Acceptance of HIV testing in medical inpatients: a local acceptability study

INTRODUCTION

The background prevalence of HIV in Croydon is 4.17 per 1000 in those aged from 15 to 59 years (Health Protection Agency, UK, 2008). ¹ In 2008 the British HIV Association (BHIVA), the British Association for Sexual Health and HIV (BASHH), and the British Infection Society released national guidelines for HIV testing in the UK. ² These guidelines recommend local evaluation of HIV testing for all medical admissions aged from 15 to 59 years in areas where the prevalence is above 2 in 1000, as late diagnosis is the most important factor related to death due to HIV in the UK. Strategies such as opt-out testing in the antenatal programme have increased HIV testing; ³ however, it is estimated that one in four people with HIV in the UK still remain undiagnosed. ⁴ An important part of the 2008 HIV testing guidelines was the identification of patients with HIV indicator illnesses. HIV indicator illnesses include conditions such as pneumonia or sepsis, which are very common in medical admissions. However, a previous audit carried out in our hospital in June 2009 by Dr R Hill-Tout showed that only 0.85% of medical admissions (9/1047) were tested for HIV. ⁵

METHODS

For a two-week period starting 05 September 2009, an acceptability study of HIV testing was performed at Mayday Hospital. A total of 104 consecutive medical admissions aged from 15 to 59 years were identified by contacting the medical team on call each day. Patients were included if they were admitted for a long enough period of time for our team to approach them about HIV testing (24 hours). Patients who were unable to consent for themselves were considered on a case-by-case basis and some were consented via an interpreter. The primary outcome measure was the number of patients who consented to have an HIV test. Other information collated included: demographic data, previous HIV testing, reasons for accepting an HIV test, reasons for declining an HIV test and the number of patients with HIV indicator illnesses. Verbal and written information was given to each patient about HIV testing and the HIV test was performed using a serum sample for antigen and antibody testing. Each patient was assigned a genitourinary (GU) medicine clinic number for the HIV test in order to help maintain patient confidentiality.

The study design was discussed with the local Research and Ethics Committee. As all HIV testing was carried out by an experienced GU medicine team using confidential GU medicine patient notes and numbers, it was deemed to be ethically appropriate and in line with recommended national guidelines.

RESULTS

A total of 104 patients aged from 15 to 59 years were approached by our team for HIV testing: 34/104 (32.7%) patients were men, 69/104 (66.3%) were women and 1/104 (1%) was transgender. The mean age was 42 years (range 16–59); 56/104 (53.9%) patients were white, 11/104 (10.6%) were black African, 10/104 (9.6%) were black British, 8/104 (7.7%) were black Caribbean and 19/104 (18.2%) were Asian.

Seventy-six of the 104 patients (73.1%) had never been tested for HIV before, 5/104 patients (4.8%) were unsure if they had been tested before and 23/104 patients (22.1%) reported having been tested previously. Eleven of those 23 reporting a previous HIV test had been tested in Mayday Hospital and 11/23 had been tested in the last three years. There were 3/104 patients (2.9%) already known to be HIV-positive and these were excluded from the study.

We offered HIV testing to the remaining 101 patients: 84/101 patients (83.2%) agreed to be tested for HIV; 43/84 (51.2%) were white, 19/84 (22.6%) were Asian, 11/84 (13.1%) were black African, 6/84 (7.1%) were black Caribbean and 5/84 (6%) were black UK. In all, 18/84 reported having taken a test before. In those 17/101 (16.8%) patients who declined testing, 12/17 (70%) were white and 5/17 reported having taken a test before. All 101 patients tested negative for HIV. In all, 43/101 patients had HIV indicator illnesses (three people had >1 indicator illness) (Table 1).

DISCUSSION

We performed the study over a two-week period (for practical reasons) and interviewed 104 patients during this time. This figure was arrived at as a result of the time constraints of a busy GU medicine team and an agreed study period. Over the course of our study, 43/104 medical admissions (41.3%) were admitted with HIV indicator illnesses as defined by BHIVA. ² In all, 3/104 patients (2.9%) were known to be HIV-positive; however, of the remaining 101 patients, 84 of them (83.2%) agreed to be tested for HIV. In a previous audit in June 2009, only 0.85% (9/1047) of all medical admissions were tested for HIV. Our study suggests that most patients would agree to have a test if offered one. Patients who were admitted to hospital for <24 hours were less likely to be consented for HIV testing as they were often discharged before a member of our clinical team could approach them: this is consistent with a previous study. ⁶ The 2008 guidelines recommend that any qualified health-care professional can consent a patient for HIV testing; however, in our study the patients were consented by a member of the HIV inpatient clinical team. It is possible that the experience that the team have in discussing issues surrounding HIV increased the numbers of patients who consented but provides valuable insight into what may be achievable when testing inpatients. The patients were told four key messages: (1) that HIV is treatable, (2) more universal HIV testing would save lives, (3) early diagnosis of HIV leads to a good prognosis and (4) HIV testing is confidential. Most patients had no questions regarding HIV testing even though the majority had never had an HIV test before (78/104 patients [75%]). None of the 101 patients tested positive for HIV.

This was an acceptability study of 104 patients and not a prevalence study. The local prevalence of HIV is 4.17 per 1000 and statistically we would have needed to have tested 240 patients in order to have obtained one positive result.

The patients who declined an HIV test made comments consistent with the documented role of HIV-related stigma as a barrier to testing, as shown in previous studies. ^7,8 Some patients declining an HIV test did not want to have an extra blood sample taken, as daily blood tests were requested by their own clinical team. Incorporation of routine HIV testing as part of other clinical investigations or the use of point of care tests may have reduced the number of patients who declined. ⁹ Many patients who accepted an HIV test did not regard HIV testing as any different from other clinical investigations and some thought that HIV testing was done routinely for every admission.

This study shows that HIV testing in medical inpatients is highly acceptable if patients are reassured that HIV is treatable with a good prognosis, and that testing is confidential. We aim to continue to increase the education and training of clinical staff regarding HIV testing and have presented the findings of this study to our colleagues in the hospital to encourage more HIV testing. Our study supports the idea that more universal HIV testing will be acceptable to medical inpatients. We are currently exploring practical ways to implement more universal HIV testing in our hospital and negotiating financial support for this with the hospital management and commissioners.