Outcome evaluation of an early psychosis program using pre-post comparison and propensity matching

• Main Author: Gibbons, Vanessa
• Format: Thesis
• Language: English
• Published: Memorial University of Newfoundland 2005
• Subjects: Newfoundland
• Online Access: https://research.library.mun.ca/10544/; https://research.library.mun.ca/10544/1/Gibbons_Vanessa.pdf

Early Psychosis Programs have been developed as a solution to reduce delays and improve outcome for first episode psychotic patients. Evaluations of the programs worldwide have found that the programs help reduce symptoms and hospitalizations and improve quality of life. The purpose of this thesis is to evaluate the overall impact of the Newfoundland and Labrador Early Psychosis Program (NL Program) for its patients. Traditionally, programs are evaluated by pre-post methodology. However, this method may have limitations since it does not use standard control groups and any changes seen in the patients from entry to completion of the program may be due to natural events, such as maturation and changes in hormones, since the patients tend to be fairly young. Therefore, this study will test a novel methodology, propensity matching, as an alternative method to evaluate the NL Program. The patients are matched to a national population from the Canadian Community Health Survey (CCHS) on several clinical and social outcomes to determine how the patients differ from the general population at entry into the NL Program and then after two years, to see if they converge back to the population after completing the NL Program. Propensity matching results are then compared to the pre-post results. Analysis of the data found the propensity matching methodology did produce similar results as the pre-post evaluation approach on social and clinical outcomes such as reducing substance use, depression, hospitalizations and suicide, and improving quality of life and vocational functioning. In conclusion, this study found the NL Program is having a major treatment effect for its patients, and propensity matching may serve as a model in future evaluations in mental health research.