BACKGROUND: The alcohol-related problems (ARPs) are a relevant issue in public health and contribute to premature deaths and avoidable disease burden. The capture-recapture (C-R) method can be a useful tool to provide reliable estimates for populations with hidden nature such as subjects with ARP. METHODS: C-R method was used to estimate the 'true count' of individuals with ARP using three independent health-related current databases in an area of northern Italy during 2007. To predict the frequency of unascertained cases, we constructed log linear models. The goodness-of-fit of a model was measured by the likelihood ratio test and the final model was selected using Akaike's Information Criterion. Confidence intervals (CIs) were calculated according to Hook and Regal. RESULTS: Altogether 1014 subjects with ARP were directly identified from the three sources using the C-R method the estimated unknown population was 2729 subjects, giving a total of 3743 subjects with ARP (95% CI 3148-4504) and a prevalence of 8.24 (95% CI 7.97-8.50) per 1000 inhabitants aged >15 years. The analyses stratified for gender estimated 12.31/1000 (95% CI 11.85-12.77) men and 4.86/1000 (95% CI 4.58-5.14) women with ARP. Besides, the analysis calculated a prevalence of 14.99 per 1000 (95% CI 14.29-15.69) for males <50 years (1731), corresponding to the majority of subjects with ARP. CONCLUSION: The C-R technique is useful to provide a more realistic picture of the size of ARP population. This has important implications both for future planning of service provision and for the way in which the impact of ARP interventions are evaluated.
BACKGROUND: The alcohol-related problems (ARPs) are a relevant issue in public health and contribute to premature deaths and avoidable disease burden. The capture-recapture (C-R) method can be a useful tool to provide reliable estimates for populations with hidden nature such as subjects with ARP. METHODS: C-R method was used to estimate the 'true count' of individuals with ARP using three independent health-related current databases in an area of northern Italy during 2007. To predict the frequency of unascertained cases, we constructed log linear models. The goodness-of-fit of a model was measured by the likelihood ratio test and the final model was selected using Akaike's Information Criterion. Confidence intervals (CIs) were calculated according to Hook and Regal. RESULTS: Altogether 1014 subjects with ARP were directly identified from the three sources using the C-R method the estimated unknown population was 2729 subjects, giving a total of 3743 subjects with ARP (95% CI 3148-4504) and a prevalence of 8.24 (95% CI 7.97-8.50) per 1000 inhabitants aged >15 years. The analyses stratified for gender estimated 12.31/1000 (95% CI 11.85-12.77) men and 4.86/1000 (95% CI 4.58-5.14) women with ARP. Besides, the analysis calculated a prevalence of 14.99 per 1000 (95% CI 14.29-15.69) for males <50 years (1731), corresponding to the majority of subjects with ARP. CONCLUSION: The C-R technique is useful to provide a more realistic picture of the size of ARP population. This has important implications both for future planning of service provision and for the way in which the impact of ARP interventions are evaluated.
Authors: Fenglian Xu; Elizabeth A Sullivan; Richard C Madden; Deborah Black; Lisa R Jackson Pulver Journal: BMC Med Res Methodol Date: 2012-01-30 Impact factor: 4.615