Secondary Logo

Share this article on:

No time to wait: how many HIV-infected homosexual men are diagnosed late and consequently die? (England and Wales, 1993–2002)

Chadborn, Timothy Ra; Baster, Kathleena; Delpech, Valerie Ca; Sabin, Caroline Ab; Sinka, Katya; Rice, Brian Da; Evans, Barry Ga

doi: 10.1097/01.aids.0000162340.43310.08

Objectives: To present national trends of the estimated number and proportion of late HIV diagnoses and short-term mortality following diagnosis among men who have had sex with men (MSM). To determine separately risk factors for late diagnosis and short-term mortality.

Methods: Analysis of national HIV/AIDS case reports of new diagnoses linked to CD4 cell counts from the CD4 Surveillance Scheme. Inverse probability weighting adjusted for individuals with no CD4 cell count at diagnosis. Outcomes were late diagnosis (CD4 cell count <200 × 106 cells/l at diagnosis) and short-term mortality (death within 1 year of diagnosis).

Results: Of 14 158 new diagnoses, 31% were estimated as late diagnoses. Despite a decreasing trend (P trend <0.01) an estimated 430 (25%) MSM were still diagnosed late in 2001. Late diagnosis disproportionately affected individuals diagnosed outside London, of non-white ethnicity, and of older age. There were 710 (5.0% of 14 158) deaths within a year of HIV diagnosis. Estimated short-term mortality was 14% for MSM diagnosed late and 1% for other MSM (adjusted odds ratio, 10.8; 95% confidence interval, 7.7–15.9). Short-term mortality declined concurrently with availability of highly active antiretroviral therapy and was independently associated with age and diagnosis outside London but not ethnicity.

Conclusions: The continued late diagnosis of one in four MSM means these individuals lose the option to start therapy early, miss opportunities to prevent further transmission and are approximately 10 times more likely to die within a year of diagnosis. Early diagnosis of all MSM in 2001 could have reduced short-term mortality by 84% and all mortality in that year by 22%.

From the aHIV/STI Department, Communicable Disease Surveillance Centre, Health Protection Agency's, Centre for Infections London

bDepartment of Primary Care and Population Sciences, Royal Free and University College Medical School, London, UK.

Received 16 September, 2004

Revised 26 November, 2004

Accepted 16 December, 2004

Correspondence to Tim Chadborn, HIV/STI Department, Communicable Disease Surveillance Centre, Health Protection Agency's, Centre for Infections 61 Colindale Avenue, London NW9 5EQ, UK. E-mail:

Back to Top | Article Outline


The recent annual report ‘Health Check: on the State of the Public Health 2003’ of the Chief Medical Officer (CMO) for England highlights the need to diagnose HIV infection early in the course of infection in the section ‘No time to wait: The importance of early diagnosis of HIV’ [1]. Diagnosis of HIV late in the course of infection results in late initiation of highly-active antiretroviral therapy (HAART), which has been shown to decrease morbidity and mortality [2–4]. It also increases costs [5] and reduces the probability of avoiding further transmission through clinical and behavioural preventive measures [6,7] The CMO's report therefore calls for healthcare settings to offer and recommend annual HIV testing to all men who have had sex with men (MSM) and more targeted health promotion campaigns to highlight the advantages of HIV testing. However, little is known about the trends in, and determinants of, late diagnosis among MSM or of its impact on mortality.

There was a 14% increase in the estimated prevalence of HIV infection among MSM in the United Kingdom (UK) during 2002 [8]. Between 1277 and 1660 new HIV diagnoses among MSM were reported every year between 1993 and 2002 but a substantial number and proportion of HIV-infected MSM in the UK remain undiagnosed [8,9]. Prior to diagnosis, these individuals cannot receive appropriate care, inform their sexual partners or be guided in safer sexual behaviour in knowledge of their status [10]. MSM that remain unaware of their HIV status for a long time and are diagnosed late will also lose the option to start therapy early. These factors reflect lost opportunities to reduce HIV transmission. Late-diagnosed individuals will also be at higher risk of death between diagnosis and initiation of HAART. Most of these pre-treatment deaths and deaths during the first year of HAART could presumably be prevented if diagnosis was made earlier in the course of infection.

Studies have shown that there is a low short-term risk of progression to AIDS for untreated individuals with CD4 counts above 200 × 106 cells/l although viral load and age independently increase that risk [11]. Patients with CD4 counts below 200 × 106 cells/l when starting HAART have a substantially higher risk of progression to AIDS and death [2,12]. Therefore, to prevent AIDS morbidity and mortality, current treatment guidelines recommend that all individuals initiate HAART before CD4 cell counts fall below 200 × 106 cells/l [13,14]. The ART Cohort Collaboration, which includes 13 cohort studies from Europe and North America, showed that after starting HAART the rate of progression to AIDS or death fell markedly, with most of the reduction in the first 6 months of therapy [2]. It is important, therefore, that individuals have the optimal chance of surviving the first year of HAART. Survival through the first year of treatment will be maximized by diagnosis of individuals before CD4 cell counts fall to less than 200 × 106 cells/l.

We estimate the number and proportion of MSM who were diagnosed late each year in England and Wales (E &W) and present national trends of late diagnosis and short-term mortality in MSM between 1993 and 2002. We compare the estimated short-term mortality of individuals diagnosed late with that of those individuals who were diagnosed earlier. In addition, we determine separately risk factors for late diagnosis and short-term mortality.

Back to Top | Article Outline


Data sources

Newly diagnosed AIDS cases have been confidentially reported to the Communicable Disease Surveillance Centre, London, since 1982 to provide national surveillance information on the HIV epidemic [15]. Since 1985, virology laboratories receiving specimens for HIV testing have voluntarily completed reports about individuals newly identified as positive. Since January 2000, clinicians diagnosing individuals with HIV for the first time have also voluntarily submitted reports [16]. Data are collected on the date of HIV diagnosis, patient ethnicity, probable route of infection and the site where the specimen was collected. No names are reported but soundex code of surname [17], date of birth and sex are used to link multiple reports from the same individual and limit duplication within the surveillance system. Reports of newly diagnosed HIV infections are matched to AIDS reports and to records of death certificates provided by the Office of National Statistics.

The CD4 Surveillance Scheme augments the epidemiological information collected through the reporting of HIV/AIDS cases but collects little epidemiological information. Laboratories in E &W that conduct CD4 cell counts and participate in the UK National External Quality Assessment Scheme for Leucocyte Immunophenotyping were invited to contribute data with the aim of monitoring trends in the prevalence of various levels of immunosuppression in HIV-infected adults [18]. Variables collected on each CD4 cell count include: soundex code, date of birth, initial, sex, patient number (from the clinic, hospital and/or laboratory), date CD4 cell count performed and absolute CD4 cell count.

Back to Top | Article Outline

Data preparation

The study population comprised HIV-infected MSM (aged 15 years and over) newly diagnosed in E &W from January 1993 to December 2002 (reported by the end of March 2003). Patient identifiers were used to match the epidemiological information on these case reports to CD4 cell counts of individuals recorded on the CD4 database (by the end of March 2003) and to individuals reported to the 1997 – 2002 annual Surveys of Prevalent HIV Infections Diagnosed (SOPHID) [19]. Almost 4,000 patient ethnicities from the SOPHID survey were used in the analysis where none was recorded in the HIV and AIDS patient database.

The CD4 cell count at diagnosis was defined as the count closest to the date of HIV diagnosis and within 60 days prior to or following the date of HIV diagnosis. This cut-off was chosen to maximize the number of patients included in the analysis but to minimize the possibility that HAART had affected CD4 cell counts at diagnosis [2,20]. For MSM with a CD4 cell count at diagnosis, late diagnosis was defined as a CD4 cell count below 200 × 106 cells/l. Short-term mortality was defined as the proportion of individuals newly diagnosed who were known to have died from any cause within 1 year of the date of HIV diagnosis. Deaths that were reported by the end of September 2003 were included. Data for 2002 was excluded from the regression analysis of short-term mortality because more reports are expected of deaths of individuals who were diagnosed in 2002. The pre-HAART era was considered to be 1993 to 1995 inclusive whereas the era of HAART availability was defined as 1996 to 2002 inclusive.

Back to Top | Article Outline

Statistical analysis

Stata 8.0 software (Stata Corporation, College Station, Texas, USA) was used for analysis of data. Two-sample t tests and χ2 tests were used for comparison of proportions. Inverse probability weighting (IPW) [21] was used to adjust for those without CD4 cell counts (weights were determined by the inverse of the probability of having a CD4 cell count within 60 days of diagnosis, calculated from a multivariable logistic regression analysis). This enabled estimation of the total number and proportion of MSM who were diagnosed late each year. Odds ratios were calculated using logistic regression analyses. There is no simple formula to account for the unknown variability introduced by IPW and therefore bootstrapping (sampling 1000 times with replacement) was used to calculate percentile confidence intervals [22].

Back to Top | Article Outline


Characteristics of MSM with CD4 cell counts at diagnosis

Between January 1993 and December 2002 (reports received by March 2003), 14 158 MSM were newly diagnosed with HIV in E &W with 79% (11 134) having matching records in the CD4 Surveillance Scheme database. Of the latter, 76% (8478) had a CD4 cell count at HIV diagnosis (Table 1). The characteristics of newly diagnosed MSM are shown in Table 1. Of the 1458 non-white individuals, 45% were other/mixed ethnicity, 26% black Caribbean, 12% black African, 10% black other and 8% Indian/Pakistani/Bangladeshi.

Table 1

Table 1

In univariable analysis (Table 1, percentage of MSM with counts at diagnosis), MSM were more likely to have a CD4 cell count at HIV diagnosis if they were diagnosed in London (66 versus 49%; P <0.01), were diagnosed in the era of HAART availability (64 versus 51%; P <0.01), were towards the middle of the age distribution (15–24 years, 56%; 30–34 years, 63%; 35–39 years, 64%; 45+ years, 55%; P <0.01), had not died within a year of diagnosis (61 versus 40%; P <0.01), and if they had a known ethnicity (white, 61%; non-white, 62% versus ethnicity not known, 50%; P <0.01).

Back to Top | Article Outline

Estimated trends in, and determinants of, late diagnosis

Overall, 31% (4425 of 14 158) of MSM were estimated to have been diagnosed late (Table 2). The median CD4 cell count of those diagnosed late was 61 × 106 cells/l [interquartile range (IQR), 28–120] compared with 450 × 106 cells/l (IQR, 330–610) for those not diagnosed late.

Table 2

Table 2

The estimated proportion of MSM diagnosed late declined from 38% (498 of 1428) in 1993 to 25% (347 of 1500) in 2002 (P trend <0.01) (number and proportions shown in Table 2, univariable odds ratios shown in Table 3). The estimated proportion of individuals diagnosed late increased with age from 14% of 15–24 year olds to 54% of those aged 45 and over (P trend for age < 0.01). Individuals diagnosed outside London were more likely to be diagnosed late than individuals diagnosed in London (38 versus 27%; P < 0.01). Non-white individuals were more likely to be diagnosed late than white individuals (35 versus 31%; P < 0.01).

Table 3

Table 3

Factors significantly associated with increased odds of late diagnosis in the multivariable logistic regression model were: year of diagnosis [odds ratio (OR), 0.91; 95% confidence interval (CI), 0.89–0.94], age at diagnosis (per additional decade) (OR, 1.86; 95% CI, 1.76–1.97), diagnosis outside London (OR, 1.68; 95% CI, 1.51–1.86), and ethnicity, with non-white individuals having significantly higher odds of being diagnosed late than white individuals (OR, 1.72; 95% CI, 1.45–2.00) (Table 3). We tested for two-way interactions but none were statistically significant.

Back to Top | Article Outline

Estimated trends in, and determinants of, short-term mortality

Of the 14 158 MSM diagnosed between 1993 and 2002, 5.0% (710) were known to have died within 1 year of HIV diagnosis. Less than 5% of these deaths were suicides, drug overdoses or accidents. The median CD4 cell count at diagnosis of MSM who died within a year was 40 × 106 cells/l (IQR, 11–117) compared with 350 × 106 cells/l (IQR, 170–540) of those who did not die within a year.

The estimated short-term mortality for MSM diagnosed late was 14.1% (622 of 4425) between 1993 and 2002 but only 1.1% (108 of 9733) for those not diagnosed late (P <0.01). In 2001 the short-term mortality was 9.9 and 0.5%, respectively. Overall, estimated short-term mortality declined from 9.8% in the pre-HAART era to 3.2% in the era of HAART availability (P <0.01), increased with age (P trend <0.01), was higher for those diagnosed outside London than in London (7.9 versus 3.6%; P <0.01), and lower for non-white MSM than for white MSM (4.0 versus 5.1%; P <0.01) (Table 2).

Factors significantly associated with the odds of short-term mortality (1993 to 2001) in the multivariable logistic regression model were: late diagnosis (OR, 10.76; 95% CI, 7.68–15.91), era of HAART availability (OR, 0.44; 95% CI, 0.26–0.75), age at diagnosis (per additional decade) (OR, 1.66; 95% CI, 1.46–1.90), diagnosis outside London (OR, 1.66; 95% CI, 1.26–2.18), but not ethnicity or year of diagnosis (Table 4). No two-way interactions were statistically significant.

Table 4

Table 4

When the absolute CD4 cell count was used in the multivariable logistic regression model rather than the binary variable, late diagnosis, the odds ratios for the effect of the other covariables did not change substantially.

Back to Top | Article Outline

Sensitivity analysis

The point estimates were recalculated using a 30-day and 90-day cut-off between the date of HIV diagnosis and the closest CD4 cell count. Of the 14 158 MSM newly diagnosed with HIV and 11 134 with matching records in the CD4 Surveillance Scheme database, 7629 (69%) had a CD4 cell count within 30 days and 8826 (79%) had a CD4 cell count within 90 days compared with 8478 (76%) within 60 days. Neither of these datasets produced a substantial change in the univariable or multivariable regression results.

Injection drug use, reported by 2% of MSM, was considered in the analyses but not found to be a significant risk factor.

Back to Top | Article Outline


This analysis estimates the number and proportion of MSM diagnosed late and subsequent short-term mortality. It also identifies separately the determinants of late diagnosis and short-term mortality. Most importantly, the results show that although HAART decreased the short-term mortality for all MSM, irrespective of CD4 cell counts at diagnosis, those diagnosed late were still approximately 10 times more likely to die within a year of diagnosis than comparable individuals who were not diagnosed late.

The estimated short-term mortality in 2001 was 0.5% (6) of the 1230 newly diagnosed individuals who were not diagnosed late and 10% (43) of the 430 newly diagnosed individuals who were diagnosed late. If no individuals were diagnosed late and short-term mortality was 0.5% for all individuals, only eight deaths would have been expected in 2001, an 84% (41 of 49) reduction in short-term mortality.

Estimated short-term mortality accounted for 13.7% (730 of 5338) of all deaths in MSM occurring between 1993 and 2002. Although short-term mortality declined when HAART became available, it has accounted for an increasing proportion of all deaths since 1996: 8.2% (70 of 850) in 1996 to 27.5% (49 of 178) in 2001. This analysis suggests that an 84% reduction in short-term mortality resulting from early diagnosis of all individuals could have given a 23% (41 of 178) reduction in the number of all deaths among MSM in 2001.

There was a significant decline in the odds of MSM being diagnosed late between 1993 and 2002. This means that each year a higher proportion of those diagnosed with HIV could be offered therapy early and therefore expect a longer and healthier life [2,4,23,24]. However, an estimated one in four MSM were still diagnosed late in 2001, the estimated number diagnosed late increased from 398 in 1999 to 430 in 2001 and substantial effort is still required to counter this situation. In addition to lives lost, there are financial implications of late diagnosis because direct care costs in the year following diagnosis are likely to be more than 200% higher for patients diagnosed late [25].

The ART Cohort Collaboration showed that the probability of death in the first year of treatment was lower for individuals who started HAART with a CD4 cell count greater than 199 × 106 cells/l compared with those who started HAART late [2]. The higher mortality in the era of HAART availability in our study (20% of MSM aged 45 or over who were diagnosed late outside London died within 1 year) may be due to a higher number and proportion of deaths occurring in those diagnosed late between diagnosis and treatment initiation. We aimed to investigate the consequences of late diagnosis rather than the consequences of late initiation of HAART and suggest that there are many pre-treatment deaths that could be prevented.

Although late diagnosis is less common among MSM than other HIV-infected individuals [26] our study supports findings of a continued high proportion of late diagnoses among MSM [18,27]. This is likely to be due to men not presenting for or declining HIV testing because it is likely that most actively homosexual men in the UK are aware of the risk of HIV infection [28]. As most MSM newly diagnosed in the UK were infected in the UK and have been exposed to years of health promotion, this should strengthen calls for targeted messages to highlight the benefits of early testing and for HIV testing policy to target this issue. Our analysis focuses on late diagnosis among MSM and shows that late diagnosis is a strong determinant of short-term mortality. In addition, we show that the proportion diagnosed late and the short-term mortality are higher outside London. A positive note is that there was a trend towards fewer late diagnoses nationally. Interestingly however, there is no evidence from this analysis that the availability of combination therapy in 1996 resulted in a marked increase in earlier HIV testing.

Earlier diagnosis of MSM since 1993 may be due in part to increased awareness of the benefits of early testing due to HIV testing promotion efforts [29,30]. However, increasing incidence in MSM [31], increasing age at infection, and a decreasing contribution to new diagnoses of long-term undiagnosed MSM who were infected during the peaks in incidence of HIV in 1983 and in 1989/1990 [32] could also account for part of this trend. Yet, encouragingly, MSM coming forward earlier for testing may have contributed to the recent increases in the numbers of new HIV diagnoses [33].

This analysis also demonstrates the use of inverse probability weighting as a method for dealing with missing data where the characteristics of the total population are known [21]. Our estimates were in close agreement with observed short-term mortality. We suggest that this methodology could be particularly useful for analysing surveillance data, which may be incomplete due to limitations of matching datasets.

There are limitations to our analysis. Inverse probability weighting has been used to account for missing observations, as CD4 cell counts were not known for all individuals at the time of their HIV diagnosis. However, a key assumption made is that the immunological status of individuals with known CD4 cell counts is representative of comparable individuals without known CD4 cell counts. This is supported by the close similarity of CD4 cell counts reported by clinicians when reporting a new HIV diagnosis that were not available from the CD4 Surveillance Scheme (data not shown). In addition, 60% of all individuals diagnosed in England and Wales between 1993 and 2002 are represented in this analysis and bias should be minimized because deaths and CD4 cell counts are independently collected and matched to reports of new HIV diagnoses.

An underestimation of late diagnosis and its impact on short-term mortality may have arisen if HAART had significantly increased CD4 cell counts between diagnosis and the first CD4 count. However, 27% of MSM involved in this analysis had CD4 cell counts within a day of the date of HIV diagnosis, 70% had counts within 14 days and 90% within 1 month. HAART is unlikely to be started before blood is taken for determination of the CD4 cell count and the delay between diagnosis and CD4 cell count is likely to be accounted for by confirmation of infection.

The lack of information about viral loads at HIV diagnosis may limit the determination of odds ratios of short-term mortality because viral load has been shown to be predictive of progression of infection independently of CD4 cell counts and age for treatment-naive individuals [11,34]. However, in patients starting HAART, CD4 cell counts are the dominant prognostic factor and HIV-RNA levels have relatively little effect on subsequent disease progression [2,12]. In addition, the effect of adherence could not be addressed as surveillance does not collect this information.

A high number and proportion of MSM are still diagnosed late despite the availability of HAART. Late diagnosis disproportionately affects MSM that are non-white, older and diagnosed outside London. This analysis shows that MSM diagnosed late are much more likely to die within 1 year of diagnosis than MSM not diagnosed late and suggests that many of these deaths are preventable. The Department of Health's National Strategy for Sexual Health and HIV for England [35]) in 2001 and the CMO's annual report in 2004, ‘On the State of the Public Health’ [1] prioritized the uptake of HIV testing as a core HIV prevention intervention and it will be important to monitor whether this leads to a marked decline in the proportion of MSM diagnosed late. There is a clear need to diagnose HIV-infected MSM at an earlier age and earlier in the course of infection. Clinicians should be particularly aware of the likely need for prompt initiation of HAART among older MSM, MSM from ethnic minorities and MSM diagnosed outside London.

Back to Top | Article Outline


This study would not have been possible without the continuing collaboration of clinicians, microbiologists, immunologists, clinical nurse specialists, health advisors, data management staff and other clinic staff who contribute to the surveillance of HIV in England and Wales.

Particular thanks to Dr. Andre Charlett (HPA CDSC) for statistical advice and Prof. Noel Gill (HPA CDSC) for conceptual advice. Acknowledgements also to James Carpenter and Mike Kenward for the ‘Introduction to Missing Data’ course, 2004, supported by the ESRC (course notes available from the Medical Statistics Unit, London School of Hygiene and Tropical Medicine). We are also grateful for continued input from the CD4 Surveillance Scheme Advisory Committee (additional to the authors): Dr David Barnett, Royal Hallamshire Hospital and UK NEQAS; Dr David Dunn, MRC Clinical Trials Unit, Professor Brian Gazzard, Chelsea and Westminster Hospital;, Dr. Linda Johnson-Laird, Department of Health; Dr. Linda Lazarus, Department of Health; Dr. Hilary Longhurst, St. Bartholomew's Hospital; Mrs June Cole, St. James’ University Hospital.

Sponsorship: Funding for the CD4 Surveillance Scheme is provided by the Department of Health for England. The SOPHID survey is funded by the Department of Health for England and the NHS London Specialised Commissioning Group.

Back to Top | Article Outline


1. Donaldson L. On the State of the Public Health: 2003. Annual report of the Chief Medical Officer 32189. 28-7-2004. London: Department of Health; 2004.
2. Egger M, May M, Chene G, Phillips AN, Ledergerber B, Dabis F, et al. Prognosis of HIV-1-infected patients starting highly active antiretroviral therapy: a collaborative analysis of prospective studies. Lancet 2002; 360(9327):119–129.
3. Sterling TR, Chaisson RE, Moore RD. HIV-1 RNA, CD4 T-lymphocytes, and clinical response to highly active antiretroviral therapy. AIDS 2001; 15:2251–2257.
4. Sterling TR, Chaisson RE, Keruly J, Moore RD. Improved outcomes with earlier initiation of highly active antiretroviral therapy among human immunodeficiency virus-infected patients who achieve durable virologic suppression: longer follow-up of an observational cohort study. J Infect Dis 2003; 188:1659–1665.
5. Krentz HB, Auld MC, Gill MJ. The changing direct costs of medical care for patients with HIV/AIDS, 1995–2001. CMAJ 2003; 169:106–110.
6. Weinhardt LS, Carey MP, Johnson BT, Bickham MS. Effects of HIV counseling and testing on sexual risk behaviour: a meta-analytic review of published research, 1985–1997. Am J Public Health 1999; 89:1397–1405.
7. Leaity S, Sherr L, Wells H, Evans A, Miller R, Johnson M, et al. Repeat HIV testing: high-risk behaviour or risk reduction strategy? AIDS 2000; 14:547–552.
8. Health Protection Agency, National Public Health Service for Wales, SCIEH, ISD, CDSC Northern Ireland and the UASSG. Renewing the Focus. HIV and other Sexually Transmitted Infections in the United Kingdom in 2002. An update: November 2003. London: Health Protection Agency; 2003.
9. McHenry A, Evans BG, Sinka K, Shaheem Z, Macdonald N, De Angelis D. Numbers of adults with diagnosed HIV infection 1996–2005–adjusted totals and extrapolations for England, Wales and Northern Ireland. Commun Dis Public Health 2002; 5:97–100.
10. Unlinked Anonymous Surveys Steering Group. Prevalence of HIV and Hepatitis Infections in the United Kingdom 2001. London: Department of Health; 2002.
11. Phillips A. Short-term risk of AIDS according to current CD4 cell count and viral load in antiretroviral drug-naive individuals and those treated in the monotherapy era. AIDS 2004; 18:51–58.
12. van Sighem AI, van de Wiel MA, Ghani AC, Jambroes M, Reiss P, Gyssens IC, et al. Mortality and progression to AIDS after starting highly active antiretroviral therapy. AIDS 2003; 17:2227–2236.
13. Pozniak A, Gazzard B, Anderson J, Babiker A, Churchill D, Collins S, et al. British HIV Association (BHIVA) guidelines for the treatment of HIV-infected adults with antiretroviral therapy. HIV Med 2003; 4(Suppl 1):1–41.
14. Dybul M, Fauci AS, Bartlett JG, Kaplan JE, Pau AK. Guidelines for using antiretroviral agents among HIV-infected adults and adolescents. Ann Intern Med 2002; 137(5 Pt 2):381–433.
15. Public Health Laboratory Service Communicable Disease Surveillance Centre, ICH (London), SCIEH. HIV & AIDS in the United Kingdom: An Epidemiological Review 2000. London: Public Health Laboratory Service; 2001.
16. Public Health Laboratory Service. Enhancement of HIV infection reporting. Commun Dis Rep Weekly 2000; 10:1.
17. Mortimer JY, Salathiel JA. ‘Soundex’ codes of surnames provide confidentiality and accuracy in a national HIV database. Commun Dis Rep Rev 1995; 5:R183–R186.
18. Gupta SB, Gilbert RL, Brady AR, Livingstone SJ, Evans BG. CD4 cell counts in adults with newly diagnosed HIV infection: results of surveillance in England and Wales, 1990–1998. CD4 Surveillance Scheme Advisory Group. AIDS 2000; 14:853–861.
19. McHenry A, Macdonald N, Sinka K, Mortimer J, Evans B. National assessment of prevalent diagnosed HIV infections. Commun Dis Public Health 2000; 3:277–281.
20. Smith CJ, Sabin CA, Lampe FC, Kinloch-de-Loes S, Gumley H, Carroll A, et al. The potential for CD4 cell increases in HIV-positive individuals who control viraemia with highly active antiretroviral therapy. AIDS 2003; 17:963–969.
21. Little RJ, Vartivarian S. On weighting the rates in non-response weights. Stat Med 2003; 22:1589–1599.
22. Efron B, Tibshirani RJ. An Introduction to the Bootstrap. New York: Chapman and Hall; 1993.
23. Hogg RS, Yip B, Chan KJ, Wood E, Craib KJ, O'Shaughnessy MV, et al. Rates of disease progression by baseline CD4 cell count and viral load after initiating triple-drug therapy. JAMA 2001; 286:2568–2577.
24. Sterling TR, Chaisson RE, Moore RD. Initiation of highly active antiretroviral therapy at CD4+ T lymphocyte counts of >350 cells/mm3: disease progression, treatment durability, and drug toxicity. Clin Infect Dis 2003; 36:812–815.
25. Krentz HB, Auld MC, Gill MJ. The high cost of medical care for patients who present late (CD4 <200 cells/microL) with HIV infection. HIV Med 2004; 5:93–98.
26. Sullivan AK, Curtis H, Sabin CA, Johnson MA, on behalf of BHIVA and the BHIVA audit subcommittee. National review of newly diagnosed HIV infections. BMJ 2004. (provisionally accepted for publication).
27. Easterbrook PJ, Yu LM, Goetghebeur E, Boag F, McLean K, Gazzard B. Ten-year trends in CD4 cell counts at HIV and AIDS diagnosis in a London HIV clinic. AIDS 2000; 14:561–571.
28. Henderson L, Keogh P, Weatherburn P, Reid D. Managing Uncertainty: Risk and Unprotected Anal Intercourse among Gay Men Who Do Not Know Their HIV Status. London, Sigma Research; 2001.
29. Chadborn TR, McGarrigle CA, Waight PA, Fenton KA. Trends in, and determinants of, HIV testing at genitourinary medicine clinics and general practice in England, 1990–2000. Sex Transm Infect 2004; 80:145–150.
30. Public Health Laboratory Service, DHSS & PS and the Scottish ISD(D)5 Collaborative Group. Sexually Transmitted Infections in the UK: New Episodes seen at Genitourinary Medicine Clinics, 1991 to 2001. London: Public Health Laboratory Service; 2002.
31. Murphy G, Charlett A, Brown A, Gill ON, Parry JV. Is HIV incidence increasing in homo/bisexual men attending GUM clinics in England, Wales and Northern Ireland? Commun Dis Public Health 2004; 7:11–14.
32. Report of a Working Group (Chairman: Professor N E Day) convened by the Director of the Public Health Laboratory Service on behalf of the Chief Medical Officers. The incidence and prevalence of AIDS and prevalence of other severe HIV disease in England and Wales for 1995 to 1999: projections using data to the end of 1994. CDR Review 1996; 6(1):R1–R24.
33. Public Health Laboratory Service Communicable Disease Surveillance Centre, ICH (London), SCIEH. HIV & AIDS in the United Kingdom 2001: An update November 2002. London: Public Health Laboratory Service; 2002.
34. Phair JP, Mellors JW, Detels R, Margolick JB, Munoz A. Virologic and immunologic values allowing safe deferral of antiretroviral therapy. AIDS 2002; 16:2455–2459.
35. Department of Health. The National Strategy for Sexual Health and HIV. London: Department of Health; 2001, pp. 1–53.

HIV infections; late diagnosis; mortality; surveillance; England and Wales

© 2005 Lippincott Williams & Wilkins, Inc.