National HIV Prevalence and Incidence Estimates for 1999: No Evidence of a Decline in Overall Incidence

HIV/AIDS EPI Updates

April 2003

Download all of the HIV/AIDS Epi Updates
82 pages (1,223 KB) in PDF format 
(How to download PDF documents)

National HIV Prevalence and Incidence Estimates for 1999: No Evidence of a Decline in Overall Incidence

---

Introduction

This report outlines estimates of the total number of Canadians who were living with HIV infection at the end of 1999 (prevalence) and the number of individuals who became newly infected in 1999 (incidence). Producing national estimates of HIV prevalence and incidence are an integral part of the work of the Centre for Infectious Disease Prevention and Control. They are used as a tool to monitor the HIV epidemic, to help evaluate and guide prevention efforts, and are part of on-going risk assessment and management work conducted by the Centre. It is anticipated that the next set of national estimates will pertain to the year 2002 and will be produced during 2003.

Methods

HIV prevalence and incidence estimates at the national level are estimated by a combination of different methods using data from a wide variety of sources, such as AIDS case reports, provincial HIV testing databases, population-based surveys, targeted epidemiological studies and census data.

Prevalence and incidence estimates are derived separately for the following exposure categories: men who have sex with men (MSM), injecting drug users (IDU), MSM-IDU, heterosexuals (includes sub-categories of heterosexual contact with a person at risk for HIV, origin in a country where HIV is endemic andheterosexual as the only identified risk) and other (includes recipients of clotting factor or blood products, perinatal and occupational transmission).

Two main methods are used to estimate both HIV prevalence and incidence: a direct and an indirect method.

Direct method:^1,2
HIV prevalence:

Prevalence rate: Estimated population size

HIV incidence:

Incidence rate: Estimated population at risk (total population for that group minus those already infected with HIV).

• Prevalence and incidence rates are estimated from targeted epidemiological studies;

• Population sizes are derived using a variety of methods, including projected population estimates, population-based surveys, and capture-recapture analyses (for IDU).

Indirect Method:^1,2

Two indirect methods are used together to estimate HIV prevalence. They are both based on the number of HIV diagnoses (obtained from provincial HIV testing databases) and on information about HIV testing behaviour (obtained from population surveys).

• In the first method, the cumulative number of HIV diagnoses (minus cumulative AIDS deaths) is divided by the proportion of the population who have ever been tested for HIV.

• In the second method, the number of HIV diagnoses in 1999 is divided by the proportion of the population who are tested in a one-year period, and the result is then added to the cumulative number of HIV diagnoses to the end of 1998 (minus cumulative AIDS deaths) plus the estimate for 1999 HIV incidence.

HIV incidence is derived using an indirect method by taking the difference between the 1999 and 1996 prevalence estimates and adding 1997-1999 AIDS deaths. The result is an estimate of the number of new infections occurring during 1997-99 which is divided by three to estimate the number during 1999. See the references for a further description of the methods and results.

These methods are carried out for each of British Columbia (B.C.), Alberta (Alta.), Ontario (Ont.) and Quebec (Que.). These four provinces account for 85% of the population of Canada and for 95% of HIV and AIDS diagnoses in Canada. For each exposure category in these provinces, the mean of the direct and indirect methods is taken as the point estimate for both prevalence and incidence. Ranges of uncertainty are calculated using statistical simulation procedures.

The remainder of the provinces and territories are grouped into Saskatchewan/Manitoba/Northwest Territories/Yukon/Nunavut and the Atlantic provinces (Nova Scotia, New Brunswick, Prince Edward Island, and Newfoundland). The prevalence and incidence estimates for these two regions are extrapolated using estimates from B.C., Alta., Ont. and Que. and regional distribution of reported AIDS cases and positive HIV diagnoses.

Prevalence and Incidence Results

As outlined in Table 1, there were an estimated 49,800 people in Canada living with HIV infection (includes those living with AIDS) at the end of 1999 compared with an estimated 40,100 at the end of 1996. This represents an increase of 24%.

An estimated 4,190 Canadians became newly infected with HIV in 1999 (Table 2). Although the estimated incidence in 1999 was essentially unchanged from 1996, the distribution among exposure categories changed significantly. From 1996 to 1999, there was a 30% increase in the number of new infections per year among MSM (from 1,240 to 1,610) and a 27% decline in the number of new infections among IDU (from 1,970 to 1,430).

---

Table 1: Point estimates and uncertainty ranges for number of prevalent HIV infections in Canada at the end of 1999 compared with point estimates for 1996, by exposure category

MSM MSM-IDU IDU Hetero Other Total 1999 Range 29,600 (26,000-33,400) 2,100 (1,700-2,600) 9,700 (8,100-11,800) 8,000 (6,300-10,100) 400 (330-470) 49,800 (45,000-54,600) 1996 25,300 1,700 7,100 5,500 500 40,100

---

Table 2: Point estimates and uncertainty ranges for number of incident HIV infections in Canada in 1999 compared with point estimates in 1996, by exposure category

MSM MSM-IDU IDU Hetero Other Total 1999 Range 1,610 (1,190-2,060) 270 (190-360) 1,430 (1,030-1,860) 880 (610-1,170) 0 4,190 (3,310-5,150) 1996 1,240 290 1,970 700 0 4,200

---

Tables 1 and 2 reflect a rise in the number of prevalent and incidence infections among heterosexuals. The estimated prevalent number of infections to 1999 is 8,000, 45% higher than the estimate of 5,500 in 1996. The number of new infections among heterosexuals increased 26% over this same time period, from 700 to 880 new infections per year.

An attempt was made to specifically evaluate the heterosexual subcategory 'endemic' to estimate what proportion of the epidemic can be attributed to individuals from a country where HIV is predominantly transmitted through heterosexual contact. Based on the proportions in reported HIV positive test reports and reported AIDS cases, it is estimated that this group represents approximately 20% to 30% of the heterosexual exposure category in the national prevalence and incidence estimates. It is important to note that there is considerable variation in this proportion between provinces, and that this estimate is quite uncertain since a large proportion of HIV test reports do not contain complete exposure category information.

Trends in Exposure Category among New Infections

The distributions of exposure categories among new HIV infections shown in Figure 1 clearly outline the changes in the HIV epidemic in Canada. Prior to 1999, the proportion of new infections attributed to IDUs had steadily increased from 2% during 1981-1983 to 24% between 1987-1990 to 47% in 1996. However, 1999 incidence estimates indicate that the proportion has now dropped to 34%. Conversely, the proportion of new infections attributed to MSM demonstrated a steady decline from over 80% in 1981-1983 to 30% in 1996. Recently, however, there has been a sharp increase in the proportion of new infections attributed to MSM to 38% in 1999. The proportion of new infections attributed to the heterosexual exposure category has increased steadily in the last two decades reaching 21% of new infections by 1999.

---

---

Trends among Women and Aboriginal Populations

An estimated 6,800 women were living with HIV infection (including those living with AIDS) at the end of 1999, a 48% increase from the 1996 prevalence estimate of 4,600. Thus, women accounted for 14% of the prevalent HIV infections in 1999 compared to 11% in 1996. The number of newly infected women remained just below 1,000 cases per year in both 1999 and 1996 (917 and 950 respectively) (Figure 2). In 1999, 54% of all new HIV infections among women were attributed to IDU and 46% were attributed to the heterosexual exposure category.

The estimated number of prevalent HIV infections among Aboriginal populations was 1,430 in 1996 and 2,740 in 1999, an increase of 91% (Figure 3). In 1999, 5.5% of all prevalent HIV infections in Canada were attributed to Aboriginal populations. There were 370 new infections in this group in 1999 (9% of all new HIV infections), 64% of which were attributed to the exposure category IDU, 17% were attributed to heterosexuals, 11% to MSM and 8% to MSM-IDU.

---

---

---

Comment

The methods employed to estimate HIV prevalence and incidence make maximum use of a wide variety of surveillance and research data. The resulting estimates are necessarily imprecise, but the degree of certainty is improved by using several independent methods in a combined approach. The advantages of this approach are its flexibility and its ability to readily incorporate new data and to highlight gaps in existing knowledge. The fact that the methods to estimate national trends of HIV are largely determined by data from large cities is a major limitation. These estimates do not, therefore, necessarily reflect local trends of HIV incidence and prevalence in Canada.

The 1999 estimates clearly illustrate the changing face of the Canadian HIV epidemic. They highlight a potential resurgence of the epidemic among MSM and an increasing urgency of the situation among Aboriginal populations. Although the estimates indicate a reduction in incidence among IDUs, the absolute number of new infections per year is still unacceptably high in this exposure category.

These findings will be critical for the planning and evaluation of prevention and care programs in Canada. The 1999 estimates support the need to strengthen and improve these programs and provide further evidence of the importance of enhancing existing surveillance and analysis. Furthermore, results underscore the fact that it is not yet time to reduce vigilance with respect to HIV and AIDS in Canada.

References

1. Archibald CP, Remis RS, Farley J, Sutherland D. Estimating HIV prevalence and incidence at the national level: combining direct and indirect methods with Monte-Carlo simulation. XI International Conference on AIDS, Geneva, 1998 (Abstract 43475).

2. Geduld J, Archibald C. National trends of AIDS and HIV in Canada. CCDR. 2000;26:193-201.

For more information please contact:

---

Health Canada

---