1. Introduction
Early Detection (ED) and Early intervention (EI) services were developed to provide support shortly before and during the onset of psychosis. These services are currently well-incorporated in mental health policies of some countries, such as the United Kingdom [Reference NICE1, Reference Garety, Craig, Dunn, Fornells-Ambrojo, Colbert and Rahaman2] where in 2009 they were provided by 145 community teams operating throughout the country [Reference Bird, Premkumar, Kendall, Whittington, Mitchell and Kuipers3].
ED services aim to reduce the transition to psychosis or the duration of untreated psychosis (DUP) which is crucial for the illness prognosis and treatment because it significantly affects the severity of symptoms, the risk of relapse, overall functioning as well as the response to treatment [Reference Marshall, Lewis, Lockwood, Drake, Jones and Croudace4, Reference Perkins, Gu, Boteva and Lieberman5]. ED services use media, public events, and community work to inform about early signs of psychoses and facilitate access of young people to mental health care [Reference Johannessen, McGlashan, Larsen, Horneland, Joa and Mardal6]. ED services focus on high risk subjects, i.e. people with prodromal symptoms (attenuated psychotic symptoms, full-blown psychotic symptoms that are brief and self-limiting, or a significant decrease in functioning in the context of genetic risk for schizophrenia) [Reference Randall, Vokey, Loewen, Martens, Brownell and Katz43]. A number of evaluations of ED programmes have reported the positive outcomes of ED in terms of shortening DUP [Reference Johannessen, McGlashan, Larsen, Horneland, Joa and Mardal6, Reference Lloyd-Evans, Crosby, Stockton, Pilling, Hobbs and Hinton7].
EI services provide continuous support to people at early stage of psychosis which is usually the first 2 to 5 years from the illness onset. EI services are usually based on a cooperation between a multidisciplinary team (usually including psychiatrist, clinical psychologist, psychiatric nurses and social care workers), general practitioners and families [Reference NICE1] and are built upon various services including case management, pharmacological treatment, psychological (most often cognitive behaviour therapy) and psychosocial interventions (such as supportive counselling or social skills training), family therapy and supported employment services [Reference Garety, Craig, Dunn, Fornells-Ambrojo, Colbert and Rahaman2, Reference Bird, Premkumar, Kendall, Whittington, Mitchell and Kuipers3]. Recent meta-analyses showed that EI services have (in comparison to the treatment as usual in a given setting) high potential for decreasing the hospital admission rates [Reference Bird, Premkumar, Kendall, Whittington, Mitchell and Kuipers3, Reference Marshall and Rathbone8, Reference Randall, Vokey, Loewen, Martens, Brownell and Katz9], and risk of relapse [Reference Bird, Premkumar, Kendall, Whittington, Mitchell and Kuipers3], and lowering the positive and negative symptoms of a severe mental illness [Reference Bird, Premkumar, Kendall, Whittington, Mitchell and Kuipers3] as well as the duration of untreated psychosis [Reference Larsen, Melle, Auestad, Haahr, Joa and Johannessen10] and suicide risk [Reference McCrone, Park and Knapp11]. At the same time, the recent studies have consistently found the positive impacts of EI on employment and education [Reference Marshall and Rathbone8, Reference Bond, Drake and Luciano12].
Furthermore, EI and ED programmes appeared to be cost-effective in a longer period of time, usually in two years, especially because of the reductions in the length of stay in hospitals and lost productivity [Reference McCrone, Park and Knapp11, Reference Andrew, Knapp, McCrone, Parsonage and Trachtenberg13–Reference Mihalopoulos, Harris, Henry, Harrigan and McGorry16]. However, a study from Denmark found the effect of EI services was not sustainable in a 5-year follow-up [Reference Bertelsen, Jeppesen, Petersen, Thorup, Øhlenschlæger and le Quach17]. Also, the analysis of patient journey presented in this issue [Reference Mohr, Galderisi, Boyer, Wasserman, Arteel and Ieven18] showed the identification of early symptoms and the provision of timely intervention as one of the key drivers towards better outcomes and recovery in patients with schizophrenia.
However, ED and EI services are mostly unavailable in the countries of Central and Eastern Europe and currently there is no formal evidence to support such an investment. Mental health care systems in this region are predominantly hospital-based and community services are not available to those who in need [Reference Winkler, Krupchanka, Roberts, Kondratova, Machů and Höschl19]. This leads to excessively long hospitalisations, exceeding 20 years in some cases and being over 100 days long on average [Reference Winkler, Mladá, Krupchanka, Agius, Ray and Höschl20, Reference IHIS21]. Central and Eastern Europe is also a region with high mortality rates among people with mental disorders, high suicide rates, excessive alcohol consumption, and high level of public stigma [Reference Winkler, Krupchanka, Roberts, Kondratova, Machů and Höschl19, Reference Krupchanka, Mladá, Winkler, Khazaal and Albanese22–Reference Winkler, Mladá, Janoušková, Weissová, Tušková and Csémy24]. Severe lack of health service and population research in psychiatry leads to decision-making not being based on evidence, which imposes a risk that already scarce resources are spent ineffectively [Reference Winkler, Krupchanka, Roberts, Kondratova, Machů and Höschl19, Reference WHO25]. It has been also repeatedly observed that institutionalization of people with mental disorders in regional psychiatric hospitals is often associated with non-adherence to human rights of people with disabilities [Reference WHO26].
Current mental health care reforms in the Czech Republic is focused on deinstitutionalization, destigmatization, improving the quality of care, and strengthening the evidence based mental health care development with the overall aim to improve the quality of life of people with mental health problems [Reference MHCZ and Health27]. Deinstitutionalization is considered to be a priority as it has been demonstrated to be preferred by patients and to improve the quality of life of people with severe mental illnesses while not leading to homelessness, crime, and suicidal behaviours [Reference Kunitoh28–Reference Henderson, Phelan, Loftus, Dall’Agnola and Ruggeri30]. Economic case for deinstitutionalization has also been made, and it has been suggested that community care is not more costly when the quality of care is taken into account [Reference Mansell, Knapp, Beadle-Brown and Beecham31, Reference Knapp, Beecham, McDaid, Matosevic and Smith32]. ED and EI services could be developed within the pursuit of mental health care reforms in CEE as they enable people with incipient psychosis to stay in the community and out of the psychiatric hospitals, and therefore are complementary to deinstitutionalization. In this paper, we aim to show the cost estimates based on an economic model for ED and EI services in the Czech Republic.
2. Methods
This study is a follow-up to the EBC initiatives which estimated a burden and costs associated with disorders of the brain in Europe in 2005 and in 2010 [Reference Andlin‐Sobocki, Jönsson, Wittchen and Olesen33–Reference Gustavsson, Svensson, Jacobi, Allgulander, Alonso and Beghi37]. The current EBC project was entitled “Value of Treatment” and its aims were to identify gaps in the current health care systems across Europe, and to estimate the value of addressing these gaps. Study by Mohr et al. [Reference Mohr, Galderisi, Boyer, Wasserman, Arteel and Ieven18] focused on journeys of patients with schizophrenia and identified a substantial gap in early detection and early intervention services, which result in both, missed or delayed diagnosis and a limited access to timely and adequate treatments. The present study focused on modelling cost-consequences of tackling these problems in the Czech Republic.
Decision analytical modelling is a systematic approach to inform decisions under uncertainty via defining a set of possible consequences of alternative actions [Reference Briggs, Claxton and Sculpher38]. We used a decision tree as a vehicle to estimate costs associated with adopting ED and EI services in the Czech Republic as it allowed us to model economic consequences of the alternative actions in the absence of direct local evidence on (cost-)effectiveness of ED and EI. In our case, the alternative actions were a) to do nothing, b) to introduce early detection services, c) to introduce early intervention services, and d) to introduce both, early detection and early intervention services for psychoses as defined by ICD-10′s F20-F29 codes. The target population of these services are young people experiencing first symptoms or first episode of psychoses (FEP) in the Czech Republic. The option a) refers to the treatment as usual (TAU) which is currently comprised of a treatment at outpatient settings, delivered by a psychiatrist which is usually limited to prescription of psychopharmaceuticals, a treatment in psychiatric hospitals, and rarely also assertive community treatment. From a societal perspective, however, we focused only on costs related to health and social care services and productivity lost, and excluded other costs for informal care or criminal justice system.
As described in detail below, our model relies on three sources of data: a) epidemiological data are based on the Czech all-cause hospitalizations register which was described in more detail in our previous studies [Reference Winkler, Mladá, Krupchanka, Agius, Ray and Höschl20, Reference Winkler, Mladá, Csémy, Nechanská and Höschl39]; b) probabilities were taken from meta-analyses which were identified via our meta-review (i.e. systematic review of systematic reviews and meta-analyses); c) costs based on Czech unit costs and experiences of EI and ED teams in south London.
2.1. Probabilities and epidemiological data
The key assumption is that international data would reasonably apply in the Czech context. This is a strong but necessary assumption in the absence of any local evidence based on Czech experience. To identify the best available international evidence on transition probabilities possible that would enter our model, we performed review of systematic reviews and meta-analyses (or meta-review). We have systematically searched the Web of Science, Medline, EMBASE and Cochrane Library to identify meta-analyses on ED and EI services. The following strategy was used for the Web of Science and translated to other databases: TOPIC: (early interven* or early diagnos* or early detect*) AND TOPIC: (mental health or mental disorder or mental illness or mental disease) AND TOPIC: (review or literature search or systematic review or meta-analysis or meta analysis) NOT TOPIC: (Alzheimer or Alzheimer's or autism or dementia or cardiovascular or PTSD or postpartum or eating or cancer). The full strategy is available in the Appendix 1.
Further assupmtions were as follows: People with FEP were defined as those with a first hospitalization for psychotic symptoms. There were 5478 of people with psychotic disorders hospitalized for the first time at a psychiatric outpatient care service in the Czech Republic in 2015 (i.e. in a period between 1st January 2015 and 31st December 2015) [Reference IHIS40]. According to a meta-analysis, the risk of transition to psychosis among the high-risk group, which is a potential target of ED services, is 0.22 [Reference Fusar-Poli, Bonoldi, Yung, Borgwardt, Kempton and Valmaggia41]. Another recent meta-analysis demonstrated that ED services reduce the risk of transition to psychosis in the high-risk group by 54% [Reference van der Gaag, Smit, Bechdolf, French, Linszen and Yung42].
If there were EI services available for people who made the transition from the high-risk group to FEP, the probability of hospitalization was estimated to drop from 0.74 to 0.52, which is based on the meta-analysis by Randall, Vokey [Reference Randall, Vokey, Loewen, Martens, Brownell and Katz43], and the probability of retaining employment would increase from 0.29 to 0.61, which is based on meta-analysis by Bond, Drake [Reference Bond, Drake and Luciano12].
2.2. Costs
The associated annual costs were calculated as follows. The cost of unemployment was assumed to be equal to the minimal Czech wage. The costs of ED services were estimated by assuming that the Czech ED and EI teams would have the same composition as they have in UK [Reference Valmaggia, McCrone, Knapp, Woolley, Broome and Tabraham14, Reference Park, McCrone and Knapp44, Reference Craig, Garety, Power, Rahaman, Colbert and Fornells-Ambrojo45]. The costs of ED were calculated by using costs of ED services per patient in the first year of the service provision. It was also assumed that one Czech EI team would be able to take care of 150 clients a year which is in line with experiences EI services in the UK [Reference Power, McGuire, Iacoponi, Garety, Morris and Valmaggia47] and Denmark [Reference Nordentoft, Thorup, Petersen, Øhlenschlæger, Christensen and Krarup48, Reference Petersen, Jeppesen, Thorup, Abel, Øhlenschlæger and Christensen49]
Czech unit costs were used to calculate overall costs of both, Czech EI and Czech ED team. The costs of treatment as usual was calculated as costs for:
i) outpatient psychiatrist (highly specialized services provided exclusively by psychiatrists and mostly limited to quick assessment of the patient and drug prescription)- these costs were based on the average consumption of this services among the sample of 138 patients who were followed-up in the community services for a 12-month period;
ii) inpatient care– these costs were based on the average length of hospital stay for the people with psychoses in the Czech Republic and on the related unit costs of one day of inpatient care service (including the costs for an overnights stay);
iii) Psychiatric medications– these costs were estimated as an average consumption of psychopharmaceuticals by clients of OASIS team [Reference Valmaggia, McCrone, Knapp, Woolley, Broome and Tabraham14] and costs of the corresponding psychopharmaceuticals in CZ as reported by the State Institute for Drug Control.
These costs were combined using the following formula: = yearly consumption of psychopharmaceuticals+yearly consumption of services of an outpatient psychiatrist + (cost of inpatient care per day * average length of stay in inpatient psychiatric hospitals in CZ - average length of stay in inpatient psychiatric hospitals in CZ * costs of outpatient psychiatrist) * probability of being inpatiently hospitalized.
All costs were converted to Euro in 2016 prices, with an exchange rate 27CZK per 1 Euro. All costs, data and probabilities are reported in the Table 1.
Table 1 Parameters and costs used within the model.
2.3. Sensitivity analysis
One- way sensitivity analyses were performed for anumber of key parameters, including sensitivity analysis for both, the median (rather than minimum) wage rate for the age group of 20–29 years which is when FEP usually occurs. Sensitivity analyses were also focused on shorter than average length of inpatient hospitalization for psychosis in the Czech Republic, because it might be assumed that the inpatient stay of people with FEP could be shorter than inpatient stay of those with chronic psychoses [Reference Huntley, Cho, Christman and Csernansky50]. Otherwise, each of the probabilities employed in the model was modified to explore all the possible uncertainties.
3. Results
Based on the data from the Czech registries [Reference IHIS40] and probabilities derived from the meta-analysis by Fusar-Poli, Bonoldi [Reference Fusar-Poli, Bonoldi, Yung, Borgwardt, Kempton and Valmaggia41] we estimated that there were 24,900 people with high risk of developing psychosis in CZ 2015. Considering the effects of ED programmes as estimated in the meta-analysis by van der Gaag, Smit [Reference van der Gaag, Smit, Bechdolf, French, Linszen and Yung42] we estimated that if the ED services were available to everyone in the Czech Republic, the number of people hospitalized with psychosis for the first time could have dropped from 5478 to 2520. Taking further into account the effectsof EI services as assessed in meta-analysis by Randall, Vokey [Reference Randall, Vokey, Loewen, Martens, Brownell and Katz43], out of the total 2520 (or 5478 if there were no ED services) people with the FEP, 1310 (or 2849) would be hospitalised and 1537 (or 3342) would retain their employment if there were EI services available in the country. If there were no EI services, then 1865 (or 4054) would be hospitalised and 731 (or 1589) would retain their employment.
The economic model demonstrated that costs associated with the above-mentioned scenarios are as follows. The costs of care as usual for people with FEP are estimated to be as high as 46 million Euro each year. These estimates are conservative in terms of that only health care costs and costs associated with reduced productivity, and not costs associated with other sectors, such as social care, informal care, criminal justice and others, were taken into account. It is also estimated that these costs could be reduced by 25% if ED services were adopted (policy change 1), 33% if EI services were adopted (policy change 2), and 40% if both, ED and EI services, were adopted (policy change 3) in the country (Fig. 1). This means cost savings of about 2,000–2,800-3,200 Euro per patient when introducing policy changes 1-2-3 respectively.
Sensitivity analyses demonstrated that the estimates are robust, and that only dramatically decreased effect of ED services would have influenced the overall results. Meta-analyses used in our model demonstrated the 54% reduction in transition to psychosis was associated with ED services, only if this effect would drop to approximately 30% if ED services would introduce additional costs to the Czech mental health care system (Fig. 2a, b, and c).
4. Discussion
The economic model presented in this paper suggests that adopting ED and EI services in the Czech Republic would be a cost-saving strategy for its mental health care development. This is an important finding because mental health care systems in the region are expected to transform from hospital-based towards more community-oriented ones in the near future. ED and EI centres, such as the EPPIC (Early Psychosis Prevention and Intervention Centre) in Australia, were developed in many countries globally as an alternative to hospitalisation [Reference McGorry51], and could serve as a good example to benchmark when reforming mental health care systems in the region of Central and Eastern Europe.
The economic evidence on ED and EI services is quite extensive and comes from many different cultural backgrounds [Reference McCrone, Park and Knapp11, Reference Valmaggia, McCrone, Knapp, Woolley, Broome and Tabraham14–Reference Mihalopoulos, Harris, Henry, Harrigan and McGorry16, Reference Park, McCrone and Knapp44]. A focus on reducing the duration of untreated psychosis (DUP) has been demonstrated to lead to better outcomes such as fewer and shorter hospital stays of people with psychosis [Reference McGorry51]. However, the evidence is not unanimous. Large trial in northern Italy tested multicomponent intervention added to the usual community based services for people with FEP [Reference Ruggeri, Bonetto, Lasalvia, De Girolamo, Fioritti and Rucci52]. Despite the significant improvements in symptoms, global functioning, and other outcomes, this study did not find a significant reduction in neither, number of hospital admissions nor length of inpatient stays among patients in the active group, compared to the control group [Reference Ruggeri, Bonetto, Lasalvia, Fioritti, De Girolamo and Santonastaso53]. It can be interpreted that this might be partly explained by a good-quality community care which already existed in the area and which was considered as the treatment as usual. Also, a stronger emphasis on early detection might have led to a reduced number of days in hospitalization in the intervention group. As there is a severe lack of community services in the region of CEE [Reference Winkler, Krupchanka, Roberts, Kondratova, Machů and Höschl19], we assume that adopting ED and EI services in the region might mimic the effectiveness of these services as demonstrated in the meta-analyses used for populating decision tree in this modelling study. The adoption and implementation should be conducted carefully, fidelity should be ensured and evaluation well planned and rigorously conducted, because the results will influence mental health care development in the region - if negative they will hinder further reforms, if positive they will help to justify further system changes to general public [Reference Ruggeri and Tansella54]. Strengths and limitations
The strengths of our model stem from the quality of data that we had available. First, our epidemiological data are based on the Czech national registers which reflect the current situation in health-care utilization. Second, all cost data are based on thorough calculations of Czech unit costs that were conducted by our team in collaboration with local health and social care providers. Third, all probabilities used within the model come from robust meta-analyses which were published quite recently and identified via meta-review.
Fig. 1. Economic model of adopting early detection and early intervention services in the Czech Republic (the editable Excel file is available – “Winkler_Editable Figures and Tables”).
Fig. 2. Results of one-way sensitivity analyses for all the parameters used within the model and reported as tornado plots for Early Detection (a), Early Intervention (b), and Early Detection and Early Intervention (c). The editable Excel file is available (“Winkler_Editable Figures and Tables”).
However, this study has a number of limitations. First and foremost, neither, ED nor EI teams, has ever operated in the Czech Republic, which is why we had to rely on meta-analysis rather than Czech specific data that would come from local services. Also, we relied on the assumption that the services would perform at least as good as reported by meta-analyses used in our model. This might not be necessarily true and, for instance, employment services could have different effects due to different legal and work environments in the Czech Republic. However, we did use sensitivity analyses to explore this and it has been demonstrated that ED and EI services would be cost-saving even if we would reduce the probability of employment from 0.61 to 0.05 (Figure 2abc).
The other characteristic of our model is that we assumed a perfect scenario and ED and EI were available to all people who are currently hospitalized in the Czech Republic with FEP. This means that the services would have to immediately have the same availability as outpatient psychiatrists in the Czech Republic. This would be ideal, but of course, achievable only in a longer time horizon. Furthermore, we did include neither, extra costs for setting up the early detection and early intervention services in the Czech Republic nor capital costs (costs of new or existing buildings and equipment). The earlier would mean higher costs for early detection and early intervention services in the first year of functioning, and the latter would (at least in long-term) not significantly change the differences in costs as an increase in capital costs for new services would be offset by decrease in capital costs for treatment as usual.
On the other hand our estimates might be considered conservative in a sense that only health care and employment related costs were included in the model. The cost savings could be much higher if we had been able to include also costs related to criminal justice, informal care and alike.
5. Conclusions
This study adds an economic argument to the analysis of schizophrenia patient journey [Reference Mohr, Galderisi, Boyer, Wasserman, Arteel and Ieven18]. Our results suggest that adopting ED and EI services in the Czech Republic would be cost saving due to decreases in hospitalisations and better employment outcomes of people with psychoses. These findings are in line with other studies conducted in England, Denmark, Australia and elsewhere [Reference McCrone, Park and Knapp11, Reference Valmaggia, McCrone, Knapp, Woolley, Broome and Tabraham14–Reference Mihalopoulos, Harris, Henry, Harrigan and McGorry16, Reference Park, McCrone and Knapp44] but have more informative value for the hospital-based systems in the region of Central and Eastern Europe where the development of mental health care has been hindered by a lack of epidemiological and economic evidence. The current mental health care reform in the Czech Republic utilizes European Structural and Investment Funds to finance the first phase of the reform. It is a unique opportunity which might become an example for other countries in the region because Ministry of Health of the Czech Republic, for the first time since the dissolution of communism more than a quarter of century ago, has fully committed to transform the mental health care system in a way which has been repeatedly suggested by both, mental health professionals and international organizations [Reference Pěč55, Reference Höschl, Winkler and Pěč56]. The results of our decision model, however, have to be taken with caution and full economic evaluations (cost-effectiveness and cost-utility analyses) alongside multi-centre trials are recommended before scaling up ED and EI services in those European countries where these services are still not available [Reference Maric, Raballo, Kuzman, Petrovic, Klosterkötter and Riecher-Rössler57, Reference Maric, Petrovic, Raballo, Rojnic‐Kuzman, Klosterkötter and Riecher‐Rössler58]. Czech Republic now intends to conduct such a study within the ongoing national mental health care reform; economic evidence generated within the forthcoming study might be decisive for policy and practice in the country.
Declarations of interest
None.
Funding
This work on schizophrenia is part of a series of case studies covering nine neurological and psychiatric conditions, conducted within the “Value of Treatment for Brain Disorders” research project of the European Brain Council. PM and PW were supported by the Ministry of Education, Youth and Sports of the Czech Republic (project NPU4NUDZ: LO1611); PM was also supported by the Ministry of Health of the Czech Republic (grant No. AZV 15-28998A).
Acknowledgement
We thank to Leonardo Koeser for thoughtful discussions on the design of analyses.
Comments
No Comments have been published for this article.