Cochrane Reviews
Alcohol Misuse Prevention
(NEW)Universal school-based prevention programs for alcohol misuse in young people
Summary
Alcohol misuse involves consuming alcohol to a point that causes either short- or long-term mental, physical, or social problems. Alcohol misuse is a particular health concern among young people, with alcohol causing 25% of male deaths and 10% of female deaths in those aged 15-24. Broad school-based prevention programs aim to discourage and prevent alcohol misuse among this group.
In their review, "Universal school-based prevention programs for alcohol misuse in young people", Foxcroft and Tsertsvadze studied whether such programs worked in children and young people under the age of 18. They included both alcohol-specific programs and generic programs (addressing a broader set of behaviours, such as tobacco and drug use and anti-social behaviour). The review included 53 trials. The trials included different combinations of the following intervention components: problem-solving, peer resistance, or decision-making skills training; education about the benefits, harms, and consequences of alcohol misuse; resilient behaviour; changing normative attitudes and beliefs; self-esteem; and social-networking skills.
Overall, prevention interventions that focused on alcohol misuse behaviour exclusively made some improvements compared to the standard curriculum; six of 11 trials reported some benefit. Amongst the 39 trials of generic program interventions aiming to prevent multiple problem behaviours, a positive impact on alcohol misuse was observed in some cases, but not all (14 of 39 showed benefits).
How were sex and gender considered?
The review authors reported that two trials were male-only, while in the mixed-sex trials, the proportion of males ranged from 36.5%-62%. Ten trials did not report the sex or gender distribution of participants.
Because of heterogeneity in populations, outcomes, and interventions, the authors did not conduct a meta-analysis and were thus not able to do a pre-specified subgroup analysis by sex. Rather, they reported their results narratively (i.e. using words to summarize, rather than statistics).
In one of the alcohol-specific trials, the intervention was significantly associated with lower rates of mean weekly alcohol use in the intervention compared with the control groups. Subgroup analyses showed reduced weekly alcohol use in the female subgroup but not in males. In another trial assessing a generic intervention aimed at multiple behaviors, the prevention programs did not significantly improve alcohol misuse at the whole group level, but did have a significant effect on drunkenness in males (OR = 0.64; 95% CI 0.49, 0.85), but not in females (OR = 0.86; 95% CI 0.63, 1.18).
The review authors discussed the observed sex differences in the trials, noting that "because examining smaller samples reduces statistical power, potential important effects in some subgroups may not have reached statistical significance (e.g. weaker effects in females)" (p. 14), and that "subgroup analyses through multiple testing may result in type I error and spuriously significant associations (e.g., positive effects shown in males but not in females)" (p. 14). They also suggest that looking only at main effects conceals potential subgroup effects such as stronger effects in males, and conclude their argument by saying that "gender and baseline alcohol use are potential effect moderators, so by not accounting for them in the analysis, subgroup effects may be missed" (p.14).
Strengths and weaknesses of the approach to sex and gender
The careful and thoughtful discussion of the statistical challenges involved in conducted subgroup analyses by sex is the greatest strength of this review. The review authors demonstrate a clear understanding of the difficulties in using averages only, and use sex and gender as the exemplars for their argument. Further, they pre-specified subgroup analyses by gender and reported sex distribution of the trial participants. The review authors highlighted the importance of gender in the review abstract: they noted that gender modified the impact of the interventions. This is an additional strength of the review, as reporting the effects of gender in the abstract helps both decision-makers and researchers identify this review as being relevant to sex and gender.
A weakness of this review is the use of "gender" in conjunction with terms referring to biological characteristics. That is, the reviewers said "gender", but they referred to "males" and "females" instead of "men" and "women". Moreover, the term "sex" is used in this review to refer to "boys" and "girls". Increased attention to standardized use of the terms sex and gender is needed in Cochrane reviews, and will strengthen their applicability and usefulness.
What do we know about sex and gender based on this review?
Implications for Policy and Practice: This review demonstrates that the sex of participants may influence the effects of school alcohol misuse prevention programs. Though the review did not discuss gender directly, it is likely that the differential effects of the prevention programs are closely tied to gender roles and the links between these roles and alcohol use and misuse.
Implications for Research: Future research is needed to evaluate the role of gender in school alcohol misuse prevention programs.
Foxcroft DR, Tsertsvadze A. Universal school-based prevention programs for alcohol misuse in young people. Cochrane Database of Systematic Reviews 2011, Issue 5. Art. No.: CD009113. DOI: 10.1002/14651858.CD009113
Violence
(NEW)School-based secondary prevention programmes for preventing violence
Summary
Violence is defined by the World Health Organization as "the intentional use of physical force or power, threatened or actual, against oneself, another person, or against a group or community, that either results in or has a high likelihood of resulting in injury, death, psychological harm, maldevelopment or deprivation".1
In their review "School-based secondary prevention programmes for preventing violence", Mytton et al. studied the effects of school-based violence prevention programmes for children and adolescents who were identified as aggressive or at risk of being aggressive. The interventions included in the review aimed to lower anger, bullying, violence, aggression, or conflict, or aimed to help youth with Conduct Disorder or Oppositional Defiant Disorder via school-based programmes. Some school-based programmes were delivered in conjunction with community interventions or skills training for parents.
Overall, this review concludes that school-based prevention programmes appear to improve behaviour at the primary and secondary school levels and this effect continues for up to 12 months after the intervention is applied.
How were sex and gender considered?
Of the 56 studies included in the review, 34 were trials suitable for meta-analysis. There were 2939 students included in these 34 studies, 12 of which were conducted solely with groups of boys and 22 with mixed groups of boys and girls. The authors undertook a preplanned subgroup analysis on the effect of sex of the participants, which showed that the interventions were effective when delivered in groups that included males-only and mixed-sex groups. Post-intervention differences in observed aggression appeared to be stronger for mixed sex groups (2234 participants, SMD = -0.45; 95% CI -0.64 to -0.26) than for males-only groups (705 participants, SMD = -0.35; 95% CI -0.61 to -0.08). However, the substantial overlap in the confidence intervals suggests that the difference may be due to chance alone. Only one study included only females and was excluded from the meta-analysis.
Strengths and weaknesses of the approach to sex and gender
A strength of this review is that the review authors conducted a preplanned subgroup analysis. Moreover, they only planned a small number of subgroup analyses. Thus, they have reduced the risk of spurious results.
A weakness of this review is that authors did not have access to sex-disaggregated data so, we have little information about the female participants. The subgroup analysis shows that the interventions work in male-only groups and in groups that include males and females. This may point to the role of gender, but this is not discussed. Further, the authors could have drawn from the substantial evidence base on the gendered nature of violence to situate their review findings.
What do we know about sex and gender based on this review?
Implications for Policy and Practice: Violence prevention interventions are effective at improving behaviour when delivered to either mixed-sex or male-only groups.
Implications for Research: Research is needed on the effects of interventions beyond 12 months and the effect of interventions for female participants.
Mytton JA, DiGuiseppi C, Gough D, Taylor RS, Logan S. School-based secondary prevention programmes for preventing violence. Cochrane Database of Systematic Reviews 2006, Issue 3. Art. No.: CD004606. DOI: 10.1002/14651858.CD004606.pub2.
Medication
Vitamin D supplementation for improving bone mineral density in children
Cochrane Journal Club
Podcast
Review summary
Relevance for Sex and/or Gender
Vitamin D has three major roles in bone health: 1) it helps our bodies absorb more calcium from the food that we eat; 2) it helps our bodies add calcium to our bones; and 3) it prevents the loss of calcium from our bodies. These functions are particularly important for children, as the bone mass developed during childhood affects the chances of fractures and osteoporosis (weak, brittle bones) in later life.
In their Cochrane review, Winzenberg et al. examine whether taking vitamin D supplements can improve bone mineral density in children. One of the objectives of the review was to examine differential effects of vitamin D by sex. The authors assessed this through subgroup analyses by sex.
The authors report the age and ethnicity of participants early in their results. However, they do not summarize the sex or gender distribution of participants here; the first mention of skewed sex distribution comes on page 15 in the discussion around heterogeneity (“There was a sole study in males”). They do, however, report the distribution of sex across studies in the tables summarizing the characteristics of the primary studies included in the review. Moreover, they report data for both sexes from each study, even when a study included only one sex (e.g. “M = 0%, F = 100%”). They are consistent and appropriate in their preference for “sex” as opposed to “gender”.
From the tables summarizing the primary study characteristics, the reader can calculate that of the six studies included in the meta-analyses (total n = 884), five of them included only female participants (n = 712), while one included only male participants (n = 172). Thus, when the data are pooled from all studies for the meta-analyses, the effects are significantly influenced by the data from females (80% of the whole). The authors identify the lack of data from males as a significant gap in the evidence.
When data from males and females were pooled, taking vitamin D did not significantly improve bone density across the entire body, nor was there a significant improvement in lumbar spine bone mineral density (BMD). However, when males and females were assessed separately, vitamin D significantly improved lumbar spine BMD for females (P = 0.04) but not males (P = 0.93).
Winzenberg TM, Powell S, Shaw KA, Jones G. Vitamin D supplementation for improving bone mineral density in children. Cochrane Database of Systematic Reviews 2010, Issue 10. Art. No.: CD006944. DOI: 10.1002/14651858.CD006944.pub2.
Pharmacotherapy for hypertension in women of different races
Relevance for Sex and/or Gender
Hypertension or high blood pressure increases the risk of stroke and heart disease. There are many risk factors for hypertension, some of which may be controlled (e.g. weight, sodium intake) and some of which cannot (e.g. sex, age, race). It may be treated with prescription medications (pharmacotherapy).
The Cochrane review on pharmacotherapy for hypertension in women of different races is somewhat unusual, in that most Cochrane reviews are not explicitly restricted to one sex—and where they are, this is biologically justified (e.g. reviews on Caesarean sections or prostate cancer). The authors justify their population choice by noting that treatment recommendations for women are based on evidence that combines results for both men and women. Moreover, they note that information is needed on the differential effects of treatments by race and age; statistical power in individual studies is often not sufficient to detect racial and age differences through subgroup analyses.
The findings from this review highlight the need to consider intersectionality among the determinants of health. For example, the authors found that treatment for hypertension reduces the risk of dying from any cause in African American women (relative risk 0.66; 95% confidence interval 0.51-0.86) and women over the age of 54 (relative risk 0.89; 95% confidence interval 0.80-1.00). However, treating hypertension in young white women (30-54 years old) did not significantly reduce the risk of dying from any cause (relative risk 1.08; 95% confidence interval 0.73-1.58).
Quan AP, Kerlikowske K, Gueyffier F, Boissel JP, INDANA Investigators. Pharmacotherapy for hypertension in women of different races. Cochrane Database of Systematic Reviews 2000, Issue 3. Art. No.: CD002146. DOI: 10.1002/14651858.CD002146.
Prevention
Psychosocial interventions for reducing risk behaviours for preventing HIV in drug users
Relevance for Sex and/or Gender
In their review, “Psychosocial interventions for reducing injection and sexual risk behaviour for preventing HIV in drug users”, Meader et al examined the effectiveness of multi-session psychosocial interventions (i.e. at least three sessions for individuals/groups combining education and skills training) on injection and sexual risk behaviours that are associated with HIV infection. Those who misused cocaine, opiates, or a combination thereof were eligible.
The reviewers concluded that “There were minimal differences identified between multi-session psychosocial interventions and standard educational interventions for both injection and sexual risk behaviour”, but that “single-gender groups may be associated with greater benefit” (p.2).
The authors reported that “Trials conducted on specific gender groups (mainly females) appeared to be slightly more effective”, but it is very difficult to determine which studies were limited to females and which studies were limited to males. A review of the Table of Included Studies shows that there were seven studies limited to females (total sample size = 1,324) and only one study limited to males (sample size = 152). It would have been helpful to report this in the Results. Further, it is surprising that the authors did not conduct subgroup analyses by gender or sex. Instead, they only conducted subgroup analyses by sex composition, reported as gender composition (mixed versus single-gender interventions). Moreover, given the proportion of female-only studies compared with male-only studies, it is not clear whether it was the single-sex factor or the female-sex factor that had a significant impact on the interventions’ effectiveness.Meader N, Li R, Des Jarlais DC, Pilling S. Psychosocial interventions for reducing injection and sexual risk behaviour for preventing HIV in drug users. Cochrane Database of Systematic Reviews 2010, Issue 1. Art. No.: CD007192. DOI:10.1002/14651858.CD007192.pub2.
Screening for abdominal aortic aneurysm
Relevance for Sex and/or Gender
An aneurysm is when a blood vessel widens or balloons outward. This weakens the blood vessel wall, which may cause it to rupture (creating a potentially life-threatening situation). In an abdominal aortic aneurysm (AAA), the blood vessel affected is the abdominal aorta, which moves oxygenated blood from the heart to the abdomen, legs, and pelvis.
In their review on screening for AAA, Cosford et al. report in the abstract that 5-10% of men aged 65-79 have AAA, but do not report the prevalence in other age groups for men or for women at all. The first discussion of women’s burden of illness comes in the methods section of the review, where the authors note that “The age-specific incidence of aortic aneurysm is lower in women than men, although the annual rate of rupture is higher (Brown 1999). This may alter the cost effectiveness of screening between men and women” (p. 3).
The authors report that four studies were included in this review, of which only one included women (127,891 men and 9,342 women). However, when the results are presented separately for men and women, the authors use sex terms (males/females). Thus, it is not clear whether they are considering sex or gender.
Meta-analyses showed that screening for AAA significantly reduced the incidence of death from AAA in males (three studies: OR 0.60, 95% CI 0.47-0.78; p=0.00011), but not in females (one study: OR 1.99, 95% CI 0.36-10.88; p = 0.43). Similarly, the incidence of progression to ruptured AAA was significantly reduced in males (one study: p=0.048), but not in females (one study: p=0.66). Neither males nor females had a significant reduction in death from all causes (three studies and one study for males and females respectively).
The reporting on sex and gender in this review is not well organized and may even be misleading. In the first line of the discussion section, the authors state that “the results appear to identify evidence of significant benefit in men” (p. 7). The statistically insignificant results for women are not explicitly noted until much later in the discussion. The authors make the point that one of the uncertainties in the review is the limited information available on the benefits and harms of screening for women [“insufficient numbers to produce any meaningful results” (p. 7)]. Notwithstanding the data limitations for effects on women, the authors later make the generalization that “the overall population benefit from screening appears to be established” (p. 7). The authors reverse this conclusion again when they report that “there is insufficient evidence to demonstrate benefit in women”, and that “the most significant gap in the current research is the balance of benefits and harms in women” (p. 7). They recommend that research in women should be a future focus.
Cosford PA, Leng GC, Thomas J. Screening for abdominal aortic aneurysm. Cochrane Database of Systematic Reviews 2007, Issue 2. Art. No.: CD002945. DOI: 10.1002/14651858.CD002945.pub2
Cranberries for preventing urinary tract infections
Relevance for Sex and/or Gender
In their review, "Cranberries for preventing urinary tract infections", Jepson et al examine the effectiveness of using cranberries to prevent urinary tract infections. They describe the sex difference (though reported as a gender difference) in population rates of urinary tract infections: these infections are 50 times more common in adult females than in adult males. They also report comparative rates for boys and girls, and give a possible biological explanation for the sex differences: the urethra in females (again, reported as "women") is shorter than the urethra in males ("men"). However, when the review authors report on one of their analyses, the sex/gender distribution of the participants is given only for two studies (women only) and is unclear for the other two.
The authors note that while cranberries can be effective in preventing urinary tract infections, this may only be the case for specific sub-populations: beneficial effects were shown in women with recurrent urinary tract infections, but were less clear in older men and women, and in people with neuropathic bladder. There is no discussion of younger men.
Jepson RG, Craig JC. Cranberries for preventing urinary tract infections. Cochrane Database of Systematic Reviews 2008, Issue 1. Art. No.: Cd001321. DOI: 10.1002/14651858.CD001321.pub4.
Primary Care
Brief alcohol interventions in primary care
Relevance for Sex and/or Gender
The authors of "Effectiveness of brief alcohol interventions in primary care populations" noted that 70% of the participants in the included studies were men. They also reported that only 8 of the 29 included studies could be included in subgroup analyses by gender; the remaining studies did not report the necessary data.
In their gender analyses, Kaner et al concluded that brief alcohol interventions delivered in primary care settings were effective for men (P = 0.00041), but not for women (P=0.63). Had the results not be presented separately, one may have concluded that brief alcohol interventions were effective for women also; the pooled result was significant (P=0.0039). The authors note that, as low statistical power may have contributed to the findings for women, more research data are needed to evaluate the effect of brief alcohol interventions in women.
Kaner EF, Dickinson HO, Beyer FR, Campbell F, Schlesinger C, Heather N, Saunders JB, Burnand B, Pienaar ED. Effectiveness of brief alcohol interventions in primary care populations. Cochrane Database of Systematic Reviews 2007, Issue 2. Art. No.: Cd004148. DOI: 10.1002/14651858.CD004148.pub3.
Contact information
Erin Ueffing
Email: erin.ueffing@uottawa.ca
Campbell and Cochrane Equity Methods Group
