Opt-out HIV testing in a hospital emergency department may cost less than managing two patients with undiagnosed HIV

Opt-out HIV testing in accident and emergency departments in areas of high HIV prevalence is likely to save NHS trusts money by avoiding expensive in-patient admissions for individuals with unexplained and deteriorating illnesses, according to presentations at the British HIV Association conference in Manchester last week.

The first analysis came from King’s College Hospital, which is located in an extremely high prevalence area of south London, where 1.2% of people have HIV, including 0.4% who are undiagnosed. Although national guidelines have recommended opt-out HIV testing in emergency departments in high prevalence areas since 2008, financial pressures have discouraged local NHS managers from implementing the policy. 

Clinicians advocating for increased HIV testing at King’s College wanted to demonstrate to the NHS that it would be a cost-saving measure. 

In 2014 a total of 34,500 people had blood tests in the hospital’s emergency department. Based on the local prevalence of undiagnosed HIV, providing HIV testing to all would have identified 138 new cases. A more realistic testing uptake of 50 or 75% would have identified 69 or 104 cases. 

There would be costs to this provision though. The cost of the test kits would have been £85,000, £127,000 or £170,000 depending on the level of uptake. 

In addition the clinicians added some fixed costs for this hypothetical programme, so as to ensure its smooth running and widespread uptake. A results administrator, project management costs, education and training, and marketing were estimated to cost £75,000 a year. Other conference delegates suggested that this may be more than is strictly necessary. 

But late diagnosis of HIV also costs money. The researchers identified 59 people who had been diagnosed with HIV in the hospital in 2014, excluding people diagnosed in sexual health or antenatal services. One quarter (15) of these individuals had in fact attended the hospital’s emergency department in the five years before their eventual diagnosis but hadn’t been offered HIV testing. When they were eventually diagnosed, most were severely immunosuppressed, with median and mean CD4 cell counts of 61 and 117.

The total cost of inpatient stays and outpatient investigations for these 15 patients, due to late presentation, was estimated to be £336,000. One man who was eventually diagnosed with a CD4 count of 34 and toxoplasmosis had had care costing £114,000 since his missed opportunity for HIV testing. Another man’s bill totalled £93,000.

The cost of this avoidable medical care was therefore greater than the cost of a testing programme. If such a programme had 75 or 100% uptake, it would cost each year a total of £203,000 or £246,000.

And the above analyses do not include calculating the potential savings from preventing HIV transmission in people with undiagnosed infection.

But how to actually achieve a high uptake of HIV testing in routine clinical practice? The problem does not generally lie with patients’ refusal to be tested, but in making healthcare professionals’ offer of HIV testing genuinely routine.

A separate presentation by staff at St Thomas’ Hospital (in the same high-prevalence area of south London as King’s) showed that it is possible if barriers to testing are removed. In the St Thomas’ emergency department computer systems were reorganised so that the order form for blood tests included an HIV test as a default option. Patients were informed, through leaflets and other means, that HIV testing is routine but can be refused.

The HIV department automatically receives an email when a test is reactive (suspected HIV-positive) and a health adviser contacts the patient directly to arrange confirmatory testing and linkage to care. Emergency department staff therefore do not need to get involved in managing reactive results.  Nonetheless their engagement is encouraged by being sent weekly, individualised feedback on their provision of HIV testing and of positive diagnoses.

Before the introduction of this programme in the summer of 2015 approximately 2% of emergency department patients were tested for HIV. Since the programme was introduced, 64% of patients have been tested, with a steady increase in testing rates over time.

Looking at data for the first 36 weeks, 172 of the 19,569 tests done were positive (0.9% prevalence). Around half of these individuals were already engaged with HIV care but 68 (0.3% prevalence) were diagnosed for the first time and 13 had previously been diagnosed but had disengaged from care. All but eight of these 81 people are now engaged with HIV care.

One in five of the newly diagnosed individuals were diagnosed with primary (recent) HIV infection. Over half of those newly diagnosed had previously attended the emergency department and so probably could have been diagnosed on an earlier occasion. Since the policy was introduced, HIV-related in-patient admissions have been reduced by 15% – confirming the conclusion of the previous study that an opt-out HIV testing policy in emergency departments can help NHS trusts avoid more expensive interventions.