A London-based study in which HIV tests were offered in a non-targeted manner to as many patients as possible between the ages of 16 and 65 has concluded that routine HIV testing in higher-prevalence areas in the UK is feasible and acceptable.
The HINTS study, which has previously reported some of its findings at UK conferences, took place between August 2009 and September 2010. During this time, non-targeted HIV tests were conducted over four different three-month periods in four medical settings in four London boroughs. These were:
An emergency (A&E) department at a large central London teaching hospital (Chelsea and Westminster).
An acute care unit for inpatients at an east London hospital in a deprived area (Homerton).
An outpatient clinic specialising in dermatology (skin conditions) at a hospital in the highest-prevalence local authority in the UK (King’s College Hospital, Lambeth).
A large primary care practice in a west London borough with wide socioeconomic inequalities (Hammersmith and Fulham).
Each of these areas has diagnosed HIV prevalence in the general population considerably in excess of the 0.2% (two cases per 1000) that is the threshold recommended by the British HIV Association (BHIVA) in 2008 for widespread, non-targeted HIV testing: prevalence varied from 0.815% in Hammersmith and Fulham to 1.33% in Lambeth.
HINTS found eight undiagnosed HIV infections out of a total of 4105 people tested, yielding an undiagnosed-HIV prevalence of 1.95 cases per 1000 people or one new diagnosis in every 513 people tested. However, prevalence varied significantly between sites. No people were diagnosed in the outpatients’ or primary care clinics, and the four detected in the A&E department translates to a undiagnosed prevalence of 0.19%, close to the BHIVA recommended threshold; but the four detected in the acute care unit, where fewer tests were conducted, represents an undiagnosed prevalence of 1.01%, five times higher.
The HINTS study was not powered to detect ‘true’ undiagnosed prevalence (it is estimated that a study of 30,000 people would be needed to do this) and the 95% confidence interval is wide (meaning that the ‘true’ undiagnosed-HIV prevalence in acute care could be anything from 0.03% to 2.05%), but it could indicate that routine testing in hospital medical admissions units would be particularly cost-effective. In addition to the eight diagnosed, two further new diagnoses were made through contact tracing.
There were seven false-positive results from the tests, five in the primary care clinic and two in A&E. These were eliminated after the quick test (which used oral fluid) was shown to be false in a second blood test.
The study issued questionnaires to staff in the different settings before conducting the study (not necessarily the staff who ended up performing the HIV tests), and to patients offered tests during the study.
Only 4% of staff disagreed that HIV tests should be offered in settings other than STI and antenatal clinics, and 83% said HIV tests should be routinely offered to all patients. However, only 54% (42% in the A&E department) said they would personally feel comfortable offering routine HIV tests, 63% said they would be concerned that patients would have questions they could not answer (72% in the inpatient unit), and 72% said they felt they would need additional training before routinely offering HIV tests (82% in the inpatient unit). A clear training need was thus identified, and additional training given to staff members who actually did perform tests.
A total of 15,042 patients between 16 and 65 were seen for care in the four units during the testing periods. Of these, 8% were ruled out because they either already had HIV or had recently been tested. Of the remaining 13,855, 7033 (51%) were actually approached for a test, of whom 839 (12% of those approached) were excluded as clinically ineligible.
The highest proportion of exclusions was in the A&E department, and the most common reasons were that the patient was too unwell, that there was a language barrier (especially in A&E), that they were too distressed, or that they had learning difficulties. Patients of black African ethnicity were actually somewhat less likely than average to be offered a test; 7% of all patients were African but only 5% of those were offered tests. It is not known why the offer rate was lower in Africans – there might have been more clinical exclusions – but it at least means healthcare staff were not trying to target patients. Patients who said they were gay on the questionnaire were no more likely to be offered tests than others.
Of 6194 offers of tests, 4105 (66%) were accepted by patients. This uptake rate ranged from 75% in the primary care clinic to 62% in the A&E department. This meant that, of the 13,855 patients eligible for a test, 30% got one. This coverage rate is in fact considerably higher than in most US studies of routine opt-out testing.
The only predictors of higher test uptake were male sex (men were twice as likely than women to accept a test), being under 27 (2.3 times more likely than average) and never having taken an HIV test before (four times more likely). Gay and black African patients were neither more nor less likely than average to take a test if offered.
About a quarter of patients were offered a questionnaire and most returned one. The most common reasons given for declining a test were “I have tested recently” (54%), “I do not think I am at risk of HIV” (47%), and “I have other health concerns today” (24% – patients could tick more than one box). The fact that most patients previously tested for HIV had had their last test more than a year ago may indicate some need for patient education about testing frequency. Nonetheless, most patients thought it acceptable to be offered a test, even 85% of those who turned the offer down.
In terms of practicability, a lower proportion of patients were tested in the A&E and acute care settings. This was because of high patient numbers in the A&E department, which required two HIV testers being on duty at all times, while in the acute care unit it was due to rapid throughput of patients, with many only staying one night.
In terms of cost-effectiveness, the researchers note that studies modelling cost-effectiveness in the UK have yet to be done: the figure for cost-effectiveness if background prevalence is over 0.2% is taken from US studies.
What the researchers do conclude, though, is that “routine HIV testing in non-specialist settings in the UK is highly acceptable to patients and staff” and that “such a strategy is feasible to deliver in a diverse range of settings”.
The challenge may now be to use this data to influence clinical practice and guidelines. Last year, a review of guidelines presented at the BHIVA conference showed that many guidelines issued for non-HIV specialists are outdated and in some cases specifically rule out HIV testing, even in some patients at higher risk or with conditions that could indicate HIV.
The HINTS researchers comment that their study “offers a substantial evidence base to underpin strong recommendations...that routine HIV testing in secondary and primary care settings in areas of high HIV prevalence ought to be developed and implemented as a matter of urgency”.
Rayment M et al. HIV testing in non-traditional settings – the HINTS study: a multi-centre observational study of feasibility and acceptability. PLoS One, 7(6):e39530. Doi:10.1371/journal.pome.0039530. 2012. Click here for open-access article.
