More information about text formats
Peters discusses some limitations in relation to our systematic
review of the prevelance of substance abuse and dependence in prisoners.
Peters states that we have not disaggregated findings obtained from
sentenced and unsentenced prisoners. In fact, in Table 3, we do present
the findings of sentenced and remand prisoners separately by gender, and
by type of substance (alcohol and drug).
Peters also states that we have reported the findings from his study
inaccurately. This depends on how the data are extracted. The Peters study
involved diagnostic interviews with 400 inmates consecutively admitted to
a state prison. Peters states that the prevalence of "drug use" was 36.8%
and "alcohol use" was 34.5% - and this uses denominators of 307 and 380
prisoners, respectively. However, we extracted the data for "dependence
disorder" rather than "use" - again, this is in order to be consistent
with the other included studies (where studies provided both data on
prevalences for "dependence" and "abuse or dependence", we used the
former). Further, we decided to use 400 inmates as the denominator. This
is based on an assumption that those who did not complete the protocols
were not substance dependent, which we believe is reasonable. Even if it
is not, as we did not meta-analyse the data, this does not alter our main
findings, and the alternative estimates from the Peters study fall within
the range of prevalences already given in our review. The alternative
prevalences for drug dependence would be 28.0% rather than 25.0%, and for
alcohol 26.3% rather than 21.5%.
Finally, Peters states that with changes in the types of drugs used
in the community, it is unclear whether the range of prevelances for drug
dependence would still be applicable. By eyeballing the figures in our
review, there does not appear to be any time trend for the prevalence
estimates in the included studies (1988-1998 in the men; 1996-2001 in the
women), periods when it is also likely for there to have been changes in
the types of drugs used in the community.
The summary notes in EBMH rightly state that the inclusion criterion
for sample size was not clearly reported in our review of reception
studies. In fact, we chose not to have a sample size criterion and all
studies - of any size - were included that met the other criteria.