Emilie Jouanjus1, Laure Pourcel, Sylvie Saivin, Laurent Molinier, Maryse Lapeyre-Mestre. 1. Equipe de Pharmacoépidémiologie, Inserm, UMR1027, Toulouse, France; Université de Toulouse III, UMR1027, Toulouse, France; Centre d'Évaluation et d'Information sur la Pharmacodépendance-Addictovigilance, Centre Hospitalier Universitaire, Toulouse, France.
Abstract
PURPOSE: Addictive behaviours are often associated with hidden characteristics that are difficult to detect by usual approaches. This study aimed to estimate the incidence of serious drug-related complications by using the capture-recapture method in defined geographical area. METHODS: Hospitalizations with mention of disorders related to drug of abuse were considered serious drug-related complications. We searched these cases in and crossed three sources of data: spontaneous reports of drug of abuse-related disorders called NotS ('Notification Spontanee') collected by the regional addictovigilance centre, computerised hospital database Programme de Medicalisation des Systemes d'Information (PMSI) and toxicological analyses (TA) carried out for hospitalized patients. RESULTS: In 2007 and 2008, 1509 distinct cases were captured. After data modelling, the estimated number of psychoactive drug-related hospitalizations was 4744 (95%CI = 4060-5429). Most frequent products were opioids (34%), cannabis (19%) and cocaine (13%). 'Multiple drugs' were observed in 26% of cases. The incidence of serious drug-related complications in the area covered should be estimated at 5.7 (95%CI = 5.5-5.9) per thousand 15- to 64-year-old inhabitants. The exhaustiveness of sources were 0.4% (95%CI = 0.2-0.6) for NotS, 11.6% (95%CI = 10.7-12.5) for TA and 22.6% (95%CI = 21.4-23.8) for PMSI. CONCLUSIONS: The 'real' number of cases far exceeds that of cases that can be identified through simple counts. In particular, it confirms the underreporting and even quantifies its magnitude. These results confirm that drug users are frequently hospitalised and require heavy medical management. Moreover, these results show the real although limited advantage of hospitalization database in detecting drug associated disorders in epidemiological studies.
PURPOSE: Addictive behaviours are often associated with hidden characteristics that are difficult to detect by usual approaches. This study aimed to estimate the incidence of serious drug-related complications by using the capture-recapture method in defined geographical area. METHODS: Hospitalizations with mention of disorders related to drug of abuse were considered serious drug-related complications. We searched these cases in and crossed three sources of data: spontaneous reports of drug of abuse-related disorders called NotS ('Notification Spontanee') collected by the regional addictovigilance centre, computerised hospital database Programme de Medicalisation des Systemes d'Information (PMSI) and toxicological analyses (TA) carried out for hospitalized patients. RESULTS: In 2007 and 2008, 1509 distinct cases were captured. After data modelling, the estimated number of psychoactive drug-related hospitalizations was 4744 (95%CI = 4060-5429). Most frequent products were opioids (34%), cannabis (19%) and cocaine (13%). 'Multiple drugs' were observed in 26% of cases. The incidence of serious drug-related complications in the area covered should be estimated at 5.7 (95%CI = 5.5-5.9) per thousand 15- to 64-year-old inhabitants. The exhaustiveness of sources were 0.4% (95%CI = 0.2-0.6) for NotS, 11.6% (95%CI = 10.7-12.5) for TA and 22.6% (95%CI = 21.4-23.8) for PMSI. CONCLUSIONS: The 'real' number of cases far exceeds that of cases that can be identified through simple counts. In particular, it confirms the underreporting and even quantifies its magnitude. These results confirm that drug users are frequently hospitalised and require heavy medical management. Moreover, these results show the real although limited advantage of hospitalization database in detecting drug associated disorders in epidemiological studies.
Authors: Chinazo O Cunningham; Joanna L Starrels; Chenshu Zhang; Marcus A Bachhuber; Nancy L Sohler; Frances R Levin; Haruka Minami; Deepika E Slawek; Julia H Arnsten Journal: BMJ Open Date: 2020-12-29 Impact factor: 2.692