HIV testing: increasing uptake among people who may have undiagnosed HIV

Updating the previous guidelines

The committee discussed the recommendations and considerations in the 2 guidelines being updated (NICE guidelines PH33 and PH34), and considered the view of the experts in the review decision. The committee agreed that the previous recommendations were still pertinent but they needed some updating to better reflect current practice. The committee removed or amended parts of the recommendations that it agreed were outdated. For more information, see update information.

The committee concluded that many of the recommendations from the previous guidelines were aimed at a broader population than men who have sex with men, and black Africans. For the more specific recommendations, the committee agreed it would be appropriate to broaden them to apply to any population with a high rate of HIV. The committee recognised that men who have sex with men, and black Africans are still the most high-risk groups for HIV in the UK and this is reflected in the guideline (recommendation 1.2.6). An additional benefit of broadening the recommendations to apply to everyone with undiagnosed HIV is that it should help to normalise HIV testing so that it is not seen differently from any other blood test. For more information on the relevance of this guideline for other groups, see the equality impact assessment.

HIV testing and HIV screening

The committee discussed the distinction between testing and screening. It was reminded that recommending screening programmes is outside NICE's role. It was aware that HIV screening in antenatal settings is currently recommended by the UK National Screening Committee and has a very high uptake. It also discussed the differences between opt-in and opt-out approaches to testing, that is whether people are asked if they want an HIV test or they are told they will be tested for HIV unless they specifically ask not to be. It agreed that it was important to make sure people understand that HIV testing is voluntary and to give everyone the opportunity to opt out of a test.

The committee discussed that HIV testing may not always be routinely undertaken among people who have conditions that might indicate HIV infection. For this reason, it agreed that it is important for national guidelines to recommend HIV testing when diagnosing or treating conditions that may indicate HIV infection (see HIV in Europe's guidance on HIV in indicator conditions).

Other guidelines

The committee agreed that the British HIV Association's UK national guidelines for HIV testing are the most up-to-date HIV guidelines available in the UK and are also accredited by NICE, so they remain in the recommendations. The only exception relates to indicator conditions, for which more up-to-date guidance is available from HIV in Europe's guidance on HIV in indicator conditions. Although this guideline has not been accredited by NICE, the committee agreed it was a useful and authoritative source of information to give people who are offering HIV testing.

High prevalence and extremely high prevalence of HIV

The committee discussed the meaning of 'high prevalence'. NICE guidelines PH33 and PH34 used the definition given by Public Health England of 2 in 1,000 people. However, as more people are being diagnosed with HIV and treated, and because they are living longer, the background prevalence of HIV is rising and the overall UK prevalence in 2014 was 1.9 per 1,000.

The committee was aware that during the development of this guideline, Public Health England carried out a new analysis of 2014 data on diagnosed HIV prevalence distribution in local authorities in England. Based on this analysis, groups and communities at high prevalence of diagnosed HIV can be defined as follows:

When this is applied to national late HIV diagnosis data, it shows that two-thirds of late HIV diagnoses occur in high-prevalence and extremely-high-prevalence local authorities. This means that if this guideline is successfully applied in high- and extremely-high-prevalence areas, it could potentially affect two-thirds of late diagnoses nationally.

The committee agreed that the definitions of high prevalence and extremely high prevalence were useful, and it incorporated them into the recommendations. The committee also agreed it was important for areas to find out their diagnosed prevalence, as well as that of surrounding areas (recommendation 1.1.1). This allows them to adapt their strategy for HIV testing using the recommendations in this guideline. These data are available from Public Health England's sexual and reproductive health profiles. Additionally, the committee was aware that local data are often available to supplement national data from Public Health England, and that local data may be more up to date and provide more granular detail about local groups and populations with a high rate of HIV.

Evidence statements not used to make recommendations

The committee did not make recommendations for all of the evidence statements. This was mainly because it did not believe, based on the evidence, that an intervention was effective; or it agreed that the intervention may not be applicable in the UK. For details of the evidence statements not used to make recommendations, see the evidence reviews section.

Current practice

Evidence for effectiveness

Evidence for cost effectiveness

The committee noted that late diagnosis of HIV is substantially more costly than early diagnosis, because of the costs of inpatient admission and treatment. The committee was aware that prompt diagnosis of HIV would be cost saving per person, and there could be further cost savings through averting transmission (recommendation 1.2.9). The committee agreed that HIV testing is most cost effective when the cost per positive test is lowest. The committee noted that the cost of an HIV test is incurred for each person tested and therefore the cost per positive test result decreases when either the cost of the HIV test decreases, or the HIV prevalence in a population increases.

The committee noted that the review of cost effectiveness identified mixed evidence as to whether it is more cost effective to test everyone for HIV, or to target certain groups. The studies identified considered the general population in the UK or the US and compared various strategies for testing everyone or only those who disclosed risk factors that identified them as a higher-risk group. Differences in populations, strategies, inputs, cost perspectives and model structure presented difficulties in comparing and interpreting the results. The committee was limited in how it could use this evidence in making recommendations.

The committee discussed that the additional cost of testing for HIV for a person already undergoing blood tests was likely to be low. The committee felt that in an areas of high and extremely high prevalence, offering HIV testing to everyone undergoing blood tests for another reason would be cost effective, because of the low incremental cost and high HIV prevalence (recommendations 1.1.6, 1.1.9).

The committee discussed that HIV testing for everyone admitted to hospital would incur an additional cost, but that this would be cost effective in areas of extremely high prevalence (recommendation 1.1.7) given the likelihood of bloods being taken for other reasons.

The committee discussed that HIV testing for everyone attending a GP surgery would incur an additional cost, but that opportunistic testing would be cost effective in areas of extremely high prevalence (recommendation 1.1.10).

In other settings or areas of lower prevalence, the committee felt that targeting specific groups was more likely to be cost effective: that is, it would result in the lowest cost per positive test (recommendations 1.1.5, 1.1.8, 1.2.6).

The committee noted an analysis of pilot projects for HIV testing in hospitals, primary care and community settings and for self-sampling, which found variation in cost per positive test. It concluded that no single setting was likely to offer the greatest value for money. The committee recommended setting up community testing in high- and extremely high-prevalence areas in addition to testing in hospitals and primary care (recommendation 1.1.14). The same analysis found that the cost per test of self-sampling was comparable to the cost per test in other settings, and so could represent a cost-effective method of diagnosing HIV (recommendations 1.1.16, 1.2.3, 1.2.4, 1.2.8).

Resource impact and implementation issues

When discussing testing in GP surgeries and emergency departments (section 1.1), the committee discussed the resource impact of the recommendations. It felt that making testing routine for all attendees in GP surgeries and emergency settings could have a substantial resource impact and could only be justified in areas of extremely high prevalence. However, the committee agreed that opportunistic testing for everyone having a blood test in these settings in extremely-high- and high-prevalence areas represents best practice, although it was mindful of the resource implications. The committee also considered the potential resource impact of self-sampling. An expert told the committee that self-sampling was likely to cost the same as traditional testing approaches involving healthcare professionals. It may also be an effective approach in harder to reach subgroups at risk, and could have potential to reduce late diagnosis in these groups. This could reduce the health and social care costs associated with late diagnosis [EP2].

The committee discussed further the role that primary care professionals (particularly GPs) could have in opportunistically offering and recommending HIV testing to a subpopulation of people at risk who are difficult to reach through methods described in recommendations 1.1.8 and 1.1.9. It discussed promoting opportunistic testing at every consultation in extremely-high-prevalence areas as a way of promoting testing to this subpopulation. The committee agreed that HIV testing for everyone attending a GP surgery would incur an additional cost. The proportion of HIV tests conducted in GP surgeries was estimated to be 3.8%. The committee reasoned that if 3.8% of all people aged 15 to 59 in areas of extremely high prevalence who do not have an HIV diagnosis had an HIV test in their GP surgery, the estimated cost would be between £2 million and £4 million depending on the cost used for GP time and laboratory costs. The committee noted that, given current HIV testing practice, this was a highly unlikely scenario and therefore the resource impact of the recommendation is likely to be much smaller. Notwithstanding, given the annual treatment costs for HIV, the life expectancy of somebody living with HIV and the potential reduction in transmission due to treatment, the committee agreed that opportunistic GP testing was good value.

Evidence for effectiveness

There was some evidence of effectiveness for media campaigns, educational videos, online social networking and information leaflets in raising awareness of HIV testing. The committee agreed that different methods of awareness raising are effective for different groups and that a range of methods should be recommended [ES1, ES2, ES6, ES7, ES8, ES9, ES10].

The committee discussed the evidence from 1 study on the effectiveness of motivational interviewing for increasing uptake of HIV testing [ES4]. The committee felt that there was not enough evidence from the study to recommend this as an intervention but that referring to NICE's guidelines on behaviour change would cover this type of approach (recommendation 1.3.1).

The committee noted that evidence suggested computerised interviews and risk assessments were unlikely to be effective at increasing uptake of HIV testing in people who may have undiagnosed HIV. However, it did not think the evidence was strong enough to recommend that they should not be used. [ES3].

There was evidence suggesting that financial incentives are effective at increasing uptake of HIV testing [ES21]. However, the committee noted a lack of evidence from UK healthcare settings. Without stronger evidence of effectiveness in the UK, the potentially significant resource impact of the intervention could not be justified.

Evidence for cost effectiveness and resource impact

The committee noted that awareness raising was a core element of any health promotion and it deemed the resource impact of the new recommendations to be minimal. The intention behind the new recommendations was to bring the principles in line with current technologies and modes of delivery of messages.

Evidence statements from 2011 guidelines

Evidence statement number E1.1 indicates that the linked statement is numbered 1.1 in review 1 for PH33, 'Review of effectiveness and cost effectiveness: increasing the uptake of HIV testing to reduce undiagnosed infection and prevent transmission among black African communities living in England'. Q1 indicates that the linked statement is numbered 1 in review 2 for PH33, 'Increasing the uptake of HIV testing to reduce undiagnosed infection and prevent transmission among black African communities living in England – barriers to HIV testing'. ES1 indicates that the linked statement is numbered 1 in the review for PH34 'Preventing and reducing HIV transmission among men who have sex with men'.

The expert report 'Time to test for HIV: expanded healthcare and community HIV testing in England. Interim report' was also used to inform the original recommendations in PH33 and PH34.

Evidence statements from 2016 reviews

ES1a.1 indicates that the linked statement is numbered 1 in review 1a for this guideline. ES1b.1 indicates that the linked statement is numbered 1 in review 1b for this guideline. ES1c.1 indicates that the linked statement is numbered 1 in review 1c for this guideline. ES2.1 indicates that the linked statement is numbered 1 in review 2 for this guideline. EP1 indicates that expert paper 'Targeted testing based on indicator conditions' is linked to a recommendation. EP2 indicates that expert paper 'HIV testing in the UK and summary of current UK practice' is linked to a recommendation.

If a recommendation is not directly taken from the evidence statements, but is inferred from the evidence, this is indicated by IDE (inference derived from the evidence).