Western society has long been concerned about the threat posed to young people by drug use and, in response, has sought to develop universal prevention programs. Traditionally, these have been school-based and education focused, but in recent years there has been increasing support for family interventions seeking to address the risk and protective factors that originate from the way a family functions. The universal prevention programs evaluated by Spoth and his colleagues  comprise a mix of school based (Life Skills Training (LST)) and family focused interventions (Iowa Strengthening Families Program (ISFP), Strengthening Families Program 10–14 (SFP 10–14)), plus a combination, comprising LST+SFP 10–14. These interventions have been subject to numerous previous evaluations that indicate their beneficial effect on the uptake of various licit and illicit substances, and, on the basis that they derive from sound science; all are represented on various US Government lists of endorsed prevention programs, [2–4]. In this respect their pedigree is impressive. This study investigates their effect on prescription drug misuse and seeks to demonstrate that sound universal drug prevention programs will also be effective in preventing uptake of these substances.
The authors make a solid case for the increasing prevalence of prescription drug misuse and the harms that accompany this practice. Clearly, it would be beneficial if existing prevention programs were also effective against the misuse of prescription drugs. Accordingly, the research is timely and potentially very valuable. However, the scientific rigour of school-based prevention programs, repeatedly identified as effective, in various lists compiled by US Government agencies and academic bodies has been questioned in a number of studies. Gandhi and colleagues  examined the evaluations of five programs consistently cited in these lists, including LST, and indicated their overall concern as to the ‘frailty of evidence’ (p. 65) underpinning listed ‘proven’ programs. Gorman  similarly raised general concerns about the scientific rigour of those school-based prevention programs consistently listed as evidence-based, best practice. In a more recent article  he and colleagues specifically critiqued the evidence-base of the combined intervention, LST+SFP 10–14, that Spoth and colleagues have evaluated in this study. They examined a previous study by Spoth and colleagues  that essentially tested the same participants, with the same suite of interventions and a similar methodology to the present study. The difference was the dependent variable: methamphetamine use. In this earlier methamphetamine study a one-tailed test was used to assess significance, in one instance with the alpha level set at 0.10. There was no pre-test methamphetamine prevalence data, so past year cannabis use was employed to examine differential attrition. The number of methamphetamine users was very small, so an intent-to-treat analysis that inflated the size of the sample analysed also magnified the effect of small differences between groups.
These criticisms can be applied in almost equal measure to the current study, and I conducted a re-analysis of the data using the same as-treated methods employed by Gorman and colleagues in their critique . The as-treated samples were considered to be the number of participants, who received one of the interventions (ISFP, PDFY, LST or LST+SFP 10–14), not the numbers assessed at follow up. Accordingly, the number of users in the respective intervention groups were reduced proportionately. For instance, in the case of 11th graders, who received LST+SFP 10–14, the as-treated sample was 137 (the number who participated in the program). This was 73% less than the number (516) assessed at follow up, so the prescription drug misuser number was reduced by a similar percentage from 20 to 5. A two tailed Fisher's Exact Test, with an alpha level of 0.05, was used to determine the significance of any difference. The results of this re-analysis of study 1 and study 2 data are respectively presented in Tables 1 and 2. In most cases, differences between the intervention and control groups were no longer significant: the one robust effect was that of ISFP on misuse of prescription narcotics.
|Follow up point, prescription drug type and outcome measure|
|12th grade past year misuse of prescription narcotics||Young adult lifetime misuse of prescription narcotics||Young adult lifetime misuse of prescription barbiturates|
|Yes||No||Total||P value 2 tailed Fisher's Exact Test||Yes||No||Total||P value 2 tailed Fisher's Exact Test||Yes||No||Total||P value 2 tailed Fisher's Exact Test|
|Follow up point and prescription drug outcome measure|
|11th grade lifetime misuse||12th grade lifetime misuse|
|Yes||No||Total||P value 2 tailed Fisher's Exact Test||Yes||No||Total||P value 2 tailed Fisher's Exact Test|
This study stirs my broader unease about the paradigm that underpins effort in this area. I feel I echo the concerns of a number of researchers [5–7, 9], who have critically examined model universal prevention programs for young people, in saying that the research base is flawed and not that useful for choosing meaningful prevention interventions. This is partly because of technical flaws in the research process, but also because of unrealistic expectations of what can be achieved by universal prevention programs. Gandhi and colleagues  question whether such programs can ever substantially prevent drug use by young people, but this does not seem to spur a reappraisal of policy or practice. Skager  states that negative findings seem to disappear into a void rather than stimulate alternative approaches, and part of this ‘paralysis of vision’ (p 581) is a fixation on universal abstinence. The clear and immediate implication of this study is the need for more research rigour. The longer-term implication is the need for broader and more pragmatic approaches that value reduction in harm as much as reduction in use.