Results of the 1:1-matching of the CTC-EFF study: A non-randomized evaluation study

• Published in: European journal of public health 2021-10, Vol.31 (Supplement_3)
• Main Authors: Röding, D, Soellner, R, Birgel, V, Groeger-Roth, F, Brender, R, Reder, M, Feierabend, N, Kleiner, C, Walter, U
• Format: Article
• Language: English
• Subjects: Chi-square test; Community development; Crime; Criminal statistics; Equivalence; Evaluation; Inhabitants; Intervention; Matching; Mean; Police; Public health; Recruitment; Rural areas; Rural communities; Shoplifting; Statistics; Towns; Unemployment; Violence; Violent crime
• ISSN: 1101-1262; 1464-360X
• DOI: 10.1093/eurpub/ckab165.409

Background In Germany, more and more communities are implementing the U.S. approach “Communities That Care (CTC)” to prevent youth problem behavior. To evaluate the effectiveness of CTC, we planned a prospective natural experiment. This required the development of a specific novel recruiting and matching procedure. In this talk, we examine the ability of our procedure to achieve base-line equivalence regarding the matching characteristics. Methods The matching procedure is based on community-level data from official statistics 2017 (www.inkar.de) and police crime statistics 2019. From April 2019 to January 2020, we recruited communities that are implementing CTC (intervention communities, IC). For each IC, we recruited a matched community in the same federal state not using CTC (comparison community, CC). To examine whether the ICs (n = 15) and the CCs (n = 15) are equivalent, chi-square tests were performed for categorical variables and Mann-Whitney U tests for metric variables. Results The following are preliminary results. Of the ICs/CCs 10%/7% are large cities, 25%/29% medium-sized cities, 40%/43% larger towns, 15%/14% smaller towns, and 10%/7% rural communities (p=.995). Regarding the community development of ICs/CCs 10%/21% are shrinking, 5%/14% stable, 55%/36% growing and 30%/29% strongly growing (p=.533). The mean fiscal power (€per capita and year) of the ICs/CCs is 1,150€/770€(p=.192). The mean long-term unemployment rate is 1.19% in ICs and 1.03% in CCs (p=.726). The average annual prevalence of violent crime per 100,000 inhabitants is 1.67 in ICs and 1.78 in CCs (p=.690). Prevalence of shoplifting per 100,000 inhabitants is 3.4 in ICs and 3.5 in CCs (p=.389). Conclusions The results show that ICs and CCs are equivalent with respect to matching characteristics. Key messages The developed novel recruiting and matching procedure is appropriate for enrolling communities in non-randomized trials. The CTC-EFF study evaluates the effectiveness of CTC in Germany for the first time.