Slow progress towards routine HIV testing in acute settings: clinician education needed

Only a minority of patients with indicators for HIV testing in hospital and primary care settings are receiving tests, according to several different surveys presented at the second BHIVA/BASHH conference in Manchester.

It is now a year and a half since national guidelines were issued, written by a consortium of professional and patient bodies including BHIVA (the British HIV Association) and BASHH (the British Association for Sexual Health and HIV).

These guidelines proposed – following on from a recommendation by the Chief Medical Officer – that HIV tests should be offered to all adults presenting for care who live in areas where HIV prevalence is more than 0.1% and/or who present with one of a list of ‘indicator diseases’ suggesting HIV infection.

No less than a quarter of presentations at the BHIVA/BASHH conference concerned HIV testing, and a large number were local surveys of testing practice. The general conclusion was that non-HIV clinicians have a long way to catch up in terms of HIV awareness. To quote some of the surveys:

• “The testing guidelines seem to have had no effect.” (Phillips)
• “Although testing is recommended as routine, clinicians still seem to be targeting it, yet failing to identify the majority of undiagnosed infection.” (Perry)
• “Some of the reasons for not testing provided by clinicians are evidence of an old-fashioned view of those at risk of HIV and associated illnesses. [One clinician said] ‘I only test if the patient is really camp’.” (Gupta)
• [In a multi-ethnic area]: “HIV testing was not discussed or performed on any patient of Black African or Caribbean ethnicity.” (Millett)

Although most presenters recommended better clinician education, one presenter from Guy's and St Thomas' NHS Foundation Trust said that their trust was considering making HIV testing a matter of clinical governance, and was logging as serious clinical incidents situations where admissions to hospital could have been avoided if they had been tested earlier (Read).

In hospital surveys, the proportion of eligible patients who were offered tests varied from 40 to 14% and varied widely between specialities and individual clinicians in each survey.

When patients were offered tests, acceptance rates were high, indicating that lack of HIV testing is more to do with clinician than patient education. Although in many settings the number of HIV diagnoses was no higher than the national average, in certain others – such as of hospital inpatients – it was over 20 times higher.

Individual surveys

A survey from Brighton (Perry) looked at 3913 acute general medicine patients admitted to hospital between August and December 2009. One in five had one of the indicator diseases suggestive of HIV infection. Only 1560 (40%) of patients were offered a test; when offered, 90% accepted one. Offer rates in individual healthcare workers varied between 20% and 60%. There were two new HIV diagnoses (0.13%).

Meanwhile an anonymous serotesting programme found that, in fact, 44 patients in acute general medicine had HIV (1.3%). While most of these were known to be HIV-positive, eleven were not and tests only detected two of these. Nine patients (20.5%) admitted with an acute condition to hospital therefore were discharged without their HIV being diagnosed.

A survey from Blackpool (Phillips) comparedMycobacterium infections including tuberculosis. Ninety-six were tested (26%). HIV testing rates actually went down after the guidelines were issued, with 31% of patients tested before their release from hospital and 20% afterwards. Forty per cent of HCV patients were offered a test but only 8% of HBV patients and 5.5% of patients with mycobacterial infections.

A survey from Basildon Hospital in Essex (Gupta) in 347 patients with a wider range of indicator diseases in the year after the guidelines were issued found a low rate of HIV testing (13.8%). Three-quarters of patients with tuberculosis were quite likely to be offered a test (one declined a test when offered), but only 22 of patients with hepatitis B, 19 with hepatitis C, 7% with lymphoma and 1.3% with CIN (cervical pre-cancer) were offered HIV tests, and, of note, none of the eight patients seen with anal cancer.

One reason for the low testing rates in this hospital was because, due possibly to unwarranted concerns about confidentiality, HIV results were not available on the computerised results database used in other tests. After this audit was presented, HIV tests were put on the database, 84% of doctors felt their practice around HIV testing had improved, and the rate of HIV tests increased by 60%.

A survey from Homerton Hospital in east London (Millett) looked at 100 consecutive admissions for acute conditions plus another 50 aged under 65 (because most acute admissions are in the elderly). It found that 33 patients (22%) of patients had a clinical indicator disease, with pneumonia the most common. Nine of these had an HIV test discussion and eight took a test, meaning that the testing rate in patients with indicator diseases was 24%. Of note, although 31% of patients were of Black ethnicity (12% Black African), none were offered a test.

A large survey (Read) looked at all HIV tests performed by the laboratory at Guy’s and St Thomas’s Hospital, which covers the highest HIV prevalence area in the UK, in 2008 (Read). The lab performed 41,095 tests for 36,392 people of whom 363 were HIV-positive. 18,872 tests came from STI clinic patients (1% positive) and 6197 from antenatal clinics (0.5% positive). Of the rest:

• 5746 came from GPs; the HIV-positive rate in these tests was actually higher than in STI patients (1.12%). Forty-three per cent of patients had indicator diseases, indicating that local GPs are becoming more aware of conditions suggesting an HIV test.
• 5303 came from hospital out-patients of whom 26 (0.5%) were HIV positive. Seventeen were new diagnoses of whom seven had attended Guy’s and St Thomas’s in the last year with indicator diseases without being tested. The average CD4 count in these seven patients was 190. Some departments where patients typically present with high rates of indicator disease ordered very few tests, for instance gastroenterology (83 tests).
• 1225 came from hospital in-patients of whom 34 were HIV positive (2.77%). Of these patients, 62% had an AIDS defining condition and the average CD4 count at testing was 62. These patients spent an average 34 days in hospital with one patient admitted for nearly a year. The average cost of each case was £36,625.

Presenter Julie Fox commented that the majority of HIV-positive patients identified would have been missed by targeting high-risk groups. HIV testing needed to be broadened to all acute settings and be truly opt-out. HIV testing rates have been made a SQUIN, an incentivised performance indictor, which should ensure more tests, but situations like the inpatients who had remained untested until admitted very ill were being logged as serious clinical incidents with a view to making non-testing an issue of clinical governance.