case Register. Analyses of mental health care usage were performed comparing the period before ACT introduction with the period thereafter. After the introduction of ACT, mental health care use increased in this group of patients, although not all patients remained under the care of ACT teams. ACT may succeed in delivering more mental health care to patients with severe mental illness and treatment needs who previously had difficulties engaging with fragmented mental health care services.
M. Drukker (&) G. Driessen J. Van Os Department of Psychiatry and Psychology, School for Mental Health and Neuroscience MHeNS, Maastricht University, P.O. Box 616, Location Vijverdal, 6200 MD Maastricht, The Netherlands e-mail: [email protected] W. Laan H. Smeets Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht, Utrecht, The Netherlands F. Dreef De Care Express, Utrecht, The Netherlands J. Van Os King’s Health Partners, Department of Psychosis Studies, Institute of Psychiatry, King’s College London, London, UK
Keywords Mental health care use Assertive Community Treatment Psychiatry Severity of illness
Introduction Assertive Community Treatment (ACT) is a method of organizing integrated mental health care for the most severely mentally ill. ACT has its origin in the USA and is designed to provide community-based mental health care to persons diagnosed with severe mental illness (Michigan Recovery Centre of Excellence 2012; National Alliance on Mental Illness 2012; The Iowa Consortium for Mental Health 2012). It is a service-delivery model including multidisciplinary teams that coordinate all care, tailor-made to the individual patient (Marshall and Lockwood 2000). Thus, ACT is ideally suited to provide treatment to patients diagnosed with severe mental illness and comorbid substance abuse who find it difficult to engage in a sustained way with mental health services (Bond and Salyers 2004). The majority of this group is homeless or has legal problems. ACT teams aim to have contact with patients on a regular basis, with an intensity as high as needed. Drop-out from treatment is avoided by assertive outreach; this implicates that when patients are reluctant or uncooperative, ACT team members keep in contact and offer services, e.g. by paying home visits (Marshall and Lockwood 2000). During periods of inpatient-care, ACT teams keep in contact with their patients, even when they are treated by a different team of mental health professionals (Marshall and Lockwood 2000). Research has established that in the Netherlands and other countries, ACT represents an effective model of service delivery (Bhugra et al. 2011; Ito et al. 2011; Marshall and Lockwood 2000; Sytema et al. 2007). In the city of Utrecht,
123
Community Ment Health J
the fourth largest city in the Netherlands (source population: 760,000), ACT was considered specifically to cater regionally for the group of patients with severe mental illness who previously had difficulties engaging with fragmented services. The background for this change in model of service delivery was that in the Netherlands, similar to many European countries, in the late 1990s, existing services such as traditional outpatient mental health care and hospital-based services failed to cater for a large group of patients diagnosed with severe mental illness and comorbid addiction. These patients typically had severe functional impairments, frequently were homeless, and presented with forensic problems. Utrecht mental health services identified three reasons why existing services failed to reach these patients. The most important reason was that patients refused help or missed appointments. Traditional services subsequently coped by excluding these patients from their care. Second, patients had a complex combination of needs while institutions typically catered for only one at the time. For example, a patient with psychotic disorder, comorbid addiction and cognitive impairment required care from four different institutions: the mental health service to treat the psychotic disorder, the addiction services, the community social worker to negotiate debts, and a rehabilitation service focussing on the cognitive impairment, in the absence of significant coordination or collaboration between these services. Third, some patients were showing aggressive and/or antisocial behaviour. For example, they threatened the mental health professional, dealt in drugs or extorted money from other patients. Existing services typically had no response to these behaviours and went on to exclude these patients from their care. In Utrecht, ACT teams were introduced in 2005 in order to provide care to the group of patients that had difficulties engaging with care and remedy the existing impasse between institutions delivering fragmented mental health care. Previous research has shown that adherence of the Utrecht ACT-teams to the ACT model, as assessed with the Dartmouth Assertive Community Treatment Scale (DACTS), was good (Van Vugt et al 2011). Given the specific purpose for the introduction of ACT, an increase in mental health care consumption in the group of patients presenting difficulties to engage with services can be considered a successful outcome. Thus, the present paper studies the impact of the Utrecht ACT teams, by assessing changes in health care use in the group of patients assigned to these teams.
Methods Setting Utrecht is the smallest of the four major cities in the Netherlands. It has about 300,000 inhabitants and is similar
123
to most towns of the same size in western European countries, most inhabitants having a white European background (CBS (Statistics Netherlands) 2012). Psychiatric Case Register Middle Netherlands The Psychiatric Case Register Middle Netherlands (PCRMN) (Smeets et al. 2011) registers the mental health care consumption of over 760,000 inhabitants from the greater Utrecht region. The PCR-MN uniquely registers inpatient care, outpatient care, emergency care, day care and assisted living on a patient-based, but anonymized level (Smeets et al. 2011). For the present analysis PCR-MN data between 1999 and 2009 were used. Preparation of the Data In order to compare health care use of patients in ACT with health care use of patients not in ACT, health care episodes were defined for each patient based upon a set of rules. Periods of health care use were defined as ACT episodes if a patient had ACT contacts within this period. Because patients could also have non-ACT contacts between ACT contacts, ACT episodes did not end at the moment the patient had any non-ACT contact, but only when a patient did not have a contact with an ACT-team for 61 days. This period of 61 days was chosen because patients should have an ACT-contact at least once a month. Thus, when a patient did not have a contact with the ACT-team for 61 days or more, the implication is that two or more consecutive ACT contacts were missed. An exception was made when patients were admitted for a period not exceeding 3 months directly after an ACT-contact, this admission was combined with the previous ACT-episode, because ACT usually continues during short-term admissions. All other periods of health care were defined as non-ACT episodes. A health care episode ended when a patient did not have contact with mental health services for 100 days or more. If a patient did not use any mental health care for 100 days or more it was assumed that the patient moved or was outside the reach of ACT (e.g. forensic services). Therefore, periods in which no care was used were not included in the analyses. In order to conduct the pre–post comparison, two types of non-ACT episodes were defined: those before the start of ACT (pre-ACT) and those after having experienced one or more ACT episodes (post-ACT). Over the course of each episode, the proportion of the different types of care, defined as inpatient care, assisted living, outpatient care, day treatment and day activities over the health care episode were calculated on the basis of the PCR-MN data by dividing the number of contacts or days per type of care by the total number of days in that episode. Given that the study focused on a group of patients with a tendency to
Community Ment Health J
drop out of care, with some patients dropping out of care within a few days, and others being transferred to ACT after a few initial non-ACT contacts, episodes shorter than 4 days were excluded from the analyses.
were skewed, permutation was performed in order to validate the results of the linear regression models.
Results Ethics and Privacy Episodes Dutch privacy law allows the use of personal (healthcare) data on behalf of scientific research if a number of demands are met. The most important one is that individual patient data cannot be traced to an individual. In the PCR-MN database, patients are identified based upon an encrypted variable using the date of birth, gender and postal code. Using this encryption it is not possible for researchers to identify individual patients. A research review commission decides whether a request of an investigative body for the delivery of data for research can be met. This decision is mainly based upon privacy, relevance, feasibility, qualification of the researchers and the participating psychiatric hospitals interests. The research review commission is formed from the board of advisors of the PCR-MN. Datasets are made available strictly under the condition of anonymity, safe data storage and inaccessibility to others. After use, the destruction of all data sets is mandatory. All authors certify responsibility for the present paper. There are no known conflicts of interest. Costs Outcomes of the present analyses were the number of hospital days or outpatient contacts. As an addition, differences in costs were estimated using these outcomes and their 2009 cost prices. The average costs of one inpatient day, one day in assisted living, one outpatient contact and 1 day in day treatment were respectively €414.00, €151.00, €171.67 and €202.50 (Hakkaart-van Roijen et al. 2011). Costs of day activities were not reported in this cost manual.
Two-hundred twenty-two patients in the PCR-MN database were identified as ACT patients. Eight were excluded from the present analysis because they only had a single ACT contact. On average, patients had 2.4 pre-ACT episodes, 1.9 ACT episodes and 1.7 post ACT episodes (Table 1). Mean duration of ACT episodes (23 months) was longer than pre- and postACT episodes (16 and 8 months respectively, p \ 0.001). Health Care Use On average, patients were 16.9 % [standard deviation (SD) 22.1 %] of one episode admitted, and they were 26.7 % (SD 21.6 %) of the time in assisted living (Table 2). Thus, on average they were admitted 5.1 days per month and spent 8.0 days in assisted living services. In addition, during 6.8 % of the days of an episode patients had outpatient contacts, which translates to one contact every 2 weeks, on average. Differences in Health Care Use Between Pre-ACT, ACT and Post-ACT Episodes In ACT episodes, patients used on average 47.0 % more assisted living (i.e. the regression coefficient was 47.0 %, p \ 0.001) than in episodes before ACT (Table 2). This
Table 1 Patient descriptives 1999–2009 Number of episodes
Length of episode (days)
Mean
5.4
525 (18 months)
37.4
SD (between)
2.3
421
11.4
Range
2–16
5b–3,099
14-76
2.7
480 (16 months)
36.1
SD = 1.2
SD = 700
SD = 11.5
1.9
684 (23 months)
38.9
SD = 0.8
SD = 508
SD = 11.6
1.7
255 (8 months)
38.6
SD = 0.7
SD = 253
SD = 12.3
Statistical Analysis
Age
Male
Total (n = 214)a
All analyses were performed using the statistical program Stata version 11 (StataCorp 2009). As all patients had at least two episodes (i.e. observations in the analysis) the assumption of independence of the observations in standard regression techniques was not met. Multilevel linear regression analyses are ideally suited for the analysis of such hierarchically structured data (Snijders and Bosker 1999). The regression coefficients obtained from multilevel linear regression analyses can be interpreted identically to the estimates obtained from standard unilevel analyses. In all analyses, the main independent variable was ACT (pre ACT, current ACT, post ACT). Dependent variables were the proportions of the various types of mental health care use. Age and gender were included in the models as confounders. Because the dependent variables
%
76
Pre-ACT (n = 213) ACT (n = 214) Post-ACT (n = 70)
c
SD standard deviation a
Eight patients excluded because they only had 1 ACT contact in total
b
Episodes B4 days are excluded, because illogical to calculate percentages
c
Fifty seven patients (27 %) end with a non-ACT episode
123
Community Ment Health J Table 2 Differences in mental health care use (percentages per type of care per episode, range 0–100 %) in before-ACT, ACT and post-ACT episodes in patients (n = 221) treated with ACT; 1999–2009 Percentage of episode Admitteda
Assisted living
Outpatient
Day treatment
Day activities
Characteristics Mean
16.9
SD betweenb 22.1 Regression results: B (95 % confidence interval)
26.7
6.8
1.0
3.0
21.6
4.5
4.0
7.3
Before ACT (reference)
0
0
0
0
0
ACT
1.4
47.0***
-1.58**
-1.62**
4.5***
(-2.2; 4.9)
(43.2; 50.8)
(-2.64;-0.52)
(-2.6; -0.6)
(3.2; 5.8)
9.2***
38.9***
-3.40***
-1.16
7.0***
Post-ACT Post-ACT versus ACT
(3.7; 14.6)
(33.0; 44.7)
(-5.00; 1.81)
(-2.6; 0.32)
(5.0; 9.0)
7.8**
-8.2**
-1.83*
0.46
2.5*
(2.30; 13.2)
(-14.0; -2.33)
(-3.46; -0.20)
(21.06; 1.98)
(0.45; 4.52)
Controlled for age and gender * p \ 0.05 ** p \ 0.01 *** p \ 0.001 a
Including sheltered housing
b
SD between individuals (multilevel data)
means they spent on average 14 days per month more in assisted living. Patients in ACT had 1.58 % fewer outpatient contacts than pre-ACT, translating to one contact each two months. If expressed in costs, results would yield approximately €2,100 per month more for each patient in ACT (14 days more assisted living and half an inpatient day more per month and half an outpatient contact fewer). In post-ACT episodes, patients also used more assisted living than before the first ACT contact (38.9 %, or 11.7 days per month more, p \ 0.001). Post-ACT patients were a larger proportion of an episode admitted (9.2 %, p \ 0.001), while this was not the case for patients in ACT care (1.4 %, p = 0.45). Post ACT, costs were approximately €572 per month higher per patient than in ACT (7.8 % more admitted is 2.5 inpatient days per month; 8.2 % fewer assisted living is 2.5 days; 1.8 % fewer outpatient = 0.5 outpatient days per month).
Discussion Summary of Findings Mental health care use in a group of patients with difficulties in engaging with mental health care increased once ACT services were introduced. In addition, duration of the episode was substantially longer if it was an ACT episode and this is suggestive for lower drop-out. These two results suggest that ACT was effective in remedying a situation of
123
fragmented care that patients had difficulties to engage with. Results additionally showed that not all patients remained in ACT care. They had on average 1.7 post-ACT episodes. ACT care aims to keep patients in ACT (Marshall and Lockwood 2000) and this was so for a small majority of the patients (n = 117; 55 %). Explanation of Findings Results of the post-ACT assessments are difficult to interpret because there are various reasons why patients are discharged from ACT. They may have been referred to less intensive treatment because symptoms improved or because they were admitted to a long-stay ward of the psychiatric hospital. It has been shown that the more severe patients remain in ACT in the long term (Mulder and Kortrijk 2012). Because the Utrecht ACT teams aimed to treat the group of patients that were not (sufficiently) served by existing fragmented services, the existence of post-ACT episodes in combination with the increase in health care use during these episodes, compared to preACT, can also be interpreted as pointing to a successful outcome of ACT. However, because the present study did not include a control group it cannot be established how many of the patients that had difficulties engaging with services would have been in care in the long term had ACT not been introduced. When interpreting the increase in health care use and thus in health care costs (€2,100) after the start of ACT,
Community Ment Health J
one has to keep in mind that the patients in the present study had difficulties engaging with care. Not receiving the care they needed was an inclusion criterion for ACT treatment. Thus, an increase in mental health care use was deemed a positive result and it is hypothesized that this increase would improve functioning in this group. Unfortunately, the present data did not include any severity of symptoms or functioning variables. Post-ACT costs increase with another €572 and this is probably the result of long-term admissions. This was to be expected as admission because of deterioration is one of the main reasons why ACT stops. Other Research The success of ACT has also been shown in previous research. A study in homeless patients with severe mental illness also showed an increase in mental health service use (Rosenheck and Dennis 2001). Another study on assertive outreach also showed that a majority of non-attenders residing in a socially deprived area and reluctant to engage with mental health services were successfully reached (Green et al. 2011). Contrary to the current study, a reduction in admissions has also been associated with ACT (O’Campo et al. 2009). This could also have been expected in the current study given that this is one of the purported aims of ACT. On the other hand, in the present study, ACT specifically targeted a group not engaged with care which a priori would make it difficult to show a reduction in admissions. Effectiveness of ACT in the Netherlands for this patient group is furthermore supported because outcomes, such as psychosocial functioning, improved in the first months after patients were newly introduced to ACT (Kortrijk et al. 2010). Mental health outcomes, including substance abuse also improved (O’Campo et al. 2009). In the long-term, effectiveness of ACT is more difficult to assess because patients remaining in ACT in the long term represent a more severe subgroup (Kortrijk et al. 2012). Strengths and Limitations Strength of the PCR-MN is that the register covers all mental health care provided in the region. Because the registration of diagnosis and psychiatric contacts is the basis for financial reimbursement, registrations are accurate and include psychiatric diagnosis. The present analysis was a pre–post comparison, within the group of patients who previously had difficulties engaging with fragmented services and were newly taken on by ACT teams. The present data did not include a control group. However, any observational control group would have different characteristics than the ‘‘intervention group’’. Therefore, we opted for the pre–post comparison.
The analyses were based on data that only refer to patients who successfully engaged with ACT. Eight subjects who only had one contact were excluded from the analysis and the number of patients who failed to engage entirely is unknown. The eight excluded subjects had fewer inpatient days and more days of treatment and day activities than pre-ACT, but the number of excluded patients is too low to draw any formal conclusions. Acknowledgments We gratefully acknowledge the financial support of the PCR-MN by the Ministry of Health, Welfare and Sport.
References Bhugra, D., Ayonrinde, O., Butler, G., Leese, M., & Thornicroft, G. (2011). A randomised controlled trial of assertive outreach vs. treatment as usual for black people with severe mental illness. Epidemiology and Psychiatric Sciences, 20(1), 83–89. Bond, G. R., & Salyers, M. P. (2004). Prediction of outcome from the Dartmouth assertive community treatment fidelity scale. CNS Spectrums, 9(12), 937–942. CBS (Statistics Netherlands). (2012). StatLine; bevolking kerncijfers. Retrieved May 15, 2012, from http://statline.cbs.nl/. Green, C. E., McGuire, P. K., Ashworth, M., & Valmaggia, L. R. (2011). Outreach and Support in South London (OASIS). Outcomes of non-attenders to a service for people at high risk of psychosis: the case for a more assertive approach to assessment. Psychological Medicine, 41(2), 243–250. Hakkaart-van Roijen, L., Tan, S., & Bouwmans, C. (2011). [Manual for costs research] Handleiding voor kostenonderzoek; methoden en standaard kostprijzen voor economische evaluaties in de gezondheidszorg. Geactualiseerde versie 2010. Diemen: Instituut voor Medical Technology Assessment, Erasmus Universiteit Rotterdam. Ito, J., Oshima, I., Nishio, M., Sono, T., Suzuki, Y., Horiuchi, K., et al. (2011). The effect of assertive community treatment in Japan. Acta Psychiatrica Scandinavica, 123(5), 398–401. Kortrijk, H. E., Mulder, C., Drukker, M., Wiersma, D., & Duivenvoorden, H. (2012). Duration of assertive community treatment and the interpretation of routine outcome data. Australian and New Zealand Journal of Psychiatry, 46(3), 240–248. Kortrijk, H. E., Mulder, C. L., Roosenschoon, B. J., & Wiersma, D. (2010). Treatment outcome in patients receiving assertive community treatment. Community Mental Health Journal, 46(4), 330–336. Marshall, M., & Lockwood, A. (2000). Assertive community treatment for people with severe mental disorders. Cochrane Database of Systematic Reviews, 2, CD001089. Michigan Recovery Centre of Excellence. (2012). Assertive community treatment (ACT/PACT). Retrieved July 11, 2012, from http://www.mirecovery.org/Recovery101/Treatments/Evidence BasedandPromisingPractices/AssertiveCommunityTreatment ACTPACT.aspx. Mulder, C. L., & Kortrijk, H. E. (2012). [Effects of treatment duration on the interpretation of ROM data in ACT patients] De invloed van de duur van behandeling op het interpreteren van ROMmetingen bij ACT. Tijdschrift voor Psychiatrie, 54(2), 191–196. National Alliance on Mental Illness. (2012). Treatment and services; assertive community treatment. Retrieved October 3, 2012, from http://www.nami.org/Template.cfm?Section=ACT-TA_Center& template=/ContentManagement/ContentDisplay.cfm&ContentID =29070.
123
Community Ment Health J O’Campo, P., Kirst, M., Schaefer-McDaniel, N., Firestone, M., Scott, A., & McShane, K. (2009). Community-based services for homeless adults experiencing concurrent mental health and substance use disorders: A realist approach to synthesizing evidence. Journal of Urban Health, 86(6), 965–989. Rosenheck, R. A., & Dennis, D. (2001). Time-limited assertive community treatment for homeless persons with severe mental illness. Archives of General Psychiatry, 58(11), 1073–1080. Smeets, H. M., Laan, W., Engelhard, I. M., Boks, M. P., Geerlings, M. I., & de Wit, N. J. (2011). The psychiatric case register middle Netherlands. BMC Psychiatry, 11, 106. Snijders, T., & Bosker, R. (1999). Multilevel analysis, an introduction to basic and advanced modeling. London: Sage.
123
StataCorp. (2009). Stata Statistical Software (Version 11). College Station, TX: Stata Corporation. Sytema, S., Wunderink, L., Bloemers, W., Roorda, L., & Wiersma, D. (2007). Assertive community treatment in the Netherlands: A randomized controlled trial. Acta Psychiatrica Scandinavica, 116(2), 105–112. The Iowa Consortium for Mental Health. (2012). ACT (Assertive Community Treatment). Retrieved July 11, 2012, from http:// www.healthcare.uiowa.edu/icmh/act/. Van Vugt, M. D., Kroon, H., Delespaul, P. A., Dreef, F. G., Nugter, A., Roosenschoon, B. J., et al. (2011). Assertive community treatment in the Netherlands: Outcome and model fidelity. Canadian Journal of Psychiatry Revue Canadienne de Psychiatrie, 56(3), 154–160.
ORIGINAL PAPER
Can Assertive Community Treatment Remedy Patients Dropping Out of Treatment Due to Fragmented Services? Marjan Drukker • Wijnand Laan • Fred Dreef Ger Driessen • Hugo Smeets • Jim Van Os
•
Received: 11 October 2012 / Accepted: 14 October 2013 Ó Springer Science+Business Media New York 2013
Abstract Previously, many patients with severe mental illness had difficulties to engage with fragmented mental health services, thus not receiving care. In a Dutch city, Assertive Community Treatment (ACT) was introduced to cater specifically for this group of patients. In a pre–post comparison, changes in mental health care consumption were examined. All mental health care contacts, ACT and non-ACT, of patients in the newly started ACT-teams were extracted from the regional Psychiatric Case Register. Analyses of mental health care usage were performed comparing the period before ACT introduction with the period thereafter. After the introduction of ACT, mental health care use increased in this group of patients, although not all patients remained under the care of ACT teams. ACT may succeed in delivering more mental health care to patients with severe mental illness and treatment needs who previously had difficulties engaging with fragmented mental health care services.
M. Drukker (&) G. Driessen J. Van Os Department of Psychiatry and Psychology, School for Mental Health and Neuroscience MHeNS, Maastricht University, P.O. Box 616, Location Vijverdal, 6200 MD Maastricht, The Netherlands e-mail: [email protected] W. Laan H. Smeets Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht, Utrecht, The Netherlands F. Dreef De Care Express, Utrecht, The Netherlands J. Van Os King’s Health Partners, Department of Psychosis Studies, Institute of Psychiatry, King’s College London, London, UK
Keywords Mental health care use Assertive Community Treatment Psychiatry Severity of illness
Introduction Assertive Community Treatment (ACT) is a method of organizing integrated mental health care for the most severely mentally ill. ACT has its origin in the USA and is designed to provide community-based mental health care to persons diagnosed with severe mental illness (Michigan Recovery Centre of Excellence 2012; National Alliance on Mental Illness 2012; The Iowa Consortium for Mental Health 2012). It is a service-delivery model including multidisciplinary teams that coordinate all care, tailor-made to the individual patient (Marshall and Lockwood 2000). Thus, ACT is ideally suited to provide treatment to patients diagnosed with severe mental illness and comorbid substance abuse who find it difficult to engage in a sustained way with mental health services (Bond and Salyers 2004). The majority of this group is homeless or has legal problems. ACT teams aim to have contact with patients on a regular basis, with an intensity as high as needed. Drop-out from treatment is avoided by assertive outreach; this implicates that when patients are reluctant or uncooperative, ACT team members keep in contact and offer services, e.g. by paying home visits (Marshall and Lockwood 2000). During periods of inpatient-care, ACT teams keep in contact with their patients, even when they are treated by a different team of mental health professionals (Marshall and Lockwood 2000). Research has established that in the Netherlands and other countries, ACT represents an effective model of service delivery (Bhugra et al. 2011; Ito et al. 2011; Marshall and Lockwood 2000; Sytema et al. 2007). In the city of Utrecht,
123
Community Ment Health J
the fourth largest city in the Netherlands (source population: 760,000), ACT was considered specifically to cater regionally for the group of patients with severe mental illness who previously had difficulties engaging with fragmented services. The background for this change in model of service delivery was that in the Netherlands, similar to many European countries, in the late 1990s, existing services such as traditional outpatient mental health care and hospital-based services failed to cater for a large group of patients diagnosed with severe mental illness and comorbid addiction. These patients typically had severe functional impairments, frequently were homeless, and presented with forensic problems. Utrecht mental health services identified three reasons why existing services failed to reach these patients. The most important reason was that patients refused help or missed appointments. Traditional services subsequently coped by excluding these patients from their care. Second, patients had a complex combination of needs while institutions typically catered for only one at the time. For example, a patient with psychotic disorder, comorbid addiction and cognitive impairment required care from four different institutions: the mental health service to treat the psychotic disorder, the addiction services, the community social worker to negotiate debts, and a rehabilitation service focussing on the cognitive impairment, in the absence of significant coordination or collaboration between these services. Third, some patients were showing aggressive and/or antisocial behaviour. For example, they threatened the mental health professional, dealt in drugs or extorted money from other patients. Existing services typically had no response to these behaviours and went on to exclude these patients from their care. In Utrecht, ACT teams were introduced in 2005 in order to provide care to the group of patients that had difficulties engaging with care and remedy the existing impasse between institutions delivering fragmented mental health care. Previous research has shown that adherence of the Utrecht ACT-teams to the ACT model, as assessed with the Dartmouth Assertive Community Treatment Scale (DACTS), was good (Van Vugt et al 2011). Given the specific purpose for the introduction of ACT, an increase in mental health care consumption in the group of patients presenting difficulties to engage with services can be considered a successful outcome. Thus, the present paper studies the impact of the Utrecht ACT teams, by assessing changes in health care use in the group of patients assigned to these teams.
Methods Setting Utrecht is the smallest of the four major cities in the Netherlands. It has about 300,000 inhabitants and is similar
123
to most towns of the same size in western European countries, most inhabitants having a white European background (CBS (Statistics Netherlands) 2012). Psychiatric Case Register Middle Netherlands The Psychiatric Case Register Middle Netherlands (PCRMN) (Smeets et al. 2011) registers the mental health care consumption of over 760,000 inhabitants from the greater Utrecht region. The PCR-MN uniquely registers inpatient care, outpatient care, emergency care, day care and assisted living on a patient-based, but anonymized level (Smeets et al. 2011). For the present analysis PCR-MN data between 1999 and 2009 were used. Preparation of the Data In order to compare health care use of patients in ACT with health care use of patients not in ACT, health care episodes were defined for each patient based upon a set of rules. Periods of health care use were defined as ACT episodes if a patient had ACT contacts within this period. Because patients could also have non-ACT contacts between ACT contacts, ACT episodes did not end at the moment the patient had any non-ACT contact, but only when a patient did not have a contact with an ACT-team for 61 days. This period of 61 days was chosen because patients should have an ACT-contact at least once a month. Thus, when a patient did not have a contact with the ACT-team for 61 days or more, the implication is that two or more consecutive ACT contacts were missed. An exception was made when patients were admitted for a period not exceeding 3 months directly after an ACT-contact, this admission was combined with the previous ACT-episode, because ACT usually continues during short-term admissions. All other periods of health care were defined as non-ACT episodes. A health care episode ended when a patient did not have contact with mental health services for 100 days or more. If a patient did not use any mental health care for 100 days or more it was assumed that the patient moved or was outside the reach of ACT (e.g. forensic services). Therefore, periods in which no care was used were not included in the analyses. In order to conduct the pre–post comparison, two types of non-ACT episodes were defined: those before the start of ACT (pre-ACT) and those after having experienced one or more ACT episodes (post-ACT). Over the course of each episode, the proportion of the different types of care, defined as inpatient care, assisted living, outpatient care, day treatment and day activities over the health care episode were calculated on the basis of the PCR-MN data by dividing the number of contacts or days per type of care by the total number of days in that episode. Given that the study focused on a group of patients with a tendency to
Community Ment Health J
drop out of care, with some patients dropping out of care within a few days, and others being transferred to ACT after a few initial non-ACT contacts, episodes shorter than 4 days were excluded from the analyses.
were skewed, permutation was performed in order to validate the results of the linear regression models.
Results Ethics and Privacy Episodes Dutch privacy law allows the use of personal (healthcare) data on behalf of scientific research if a number of demands are met. The most important one is that individual patient data cannot be traced to an individual. In the PCR-MN database, patients are identified based upon an encrypted variable using the date of birth, gender and postal code. Using this encryption it is not possible for researchers to identify individual patients. A research review commission decides whether a request of an investigative body for the delivery of data for research can be met. This decision is mainly based upon privacy, relevance, feasibility, qualification of the researchers and the participating psychiatric hospitals interests. The research review commission is formed from the board of advisors of the PCR-MN. Datasets are made available strictly under the condition of anonymity, safe data storage and inaccessibility to others. After use, the destruction of all data sets is mandatory. All authors certify responsibility for the present paper. There are no known conflicts of interest. Costs Outcomes of the present analyses were the number of hospital days or outpatient contacts. As an addition, differences in costs were estimated using these outcomes and their 2009 cost prices. The average costs of one inpatient day, one day in assisted living, one outpatient contact and 1 day in day treatment were respectively €414.00, €151.00, €171.67 and €202.50 (Hakkaart-van Roijen et al. 2011). Costs of day activities were not reported in this cost manual.
Two-hundred twenty-two patients in the PCR-MN database were identified as ACT patients. Eight were excluded from the present analysis because they only had a single ACT contact. On average, patients had 2.4 pre-ACT episodes, 1.9 ACT episodes and 1.7 post ACT episodes (Table 1). Mean duration of ACT episodes (23 months) was longer than pre- and postACT episodes (16 and 8 months respectively, p \ 0.001). Health Care Use On average, patients were 16.9 % [standard deviation (SD) 22.1 %] of one episode admitted, and they were 26.7 % (SD 21.6 %) of the time in assisted living (Table 2). Thus, on average they were admitted 5.1 days per month and spent 8.0 days in assisted living services. In addition, during 6.8 % of the days of an episode patients had outpatient contacts, which translates to one contact every 2 weeks, on average. Differences in Health Care Use Between Pre-ACT, ACT and Post-ACT Episodes In ACT episodes, patients used on average 47.0 % more assisted living (i.e. the regression coefficient was 47.0 %, p \ 0.001) than in episodes before ACT (Table 2). This
Table 1 Patient descriptives 1999–2009 Number of episodes
Length of episode (days)
Mean
5.4
525 (18 months)
37.4
SD (between)
2.3
421
11.4
Range
2–16
5b–3,099
14-76
2.7
480 (16 months)
36.1
SD = 1.2
SD = 700
SD = 11.5
1.9
684 (23 months)
38.9
SD = 0.8
SD = 508
SD = 11.6
1.7
255 (8 months)
38.6
SD = 0.7
SD = 253
SD = 12.3
Statistical Analysis
Age
Male
Total (n = 214)a
All analyses were performed using the statistical program Stata version 11 (StataCorp 2009). As all patients had at least two episodes (i.e. observations in the analysis) the assumption of independence of the observations in standard regression techniques was not met. Multilevel linear regression analyses are ideally suited for the analysis of such hierarchically structured data (Snijders and Bosker 1999). The regression coefficients obtained from multilevel linear regression analyses can be interpreted identically to the estimates obtained from standard unilevel analyses. In all analyses, the main independent variable was ACT (pre ACT, current ACT, post ACT). Dependent variables were the proportions of the various types of mental health care use. Age and gender were included in the models as confounders. Because the dependent variables
%
76
Pre-ACT (n = 213) ACT (n = 214) Post-ACT (n = 70)
c
SD standard deviation a
Eight patients excluded because they only had 1 ACT contact in total
b
Episodes B4 days are excluded, because illogical to calculate percentages
c
Fifty seven patients (27 %) end with a non-ACT episode
123
Community Ment Health J Table 2 Differences in mental health care use (percentages per type of care per episode, range 0–100 %) in before-ACT, ACT and post-ACT episodes in patients (n = 221) treated with ACT; 1999–2009 Percentage of episode Admitteda
Assisted living
Outpatient
Day treatment
Day activities
Characteristics Mean
16.9
SD betweenb 22.1 Regression results: B (95 % confidence interval)
26.7
6.8
1.0
3.0
21.6
4.5
4.0
7.3
Before ACT (reference)
0
0
0
0
0
ACT
1.4
47.0***
-1.58**
-1.62**
4.5***
(-2.2; 4.9)
(43.2; 50.8)
(-2.64;-0.52)
(-2.6; -0.6)
(3.2; 5.8)
9.2***
38.9***
-3.40***
-1.16
7.0***
Post-ACT Post-ACT versus ACT
(3.7; 14.6)
(33.0; 44.7)
(-5.00; 1.81)
(-2.6; 0.32)
(5.0; 9.0)
7.8**
-8.2**
-1.83*
0.46
2.5*
(2.30; 13.2)
(-14.0; -2.33)
(-3.46; -0.20)
(21.06; 1.98)
(0.45; 4.52)
Controlled for age and gender * p \ 0.05 ** p \ 0.01 *** p \ 0.001 a
Including sheltered housing
b
SD between individuals (multilevel data)
means they spent on average 14 days per month more in assisted living. Patients in ACT had 1.58 % fewer outpatient contacts than pre-ACT, translating to one contact each two months. If expressed in costs, results would yield approximately €2,100 per month more for each patient in ACT (14 days more assisted living and half an inpatient day more per month and half an outpatient contact fewer). In post-ACT episodes, patients also used more assisted living than before the first ACT contact (38.9 %, or 11.7 days per month more, p \ 0.001). Post-ACT patients were a larger proportion of an episode admitted (9.2 %, p \ 0.001), while this was not the case for patients in ACT care (1.4 %, p = 0.45). Post ACT, costs were approximately €572 per month higher per patient than in ACT (7.8 % more admitted is 2.5 inpatient days per month; 8.2 % fewer assisted living is 2.5 days; 1.8 % fewer outpatient = 0.5 outpatient days per month).
Discussion Summary of Findings Mental health care use in a group of patients with difficulties in engaging with mental health care increased once ACT services were introduced. In addition, duration of the episode was substantially longer if it was an ACT episode and this is suggestive for lower drop-out. These two results suggest that ACT was effective in remedying a situation of
123
fragmented care that patients had difficulties to engage with. Results additionally showed that not all patients remained in ACT care. They had on average 1.7 post-ACT episodes. ACT care aims to keep patients in ACT (Marshall and Lockwood 2000) and this was so for a small majority of the patients (n = 117; 55 %). Explanation of Findings Results of the post-ACT assessments are difficult to interpret because there are various reasons why patients are discharged from ACT. They may have been referred to less intensive treatment because symptoms improved or because they were admitted to a long-stay ward of the psychiatric hospital. It has been shown that the more severe patients remain in ACT in the long term (Mulder and Kortrijk 2012). Because the Utrecht ACT teams aimed to treat the group of patients that were not (sufficiently) served by existing fragmented services, the existence of post-ACT episodes in combination with the increase in health care use during these episodes, compared to preACT, can also be interpreted as pointing to a successful outcome of ACT. However, because the present study did not include a control group it cannot be established how many of the patients that had difficulties engaging with services would have been in care in the long term had ACT not been introduced. When interpreting the increase in health care use and thus in health care costs (€2,100) after the start of ACT,
Community Ment Health J
one has to keep in mind that the patients in the present study had difficulties engaging with care. Not receiving the care they needed was an inclusion criterion for ACT treatment. Thus, an increase in mental health care use was deemed a positive result and it is hypothesized that this increase would improve functioning in this group. Unfortunately, the present data did not include any severity of symptoms or functioning variables. Post-ACT costs increase with another €572 and this is probably the result of long-term admissions. This was to be expected as admission because of deterioration is one of the main reasons why ACT stops. Other Research The success of ACT has also been shown in previous research. A study in homeless patients with severe mental illness also showed an increase in mental health service use (Rosenheck and Dennis 2001). Another study on assertive outreach also showed that a majority of non-attenders residing in a socially deprived area and reluctant to engage with mental health services were successfully reached (Green et al. 2011). Contrary to the current study, a reduction in admissions has also been associated with ACT (O’Campo et al. 2009). This could also have been expected in the current study given that this is one of the purported aims of ACT. On the other hand, in the present study, ACT specifically targeted a group not engaged with care which a priori would make it difficult to show a reduction in admissions. Effectiveness of ACT in the Netherlands for this patient group is furthermore supported because outcomes, such as psychosocial functioning, improved in the first months after patients were newly introduced to ACT (Kortrijk et al. 2010). Mental health outcomes, including substance abuse also improved (O’Campo et al. 2009). In the long-term, effectiveness of ACT is more difficult to assess because patients remaining in ACT in the long term represent a more severe subgroup (Kortrijk et al. 2012). Strengths and Limitations Strength of the PCR-MN is that the register covers all mental health care provided in the region. Because the registration of diagnosis and psychiatric contacts is the basis for financial reimbursement, registrations are accurate and include psychiatric diagnosis. The present analysis was a pre–post comparison, within the group of patients who previously had difficulties engaging with fragmented services and were newly taken on by ACT teams. The present data did not include a control group. However, any observational control group would have different characteristics than the ‘‘intervention group’’. Therefore, we opted for the pre–post comparison.
The analyses were based on data that only refer to patients who successfully engaged with ACT. Eight subjects who only had one contact were excluded from the analysis and the number of patients who failed to engage entirely is unknown. The eight excluded subjects had fewer inpatient days and more days of treatment and day activities than pre-ACT, but the number of excluded patients is too low to draw any formal conclusions. Acknowledgments We gratefully acknowledge the financial support of the PCR-MN by the Ministry of Health, Welfare and Sport.
References Bhugra, D., Ayonrinde, O., Butler, G., Leese, M., & Thornicroft, G. (2011). A randomised controlled trial of assertive outreach vs. treatment as usual for black people with severe mental illness. Epidemiology and Psychiatric Sciences, 20(1), 83–89. Bond, G. R., & Salyers, M. P. (2004). Prediction of outcome from the Dartmouth assertive community treatment fidelity scale. CNS Spectrums, 9(12), 937–942. CBS (Statistics Netherlands). (2012). StatLine; bevolking kerncijfers. Retrieved May 15, 2012, from http://statline.cbs.nl/. Green, C. E., McGuire, P. K., Ashworth, M., & Valmaggia, L. R. (2011). Outreach and Support in South London (OASIS). Outcomes of non-attenders to a service for people at high risk of psychosis: the case for a more assertive approach to assessment. Psychological Medicine, 41(2), 243–250. Hakkaart-van Roijen, L., Tan, S., & Bouwmans, C. (2011). [Manual for costs research] Handleiding voor kostenonderzoek; methoden en standaard kostprijzen voor economische evaluaties in de gezondheidszorg. Geactualiseerde versie 2010. Diemen: Instituut voor Medical Technology Assessment, Erasmus Universiteit Rotterdam. Ito, J., Oshima, I., Nishio, M., Sono, T., Suzuki, Y., Horiuchi, K., et al. (2011). The effect of assertive community treatment in Japan. Acta Psychiatrica Scandinavica, 123(5), 398–401. Kortrijk, H. E., Mulder, C., Drukker, M., Wiersma, D., & Duivenvoorden, H. (2012). Duration of assertive community treatment and the interpretation of routine outcome data. Australian and New Zealand Journal of Psychiatry, 46(3), 240–248. Kortrijk, H. E., Mulder, C. L., Roosenschoon, B. J., & Wiersma, D. (2010). Treatment outcome in patients receiving assertive community treatment. Community Mental Health Journal, 46(4), 330–336. Marshall, M., & Lockwood, A. (2000). Assertive community treatment for people with severe mental disorders. Cochrane Database of Systematic Reviews, 2, CD001089. Michigan Recovery Centre of Excellence. (2012). Assertive community treatment (ACT/PACT). Retrieved July 11, 2012, from http://www.mirecovery.org/Recovery101/Treatments/Evidence BasedandPromisingPractices/AssertiveCommunityTreatment ACTPACT.aspx. Mulder, C. L., & Kortrijk, H. E. (2012). [Effects of treatment duration on the interpretation of ROM data in ACT patients] De invloed van de duur van behandeling op het interpreteren van ROMmetingen bij ACT. Tijdschrift voor Psychiatrie, 54(2), 191–196. National Alliance on Mental Illness. (2012). Treatment and services; assertive community treatment. Retrieved October 3, 2012, from http://www.nami.org/Template.cfm?Section=ACT-TA_Center& template=/ContentManagement/ContentDisplay.cfm&ContentID =29070.
123
Community Ment Health J O’Campo, P., Kirst, M., Schaefer-McDaniel, N., Firestone, M., Scott, A., & McShane, K. (2009). Community-based services for homeless adults experiencing concurrent mental health and substance use disorders: A realist approach to synthesizing evidence. Journal of Urban Health, 86(6), 965–989. Rosenheck, R. A., & Dennis, D. (2001). Time-limited assertive community treatment for homeless persons with severe mental illness. Archives of General Psychiatry, 58(11), 1073–1080. Smeets, H. M., Laan, W., Engelhard, I. M., Boks, M. P., Geerlings, M. I., & de Wit, N. J. (2011). The psychiatric case register middle Netherlands. BMC Psychiatry, 11, 106. Snijders, T., & Bosker, R. (1999). Multilevel analysis, an introduction to basic and advanced modeling. London: Sage.
123
StataCorp. (2009). Stata Statistical Software (Version 11). College Station, TX: Stata Corporation. Sytema, S., Wunderink, L., Bloemers, W., Roorda, L., & Wiersma, D. (2007). Assertive community treatment in the Netherlands: A randomized controlled trial. Acta Psychiatrica Scandinavica, 116(2), 105–112. The Iowa Consortium for Mental Health. (2012). ACT (Assertive Community Treatment). Retrieved July 11, 2012, from http:// www.healthcare.uiowa.edu/icmh/act/. Van Vugt, M. D., Kroon, H., Delespaul, P. A., Dreef, F. G., Nugter, A., Roosenschoon, B. J., et al. (2011). Assertive community treatment in the Netherlands: Outcome and model fidelity. Canadian Journal of Psychiatry Revue Canadienne de Psychiatrie, 56(3), 154–160.
