HIV Testing Strategies Employed in Health Care Settings in the European Union/European Economic Area (EU/EEA)

Evidence From a Systematic Review

S Desai; L Tavoschi; AK Sullivan; L Combs; D Raben; V Delpech; SF Jakobsen; AJ Amato-Gauci; S Croxford


HIV Medicine. 2020;21(3):163-179. 

In This Article


Study Identification and Overview

The searches yielded 15 004 records after de-duplication; after full-text review, 368 studies were included in the overall systematic review (Figure 1). Of the 368 studies, 130 are described in this paper, exploring interventions and feasibility of HIV testing in health care settings in Europe, including 84 peer-reviewed articles and 46 conference proceedings (Table S8).[18–147]

Figure 1.

Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram. EU/EEA, European Union/European Economic Area.

Studies were from 13 of the 30 EU/EEA countries. Most studies were from Northern Europe (n = 78; 64%), followed by Western (n = 23) and Southern Europe (n = 24). There was only one study from Eastern Europe[109] and four studies set across multiple European countries.[110,114,127,128] The majority of the studies from Northern Europe were from the UK (92%). Other than the UK, there were two countries with more than five studies (Spain and France).

Studies were set in a range of health care facilities, the most common being primary care (n = 45) followed by inpatient services (n = 25), STI clinics (n = 24) and emergency departments (n = 23). Other health care testing sites for HIV included outpatient services (n = 16), prisons (n = 4) and pharmacies (n = 4). Almost a fifth of studies were conducted in more than one setting type.

There were a number of interventions implemented to increase HIV testing in health care settings, including innovative/improved testing provision (n = 94), use of testing campaigns (n = 14), use of communication technologies (n = 4), education and training for staff and patients (n = 20), use of tools to aid clinical decision-making (n = 10) and relocation of a clinic to a higher men who have sex with men (MSM) density area (n = 1). Twenty-two studies applied strategies with multiple interventions to increase testing.

The quality of the peer-reviewed studies was variable; of the 84 articles, 70% were of high quality, 19 were of medium quality (23%) and six were of low quality (7%) (Table S8). Risk of bias was low in 42 studies (50%), medium in 36 studies (43%) and high in six studies (7%) (Table S8). Risk of bias was low in 42 studies (50%), medium in 36 studies (43%) and high in six studies (7%) (Table S8).

Testing Provision Strategies

Novel HIV testing technologies were employed by 40 studies in a wide range of clinical settings to increase testing coverage; the majority utilized rapid testing (n = 36)[18,19,22,24,25,34,36,45,46,48–50,56,57,59,65,75,77,81,83,85,88,89,99–101,107,113,117,129,132,134,137,139,141] while four utilized self-sampling (n = 3)[44,52,102] and self-testing strategies (n = 1).[108] Two of the four self-sampling studies used oral fluid sampling while the self-test required a blood sample. Novel testing approaches were particularly applied to improve testing coverage in HIV risk groups including MSM (n = 9),[44–46,52,57,81,85,129,134] migrants (n = 7),[22,45,46,57,85,129,132] and people who use/inject drugs (PWUD/PWID) (n = 5).[50,57,85,89,129]

Other testing strategies included routine testing (n = 32),[20,21,24,25,32,35,36,43,61,64,65,69,74,77,78,92–96,98,104,111,115–117,120,125,130,135,137,147] provision of HIV testing as a component of an integrated testing programme (n = 29),[22,23,32,35,37,38,47,50,52,55,62,63,70,72,74,79,92–94,96,97,102,103,119,122,123,134,137,143] IC testing (n = 14)[26,38,47,57,88,110,114,116,120,125–128,135] and partner notification (n = 4).[54,60,104,133] Routine testing was most commonly implemented in hospital departments including emergency departments (n = 15),[20,21,25,35,36,64,69,78,92,93,115–117,137,147] inpatient units (n = 10)[24,43,61,95,98,104,111,125,130,135] and outpatient departments (n = 4).[21,74,113,116] Similarly, IC testing programmes were predominantly instigated in hospitals (n = 8)[109,114,116,125–128,135] and primary care (n = 10).[26,38,47,57,88,110,114,116,127,128] In contrast, integrated testing for HIV with other infections such as hepatitis B and C and STIs was adopted in diverse settings including prisons (n = 3),[70,103,122] STI clinics (n = 4),[23,52,63,134] drug services (n = 2)[50,79] and pharmacies (n = 1).[102]

The majority of the 94 testing interventions were directed to the general population (74%). Testing strategies directed to risk groups included studies among migrants and black and minority ethnic groups (n = 12),[22,23,32,45,46,57,62,80,85,129,132,143] MSM (n = 11),[44–46,52,57,63,81,85,129,134,136] young people (n = 1),[102] PWUD/PWID (n = 7),[50,57,79,85,89,103,129] and mental health patients (n = 1).[123] Often, these studies targeted multiple risk groups without presenting group-specific results.

HIV test coverage and positivity differed considerably between health care settings (Table 1). HIV test coverage varied from 2.9 to 94% in primary care and from 3.9 to 66% in emergency departments. HIV positivity ranged from 0 to 25% in STI clinics, with the higher rates achieved when partner notification was used to identify cases. In general, positivity rates were lower in studies set in emergency departments (0–1.3%) and antenatal services (0–0.05%) than in those set in other hospital departments (e.g. up to 5.3% in inpatient units).

Testing Provision Strategies That Increase Testing

Thirty studies evaluated the impact of a testing intervention by comparing the intervention data with baseline data (n = 24)[20,26,34,44,47,65,72,75,78,80,83,87,95,96,98,102,111,113,125,126,130,133,135,141] or with a control group (n = 6)[22,77,81,82,120,136] (Table 2). Twelve studies employed novel testing (10 rapid testing and two self-sampling) in diverse settings, of which one reported an increase in HIV test coverage from 2% before the intervention to 45% after,[65] while others reported increases in HIV diagnoses,[77,141] testing[44,75,83,102] and test acceptance[81] and higher positivity rates[34] after the intervention. The use of rapid tests also resulted in 98% of people obtaining their results compared to 64% in the standard serology group.[22] One study reported a decline in numbers of tests performed.[113] A further six studies conducted in inpatient services, TB services and primary care reported the impact of IC testing: HIV test coverage changed from 3.9–72% before to 12–85% after its implementation, with most studies reporting a 10–20% increase[26,120,125,126,135] and the median number of tests also increased.[47] Twelve studies, of which half were set in inpatient services, examined the impact of universal routine testing on test coverage. These studies reported an increase in coverage from 2–28% before to 17–80% after the intervention[65,78,98,111,125,130,135] and higher coverage in the intervention group compared to the control group (85% versus 72%, respectively).[120] Other indicators included increases in test acceptance,[20] numbers tested (although the increase was small)[95] and numbers diagnosed[77] and a reduction in vertical transmission.[96] Only six studies measured the impact of the intervention in at least one risk group; five in MSM,[44,81,133,136,141] one in young people[102] and one in migrants.[141]

Other HIV Testing Strategies Targeted to Providers

There were 30 studies using other strategies (campaigns, education and use of clinical decision-making tools) directed to providers. Three campaigns targeted providers to increase awareness using posters, social media (e.g. Twitter) and promotional materials.[21,111,146] A significant number of educational intervention studies targeted providers (n = 19), with the majority providing HIV testing training sessions to health care professionals including hospital doctors, general practitioners (GPs), medical students and nurses[30,47,67,68,77,82,83,95,105,112,126,133,138,139,146] and pharmacists.[48,49] One employed the plan, do, study, act (PDSA) methodology, which increased physicians' willingness to test but did not increase testing.[84] Another used serious incident reporting to improve testing awareness within clinics.[140] Where assessed, education provision resulted in an increase in the offer of HIV testing from 2 to 11%[146] and in the number of individuals tested from 11–13 before to 16–20 after the intervention in two smaller studies, and from 420–1056 before to 676–2333 after the intervention in two larger studies.[83,105,112,126] Two studies reported a decline in test offer from 8–15% to 0–10%, which may have been attributable to the small numbers of patients included in the studies (n = 4–26).[68,95] There were eight studies using clinical decision-making tools to aid providers in identifying populations that should be tested for HIV. A variety of tools were developed: addition of HIV tests to the blood test 'set' requests or checklist,[31,39,76,77] computer prompts for higher risk populations[27,38] and risk assessments.[118,144] Where recorded, the above interventions were successful at increasing testing.[27,31,39]

Other HIV Testing Strategies Targeted to Patients

Other than interventions where testing was provided, there were 23 studies using other interventions directed to patients. The majority of these interventions were campaigns that promoted local testing using social media, posters, digital media and websites,[19,78,83,89,146] campaigns to promote National HIV Testing Week[21,71,111,124] and other regional campaigns to promote testing.[48,49,85,97,142] National HIV Testing Week increased testing from 4–9% before to 8–28% during the week,[111,124] and it also resulted in half of those having blood samples collected at a hospital being tested for HIV.[21] During the regional Go Viral campaign, 27% of patients were tested for blood-borne viruses (BBVs).[97]

There were two educational interventions targeting patients; one for pregnant women,[90] which resulted in an increase in testing coverage (from 87 to 92%) after provision of a patient information leaflet and one for patients admitted to a hospital inpatient unit, where there was a decline in test offer (from 8 to 0%).[95] All four communication technology studies were directed to patients, with two providing videos on HIV testing,[24,106] one utilizing text messages to recall MSM for testing[136] and one using online platforms for partner notification.[60] The only study measuring intervention impact reported an increase in the re-testing rate among MSM who were actively recalled (from 19 to 44%).[136] Two studies adopted decision-making tools that helped individuals determine if an HIV test would be recommended.[42,118] One of these examined whether computer-assisted self-interviewing resulted in increased HIV testing when compared to interviews with clinicians[118] and found significantly less testing in the self-interviewing population (63% versus 69%, respectively). Finally, in one study, the STI clinic moved location to be in a higher MSM density area, which resulted in a large increase in the number of HIV diagnoses from 175 in 2008 to 381 in 2013 after relocation.[73]

Feasibility and Acceptability

HIV testing interventions were generally acceptable to patients and providers in health care settings (Table 3). Results also suggest that rapid testing is acceptable to both groups (patient studies, n = 5; provider studies, n = 6), with providers willing to use rapid tests[24,53,56,57,107] and finding their interpretation easy.[50] Some feasibility studies highlighted that nontraditional health care settings can target populations not previously tested for HIV (n = 4),[18,85,132,145] with the reported percentage of first-time testers in such settings ranging from 51 to 75%. One study suggested that provider-initiated testing is unlikely to be acceptable when specific populations are targeted (in this case, sub-Saharan African patients).[86]