Sexually transmitted infections: condom distribution schemes

Outcomes

Most included studies reported intermediate outcomes, such as intention to use condoms, condom use at last intercourse or attitudinal measures. Few reported STI outcomes – those that did were poor quality studies.

The committee was aware that the focus of this guideline was condom provision to prevent STIs and that avoiding pregnancy was outside the scope. But it was also clear that increasing condom use would help avoid some pregnancies. Indeed, many of the proxy measures mentioned are as relevant to preventing pregnancy as to STI prevention.

The costs of these avoided pregnancies were included in the economic analysis. In addition, the review database was checked to ensure that no studies with pregnancy outcomes alone had been overlooked and the committee was confident that a body of evidence had not been missed.

The committee was aware of a recently published meta-analysis indicating that: 'interventions increasing the availability of, or accessibility to, condoms were shown to be efficacious in increasing condom use behaviours and that condom schemes provide an efficacious means of HIV/STI prevention'. Although most papers in the meta-analysis did not meet the inclusion criteria for this guideline (for example, two-thirds were in Africa or Asia) the committee was reassured that this review supported its conclusions.

Efficacy of structural-level condom distribution interventions: a meta-analysis of U.S. and international studies (Charania MR, Crepaz N, Guenther-Gray C et al. 2011)

Unintended consequences

The included studies clearly showed that condom schemes do not increase levels of sexual activity among young people, nor do they reduce the age at which young people become sexually active [ES1, ES5].

Key gaps in the evidence

The committee was clear that an understanding of behaviour change must underpin condom schemes, but there was no specific evidence on the techniques used to deliver any of the evaluated schemes. The committee agreed that schemes should be delivered in line with the NICE guideline on behaviour change: general approaches, behaviour change: individual approaches and patient experience in adult NHS services.

The committee discussed the importance of collecting data from UK-based condom schemes. An expert told the committee that one of the largest sources of evidence could be existing C‑Card schemes. Many of these undertake regular, extensive monitoring and evaluation of their programmes, as recommended in the C‑Card guidance [EP1].

The committee agreed that a standardised approach to assessing the effectiveness of local schemes would be extremely beneficial and enable a national evidence synthesis to explore the effectiveness of C‑Card schemes in STI prevention. It agreed this should be strongly reflected in the research recommendations.

The committee discussed the possibility that online services could address some equity issues for people in rural areas. But the evidence was poor so the committee made a research recommendation on this.

Some GP surgeries provide condoms as part of their reproductive health role, but the committee questioned whether they would want to get involved in condom schemes aimed specifically at preventing STIs. Stakeholder comments reassured them that GPs might be interested in delivering such schemes. The committee agreed that this approach could be particularly useful for GP practices used by those most at risk, for example, in universities. But because there was little evidence, the committee made a research recommendation on this.

Cost effectiveness

The committee heard expert testimony on the cost effectiveness of condom schemes. This showed that they are most cost effective and sustainable if they target people at most risk of STIs and are embedded into existing services. It was also told that: "Commissioning is currently taking place within the context of a challenging economic climate. Local authority budgets, in particular, are reducing, which results in less funding available for prevention work … [such as] for condom and lube schemes."

The expert noted that commissioners need to work collaboratively and commission services for communities of people who share a common interest, belief or other characteristic, not just communities linked by geographical area. Another expert told the committee that in recent years some schemes have been commissioned to cover multiple local authority areas. Examples include the Come Correct scheme in London, which is funded by more than 20 local authorities. This enables local areas to buy into a pre-designed scheme and potentially benefit from economies of scale [ES10, EA, EP1, EP2].

Inequalities

The committee recognised that targeting schemes at different population groups or geographical areas could lead to inequalities (for example, people living outside cities may not have access to city-based services). It also noted the lack of evidence of effectiveness for some groups, for example, people with learning disabilities. For this reason, the committee kept its recommendations as broad as possible.

It also agreed that, although the evidence for selling condoms at cost price was lacking, it could help offset some of the potential inequalities generated by targeted schemes. It agreed that web-based postal systems, in particular, might help to overcome inequalities related to geographical isolation or stigma [ES7]. But, based on the economic modelling, these schemes would need to be very low cost.

Cost effectiveness

The economic analysis used a model scheme that provided education, condoms (using a credit card type C‑Card) and telephone counselling, because these are common elements. Effectiveness data came from a multicomponent scheme for school students (aged 17) who were 1.23 times more likely to use a condom than students in a school without a scheme. It cost £0.48 per person per year, calculated from cost data from 4 UK C‑Card schemes. This included costs for: condoms and lubricants, staff time for training and administration, website, advertising and the C‑Card.

STIs included in the model were chlamydia, gonorrhoea, HIV and syphilis. STI diagnosis rates were used to judge the initial prevalence of these STIs. STI incidence in the model was influenced by initial prevalence, transmission rates, sexual activity levels and condom failure rates, as well as condom use. The incidence of each STI was associated with a quality-adjusted life year (QALY) loss and a cost.

For a target population aged 13 to 18, the model showed a condom distribution scheme prevented 1,373 STIs. This led to savings on STI-related costs of £758,947. Each person gained 17 QALYs. The scheme cost £1,538,499 and the incremental cost effectiveness ratio (ICER) of using condom schemes to prevent STIs only was £45,856.

The committee was clear that increasing condom use would also help avoid some pregnancies, so an economic analysis was conducted on preventing pregnancies for a population aged 14 to 18. This used an economic model from the NICE guideline on contraceptive services for under 25s. The model assumed that all pregnancies in this age group were unintended and that increased condom use would either delay or prevent pregnancy, as well as preventing STIs.

The committee noted that for a population of 100,000 people aged 14 to 18, increasing condom use by 22% would lead to pregnancy-related savings of over £11 million. This would make condom schemes highly cost saving.

Analysis conducted for a target population aged 13 to 25, considering the effect of condom schemes on STIs only, resulted in an ICER of £17,411. Condom schemes were more cost effective for this broader age group because the rates of both sexual activity and STI prevalence were higher. In this age group, an analysis of the effect of increasing HIV prevalence to 0.19% (the UK average) showed that this would make the condom scheme cost saving. So it would be more effective and cost less than not providing a scheme. A scenario analysis considered that training provided by multicomponent schemes may reduce condom breakage. This reduced the ICER to £14,469, potentially demonstrating the importance of including training in condom schemes.

The committee noted a threshold analysis in which effectiveness (relative risk of condom use) and cost per person in the target population were varied. This suggested that schemes can be cost effective even without considering pregnancy effects, as long as costs and condom use effects can be balanced. If a scheme did not achieve any change in condom use it would not be cost effective because it would accrue the cost for no health benefits.

Additionally, the committee heard that the use of static models and short time horizons are likely to have underestimated the cost effectiveness of schemes.

Cost effectiveness

The committee noted that specific cost and effectiveness data were not available for condom schemes aimed at adults most at risk of STIs. Cost-effectiveness analysis was conducted for 2 groups that include people most at risk of HIV: men who have sex with men, and black Africans. This showed that distributing free condoms to people at most risk is highly cost effective or cost saving, even with high scheme costs and relatively small effects. That is because, among people most at risk of an STI, a small increase in condom usage can avert numerous HIV cases, saving £100,000 and 4.5 QALYs per case [EA].

Cost effectiveness

The committee agreed that programmes selling condoms at cost or reduced price, particularly large-scale programmes such as a national web-based scheme, could be cost effective. They would have the added advantage of diminishing some of the potential inequalities in service provision as a result of specific targeting of condom schemes – possibly reaching people who would not otherwise be able to access schemes. This was particularly felt to be the case for web-based postal schemes [ES7].

The committee considered the cost effectiveness of condom schemes for the general population. It noted that when using diagnosis rates for HIV prevalence, condom schemes would have to increase condom use by 20% and cost less than £0.20 per person per year. It noted that any reduction in unplanned pregnancies would increase the cost effectiveness of schemes.

In an analysis that increased HIV prevalence to an average of 0.19%, schemes that cost £5 per person would be cost effective if they increased condom use by more than 50%.

The committee discussed the fact that ICERs are higher for the general population because of the relatively low prevalence of STIs, and that schemes targeting those most at risk would be more cost effective.

In an analysis that increased HIV prevalence to 0.4%, condom schemes costing up to £2 per person per year would be cost effective if they increased condom use by 10%. Those costing £5 would be cost effective if they increased use by 24%. If a scheme did not achieve any change in condom use it would not be cost effective because it would accrue the cost for no health benefits.

If cost-price sales schemes can recover any administrative costs, through charging for condoms and postage and packaging, they could be cost‑neutral.