OUP user menu

AIDS mortality before and after the introduction of highly active antiretroviral therapy: does it vary with socioeconomic group in a country with a National Health System?

Carme Borrell , Maica Rodríguez-Sanz , M. Isabel Pasarín , M. Teresa Brugal , Patricia García-de-Olalla , Marc Marí-Dell'Olmo , Joan Caylà
DOI: http://dx.doi.org/10.1093/eurpub/ckl062 601-608 First published online: 12 May 2006

Abstract

Background: The aim of this study is to determine whether socioeconomic AIDS mortality inequalities before and after the introduction of highly active antiretroviral therapy (HAART) have increased or decreased in a Spanish city where HAART is free. Methods: The study used a trend design, including all Barcelona residents older than 19 years of age. All AIDS deaths, which occurred among these residents between 1991 and 2001 were included. The variables studied were age, sex, socioeconomic (SES) group and HIV transmission group. AIDS age-standardized mortality rates for each year were estimated. Poisson regression models were fitted to obtain the relative risk (RR) of AIDS death for each socioeconomic group with respect to the reference group. Results: AIDS mortality increased up until 1995 and subsequently decreased due to the introduction of HAART. The increase in AIDS mortality was greater in the lowest SES group, which had higher rates and a RR of dying larger than that of the highest SES group, fact that remained fairly stable over the whole period. A similar pattern was observed in intravenous drug users. In the homosexual transmission group, rates for the lowest SES group were higher for the whole period and increased until 1996, while rates for the other SES groups were lower and decreased over the entire period. Conclusions: The fact that inequalities in AIDS mortality by SES group remained fairly stable for the whole period suggests that perhaps access to HAART, or adherence, is lower than desirable, in people of lower SES groups. These results ought to be taken into account when implementing treatment and prevention strategies.

  • AIDS mortality
  • HAART therapy
  • socioeconomic inequalities
  • trends
  • urban area
  • Southern Europe

The pattern of AIDS mortality changed after 1996 in the developed countries. Previously, it had been increasing since the appearance of the disease, having become the leading cause of death among young adults in Spain; after that year it started to decline.14 This change was due to the introduction of highly active antiretroviral therapy (HAART) and also to the utilization of preventive measures for opportunistic diseases.57 However, HAART did not only diminish mortality, it also improved the survival and the quality of life of persons with AIDS as has been reported by several authors.810

HIV transmission and AIDS have been related to social and material deprivation such as poverty, use of drugs, prostitution and immigration.11,12 Nowadays, AIDS is concentrated in the most socially vulnerable groups and countries,1317 inner city urban areas generally having high incidence. It has been reported that in the city of Barcelona (Spain) there are social inequalities in the AIDS distribution among the neighbourhoods of the city. The inequality pattern was different in the different transmission categories: material and social deprivation was related with AIDS in intravenous drug users (IDUs) and heterosexual contacts, but inversely related in the case of homosexual contacts.18

Several studies have shown that, prior to the introduction of HAART, AIDS mortality was increasing more in areas with greater deprivation than in other areas.19,20 Furthermore, some authors have reported that AIDS survival was longer for people having a higher socioeconomic status,21,22 although other studies have not found any differences.23 However, relatively few studies have tried to assess the effect on social inequalities of the introduction of this therapy.24,25

In Spain HAART was introduced in 1996 and its effectiveness has increased with the introduction of new drugs in the therapy.10 HAART is free of charge for all patients due to the existence of a National Health System. For this reason we may hypothesize that AIDS mortality inequalities after the introduction of HAART should narrow. Therefore, our aim is to determine whether socioeconomic AIDS mortality inequalities before and after the introduction of HAART have increased or decreased in a Spanish city.

Methods

Design and study population

Barcelona, the second largest city in Spain (1 508 805 inhabitants in 1996), is located on the north eastern coast. The study used a trend design, including all Barcelona residents older than 19 years of age. All AIDS-related deaths in the mortality register, which occurred among these residents between 1991 and 2001 were included. The underlying causes of death were coded using the International Classification of Diseases (ICD) 9th revision until 1999, the codes attributed to AIDS/HIV in Spain being: 279.5 (AIDS) and 795.8 (HIV without other specified causes). After 1999 the codes used were B20–B24 of the ICD-10.

The educational level and the neighbourhood of the deceased were obtained through a record linkage between the mortality register and the municipal census (87.4% of cases were linked), and transmission group was obtained through a record linkage with the AIDS cases register of Barcelona (85.6% of cases were linked). These are confidential probabilistic linkages based on the name and the surname and date of birth of the deceased.26

Information on the population at risk, including age, sex, the highest completed level of education and neighbourhood of residence comes from the 1991 and 1996 Municipal Census, an administrative register not subject to statistical secrecy; we did not have data of the census by educational level for other years. Every resident in Barcelona is registered in the municipal census. This census is continually updated to incorporate data on migration, births and deaths. For the years between the two censuses an estimate of the intercensal population was obtained based on a geometric curve. The denominators for the years 1997–2001 were based on the 1996 census.

Variables and indicators analysed

The variables studied were age, sex, highest level of education completed (educational level), neighbourhood of residence, and HIV transmission group.

Educational level was categorized in two groups: (i) Low educational level: includes illiterate, people with no education (no schooling) who have 0–4 years of schooling and people with primary studies not finished (5–6 years of schooling) and (ii) High educational level: includes people with more than primary studies (≥7 years of schooling). These two groups were chosen after inspecting the distribution of AIDS mortality rates in five educational levels (illiterate and no education, primary not finished, primary finished, secondary and university studies), it being found that AIDS rates in the first two groups were similar and likewise the rates in the other three groups.

Barcelona has 38 neighbourhoods, their populations varying, in 1996, between 1081 and 95 382 inhabitants (mean: 39 705 inhabitants per neighbourhood). We grouped the neighbourhoods following our prior analysis that divided them in two groups20: one group contained the low socioeconomic (SES) neighbourhoods while the other included all remaining neighbourhoods. The method employed was cluster analysis using SES variables (unemployment and illiteracy) obtained from local censuses corresponding to the years 1981, 1986 and 1991. Neighbourhoods were divided in two groups due to the small numbers of deaths in each neighbourhood, and also to the fact that there is an important degree of homogeneity in the group of neighbourhoods with high SES level, both in terms of SES indicators and in terms of mortality . The grouping involved 11 deprived neighbourhoods with high unemployment and illiteracy (and a population of 262 921 inhabitants in 1996), and included the older inner-city neighbourhoods, as well as a few new peripheral neighbourhoods; the other group consisted of 27 neighbourhoods, with a total population of 1 245 884 inhabitants in 1996 with lower values of the SES indicators (unemployment and illiteracy).

Using the two educational levels and the two groups of neighbourhoods we obtained 4 SES groups, hence subjects were classified as: (i) Living in a non-deprived neighbourhood and having a high educational level, (ii) Living in a non-deprived neighbourhood and having a low educational level, (iii) Living in a deprived neighbourhood and having a high educational level, (iv) Living in a deprived neighbourhood and having a low educational level.

Three HIV transmission groups were studied: homosexuals and bisexuals who were not IDUs, heterosexuals who were not IDUs and IDUs (i.e. includes homosexuals, bisexuals and heterosexuals who were IDUs).

Data analysis

All the analyses were performed for each sex and transmission group. Age-standardized AIDS mortality rates for each year, standardized through the direct method, were estimated using the 1996 population of Barcelona as the reference population. For IDUs we calculated the standardized rates for the 20–49 age-group because there were almost no cases older than 49 years.

Poisson regression models27 were fitted to obtain the relative risk (RR) of AIDS death for each SES group with respect to the reference group (living in a non-deprived neighbourhood and having a high educational level), adjusted for age for the years 1991, 1993, 1995, 1996, 1998 and 2000. The dependent variable was the logarithm of the AIDS mortality rate, and the independent variables were SES group and age (grouped as: 20–34 years and ≥35 years). For IDUs the models were estimated for the 20–49 age-group.

Results

We included 2087 cases of male AIDS death, and 488 female. Table 1 shows the distribution of these cases of death by year, transmission group, SES group and age-group. The majority of AIDS deaths occurred in the IDUs group, followed by the male homosexual transmission group.

View this table:
Table 1

Distribution of AIDS deaths by year of death, HIV transmission group, socioeconomic group and age-group. Men and women, Barcelona 1991–2001

MenWomen
N%N%
Year of death
    19911929.2326.5
    199224211.65110.5
    199327012.96813.9
    199430414.68717.8
    199534416.57415.2
    199631715.26713.7
    19971215.8316.4
    1998783.7234.7
    1999663.2163.3
    2000823.9173.5
    2001713.4224.5
HIV transmission group
    Intravenous drug user94745.426153.5
    Homosexual63630.5
    Heterosexual2059.815531.8
    Missing data29914.37214.7
Socio-economic group
    Non-deprived neighbourhood and high educational level44221.213126.8
    Non-deprived neighbourhood and low educational level41920.111223.0
    Deprived neighbourhood and high educational level25512.25711.7
    Deprived neighbourhood and low educational level69433.314028.7
    Missing data27713.2489.8
Age-group
    20–34 years88342.331163.7
    ≥35 years120457.717736.3
Total2087100.0488100.0

Figure 1 shows total age-standardized AIDS mortality rates during the whole study period by sex. For men, the increase until 1995 was higher among those living in deprived neighbourhoods; both these SES groups experienced a decrease which ended in 1998, after which rates either increased again or remained stable. For women, the pattern was similar. After 1998 the rates among women living in deprived neighbourhoods increased again, showing a different pattern compared to women living in non-deprived neighbourhoods.

Figure 1

Age-standardized AIDS mortality rate per 100 000 inhabitants aged ≥20 years by socioeconomic group. Barcelona, men and women, 1991–2001

Tables 2 and 3 show, for men and women, the RR of dying of each SES group comparing with the most privileged one, for the different years and for the different transmission groups. For total AIDS mortality, it is possible to observe how the RR increase mainly in the lowest SES in both sexes for the whole period: the RR were higher in 1995, 1998 and 2000 for men and in 1991, 1993 and 2000 for women, although the confidence intervals are wide. But it has to be mentioned that absolute differences are higher in 1995 than in the other years, mainly due to the higher rates. For example, the absolute difference for men between the lowest SES group and the highest in 1995 is 200.04 per 100 000 inhabitants and in 2000 is 59.9, for women these differences are 51.4 and 21.7.

View this table:
Table 2

Relative risk of dying of AIDS by socioeconomic group. Different transmission groups in men aged ≥20 years. Barcelona, 1991–2000

Transmission group and yearSocioeconomic group
Non deprived neighbourhoodDeprived neighbourhood
High educational levelLow educational levelHigh educational levelLow educational level
RRRR95% CIRR95% CIRR95% CI
Total
    199111.91.2–2.81.30.7–2.33.72.4–5.8
    199311.30.9–1.81.40.9–2.22.61.8–3.8
    199511.61.2–2.32.41.7–3.45.74.2–7.6
    199611.41.0–1.92.41.7–3.53.22.3–4.4
    199812.01.0–3.83.91.9–8.06.63.5–12.2
    200012.91.4–5.96.32.9–13.910.05.0–20.0
Intravenous drug usera
    199114.62.5–8.51.40.6–3.49.65.1–18.0
    199312.61.5–4.41.40.7–2.75.53.2–9.4
    199514.12.7–6.43.42.1–5.513.38.9–19.8
    199614.42.8–6.73.92.4–6.28.75.6–13.6
    199814.61.8–11.64.01.4–11.47.82.9–21.2
    200014.81.6–14.34.41.2–15.515.15.5–41.7
Homosexual
    199111.00.5–1.71.20.5–3.10.90.4–2.1
    199310.80.5–1.51.20.5–2.61.40.7–2.7
    199510.60.4–1.21.80.9–3.52.01.2–3.6
    199610.70.4–1.21.91.0–3.82.01.1–3.5
    199811.70.3–8.55.00.8–30.27.91.8–33.8
    200010.90.2–3.83.30.6–16.91.70.3–8.8
Heterosexual
    199110.90.2–5.01.80.2–16.43.70.8–17.1
    199311.00.4–2.82.30.6–8.32.81.0–7.8
    199511.00.3–3.01.80.5–6.35.02.0–12.4
    199610.60.3–1.51.00.2–4.11.00.3–3.0
    199810.60.1–3.11.60.2–14.04.31.2–15.0
    200010b0–∞b8.80.6–140.97.90.7–86.6
  • RR: Age-adjusted relative risk; 95% CI: 95% confidence interval of RR

  • a: For intravenous drug users it refers to 20–49 age-group

  • b: The number of cases in this socio-economic group is 0

View this table:
Table 3

Relative risk of dying of AIDS by socioeconomic group. Different transmission groups in women aged ≥20 years. Barcelona, 1991–2000

Transmission group and yearSocioeconomic group
Non deprived neighbourhoodDeprived neighbourhood
High educational levelLow educational levelHigh educational levelLow educational level
RRRR95% CIRR95% CIRR95% CI
Total
    199116.32.3–17.31.30.3–6.313.14.7–36.5
    199314.22.0–8.73.11.4–7.112.16.0–24.6
    199511.70.9–3.61.90.8–4.17.44.0–13.8
    199612.01.0–4.12.10.9–4.66.23.2–12.1
    199812.60.8–8.40.90.1–7.05.71.7–19.4
    200014.00.9–18.72.00.2–19.520.25.0–81.9
Intravenous drug usera
    199117.92.3–27.12.10.4–11.013.63.7–50.6
    199314.61.7–12.12.80.9–8.113.45.4–33.1
    199515.62.3–13.62.30.7–7.418.68.2–42.2
    199612.71.1–6.92.81.1–4.39.44.1–21.8
    199813.10.5–19.52.10.2–20.011.62.3–59.0
    200014.60.3–75.46.30.4–100.856.96.5–497.9
Heterosexual
    199113.40.4–28.50b0–∞b14.92.3–98.0
    199315.61.5–21.42.90.5–15.811.83.0–46.7
    199510.70.2–2.20.90.2–3.83.21.1–8.8
    199612.10.7–6.20.80.1–6.44.71.5–14.5
    199814.50.8–25.90b0–∞b4.00.4–40.6
    20001c0–∞cc0–∞cc0–∞c
  • RR: Age-adjusted relative risk; 95% CI: 95% confidence interval of RR

  • a: For intravenous drug users it refers to 20–49 age-group

  • b: The number of cases in this socioeconomic group is 0

  • c: The number of cases in the reference group is 0

Figure 2a shows the pattern of the evolution of age standardized AIDS mortality rates in IDUs. The pattern is very similar to that for all deaths, increasing until 1995, decreasing to 1998 and then increasing again mainly in people living in deprived neighbourhoods. RR are higher than for deaths due to all causes, mainly in the lowest SES group all over the years, although in the years 1995 and 2000 the RR are higher for men and in 2000 for women (tables 2 and 3). Again, the absolute differences are higher in 1995 (figure 2a).

Figure 2

Age-standardized AIDS mortality rate per 100 000 inhabitants aged ≥20 years by socioeconomic group and transmission group. Barcelona, men and women, 1991–2001. (a) Intravenous drug users (20–49 age-group) (b) Homosexual men (c) Heterosexual

In the homosexual transmission group, the trends in men are quite different from the trends in total AIDS mortality (figure 2b). Rates for the lowest SES group increased until 1996, while rates for the other SES groups decreased over the entire period. Table 2 shows the higher RR for the lowest SES group, with a peak in 1998.

Figure 2c shows trends in the heterosexual transmission group: for the lowest SES group, rates are higher and increase more for women than for men. For men, RR across SES groups are statistically significant particularly for the lowest SES group for almost all the years (table 2). For women, the small number of cases in this group makes the estimations difficult, but the lowest SES had high RR mainly in the first period (1991 and 1993) (table 3).

Discussion

These results show how AIDS mortality rates which were increasing prior to the introduction of HAART in 1996 decreased after that year. They also show that although the evolution of AIDS mortality for different SES groups, sexes and transmission groups was slightly different, AIDS mortality inequalities by SES group remained quite stable, in the sense that lower SES groups had higher mortality than the more privileged one's, and inequalities did not narrow after the introduction of HAART as we had hypothesized. These results refer to a Southern European urban area, in a country with a National Health System where HAART therapy is free and therefore accessible to everyone. To our knowledge this relationship has not been described before in any Southern European country, despite being the region where the highest AIDS rates among developed countries have been experienced.

The pattern of trends in AIDS mortality overall is very similar to that observed in IDUs, with rates being higher in males and in deprived neighbourhoods. In IDUs, mortality rates had already begun to decrease before the introduction of the HAART therapy in the majority of SES groups, probably due to the intervention programs in the AIDS field addressed to this group of people.28 These programs were designed to assist drug users, dealing not only with specific health problems (infectious disease, mental health…) but also with practical needs (housing, food…). Moreover, the harm reduction programs include active case-finding and treatment of organic problems (such as tuberculosis), methadone maintenance programs,29 syringe exchange, outreach programs, legal and occupational support. At the end of the period rates in IDUs tended to increase again, mainly those of the deprived neighbourhoods.

In the homosexual and heterosexual transmission groups, rates were higher in the lowest SES group. AIDS mortality trends in these groups have tended to decrease since 1996.

Other papers have reported AIDS evolution in terms of SES groups using both individual and aggregated data. Rapiti et al.30 studied AIDS survival by neighbourhood SES status before (1993–1995) and after (1996–1997) the introduction of HAART therapy in Rome in a retrospective cohort of AIDS cases. For persons with AIDS diagnosed in the 1993–1995 period they found little difference in the risk of death by neighbourhood SES status; but for 1996–1997, the risk of death was greater for persons with lower neighbourhood SES status, after having adjusted for age, sex, intravenous drug use, CD4 cell count diagnosis, AIDS defining disease and hospital diagnosis. The authors discuss the roles which health-care access, medical management, or poor adherence to treatment may play in explaining these inequalities.

Several studies, in American countries, have shown how certain trends in inequalities are related to the lack of access to HAART therapy. Wallace24 showed that decreases in AIDS mortality in areas of New York (USA), defined by zip code, were related to an array of SES and community stress variables, the areas with indicators of less deprivation (white population, college degree, higher income) having a greater decrease, while more deprived areas had a lower decrease. In San Franciso (USA), persons living in poorer neighbourhoods were less likely to have undergone HAART at any time in the past compared with persons in wealthier neighbourhoods, leading to poorer AIDS survival for people living in deprived areas.25 Blair et al.31 examined the evolution of AIDS mortality in men who have sex with men in the USA by racial/ethnic group: blacks and Hispanics had higher rates and smaller declines than whites. However one study, in Sao Paolo (Brasil), which found no relation between the decline of AIDS mortality rates and deprivation of the neighbourhoods, the authors hypothesize that this may be due to programs of free treatment and prevention.32

In the present study it was not possible to investigate differences in plasma viral load, CD4 cell count, comorbidity, resistance to treatment or behaviours related to AIDS (such as syringe sharing or risky sexual behaviour) between different SES groups. We only can say that late diagnosis of AIDS (which can be a proxy of CD4 cell counts) was not different among SES groups in the AIDS death cases where we had this variable available through the record linkage with the AIDS cases register (88.4% of death cases). One study conducted in British Columbia (Canada) where HAART therapy is also free, found that, after adjusting for all the factors mentioned above (except behaviours), individuals of lower SES groups had higher mortality and were less likely to receive triple therapy.33

The persistence of SES-group related mortality inequalities in Barcelona following the introduction of HAART may be explained by different factors.16 First of all, factors related to health care such as lack of access to treatment for people in the lowest SES groups,34 even though Spain has a National Health System and the therapy is free. In this sense, one Spanish study has reported that 8.4% of IDUs have not had an HIV test, which probably leads to a late diagnosis of AIDS and late treatment. The same study also reports that 36.8% of IDUs who are HIV positive have never taken HAART therapy.35 Other studies in Spain have found social class inequalities in access to preventive practices related to cancer (mammography, cytology, etc.) despite their being free; people of advantaged classes undertake these practices more frequently.36 Another explanation may involve differences in adherence to treatments because HAART is a long and complex treatment; people of low SES may have more difficulties in following it.16 Also, comorbidity may play a role: the presence of other diseases mainly in the lowest SES groups, such as hepatitis C, pulmonary infections, tuberculosis, etc. may affect HIV prognosis. Finally, differences in living conditions, life style and health behaviours may have consequences that affect the course of HIV infection.16

One of the limitations of the present study is that we only included death cases having AIDS as the underlying cause of death in the mortality register. In the pre-HAART era the majority of deaths in people with AIDS were AIDS-related, this cause of death having high validity,37 but after the introduction of HAART the causes of death probably changed among HIV infected individuals, diminishing the AIDS-related deaths as has been described by other authors.38 We have to take into account that this study did not collect the deaths due to other causes. Another limitation could be the small numbers of AIDS death cases for the heterosexual transmission group, particularly women in the last few years. Finally, we have to recognize the presence of some missing values in our variables (see table 1), these missing values were not however related to the other variables, except in the HIV transmission group which had more missing values for the years 2000 and 2001.

This study has described how SES group related inequalities in AIDS mortality did not narrow after the introduction of HAART, possibly suggesting that access, or adherence, to HAART among people in low SES groups is rather less than desirable. These SES inequalities must be considered when prevention and treatment strategies are implemented,39,40 taking into account that some studies have shown that it is possible to administer the HAART therapy to poor populations.41,42

Key points

  • This study pretends to determine whether socio-economic AIDS mortality inequalities before and after the introduction of highly active antiretroviral therapy (HAART) have increased or decreased in Barcelona.

  • The evolution of AIDS mortality for different socioeconomic (SES) groups, sexes and transmission groups was slightly different.

  • AIDS mortality inequalities by SES group remained quite stable, therefore inequalities did not narrow after the introduction of HAART. A similar pattern was observed in intravenous drug users.

  • These results ought to be taken into account when implementing treatment and prevention strategies in a country where HAART therapy is free.

Acknowledgments

This study was made possible by financial support from FIPSE (grant number 2425/01) and FISS C03/09, G03/05 (Cooperative Investigation Networks).

References

View Abstract