Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Factors associated with late HIV diagnosis in North-East Scotland: a six-year retrospective study Noble G1, Okpo E 1, 2, Tonna I 3, Fielding S 1 1 Institute of Applied Health Sciences, University of Aberdeen, Aberdeen, Scotland, UK, 2 NHS Grampian, Public Health Department, Summerfield House, 2 Eday Road, Aberdeen, AB15 6RE 3 Aberdeen Royal Infirmary, Infectious Diseases Unit, Emergency Care Centre, Aberdeen, UK Keywords: HIV; late diagnosis; missed diagnosis; CD4 cell count Corresponding author: Dr Emmanuel Okpo, Consultant in Public Health Medicine, NHS Grampian, Public Health Department, Summerfield House, 2 Eday Road, Aberdeen, AB15 6RE Email: [email protected], Telephone: 01224 558429 1 Abstract Objectives Late HIV diagnosis is associated with increased morbidity and mortality, increased risk of transmission, impaired response to antiretroviral therapy and increased healthcare costs. The aim of this study was to determine the factors associated with late HIV diagnosis in Grampian, North-East Scotland. Study design A population based retrospective database analysis. Methods All newly diagnosed HIV positive individuals in Grampian, North – East Scotland between 2009 and 2014 were included in the study. Participants were classified as having a late diagnosis if the CD4 cell count at presentation was less than 350 cells/mm3. Socio-economic and demographic factors were investigated in relation to outcome (late diagnosis) using Chi-squared and Mann-Whitney Tests. Results CD cell count results were available for 111 (89.5%) of the 124 newly diagnosed individuals during the study Period. The prevalence of late diagnosis was 53.2% (n=59). Those infected via heterosexual mode of transmission had a 2.83 times higher odds of late diagnosis (OR 2.83 (95% CI; 1.10, 7.32)) than men who have sex with men (MSM) and those with no previous HIV testing had a 5.46 increased odds of late diagnosis (OR 5.46 (95% CI; 1.89, 15.81)) compared to those who had previously been tested. Missed opportunities for HIV diagnosis were identified in 16.3% (n=15) of participants. Conclusion Heterosexual individuals and those with no previous HIV testing were more likely to be diagnosed late. Targeted initiatives to increase perception of HIV risk and uptake of testing in these risk groups are recommended. 2 Introduction Human immunodeficiency virus (HIV) infection is now a treatable medical condition with the introduction of effective antiretroviral therapy (ART) which has improved clinical outcomes for people living with the infection. Despite this, there continues to be avoidable morbidity and mortality associated with HIV in the United Kingdom (UK) [1]. Undiagnosed, untreated and advanced infection facilitates onward HIV transmission and impinges upon both individual and wider population health. In 2013, around 107,800 people were estimated to be living with HIV in the UK and it is estimated that about a quarter (24%) of these people are unaware of their infection status [2]. All persons living with HIV are at some risk of transmitting HIV to others; those that are unaware of their infection are particularly important because they unknowing could infect others without realising [2]. Late HIV diagnosis has been associated with increased morbidity [3] and mortality [4], blunted response to ART [5] and increased healthcare costs [6, 7]. Late diagnosis has been identified as a major risk factor for early intensive care unit admission in HIV infected patients [8] and is the largest remediable factor for HIV-associated deaths in the UK [4]. A national audit of deaths amongst HIVpositive adults determined that 24% of deaths were directly attributable to late diagnosis of HIV [4]. The UK Collaborative HIV Cohort (UK CHIC) study found that commencing ART at a CD4 cell count of less than 350 cells/mm3 resulted in up to 15 years loss of life [9]. Healthcare costs in the year preceding diagnosis have been found to be 200% higher for those who present late [6]. Furthermore, knowledge of HIV status has been linked to a reduction in risk-taking behaviour and thus it has been suggested that the HIV/AIDS epidemic could be reduced substantially by increasing the number of HIV-positive individuals who are aware of their status [10]. Current literature shows that many ‘late presenters’ were seen in the recent past by healthcare professionals but failed to be diagnosed [11] and a national audit conducted in 2010 by the British HIV Association identified that a quarter of patients had a missed opportunity for diagnosis [12]. 3 In 2014 UNAIDS established a new global 90-90-90 treatment target for HIV; by 2020, 90% of all people living with HIV will know their HIV status, 90% of all people with diagnosed HIV infection will receive sustained ART and 90% of all people receiving ART will be suppressed [13]. Given recent reports that suggest that around a quarter of people living with HIV in the UK are currently unaware of their infection status [2], it is clear that late diagnosis of HIV is an important healthcare issue requiring address in order to meet the 90-90-90 target and improve health outcomes for people living with HIV. Whilst several studies have been conducted worldwide to investigate late or missed diagnosis of HIV, limited studies have been conducted within the UK. An observational study of 1,536 newly diagnosed HIV patients in Brighton found that older adults, aged 50 years or older, were more likely to present late and had a higher mortality. The authors additionally identified African origin and being a male heterosexual as factors associated with late presentation [14]. Within North-East Scotland only one relevant study has been published to date [15]. However, this study focused primarily on missed HIV diagnoses and compared the diagnostic pathways of two earlier cohorts of patients; 1995 to 2000 and 2004 to 2009. A recent search of PubMed did not reveal any study in the North-East of Scotland that explored factors associated with late HIV diagnosis. Thus, this study aims to examine the factors associated with late diagnosis of HIV in Grampian with the aim to provide research evidence and basis for the development of targeted interventions and educational programmes for earlier diagnosis of HIV in North-East Scotland. 4 Methods Study design and setting The study was a population based retrospective database analysis. It involved the analysis of routinely collected HIV surveillance data from the Health Protection Scotland (HPS) HIV database and the review of case notes of all newly diagnosed HIV patients’ during the study period. The HPS HIV database was set up in the early 1980s to collate data on all newly diagnosed HIV antibody positive individuals and AIDS cases to monitor trends in diagnosed HIV infection and AIDS cases among the Scottish population and to provide timely and useful information for the targeting of health promotion, the evaluation of preventive measures, and the planning of medical and social services for those affected by HIV. Information on newly diagnosed HIV/AIDS infection come from reporting of all newly diagnosed HIV infections by virology laboratories in Scotland and AIDS diagnoses by clinicians. In addition, the General Register Office for Scotland (GROS) reports all deaths that record AIDS or HIV among the causes of death to HPS. Records of HIV diagnosis, AIDS and death, which are regarded as relating to the same individual, are merged to create one record in the database [16]. Setting The setting of the study was Grampian, North-East Scotland. In 2015, the population of Grampian was estimated to be 587, 820. Around 10% of the workforce is employed directly by the oil and gas industry. The oil and gas workforce is very mobile with frequent contacts and activities in countries where the prevalence of HIV is high. In addition, Aberdeen has two prestigious universities that offer postgraduate degree programmes in oil and gas engineering and law that are popular among students from countries where the prevalence of HIV is high [17]. 5 HIV care within the National Health Service (NHS) Grampian Health Board is shared between the Genitourinary Medicine (GUM) and Infectious Diseases (ID) departments; inpatient care is provided in Aberdeen Royal Infirmary’s regional ID unit and outpatient care is delivered by both GUM and ID. Definitions The primary outcome of interest was late HIV diagnosis which, is defined by The European Late Presenter Consensus Working Group as; persons diagnosed or presenting for HIV care with a CD4 count below 350 cells/mm3 (within 3 months of diagnosis) or presenting with an AIDS-defining event, regardless of CD4 count [18]. Secondary outcomes were missed opportunities for diagnosis of HIV. This was defined as failure to diagnose HIV within three months of the patient presenting with a clinical indicator disease or AIDS-defining illness. Definitions of clinical indicator disease and AIDS-defining illness were adopted from the British HIV Association (BHIVA) UK National Guidelines for HIV testing [1]. Definitions for ethnicity were adopted from the Scottish Census [19]. Patient postcodes were utilised to ascertain the appropriate 2012 Scottish Index of Multiple Deprivation (SIMD) quintile for each patient [20] SIMD quintile was used as a proxy for socioeconomic status in the study with SIMD 1 representing the most deprived and SIMD 5 representing the least deprived areas. Study population and data collection All HIV positive patients aged 16 years and over, newly diagnosed in Grampian between January 2009 and December 2014, were included in the study. Patients were identified through the HPS HIV database. Patients were excluded if they were aged less than 16 years. One HIV positive individual in the custody of Her Majesty’s Prison Service was excluded on ethical grounds. 6 Ethical approval The study has been carried out in compliance with the Declaration of Helsinki [21]. Ethical approval was granted by the North of Scotland Research Ethics Committee (REC reference 15/NS/0056). Statistical analysis Chi-squared and Mann-Whitney tests were used to identify associations between possible predictor variables and late HIV diagnosis. These included gender, age at diagnosis, SIMD, ethnicity, migrant status, probable mode of transmission, region of exposure, registration with GP, contact with healthcare professionals in the year preceding HIV diagnosis, previous HIV testing, clinical indicator diseases in the five years preceding HIV diagnosis, hepatitis B/C co-infection and missed opportunity for diagnosis. Multivariate logistic regression was then used to ascertain which of the identified variables in combination were predictive for late HIV diagnosis. Variables were entered into a multivariate model if p<0.05 in univariate analysis. There were a small number of missing data in this study but to check the robustness of results we undertook a couple of sensitivity analyses which assumed extremes for some variables and for others we made an informed assumption. For example in sensitivity analysis 1, for missing ethnicity data, one person was assumed to be Black-African using knowledge of their migrant status, and the remainder assumed to be white British. For SIMD, in both sensitivity analyses we assumed that all missing data were SIMD 3. The number of missing data and assumptions made in both analyses are listed in table 1. All statistical analyses were carried out using IBM SPSS Statistics 22.0 [22]. Table 1: Assumptions used for sensitivity analyses 7 Results Study Population Of the 124 patients with a diagnosis of HIV during the study period, 111 (89.5%) had a valid CD4 count recorded within 3 months of diagnosis and were included in the analysis. Of these; 68.5% (n=85) were male, 25.0% (n=31) were aged 16-29 years, 25.0% (n=31) belonged to the least deprived socioeconomic group, 51.6% (n=64) were of White-British ethnicity, 51.6% (n=64) were UK Nationals, 51.6% (n=64) had a heterosexual probable mode of transmission and 42.7% (n=53) were exposed in the UK and Ireland. Characteristics of included patients are presented in Table 2. Table 2: Socio-demographic characteristics of included participant at baseline and comparison of late versus early diagnosis Prevalence and factors associated with late HIV diagnosis Overall, the prevalence of late diagnosis was 53.2% (n=59). In the univariate analysis, late diagnosis of HIV was significantly associated with female gender (p=0.036), Black-African ethnicity (p=0.014), being an immigrant (p=0.014), heterosexual mode of transmission (p=0.002), exposure in sub-Saharan Africa (p=0.015), having undergone no previous HIV testing (p<0.001), having a number of clinical indicator diseases (p=0.034) and co-infection with hepatitis B and/or C (p=0.013) ( Table 2). In multivariate logistic regression, individuals with no previous HIV testing were more likely to be diagnosed late compared to those who had previously been tested (adjusted OR 5.46 (95% CI; 1.89, 15.81)). Moreover, heterosexuals were 2.83 times more likely to be diagnosed late compared to MSM (adjusted OR 2.83 (95% CI; 1.10, 7.32). Unadjusted and adjusted odds ratios and 95% confidence intervals are presented in Table 3. The findings from the sensitivity analyses were consistent and similar to the results of the original univariate and multivariate analyses. 8 Clinical indicator disease Late diagnosis was observed in 59% of individuals in whom clinical indicator disease in the 5 years preceding diagnosis was present (n=61) compared to 45.2% in those with no clinical indicator disease in the 5 years preceding diagnosis (p=0.298). Missed opportunity for HIV diagnosis Missed opportunities for diagnosis were identified in 16.3% (n=15) of included participants. There were more people diagnosed late and who had missed opportunities (73.3%) than those without missed opportunities (50.6%); however this was not statistically significant (p=0.183). Table 3: Factors associated with late diagnosis 9 Discussion Our study showed that over half (53.2%) of the newly diagnosed HIV positive individuals in the 5-year period between January 2009 and December 2014 had a CD4 cell count of less than 350 cells/mm3 at the time of diagnosis or within 3 months of diagnosis. This proportion of late diagnosis appears to be consistent with those found in other UK studies [12, 14-15]. In this study, we found that heterosexual mode of transmission and no previous HIV testing were independently associated with late HIV diagnosis. Several studies [23-28] have shown that heterosexual mode of transmission is associated with late HIV diagnosis. In contrast; only one existing study identified previous HIV testing as an important factor in late HIV diagnosis. This study, carried out in China, found that late HIV diagnosis was more common in individuals with no prior consideration of HIV testing [29]. Interestingly in the univariate analysis, female gender was associated with late diagnosis contrary to existing literature which suggests that men are more likely to be diagnosed late. Only one other existing study from Belgium identified female gender as a demographic feature of late diagnosis in univariate analysis [30].This finding could be explained by the notion that late diagnosis is generally more prevalent in individuals who are not perceived or do not perceive themselves to be at high risk of HIV acquisition [31]. Furthermore, the fear of stigma and discrimination which still remains a significant issue particularly among minority ethnic groups could be a barrier to testing [32]. In addition inadequate partner notification may contribute to late diagnosis in females observed in this study. A service evaluation conducted in NHS Grampian as part of an HIV Needs Assessment found that some individuals reported the person from whom they most likely contracted HIV from had been aware they were HIV positive but had not disclosed this information [33]. Although partner 10 notification has long been recognised as an important HIV prevention intervention there are a number of challenges preventing its effectiveness [34]. Unlike other recent studies, this study found no evidence that injecting drug use was associated with late diagnosis. This could be due to a well established needle exchange programme in Grampian thus resulting in little or no HIV transmission in that risk group. The finding of this study suggests that migrant populations, particularly those originating from subSaharan Africa (SSA), are an important demographic group to consider when addressing late HIV diagnosis. This is consistent with surveillance data which indicates that migrants from HIV endemic countries, particularly individuals from SSA, account for a disproportionate and rising number of HIV infections in Western Europe [35] and majority have acquired their HIV infection in their country of origin [36]. In this study, 16.3% of patients had a missed opportunity for diagnosis and late diagnosis was associated with a higher number of clinical indicator diseases and co-existing hepatitis B/C infection. Current evidence suggests that testing for HIV following diagnosis of an indicator disease significantly decreases the risk of late HIV diagnosis [37]. Similar numerous missed previous opportunities for testing in patients diagnosed late were reported in an audit carried out in Central Scotland [38] and a review published in 2014 [39] reported non adherence to the 2008 UK national HIV testing guidelines recommending a test for all patients with HIV indicator disease. Thus the finding of our study calls for a renewed effort to raise awareness among clinicians regarding HIV indicator disease and late diagnosis. In existing literature, age at diagnosis has repeatedly been identified as a factor associated with late diagnosis; with older individuals more likely to be diagnosed late [14, 23-25, 27-29, 40-41]. Notably in the current study, age at diagnosis was not found to be predictive of late HIV diagnosis either at 11 univariate or multivariate analysis. Proportions of late diagnosis were relatively similar across the age groups with the lowest proportion of late diagnoses (45.5%) observed in the 40-44 year age group and the highest proportion (62.5%) observed in the 35-39 year age group. It is possible that this finding is a result of the sample size and if more patients were included in the analysis greater disparities between the age groups may be detected. To address some the findings in this study, a number of interventions have already been implemented in Grampian including a late diagnosis multidisciplinary subgroup [HIV clinical team, Public Health, Primary care] to review all new cases of late diagnoses to identify any missed opportunity and feedback to the service where the missed opportunity occurred, raising awareness of HIV clinical indicators in primary care through educational programmes, opportunistic testing in a university setting during ‘fresher’s week’ and offering of HIV test during new patient registration in primary care. This study has some limitations. Firstly, it was a retrospective study with an aspect of case note review and interrogation of several patient electronic management systems with incomplete and missing data which could affect the result. Secondly, the sample size was small which limits comparability. Conclusion In conclusion, a good proportion of HIV positive patients continue to be diagnosed late. Heterosexual route of transmission and having no previous HIV test were independently associated with late HIV diagnosis. There is need to highlight the importance of early diagnosis among the public and healthcare workers. Offering of testing in non-NHS (community-based) settings such as in universities and faith-based organisations settings is recommended to increase uptake of HIV testing in hard to reach groups and in those who have never had a test. Acknowledgements 12 The authors would like to thank Dr’s Steve Baguley, Ambreen Butt, and Daniela Brawley for their help with access to patient notes in GUM. Funding None Author Contributions This paper is an abridged version of a master’s thesis presented for the degree of Master of Science (MSc) in Global Health and Management, University of Aberdeen by GN. EO, IT, SF supervised the thesis. EO and IT developed the concept for the study. GN and SF conducted the statistical analysis. All authors contributed significantly at every stage of the paper, read and approved the final manuscript. Conflict of interest The authors have no competing interest to declare 13 References 1. British HIV Association (BHIVA). UK National Guidelines for HIV Testing 2008 [Internet]. 2008 [cited 2016 May 14]. Available from: http://www.bhiva.org/documents/Guidelines/Testing/GlinesHIVTest08.pdf. 2. Public Health England (PHE). HIV in the United Kingdom: 2014 Report [Internet].2014 [cited 2016 May 14]. Available from: https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/401662/2 014_PHE_HIV_annual_report_draft_Final_07-01-2015.pdf. 3. Sabin CA, Smith CJ, Gumley H, Murphy G, Lampe FC, Phillips AN, et al. Late presenters in the era of highly active antiretroviral therapy: uptake of and responses to antiretroviral therapy. AIDS. 2004;18(16):2145-51. 4. Lucas SB, Curtis H and Johnson MA. National review of deaths among HIV-infected adults. Clin Med. 2008; 8(3):250-2. 5. Stӧhr W, Dunn DT, Porter K, Hill T, Gazzard B, Walsh J, et al. CD4+ count and initiation of antiretroviral therapy: trends in seven UK centres, 1997-2003. HIV Med. 2007; 8:135-141. 6. Krentz HB, Auld MC and Gill MJ. The high cost of medical care for patients who present late (CD4+<200 cells/µL) with HIV infection. HIV Med. 2004; 5:93-8. 7. Fleishman JA, Yehia BR, Moore RD, Gebo KA and HIV Research Network. The economic burden of late entry into medical care for patients with HIV infection. Med Care. 2010; 48(12):10719. 8. Shrosbree J, Campbell LJ, Ibrahim F, Hopkins P, Vizcaychipi M, Strachan S, et al. Late HIV diagnosis is a major risk factor for intensive care unit admission in HIV-positive patients: a single centre observational cohort study. BMC Infect Dis. 2013; 13:23. 14 9. May M, Gompels M, Delpech V, Porter K, Post F, Johnson M, et al. Impact of late diagnosis and treatment on life expectancy in people with HIV-1: UK Collaborative HIV Cohort (UK CHIC) Study. BMJ. 2011; 343:d6016. 10. Marks G, Crepaz N and Janssen RS. Estimating sexual transmission of HIV from persons aware and unaware that they are infected with the virus in the USA. AIDS. 2006; 20:1447-50. 11. Sullivan AK, Curtis H, Sabin CA and Johnson MA. Newly diagnosed HIV infections: review in UK and Ireland. BMJ. 2005; 330:1301-2. 12. Ellis S, Curtis H and Ong ELC. HIV diagnoses and missed opportunities. Results of the British HIV Association (BHIVA) National Audit 2010. Clin Med. 2012; 12(5):430-4. 13. UNAIDS. 90-90-90: An ambitious treatment target to help end the AIDS epidemic [Internet]. 2014 [cited 2015 Jun 12]. Available from: http://www.unaids.org/sites/default/files/media_asset/90-90-90_en_0.pdf. 14. Iwuji CC, Churchill D, Gilleece Y, Weiss HA and Fisher M. Older HIV-infected individuals present later and have a higher mortality: Brighton, UK cohort study. BMC Public Health. 2013; 23:397. 15. Wohlgemut J, Lawes T and Laing RBS. Trends in missed presentations and late HIV diagnosis in a UK teaching hospital: a retrospective comparative cohort study. BMC Infect Dis. 2012;12:72 16. Health Protection Scotland BBV/STI Surveillance System Details [Cited 2016 April 12] Available from: http://www.hps.scot.nhs.uk/bbvsti/ssdetail.aspx?id=27 17. Mackenzie AR, Laing RB, Urbaniak SJ, Molyneaux PJ, Douglas JG and Smith CC. Epidemiology and outcome of HIV infection in North-East Scotland (1985-1997). J Infect. 1999; 38(2):10710. 18. Antinori A, Coenen T, Costagiola D, Dedes N, Ellefson M, Gatell J, et al. Late presentation of HIV: a consensus definition. HIV Med. 2011; 12:61-4. 15 19. The Scottish Government. Summary: Ethnic Group Demographics [Internet]. 2014 [cited 2015 May 15]. Available from: http://www.gov.scot/Topics/People/Equality/Equalities/DataGrid/Ethnicity/EthPopMig. 20. The Scottish Government. Scottish Index of Multiple Deprivation [Internet]. 2012 [cited 2015 May 15]. Available from: http://www.sns.gov.uk/Simd/Simd.aspx. 21. World Health Organization (WHO). World Medical Association Declaration of Helsinki: Ethical Principles for Medical Research Involving Human Subjects. Bull World Health Organ. 2001; 79 (4):373-4. 22. IBM Corp. Released 2013. IBM SPSS Statistics for Windows, Version 22.0. Armonk, NY: IBM Corp.) 23. Ndiaye B, Salleron J, Vincent A, Bataille P, Bonnevie F, Choisy P, et al. Factors associated with presentation to care with advanced HIV disease in Brussels and Northern France: 1997-2007. BMC Infect Dis. 2011; 11:11. 24. Camoni L, Raimondo M, Regine V, Salfa MC and Suligoi B. Late presenters among persons with a new HIV diagnosis in Italy, 2010-2011. BMC Public Health. 2013;13:281 25. Metallidis S, Pilalas D, Skoura L, Haidich A, Tsachouridou O, Papaioannou M, et al. Time trends and correlates of late presentation for HIV care in Northern Greece during the decade 2000 to 2010. J Int AIDS Soc. 2012; 15:17395. 26. Arbune M. Characteristics of HIV late presenters in south-eastern Romania. Acta Medica Mediterr. 2013; 29:71-5. 27. Mojumdar K, Vajpayee M, Chauhan NK and Mendiratta S. Late presenters to HIV care and treatment, identification of associated risk factors in HIV-1 infected Indian population. BMC Public Health. 2010; 10:416. 28. Lee JH, Kim GJ, Choi BS, Hong KJ, Heo MK, Kim SS, et al. Increasing late diagnosis in HIV infection in South Korea: 2000-2007. BMC Public Health. 2010; 10:411. 16 29. Dai SY, Liu JJ, Fan YG, Shan GS, Zhang HB, Li MQ, et al. Prevalence and factors associated with late HIV diagnosis. J Med Virol. 2015; 87:970-7. 30. Yombi JC, Jonckheere S, Vincent A, Wilmes D, Vandercam B and Belkhir L. Late presentation for human immunodeficiency virus (HIV) diagnosis: results of a Belgian single centre. Acta Clin Belg. 2014; 69(1):33-9. 31. Girardi E, Sabin CA and d’Arminio Monforte A. Late diagnosis of HIV infection: epidemiological features, consequences and strategies to encourage earlier testing. J Acquir Immune Defic Syndr. 2007; 46(1): suppl 1. 32. A. Adler , S. Mounier-Jack & R.J. Coker (2009) Late diagnosis of HIV in Europe: definitional and public health challenges, AIDS Care, 21:3, 284-293 33. HIV Working Group. Grampian Human Immunodeficiency (HIV) Needs Assessment 2013. NHS Grampian; 2013 (Unpublished). 34. National AIDS Trust (NAT) Report, 2012. HIV Partner notification: a missed opportunity? [Cited 2016 May 16]. Available from: http://www.nat.org.uk/media/Files/Publications/May-2012HIV-Partner-Notification.pdf 35. Hamers FF. The changing face of the HIV epidemic in Western Europe: what are the implications for public health policies? Lancet. 2004; 364(9428):83-94. 36. Alvarez-del Arco D, Monge S, Azcoaga A, Azcoaga A, Rio I, Hernando V, et al. HIV testing and counselling for migrant populations living in high-income countries: a systematic review. Eur J Public Health. 2012; 23(6):1039-45. 37. Scognamiglio P, Chiaradia G, De Carli G, Giuliani M, Mastroianni CM, Barbacci SA, et al. The potential impact of routine testing of individuals with HIV indicator diseases in order to prevent late HIV diagnosis. BMC Infect Dis. 2013; 13:473. 38. Acquah RR, Baggott A, McGoldrick C, Kennedy N, HIV testing in Lanarkshire 2014 J R Coll Physicians Edinb 2014; 44: 278–82 17 39. Elmahdi R, Gerver SM, Gomez GG, Fidler S, Cooke G, Ward H. Low levels of HIV test coverage in clinical settings in the UK: a systematic review of adherence to 2008 guidelines. Sex Transm Infect 2014. 40. Dickson NP, McAllistar S, Sharples K and Paul C. Late presentation of HIV infection among adults in New Zealand: 2005-2010. HIV Med. 2011; 13(3):182-9. 41. Mukolo A, Villegas R, Aliyu M and Wallston KA. Predictors of late presentation for HIV diagnosis: a literature review and suggested way forward. AIDS Behav. 2013; 17:5-30. 18 Table 1: Sensitivity analysis assumptions Variable Missing data Gender Age at diagnosis SIMD Sensitivity analysis 1 Sensitivity analysis 2 No missing data - - No missing data - - 4 Missing data assumed to be SIMD 3 Missing data assumed to be SIMD 3 3 n=1 assumed to be BlackAfrican based on migrant status.* Remaining missing data assumed as WhiteBritish n=1 assumed to be BlackAfrican based on migrant status.* Remaining missing data also assumed as BlackAfrican 3 n=1 assumed to be a migrant based on ethnicity.** Remaining missing data assumed as UK National n=1 assumed to be a migrant based on ethnicity.** Remaining missing data assumed as migrant 4 Missing data assumed to be heterosexual transmission n=1 assumed to be heterosexual as female. Remaining missing data assumed to be MSM n=1 assumed to be exposed in SSA based on ethnicity.*** Remaining missing data assumed as UK and Ireland n=2 assumed to be exposed in the UK and Ireland and n=1 assumed to be exposed in SSA based on ethnicity# Remaining missing data assumed as SSA Missing data assumed as not registered with GP Missing data assumed as registered with GP Missing data assumed as no contact with healthcare professionals Missing data assumed as contact with healthcare professionals Missing data assumed to be no previous HIV testing Missing data assumed as previous HIV testing Ethnicity Migrant status Probable mode of transmission Region of exposure 5 Registration with GP Contact with healthcare professionals in the year preceding diagnosis Previous HIV testing 5 27 8 19 Clinical indicator disease (present or absent) in the five years preceding diagnosis Co-existing hepatitis B/C infection Missed opportunity for diagnosis 19 Missing data assumed as no clinical indicator disease Missing data assumed as clinical indicator disease present 8 Missing data assumed as no hepatitis B/C coinfection Missing data assumed as hepatitis B/C co-infection 19 Missing data assumed to be no missed opportunity for diagnosis Missing data assumed to be a missed opportunity for diagnosis *n=1 assumed to be Black-African as cross tabulation of migrant status against ethnicity revealed that the majority of migrants (69.8%) were of Black-African ethnicity. **n=1 assumed to be a migrant based on results of ethnicity as cross tabulation of ethnicity against migrant status showed that all (100.0%) patients of Black-African ethnicity were migrants. ***n=1 assumed to be exposed in SSA as cross tabulation of region of exposure and ethnicity revealed the majority of Black-African patients (83.3%) were exposed in SSA. # n=2 assumed to be exposed in the UK and Ireland as cross tabulation of region of exposure and ethnicity demonstrated that the majority of White-British patients (83.9%) were exposed in the UK and Ireland. 20 Table 2: Socio-demographic characteristics of included participant female Male N (%) 111 (100) 26 (23.4) 85 (76.6) Early Diagnosis N (%) 52 (46.8) 7 (26.9) 45 (52 ) Late Diagnosis N (%) 59 (53.2) 19 (73.1) 40 (47.1) 16-29 30-34 35-39 40-44 45-49 31 (27.9) 22 (19.8) 16 (14.4) 11 (9.9) 14 (12.6) 16 (51.6) 9 (40.9) 6 (37.5) 6 (54.5) 7 (50.0) 15 (48.4) 13 (59.1) 10 (62.5) 5 (45.5) 7 (50.0) ≥50 17 (15.3) 8 (47.1) 9 (52.9) Variable CD4 Cell count Gender (n=111) Age (years) at Diagnosis (n=111) SIMD Quintile (n=97) Ethnicity (n=108) Migrant status (n=108) Probable mode of transmission (n=107) Probable region of exposure (n=106) Registration with GP (n=106) 1 (Most deprived) 2 3 4 5 (Least deprived) Black-African White British White Other UK National P-value p= 0.036* p= 0.984** p= 0.263** 15 (15.5) 16 (16.5) 18 (18.6) 17 (17.5) 7 (46.7) 4 (25.0) 8 (44.4) 9 (52.9) 8 (53.3) 12 (75.0) 10 (55.6) 8 (47.1) 31 (32.0) 31 (28.7) 64 (59.3) 13 (12.0) 64 (59.3) 16 (51.6) 8 (25.8) 37 (57.8) 6 (46.2) 37 (57.8) 15 (48.4) 23 (74.2) 27 (42.2) 7 (53.8) 27 (42.2) Migrant 44 (40.7) 14 (31.8) 30 (68.2) Heterosexual MSM 64 (59.8) 42 (39.3) 23 (35.9) 29 (69.0) 41 (64.1) 13 (31.0) IDU 1 (0.9) Excluded Excluded UK and Ireland Sub-Saharan Africa 53 (50.0) 32 (60.4) 21 (39.6) 26 (24.5) 7 (26.9) 19 (73.1) All other regions 27 (25.5) 11 (40.7) 16 (59.3) p = 0.014*** p = 0.014* p= 0.002* p= 0.015*** p= 0.335# Registered Unregistered 96 (90.6) 10 (9.4) 46 (47.9) 3 (30.0) 50 (52.1) 7 (70.0) p= 1.000* Contact with healthcare professional in the year preceding diagnosis (n=84) No contact 26 (31.0) 11 (42.3) 15 (57.7) Contact 58 (69.0) 25 (43.1) 33 (56.9) 21 Number of contact with healthcare professional in the year preceding diagnosis p= 0.508## Median (IQR) 84 1 (0-1) 1 (0-2) p= 0.298* Clinical indicator disease in the 5 years preceding diagnosis (n=92) Absent 31 (33.7) 17 (54.8) 14 (45.2) Present 61 (66.3) 25 (41.0) 36 (59.0) P= 0.034## Number of Clinical indicator disease in the 5 years preceding diagnosis Previous HIV testing (n=103) Co-existing Hepatitis B or C infection (n=103) Median (IQR) No previous HIV test Previous HIV test 92 1 (0-1) 1 (0-1.25) p<0.001* 74 (71.8) 24 (32.4) 50 (67.6) 29 (28.2) 23 (79.3) 6 (20.7) Co-infection 21 (20.4) 4 (19.0) 17 (81.0) No co-infection 82 (79.6) 43 (52.4) 39 (47.6) p=0.013* p=0.183* Missed opportunity for diagnosis (n=92) Missed opportunity 15 (16.3) 4 (26.7) 11 (73.3) No missed opportunity 77 (83.7) 38 (49.4) 39 (50.6) *Continuity Correction, **Linear by Linear, ***Pearson Chi Square, #Fishers Exact test, ##Mann-Whitney 22 Table 3: Factors associated with Late diagnosis of HIV Unadjusted Factors OR 95% CI OR Gender (n=111) Female 1.00 Male 3.05 1.16, 8.02 Ethnicity (n=108) White-British 1.00 Black-African 3.94 1.53, 10.14 White-Other 1.60 0.48, 5.30 Migrant status (n=108) UK National Migrant 1.00 2.94 1.31, 6.57 Probable mode of transmission (n=106) MSM Heterosexual 1.00 3.98 1.73, 9.12 Region of exposure (n=106) UK and Ireland Sub-Saharan Africa All other regions 1.00 4.14 1.48, 11.55 2.22 0.86, 5.70 Previous HIV testing (n=103) Testing No testing 1.00 7.99 2.87, 22.19 Number of clinical indicator diseases in five years preceding diagnosis (n=92) 1.91 1.11, 3.30 Co-existing Hepatitis B/C infection (n=103) Co-infection-NO Co-infection -YES 1.00 4.69 1.45, 15.13 Adjusted 95% CI p 1.00 2.83 1.10, 7.32 p=0.031 1.00 5.46 1.89, 15.81 p=0.002 23