Changes in HIV testing rates among patients with tuberculosis in a large multiethnic city in the UK

INTRODUCTION

The UK surveillance data show that approximately one-third of adult HIV infections remain undiagnosed. The British HIV Association (BHIVA) published HIV testing guidelines in 2008 to encourage ‘normalization’ of HIV testing in conditions where HIV infection enters the differential diagnosis. ¹ We undertook a retrospective audit to evaluate HIV testing rates among tuberculosis (TB) patients in 2008/09 before and after the implementation of BHIVA HIV testing guidelines (2008) and compared the findings with a similar audit done in 2005.

METHODS

For both audits, TB patients were identified using the Birmingham TB Aftercare Database, which includes data from the Birmingham and Solihull Health Protection Agency (HPA) catchment area. Data collection periods were 1 January to 31 December 2005 and 1 April 2008 to 31 March 2009.

Data collected included age, gender, ethnicity, country of birth, notification date, respiratory/non-respiratory diagnosis and hospital attended. Patients aged under 18 years, private patients, on chemoprophylaxis, non-TB mycobacteria and diagnoses outside the catchment area were excluded.

HIV testing was determined through hospital laboratory systems in the six months prior to and following TB notification. As the primary outcome of the audit was to determine the HIV testing rates, the rate of positive HIV results was not analysed. Data were anonymized before analysis using SPSS software (SPSS Inc, Chicago, IL, USA). Differences between groups were determined by calculating the odds ratio (OR) and 95% confidence interval (CI) for each variable. A separate analysis was done on patient characteristics pre- and post-guidelines to account for any variations.

RESULTS

Following exclusions, 371 patients in 2005 and 407 patients in 2008/09 were included in the study. In both audits, the majority were in the 18–34 years age group (318, 41%) and were South Asian (471, 61%) with a roughly equal men to women ratio (54% versus 46%). Most (576, 74%) were born outside the UK. Total TB notifications increased by 10% but respiratory notifications decreased by 22% while non-respiratory notifications increased by 34% since 2005.

HIV TESTING

Overall, HIV testing increased from 14% in 2005 to 43% in 2008/09 (P < 0.001). In 2005, patients aged >55 years and of Asian origin were significantly less likely to have been tested compared with those aged 18–34 years and those of white ethnicity. In 2008/09, age groups 35–54 and >55 years were significantly less likely to be tested. No other significant differences were found in patient characteristics for both cohorts (Tables 1 and 2).

Comparing both audit cohorts, patients were significantly more likely to have been tested in all patient characteristic groups in 2008–09 compared with 2005–06, except for two ethnic groups – black African and African Caribbean – where no significant differences were observed across the two periods (Table 3).

No statistically significant difference in testing rates was observed before and after the BHIVA guideline implementation (OR 1.3, 0.89–1.93). Analysing the case-mix profile before and after the guideline implementation, no significant differences were noted except in the age group distribution (18–34 years – 103 versus 69, P < 0.001; 35–54 years – 60 versus 90, P = 0.03; >55 years – 37 versus 66, P < 0.01), with more patients in >18 years in the post-guidelines cohort. In 2008/09, 230 (57%) patients with TB were not tested for HIV, of which 103 (53%) were seen after the new guideline implementation.

DISCUSSION

The audits have shown a significant increase in the rates of HIV testing among patients with TB in the Birmingham and Solihull HPA region between 2005 and 2008/09, but failed to show any change in testing rates since the implementation of the new BHIVA HIV testing guidelines. The increased rate of testing seen in all patient categories, including the ethnic groups, in the 2008/09 audit suggests that clinicians are moving towards a universal approach rather than one based on ethnicity alone. However, a significant difference between age groups in both audits, with patients aged over 35 years less likely to be tested in 2008/09, suggests age may influence offer of testing. The lack of significant difference in testing rates before and after guidelines could be explained by several factors, such as the audit being done too soon after issue of the guidelines, seasonal variations, differences in the age group distribution, language barriers and non-availability of professional translators within ethnic minorities, which could limit offer and uptake of HIV testing. Equally important could be low guideline awareness among clinicians, as TB patients are usually looked after by non-HIV physicians (respiratory/infectious diseases). A recent HIV and TB audit by BHIVA also showed that only around 25% of the TB and integrated services offered HIV testing as a routine for all adults. ² These findings highlight the importance of publicizing the BHIVA testing guidelines beyond the genitourinary medicine/HIV specialty. It is unlikely that a large number of patients were offered testing but declined in our audits, as a recent study in London showed that 73% with TB accepted HIV testing when offered. ³

Although improved HIV testing rates were found between the two audits, it is a major concern that 57% of patients with TB in 2008/09 and 53% since publication of the BHIVA guidelines were not tested for HIV. This study highlights the persisting gaps in HIV testing among TB patients and underlines the importance of continued efforts at various levels towards achieving a universal testing for HIV in TB services in the UK.