Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Increasing medication adherence and income assistance access for first-episode psychosis patients

  • Jason Randall,

    Affiliations Department of Community Health Sciences, Max Rady College of Medicine, Rady Faculty of Health Sciences, University of Manitoba, Winnipeg, Manitoba, Canada, Manitoba Centre for Health Policy, Department of Community Health Sciences, Max Rady College of Medicine, Rady Faculty of Health Sciences, University of Manitoba, Winnipeg, Manitoba, Canada

  • Dan Chateau ,

    Dan_Chateau@cpe.umanitoba.ca

    Affiliations Department of Community Health Sciences, Max Rady College of Medicine, Rady Faculty of Health Sciences, University of Manitoba, Winnipeg, Manitoba, Canada, Manitoba Centre for Health Policy, Department of Community Health Sciences, Max Rady College of Medicine, Rady Faculty of Health Sciences, University of Manitoba, Winnipeg, Manitoba, Canada

  • James M. Bolton,

    Affiliations Department of Community Health Sciences, Max Rady College of Medicine, Rady Faculty of Health Sciences, University of Manitoba, Winnipeg, Manitoba, Canada, Manitoba Centre for Health Policy, Department of Community Health Sciences, Max Rady College of Medicine, Rady Faculty of Health Sciences, University of Manitoba, Winnipeg, Manitoba, Canada, Department of Psychiatry, Max Rady College of Medicine, Rady Faculty of Health Sciences, University of Manitoba, Winnipeg, Manitoba, Canada

  • Mark Smith,

    Affiliation Manitoba Centre for Health Policy, Department of Community Health Sciences, Max Rady College of Medicine, Rady Faculty of Health Sciences, University of Manitoba, Winnipeg, Manitoba, Canada

  • Laurence Katz,

    Affiliations Manitoba Centre for Health Policy, Department of Community Health Sciences, Max Rady College of Medicine, Rady Faculty of Health Sciences, University of Manitoba, Winnipeg, Manitoba, Canada, Department of Psychiatry, Max Rady College of Medicine, Rady Faculty of Health Sciences, University of Manitoba, Winnipeg, Manitoba, Canada

  • Elaine Burland,

    Affiliation Manitoba Centre for Health Policy, Department of Community Health Sciences, Max Rady College of Medicine, Rady Faculty of Health Sciences, University of Manitoba, Winnipeg, Manitoba, Canada

  • Carole Taylor,

    Affiliation Manitoba Centre for Health Policy, Department of Community Health Sciences, Max Rady College of Medicine, Rady Faculty of Health Sciences, University of Manitoba, Winnipeg, Manitoba, Canada

  • Nathan C. Nickel,

    Affiliations Department of Community Health Sciences, Max Rady College of Medicine, Rady Faculty of Health Sciences, University of Manitoba, Winnipeg, Manitoba, Canada, Manitoba Centre for Health Policy, Department of Community Health Sciences, Max Rady College of Medicine, Rady Faculty of Health Sciences, University of Manitoba, Winnipeg, Manitoba, Canada

  • Jennifer Enns,

    Affiliation Manitoba Centre for Health Policy, Department of Community Health Sciences, Max Rady College of Medicine, Rady Faculty of Health Sciences, University of Manitoba, Winnipeg, Manitoba, Canada

  • Alan Katz,

    Affiliations Department of Community Health Sciences, Max Rady College of Medicine, Rady Faculty of Health Sciences, University of Manitoba, Winnipeg, Manitoba, Canada, Manitoba Centre for Health Policy, Department of Community Health Sciences, Max Rady College of Medicine, Rady Faculty of Health Sciences, University of Manitoba, Winnipeg, Manitoba, Canada, Department of Family Medicine, Max Rady College of Medicine, Rady Faculty of Health Sciences, University of Manitoba, Winnipeg, Manitoba, Canada

  • Marni Brownell,

    Affiliations Department of Community Health Sciences, Max Rady College of Medicine, Rady Faculty of Health Sciences, University of Manitoba, Winnipeg, Manitoba, Canada, Manitoba Centre for Health Policy, Department of Community Health Sciences, Max Rady College of Medicine, Rady Faculty of Health Sciences, University of Manitoba, Winnipeg, Manitoba, Canada

  • the PATHS Equity Team

    Membership of the PATHS Equity Team is provided in the Acknowledgments.

Increasing medication adherence and income assistance access for first-episode psychosis patients

  • Jason Randall, 
  • Dan Chateau, 
  • James M. Bolton, 
  • Mark Smith, 
  • Laurence Katz, 
  • Elaine Burland, 
  • Carole Taylor, 
  • Nathan C. Nickel, 
  • Jennifer Enns, 
  • Alan Katz
PLOS
x

Abstract

Background

Assertive community treatment for first-episode psychosis programs have been shown to improve symptoms and reduce service use. There is little or no evidence on whether these programs can increase access to income assistance and improve medication adherence in first episode psychosis patients. This research examines the impact of the Early Psychosis Prevention and Intervention Service (EPPIS) on these outcomes.

Methods

We extracted data on EPPIS patients held in the Data Repository at the Manitoba Centre for Health Policy. The Repository is a comprehensive collection of person-level de-identified administrative records, including data from Manitoba’s health services. We compared income assistance use and antipsychotic medication adherence in EPPIS patients to a historical cohort matched on pattern of diagnosis. Confounders were adjusted through propensity-score weighting with asymmetrical trimming. Odds ratios (OR), hazard ratios (HR) and 95% confidence intervals were calculated.

Results

We identified a matched sample of 244 patients and 449 controls. EPPIS patients had a higher rate of income assistance use during the program (67·4% vs. 38·7%; p< 0·0001). EPPIS patients were more likely to have been prescribed at least one antipsychotic medication than the control cohort, both during the program (OR = 15·05; 95%CI 10·81 to 20·94) and after the program ended (OR = 5·20; 95%CI: 4·50 to 6·02). Patients in EPPIS were also more likely to adhere to their medication during the program (OR = 4·71; 95%CI 3·75 to 5·92), and after the program (OR = 2·54; 95%CI 2·04 to 3·16).

Conclusion

Enrolment in the EPPIS program was associated with increased adherence to antipsychotic medication treatment and improved uptake of income assistance.

Introduction

Early intervention for psychosis and Assertive Community Treatment (ACT) programs have become increasing popular methods in the treatment of psychotic disorders [17]. These programs tend to focus on individuals with a first episode of psychosis and use a comprehensive treatment approach that combines medication, psychotherapy, and other resources. Generally, the approach in assertive community treatment is to engage patients in active treatment outside of the hospital [8,9]. Family members are often involved in the treatment process in order to increase their awareness and ability to cope with and respond to the symptoms of the patient. Previous research has provided evidence that these programs are effective at improving symptoms and prognosis [4,7,1014].

Lack of medication adherence among those with psychotic disorders is associated with poorer outcomes [1517]. Improving medication adherence is one way that first-episode interventions can improve prognosis. Individuals with schizophrenia are disproportionally of low socio-economic status [18]. The cause of this has not yet be conclusively determined, but schizophrenia may cause decreased SES due to the work and social impairments caused by the disorder [19]. Low SES can prevent people from affording their medication. However, income assistance and other welfare programs that subsidize drug costs may increase adherence to medication. Therefore, research assessing how well these programs impact on adherence and access to government programs that help provide medications is needed.

Early Psychosis Prevention and Intervention Service (EPPIS)

The Early Psychosis Prevention and Intervention Service (EPPIS) was established in Winnipeg, Manitoba, Canada in 2003 as an ACT treatment program for first-episode psychosis patients. This program targets individuals aged 13 to 35 with a recent onset of psychotic symptoms within a single-payer integrated healthcare system with a population of about 650 000 people. Similar to other first-episode programs, EPPIS emphasizes adherence to medication and psychotherapy (including group therapy with family members and other clients), with the goal of promoting and maintaining social connectedness. EPPIS staff also work with guidance counsellors, employers, and other key individuals in the client’s environment, to assist with re-integration of clients into their communities. This may involve community psychoeducation to reduce the stigma of psychotic disorders, and/or developing flexible work and educational environments.

Referrals are made directly to the EPPIS program by clinicians, at which time a Mental Health Clinician is assigned. If further follow-up is required, clients are assigned to a psychiatrist within one working day. An initial interview is conducted within 7 days of referral; however, cases considered urgent are seen within 3 working days. Cases of immediate risk are referred to the emergency department or a mobile crisis team. The initial interview is conducted in a non-clinical setting if necessary. At the initial interview, information is collected about basic demographics, personal history and experience, and assessment of general psychiatric symptoms. The Mental Health Clinician assigned at referral remains the clinician throughout the remainder of treatment in the program. Antipsychotic medication is considered a primary form of therapy in this program. However, clients retain the ability to refuse medication and still remain in the program. The program not only treats the individual client but aims to provide support for their family as well.

This study used a unique population-based dataset of social and medical records to examine whether EPPIS affected medication use and access to income assistance. The specific outcomes assessed were medication prescriptions received, medication adherence, and receipt of income assistance. We hypothesized that the assertive nature of this program would improve medication outcomes and increase the use of available social assistance programs.

Methods

Data source

This study used data from the PATHS Data Resource in the Population Health Research Data Repository housed at the Manitoba Centre for Health Policy (MCHP)[20]. The Repository contains the de-identified administrative health records for Manitoba’s publicly-funded health system representing virtually the entire provincial population with the exception of individuals who are incarcerated or federal military officials. The Repository data are linkable at the individual level across multiple service areas, including hospital discharge abstracts, physician billing submissions, pharmaceutical prescriptions from the Drug Programs Information Network (DPIN), and services provided by the Manitoba Adolescent Treatment Centre (MATC)–which contains data for the Winnipeg EPPIS program. Data for inpatient service use is coded in ICD—9-CM up to March 31, 2004, and in ICD-10-CA after this date. Physicians’ services are billed using ICD-9-CM. These data contain information on physician treatment both inside and outside of hospital Pharmaceutical prescriptions are coded using Anatomical Therapeutic Chemical (ATC) classification codes or Generic Drug names. SES is derived using the average income of each census dissemination area (DA). Each DA contains between 400 and 700 people, providing neighborhood-level income. Data on income assistance was collected through the Social Assistance Management Information Network. People in Manitoba are eligible for income assistance if they need help meeting basic personal and family needs; this includes those who have a mental illness that is likely to last more than 90 days which prevents them from meeting their basic needs.

Study design

A matched-cohort design was used in this study. Individuals enrolled in the EPPIS program were identified using data provided by MATC which was responsible for administrating the program for the period assessed by this study. Individuals with 3+ visits in the program were considered to have been treated by the program. Data were available for the calendar years 2003 to 2012. Controls were then identified through matching for diagnostic pattern of psychotic disorders. EPPIS patients’ records were assessed for two years prior to program entrance for the occurrence of psychosis diagnoses. Matches were required to have received all of the same psychosis diagnosis codes but not have been treated by the program. In order to reduce differences between controls and cases, propensity scores were calculated using logistic regression and conditional probabilities for treatment were derived [21]. These conditional probabilities were used to determine inverse probability of treatment weights [21]. Asymmetric trimming was performed at the 2·5th and 97·5th percentiles to reduce the effect of individuals with extreme weights [22]. Trimming is necessary to remove outliers from the EPPIS and control groups that could bias the results. Confounders were selected using the High Dimensional Propensity Score algorithm that detected variables that were associated with both treatment and outcome [23]. Standardized differences were assessed to determine whether confounder variables were successfully adjusted through weighting [24]. Treated individuals were compared to the matched cohort during the period of treatment (up to 24 months) and after treatment. The control sample was drawn from a time period before the EPPIS program (1998–2001). We were unable to find a control group of sufficient size that was similar to the client group within the EPPIS program period, possibly due to the program successfully enrolling its entire target population. Regions of the province not covered by EPPIS are not directly comparable due to large differences in health service availability and the lack of another major city in the province.

Descriptive statistics

Descriptive statistics for age, sex, SES (grouped into quintiles), and follow-up length (years) were derived for both cohorts. For those in the program, information on total time in treatment (days) and number of visits to the treatment program were derived T-tests were used to determine if treatment days and visits differed significantly between low (quintiles 1 and 2) and high (quintiles 3–5) income groups.

Outcomes

The first outcome was the use of income assistance. A binary variable was derived from these data for each group for the two-year period before treatment, during treatment and for a two-year period following the end of treatment, which identified individuals who were receiving income assistance at some point during those periods. A binary variable for ever being on income assistance after treatment was also derived. Time on income assistance was derived by measuring the years on income assistance per person year.

The other outcomes pertained to the use of and adherence to antipsychotic medications (including depot medications) and included any first-or-second generation antipsychotic medications (ATCs -'N05AH', 'N05AX', 'N05AD', 'N05AB', 'N05AF', 'N05AA') at any dosage. The DPIN database was the data source for these outcomes We determined (i) whether individuals had ‘received any prescription for an antipsychotic medication’, and ii) whether individuals had ‘antipsychotic medication adherence’; which was defined as having a medication possession ratio (i.e., days of medication received divided by days in treatment period) of at least 80% in a year [25]. These outcomes both refer specifically to outpatient dispensation of antipsychotics. Medications administered within inpatient units are not included in this study.

Statistical analysis

Logistic regression and Cox proportional hazards regression analyses were used to estimate odds ratios (OR), hazard ratios (HR) and 95% confidence intervals (95%CI). Analyses were performed for the treatment period, and during post-treatment. The post-treatment analysis was performed using two methods. First, ORs were determined for a two-year period following discharge from the program. Then HRs were estimated using all available follow-up years for each individual, without restriction to the two-year period. Age was included as a covariate in all of the regression models. Baseline use of income assistance was also included as a covariate for the income assistance analysis. Inverse probability of treatment weighting was applied so that the results would estimate the average treatment effect (ATE). Statistical analysis was performed using SAS 9.3 (SAS Institute, Cary NC).

Ethics

Approval was obtained from the University of Manitoba’s Health Research Ethics Board (HREB #: H2011:294) and the Government of Manitoba’s Health Information Privacy Commission (HIPC #- 2013/2014–22). The study used de-identified administrative health data to protect the health information of those in the sample.

Results

MATC records contained data on 284 EPPIS patients, and an initial matched control group of 1036 individuals was identified based on diagnostic pattern. Trimming resulted in a sample of 244 cases and 449 controls. See Table 1 for the ICD codes and patient characteristics that were identified as potential confounders through their correlation with exposure status and the outcome, and controlled through propensity score weighting. These conditions are expected to be related to first-episode psychosis and poor mental health. Skin disorders and use of ectoparasiticides, for example, might be coded due to the development of delusions around parasite infestations. Refraction/accommodation disorders might be coded when individuals report vague visual disturbances. After weighting, the standardized differences for these variables were all below 0·1 with the exception of macrolides and lincosamides which was 0·10. This indicates good adjustment for these covariates [26].

Descriptive statistics on controls and EPPIS treatment group are shown in Table 2. The sample was 77.5% males with an average age of 18.8. Approximately three-quarters of the sample was between 17 and 24 and only 9 individuals were older than 24. The average follow-up time available was considerably longer for the control group (10·2 years versus 3·1 years due to the use of a historical control with more years of potential administrative data to use). Clients of the program had a mean duration of treatment of 510.4 days (median = 467) with a mean of 98.9 visits. Mean duration of treatment was significantly longer for high income individuals (555.0 days) than low income individuals (452.8; p = 0·019).

thumbnail
Table 2. EPPIS and control group demographics and treatment statistics.

https://doi.org/10.1371/journal.pone.0179089.t002

Table 3 contains descriptive data on the use of income assistance. Before initiation of treatment, the EPPIS group had a slightly higher rate of income assistance (50% versus 41·2%), but this difference increased during the treatment period (67·4% versus 38·7%). Table 4 shows the results of the regression analyses for income assistance use post-treatment. The EPPIS program was associated with a significant increase in the use of income assistance after the treatment period (HR = 2·38, 95%CI 2·06 to 2·74).

Table 5 shows the effect of the intervention on outpatient medication use and adherence. During treatment, those in the program were significantly more likely to (i) have used at least 1 antipsychotic prescription (OR = 15·05, 95%CI: 10·81 to 20·94), and (ii) adhered to that medication (OR = 4·71, 95%CI: 3·75 to 5·92). These effects diminished, but were still significant, during the post-treatment period.

thumbnail
Table 5. Effect of EPPIS treatment on receiving outpatient medications and medication adherence.

https://doi.org/10.1371/journal.pone.0179089.t005

Discussion

In this study, we evaluated the EPPIS program for its effect on access to income assistance and medication use and adherence. During and after the EPPIS program, clients were more likely than controls to be receiving income assistance. They were also more likely to be prescribed medication and adhere to that medication. Individuals from high-income areas stayed in the treatment program for a longer time period, but did not make more visits to the program than their lower income counterparts.

Previous research on first-episode intervention programs has largely focused on clinical and health service outcomes. Some research has supported these programs as a method of reducing the severity of symptoms during the treatment phase [5]. Individual studies have not consistently found a significant reduction in hospital service use, although a meta-analysis showed reduced hospitalization [6]. There is currently less evidence around the effect of these programs on medication adherence, social service use/income assistance, and their effect on reducing health disparity.

In our study, the use of antipsychotic medication was higher in the treatment group and remained higher after clients left the program. This suggests the program was effective in its goal of initiating and maintaining medication adherence. Poor medication adherence is a risk factor for poor prognosis among individuals with psychotic disorders [15,16]. Adherence to medication has been found to increase the odds of remission and to reduce the time to remission in first episode psychosis patients [15]. Reaching patients that disengage from treatment is an important goal of assertive community treatment models. Being able to demonstrate that patients in the program are using medications at an increased rate is an indication that the program is effective at achieving this goal. Some studies have examined the effect of the program on the use of antipsychotics. Chen et al (2011) found that their first-episode intervention was associated with increased use of second generation antipsychotics [13]. Other studies have found mixed or negative results [27,28]. The finding that EPPIS increased the use of antipsychotics after program discharge is significant and indicates that these programs can alter long-term adherence to medications.

Individuals with psychotic disorders not only have issues maintaining employment, but can also experience difficulty accessing social services available to them. The EPPIS program was associated with an increased use of income assistance. Enrolling in income assistance can not only help the patient in maintaining housing and other personal needs, but also improves their ability to adhere to their pharmaceutical treatment.

The assessment of length of time treated and number of visits suggests that there could still be health disparity among patients of the program. Because we were unable to find a concurrent sample of controls, we cannot determine whether the program actually reduced disparity by income.

Strengths and limitations

The strengths of this study include the use of a large population-based sample, adjustment for a large variety of confounding factors, objective outcome measures, low levels of attrition, and capture of the entire population treated by EPPIS during the study years. These strengths reduce the likelihood of bias due to selection, attrition, or outcome detection (whether outcome events are more/less likely to be detected based on treatment group). We were also able to link the data across multiple sectors (health, social services) and follow individuals over long periods of time. This improves our ability to control covariates and to detect differences in outcomes.

Limitations include the lack of information about specific symptoms experienced by individual patients, both during the initial assessment of these patients and during treatment. This prevents us from adjusting for the severity of illness through the propensity-score weighting. It is possible that patients with more severe psychotic symptoms were more likely to be enrolled in the EPPIS program. This might have been why finding concurrent controls was difficult. Matching was done based on diagnostic patterns in an attempt to reduce the likelihood of there being a bias due to illness type and severity. However residual confounding may still exist. We were unable to find a group of comparable patients for matching in the time frame that EPPIS was operating. This could bias the results through the influence of secular trends, but expert clinicians in the Winnipeg area believe that EPPIS was the only significant change in treatment during this period. Our inability to locate a concurrent sample of controls may be a limitation to the study, but is also a sign that the program was effective at locating and enrolling its target population. Had there been a large number of eligible patients that were missed during the treatment period, then we should have been able to locate them through similarities in their medical records. Income groups were determined based on census dissemination area and are therefore not assessed individually. This reduces the accuracy of the income quintiles with respect to the individuals involved in this study. There was a parallel income assistance program run by the city of Winnipeg until April of 1999. Information on this program and its clients is not available and may be the cause of the initially low levels of assistance for the control group. However, any potential bias in the regression results due to this would be expected to be towards the null. Adherence and medication use measures in the administrative data cannot confirm that medications were taken properly, only that the medications were dispensed to the patients. The use of expensive antipsychotic medications which are covered by income assistance increases the likelihood that those individuals will be on income assistance. This may introduce some confounding for those outcomes.

In conclusion, this study showed improvements in the treatment measures that we expected the EPPIS program to affect. This study provides additional support for the adoption of assertive treatment services targeted toward first-psychosis patients.

Acknowledgments

The authors acknowledge the Manitoba Centre for Health Policy (MCHP) for use of data contained in the Population Health Research Data Repository under project # 2012–010 (HIPC #2013/2014–22). The results and conclusions are those of the authors and no official endorsement by MCHP, Manitoba Health, Seniors and Active Living (MHSAL), or other data providers is intended or should be inferred. Data used in this study were provided by MHSAL; Vital Statistics; Manitoba Jobs and the Economy; Family Services; and the Manitoba Adolescent Treatment Centre. Authors would also like to thank the PATHS Equity Team for its support. PATHS Equity Team members: James Bolton; Marni Brownell; Charles Burchill; Elaine Burland; Mariette Chartier; Dan Chateau; Malcolm Doupe; Jennifer Enns; Greg Finlayson; Randall Fransoo; Chun Yan Goh; Milton Hu; Doug Jutte; Alan Katz; Laurence Katz; Lisa Lix; Patricia J. Martens (deceased); Colleen Metge; Nathan C. Nickel; Farzana Qudus; Colette Raymond; Les Roos; Noralou Roos; Rob Santos; Joykrishna Sarkar; Mark Smith; Carole Taylor; Randy Walld. All members of the PATHS Equity team are affiliated with the Max Rady College of Medicine at the University of Manitoba. The principal investigator of the PATHs team is Marni Brownell (Marni_Brownell@cpe.umanitoba.ca).

Author Contributions

  1. Conceptualization: MB DC JMB MS LK AK.
  2. Data curation: CT.
  3. Formal analysis: CT.
  4. Funding acquisition: MB DC.
  5. Methodology: JR DC JMB MS LK EB CT NCN JE AK MB.
  6. Project administration: EB DC.
  7. Writing – original draft: JR.
  8. Writing – review & editing: DC JMB MS LK EB CT NCN JE AK MB.

References

  1. 1. McGorry P (1993) Early psychosis prevention and intervention centre. Australas Psychiatry 1: 32–34. pmid:21275861
  2. 2. McGorry PD, Edwards J, Mihalopoulos C, Harrigan SM, Jackson HJ (1996) EPPIC: an evolving system of early detection and optimal management. Schizophr Bull 22: 305–326. pmid:8782288
  3. 3. Addington J, Addington D (2001) Early intervention for psychosis: the Calgary Early Psychosis treatment & Prevention Program. Can Psychiatric Assoc Bull 33: 11–16.
  4. 4. Bertelsen M, Jeppesen P, Petersen L, Thorup A, Ohlenschlaeger J, le QP, Christensen TO, Krarup G, Jorgensen P, Nordentoft M (2008) Five-year follow-up of a randomized multicenter trial of intensive early intervention vs standard treatment for patients with a first episode of psychotic illness: the OPUS trial. Arch Gen Psychiatry 65: 762–771. pmid:18606949
  5. 5. Boden R, Sundstrom J, Lindstrom E, Wieselgren IM, Lindstrom L (2010) Five-year outcome of first-episode psychosis before and after the implementation of a modified assertive community treatment programme. Soc Psychiatry Psychiatr Epidemiol 45: 665–674. pmid:19652896
  6. 6. Randall JR, Vokey S, Loewen H, Martens PJ, Brownell M, Katz A, Nickel NC, Burland E, Chateau D (2015) A Systematic Review of the Effect of Early Interventions for Psychosis on the Usage of Inpatient Services. Schizophr Bull 41: 1379–1386. pmid:25745034
  7. 7. Randall JR, Chateau D, Smith M, Taylor C, Bolton J, Katz L, Burland E, Katz A, Nickel NC, Enns J, Brownell M (2016) An early intervention for psychosis and its effect on criminal accusations and suicidal behaviour using a matched-cohort design. Schizophr Res 176: 307–311. pmid:27262385
  8. 8. Bond GR, Witheridge TF, Dincin J, Wasmer D (1991) Assertive community treatment: correcting some misconceptions. Am J Community Psychol 19: 41–51. pmid:1867150
  9. 9. Stein LI, Test MA (1980) Alternative to mental hospital treatment. I. Conceptual model, treatment program, and clinical evaluation. Arch Gen Psychiatry 37: 392–397. pmid:7362425
  10. 10. Craig TK, Garety P, Power P, Rahaman N, Colbert S, Fornells-Ambrojo M, Dunn G (2004) The Lambeth Early Onset (LEO) Team: randomised controlled trial of the effectiveness of specialised care for early psychosis. BMJ 329: 1067. pmid:15485934
  11. 11. Goldberg K, Norman R, Hoch JS, Schmitz N, Windell D, Brown N, Malla A (2006) Impact of a specialized early intervention service for psychotic disorders on patient characteristics, service use, and hospital costs in a defined catchment area. Can J Psychiatry 51: 895–903. pmid:17249632
  12. 12. Dodgson G, Crebbin K, Pickering C, Mitford E, Brabban A, Paxton R (2008) Early intervention in psychosis service and psychiatric admissions. Psychiatric Bulletin 32: 413–416.
  13. 13. Chen EY, Tang JY, Hui CL, Chiu CP, Lam MM, Law CW, Yew CW, Wong GH, Chung DW, Tso S, Chan KP, Yip KC, Hung SF, Honer WG (2011) Three-year outcome of phase-specific early intervention for first-episode psychosis: a cohort study in Hong Kong. Early Interv Psychiatry 5: 315–323. pmid:21726421
  14. 14. Fowler D, Hodgekins J, Howells L, Millward M, Ivins A, Taylor G, Hackmann C, Hill K, Bishop N, Macmillan I (2009) Can targeted early intervention improve functional recovery in psychosis? A historical control evaluation of the effectiveness of different models of early intervention service provision in Norfolk 1998–2007. Early Interv Psychiatry 3: 282–288. pmid:22642731
  15. 15. Malla A, Norman R, Schmitz N, Manchanda R, Bechard-Evans L, Takhar J, Haricharan R (2006) Predictors of rate and time to remission in first-episode psychosis: a two-year outcome study. Psychological Medicine 36: 649–658. pmid:16515734
  16. 16. Fenton WS, Blyler CR, Heinssen RK (1997) Determinants of medication compliance in schizophrenia: empirical and clinical findings. Schizophr Bull 23: 637–651. pmid:9366000
  17. 17. Robinson D, Woerner MG, Alvir JM, Bilder R, Goldman R, Geisler S, Koreen A, Sheitman B, Chakos M, Mayerhoff D, Lieberman JA (1999) Predictors of relapse following response from a first episode of schizophrenia or schizoaffective disorder. Arch Gen Psychiatry 56: 241–247. pmid:10078501
  18. 18. Dohrenwend BP (1990) Socioeconomic status (SES) and psychiatric disorders: are the issues still compelling? Psychiatr Epidemiol 25: 41–47.
  19. 19. Aro S, Aro H, Keskimaki I (1995) Socio-economic mobility among patients with schizophrenia or major affective disorder. A 17-year retrospective follow-up. Br J Psychiatry 166: 759–767. pmid:7663824
  20. 20. Nickel NC, Chateau DG, Martens PJ, Brownell MD, Katz A, Burland EM, Walld R, Hu M, Taylor CR, Sarkar J, Goh CY (2014) Data resource profile: Pathways to Health and Social Equity for Children (PATHS Equity for Children). Int J Epidemiol 43: 1438–1449. pmid:25212478
  21. 21. Guo S and Fraser MW (2009) Propensity Score Analysis: Statistical Methods and Applications. Sage Publications.
  22. 22. Stürmer T, Rothman KJ, Avom J, Glynn RJ (2010) Treatment Effects in the Presence of Unmeasured Confounding: Dealing with Observations in the Tails of the Propensity Score Distribution- A Simulation Study. American Journal of Epidemiology 172: 843–854. pmid:20716704
  23. 23. Schneeweiss S, Rassen JA, Glynn RJ, Avorn J, Mogun H, Brookhart MA (2009) High-dimensional propensity score adjustment in studies of treatment effects using health care claims data. Epidemiology 20: 512–522. pmid:19487948
  24. 24. Austin PC (2009) Balance diagnostics for comparing the distribution of baseline covariates between treatment groups in propensity-score matched samples. Stat Med 28: 3083–3107. pmid:19757444
  25. 25. Gilmer TP, Dolder CR, Lacro JP, Folsom DP, Lindamer L, Garcia P, Jeste DV (2004) Adherence to treatment with antipsychotic medication and health care costs among Medicaid beneficiaries with schizophrenia. Am J Psychiatry 161: 692–699. pmid:15056516
  26. 26. Normand ST, Landrum MB, Guadagnoli E, Ayanian JZ, Ryan TJ, Cleary PD, McNeil BJ (2001) Validating recommendations for coronary angiography following acute myocardial infarction in the elderly: a matched analysis using propensity scores. J Clin Epidemiol 54: 387–398. pmid:11297888
  27. 27. Cullberg J, Levander S, Holmqvist R, Mattsson M, Wieselgren IM (2002) One-year outcome in first episode psychosis patients in the Swedish Parachute project. Acta Psychiatr Scand 106: 276–285. pmid:12225494
  28. 28. Sanbrook M, Harris A, Parada R, Young P (2003) The effectiveness of an early intervention in the treatment of first-episode psychosis. AeJAMH 2.