HIV risk-related behaviors among injection drug users in Rome: differences between 1990 and 1992

A mathematical model for HIV and hepatitis C co-infection and its assessment from a statistical perspective

The hepatitis C virus (HCV) and the human immunodeficiency virus (HIV) are a clear threat for public health, with high prevalences especially in high risk groups such as injecting drug users. People with HIV infection who are also infected by HCV suffer from a more rapid progression to HCV-related liver disease and have an increased risk for cirrhosis and liver cancer. Quantifying the impact of HIV and HCV co-infection is therefore of great importance. We propose a new joint mathematical model accounting for co-infection with the two viruses in the context of injecting drug users (IDUs). Statistical concepts and methods are used to assess the model from a statistical perspective, in order to get further insights in: (i) the comparison and selection of optional model components, (ii) the unknown values of the numerous model parameters, (iii) the parameters to which the model is most 'sensitive' and (iv) the combinations or patterns of values in the high-dimensional parameter space which are most supported by the data. Data from a longitudinal study of heroin users in Italy are used to illustrate the application of the proposed joint model and its statistical assessment. The parameters associated with contact rates (sharing syringes) and the transmission rates per syringe-sharing event are shown to play a major role.

Health systems' responses to 25 years of HIV in Europe: inequities persist and challenges remain

Europe is currently experiencing the fastest rate of growth of HIV of any region of the world. An analysis of policy and health system responses to the HIV epidemic in Europe and central Asia (hereafter referred to as Europe) over the last 25 years reveals considerable heterogeneity. In general, while noting hazards of broad generalisations and the differences that exist across countries in a particular grouping, effective policies to control HIV have been implemented more widely in western than in central and eastern Europe. However, the evidence suggests persistence of inequalities in access to preventive and treatment services, with those at highest risk, such as commercial sex workers, prisoners, intravenous drug users, and migrants often particularly disadvantaged, despite many targeted programmes. Responses in individual countries, especially in the early stages of the epidemic, were influenced by specific cultural and political factors. Strong leadership and active involvement by civil society organisations emerge as important factors for success but also a limiting factor to the response observed in eastern Europe, where civil society or NGO culture is weak as compared to western Europe. Scaling up of effective responses in many countries in eastern Europe will be challenging-where increased financial resources will have to be accompanied by broader changes to health system organization with greater involvement of the civil society in planning and delivery of client-focused services.

Intensity of drug injection as a determinant of sustained injection cessation among chronic drug users: the interface with social factors and service utilization

AIMS

The objective of this study was to identify factors associated with sustained injection cessation and to examine further the relationship between the occurrence of sustained injection cessation of injection drug users (IDUs) and prior injection frequency.

DESIGN AND SETTING

IDUs in the Montreal St Luc Cohort who had at least three consecutive interviews between 1995 and 1999 were included. Sustained injection cessation was defined as a period of at least 7 consecutive months without injection. All IDUs completed interview-administered questionnaires on socio-demographic characteristics, drug and sexual behaviours and health-related issues. Logistic regression was used for analyses.

FINDINGS

A total of 186/1004 (18.5%) IDUs reported a period of sustained injection cessation during the study period. In multivariate analysis, HIV-positive status, 'booting' and cumulative time spent in prison were negatively associated with injection cessation, while injection initiation after 35 years of age and frequent crack use were positively associated with injection cessation. We found a negative association between the occurrence of injection cessation and the frequency of injection; the odds ratios (OR) for cessation were 0.49 [95% confidence interval (CI): 0.03, 0.78] for IDUs who injected 30-100 times and 0.21 (95% CI: 0.10, 0.46) for IDUs who injected more than 100 times in the previous month. Attending needle exchange programmes (NEPs) or pharmacies appeared to be a modifier of the relation between cessation and prior injection frequency. The OR was 0.68 (95% CI: 0.42, 1.12) for IDUs who injected 30-100 times prior to injection and attended NEPs or pharmacies and was 0.07 (0.01, 0.30) for IDUs who did not use these services.

CONCLUSIONS

Overall, a fifth of IDUs experienced at least one episode of injection cessation of 7 months or more during a period of 4.5 years. Our data suggest that NEPs and pharmacies may have played a role in inducing injection cessation episodes in a subgroup of IDUs. Research is needed to better identify the characteristics of IDUs who could benefit from an injection cessation intervention strategy. This information is important for social and health policy planning.

Who's getting the message? Intervention response rates among women who inject drugs and/or smoke crack cocaine

BACKGROUND

Injection drug users (IDUs) who also smoke crack may be at greatest risk for infection with HIV as well as other blood-borne and sexually transmitted infections and in most need of positive behavioral changes.

METHODS

Three hundred and thirty-three women (aged 18-59 years) were randomly assigned to one of two enhanced gender- and culturally specific HIV intervention conditions or to the NIDA standard condition. Of primary interest in this study were baseline risk and intervention response rates among three groups of drug users--IDUs who did not smoke crack, IDUs who did smoke crack, and crack smokers who did not inject. Univariate and multivariate methods were utilized, including generalized estimating equations.

RESULTS

The intervention produced positive behavioral changes over time, but response rates varied according to drug using group. Overall, women falling into the crack smoking IDU category appeared to be less responsive to the intervention than those in the other drug using groups, and participants in the crack smoking only group were less responsive than those in the IDU only group.

CONCLUSIONS

Results presented here indicate the continuing need to develop and target effective interventions to particular subgroups of high-risk individuals who may be most resistant to change.

Sexual risk behaviors of male current and ex-opiate users in Chiang Rai, Thailand

Injecting drug use and unsafe sexual practice are both considered major risk factors for HIV infection. This study reports sexual behaviors among male "current" (i.e. using heroin and/or opium at least once in the past 3 months) and "ex-" opiate users in Chiang Rai province in Northern Thailand. Between January 1999 and August 2000, 206 male opiate users were recruited by mail callback. Of the 206 drug users, 89 (43.2%) could be classified as current users. Current users did not differ from ex-users, except for educational level and ethnicity. Current and ex-opiate users showed no difference in number of regular sexual partners, proportion of having sex with commercial and non-commercial sex partners, and reported histories of sexually transmitted diseases. This study suggests that the importance of sexual risk behaviors in HIV transmission cannot be ignored in both current and ex-opiate users.

Sexual behavior of drug-users: is it different?

BACKGROUND

Across the world, drug-users (DU) have been shown to be at greater risk of contracting and transmitting sexually transmitted diseases and HIV infection. However, in the metropolitan cities of India, where drug-use patterns may uniquely vary between injection and noninjection, depending on drug availability, little is known about the risk behavior of drug-users.

OBJECTIVE

The objective of the study was to find out whether drug-users are more likely to practice high-risk sexual behavior compared to non-drug users.

METHODS

The sexual behavior of 70 drug-users, selected by snowball sampling, was compared with 128 age- and economic status-matched controls (non-drug-users), randomly selected from the same community.

RESULTS

Seventy-eight percent sexually active drug-users and 43% sexually active non-drug-users had multiple sex partners (P < 0.02). The average number of sex partners was significantly higher among the drug-users (P = 0.05). Anal intercourse was practiced by 25.7% drug-users and 8.6% non-drug-users (OR = 3.7; P < 0.01). The drug-users were 6.7 times more likely to visit commercial sex workers (P < 0.0001). Sexually transmitted disease prevalence was higher among the drug-users (P < 0.003).

CONCLUSIONS

Notwithstanding the small sample size, this study shows that drug-users practice high-risk sexual behavior more frequently. Interventions, in the form of awareness generation and prevention education programs, are required in order to reduce risk of sexually transmitted diseases and HIV transmission among drug-users.

Nosocomial bloodstream infections among human immunodeficiency virus-infected patients: incidence and risk factors

To assess the incidence of nosocomial bloodstream infections (NBSIs) in human immunodeficiency virus (HIV)-infected patients, and to analyze the main associated risk factors, we performed a 1-year multicenter prospective study of patients with advanced HIV infection who were consecutively admitted to 17 Italian infectious diseases wards. As of May 1999, a total of 65 NBSIs (4.7%) occurred in 1379 admissions, for an incidence of 2.45 NBSIs per 1000 patient-days. Twenty-nine NBSIs were catheter-related bloodstream infections, with a rate of 9.6 central venous catheter-associated infections per 1000 device-days. Multivariate analysis indicated that variables independently associated with NBSIs included active injection drug use, a Karnofsky Performance Status score of <40, presence of a central venous catheter, and length of hospital stay. Mortality rates were 24.6% and 7.2% among patients with and without NBSIs, respectively (P<.00001). In the era of highly active antiretroviral therapy, NBSIs continue to occur frequently and remain severe and life-threatening manifestations.

HIV-positive notification and behavior changes in Montreal injection drug users

This study examined the effect of an HIV-positive notification on the sexual and injection behaviors, living conditions, and medical demands of injection drug users (IDUs) in Montreal. The behavior changes of 73 IDUs who received an HIV-positive test result, and 219 IDUs who tested negative were compared, and the net difference (ND) in the proportion of IDUs adopting the particular change was determined. No significant changes were found in drug use or needle sharing practices. A higher proportion of IDUs who received an HIV-positive notification acquired unstable living conditions, ND = 20.7% (95% CI = 3.3, 38.1), began medical follow up, ND = 34.4% (95% CI = 20.8, 48.7), and increased needle exchange program (NEP) utilization, ND = 20.5% (95% CI = 8.3, 32.8). Compared with HIV-negative males, more HIV-positive male IDUs stopped sexual relations, ND = 24.6% (95% CI = 0.4, 48.9), and sex work, ND = 31.8% (95% CI = 12.4, 51.3), and fewer began new relations, ND = -38.2% (95% CI = -52.6, -23.9). The medical community and NEPs have an important role in providing support for newly diagnosed IDUs.

Sexual behaviour of heterosexual individuals with HIV infection naive for antiretroviral therapy in Italy

BACKGROUND

Specific information about determinants of sexual behaviour of HIV infected heterosexuals, like injecting drug use (IDU), are essential to design interventions aimed at promoting safer sex practices.

METHODS

We analysed data on sexual behaviour collected, between March 1997 and March 1999, through a self administered questionnaire among 1050 IDUs and 642 non-IDU heterosexuals enrolled in a prospective multicentre cohort study on the natural history of HIV infection.

RESULTS

Among non-IDU heterosexuals, more women (48.5%) than men (25.1%) (p<0.001) reported that they were infected by HIV positive regular partners whose HIV status they were not aware of. Among the 1119 heterosexual males, one fifth reported having had more than 25 sexual partners during their lifetime. Condom use in the last sexual intercourse was more common among heterosexual IDUs (64.9%) than among non-IDU heterosexual males (58.3%) (p=0.05). Heterosexual IDU males were more likely (66.7%) than non-IDU heterosexuals (50.6%) to have an HIV negative partner (p<0.001). Of the 573 heterosexual females studied, 10.2% reported having had more than 25 lifetime sex partners. This proportion was higher among heterosexual IDUs (18.8%) than among non-IDU heterosexuals (4.3%) (p<0.001). Nearly 50% of the women in both groups reported having used a condom in the last intercourse. Almost 57% of heterosexual IDUs had a current HIV negative partner, compared with 34.9% non-IDU heterosexuals (p<0.001). In both sexes, the findings from univariate analysis were confirmed by multiple logistic regression analysis.

CONCLUSIONS

This study identified some important differences, in both males and females, in sexual lifestyles according to injecting drug use (for example, in terms of HIV negative partners). This observation indicates the need to tailor HIV prevention messages according to history of injecting drug use.

Risk behaviour change and HIV infection among injection drug users in Montreal

OBJECTIVE

To investigate the independent association between changes in risk behaviour and HIV seroconversion risk among Montreal injection drug users (IDU).

DESIGN

A longitudinal study of risk behaviour change and the maintenance of low-risk practices. At baseline and semi-annually, subjects were tested for HIV, and questionnaires on risk behaviour were completed.

RESULTS

A total of 833 IDU were recruited from January 1992 to June 1998, and completed a minimum of three visits. Large fluctuations in risk behaviour were observed, and the risk of HIV infection appeared to be dependent upon the consistency of risk behaviour practised. IDU who consistently engaged in risky behaviour were at high risk of HIV infection. IDU who attempted to practise low-risk behaviour but experienced relapses to risky behaviour were also at considerable risk of infection. IDU who managed to maintain low-risk practices were at minimal risk. Using Cox regression analysis, the hazard ratio (HR) of HIV seroconversion among IDU who consistently and inconsistently shared needles with an HIV-positive partner was 8.17 (95% CI 3.59-18.59) and 2.63 (95% CI 1.33-5.17), respectively, relative to non-needle sharers. Corresponding HIV incidence rates were 30.42 per 100 person-years (py) among consistent sharers, 13.78 per 100 py among inconsistent sharers and 2.51 per 100 py among non-sharers.

CONCLUSION

Although some HIV risk reduction was evident, behaviour change seems to be effective only in IDU who adopt and maintain low-risk practices. Additional strategies may be needed to assist IDU in the maintenance of low-risk practices.

[Evolution of the prevalence of HIV infection and risk behavior of drug injection among intravenous drug users infected or not infected by HIV in 3 Spanish cities. Working Group of Physicians of the World for the monitoring of HIV infection and risk behaviors in intravenous drug users]

OBJECTIVE

To know the evolution of HIV prevalence and risk behavior practices of drug injection in three spanish cities on the basis of the serologic status of injecting drug users.

METHODS

Cross-sectional surveys among users in syringe interchange programmes.

RESULTS

HIV prevalence (self-reported) decreased in Madrid from 50% (1992) down to 41% (1996) (p = 0.01) and did not change in Seville and Valencia from 1994 to 1996. Among HIV-positive injecting drug users, a decrease was observed in the practice of giving and taking used syringes in the three cities, although this decrease reached a statistically significance (p = 0.006) only for giving in Valencia from 1994 to 1996 and was almost significant (p = 0.08) for taking in Madrid from 1993 to 1996. Among HIV-negative injecting drug users, no decreases were observed in the practice of giving and taking used syringes and even a significant increase in giving syringes in Seville was recorded (p = 0.01) from 1994 to 1996.

CONCLUSIONS

HIV prevalence among injecting drug users is stabilized or decreasing in the three studied cities. The prevalence of injecting risk behavior evolves differently according to the serologic status.

Longitudinal pathways to condom use: a psychosocial study of male IDUs

This longitudinal study examined the psychosocial risk and protective factors involved in condom use among 265 male IDUs. Subjects were individually interviewed at two points in time using a structured questionnaire, which included psychosocial measures and questions about drug use and condom use. Data were analyzed using Pearson correlations and hierarchical regression analyses. T1 condom use was the most powerful predictor of T2 condom use, with or without control on other variables. T1 family support for condom use, friends' support for condom use, adaptive coping with AIDS or the threat of AIDS, and planning to use condoms were significantly related to T2 condom use. Protective father factors enhanced the effects of other protective psychosocial factors, increasing T2 condom use. The results suggest several approaches for changing risk-taking sexual behavior in male IDUs: enhancing coping abilities, selecting peers who take fewer risks, and planning to use condoms.

Longitudinal changes in sexual risk behavior among HIV+ and HIV- male injecting drug users

Injecting drug users (IDUs) play a prominent role in the transmission of human immunodeficiency virus (HIV), particularly in urban areas such as New York City, where they comprise nearly half of all adult acquired immunodeficiency syndrome (AIDS) cases. Intervention studies have demonstrated that IDUs are responsive to safer sex messages, but sexual behavior appears to be more resistant to change than drug use behavior. This multidisciplinary study (without an intervention component) assesses changes in sexual risk behavior as a function of time, HIV status, and disease progression in a cohort of HIV+ and HIV- male IDUs (N = 144) for 4 years.

RESULTS

For HIV+ and HIV- men, there were increases in abstinence and monogamy, with decreases in the frequency of unprotected vaginal/anal sex and sexual risk index scores. With the exception of monogamy, HIV+ men reported lower levels of risk. Although there was also a decline in substance use, this accounted for only some of the decline in sexual risk behavior. Among the HIV+ men, a CD4 level below 200 was associated with more abstinence and monogamy. HIV-related medical symptoms were associated with increased abstinence, less unprotected sex, and lower sexual risk index scores. Lower neuropsychological memory test scores were associated with increased abstinence and lower sexual risk index scores. Neurological impairment and depression were not associated with sexual risk behavior.

CONCLUSION

IDU men in New York City have modified their sexual behavior toward safer practices. Lower levels of risk are found among HIV+ men, particularly those with more progressed HIV illness. Nevertheless, a substantial amount of sexual risk behavior remained in this cohort, indicating the continued need for education and intervention.

[Prevalence of human immunodeficiency virus and risk behaviours among opioid users seen in an emergency room]

INTRODUCTION

As prevalence of HIV-1 among drug users in Spain is high and it is difficult to contact them because of their hidden behaviors, emergency rooms are one of the health facilities where they can be located. A cross-sectional interview study was planned. The aims of the study were to estimate prevalence of HIV-1 and to describe risk behaviors.

METHODS

The sample included all patients that in spring 1992 were detected and referred by the doctor as being current opiate users, defined as any use in the 30 days prior the interview. Drug users who did not know their HIV status or were negative for more than six months were asked to provide a urine sample to test HIV serology. A descriptive analysis with simple stratification was carried out. Row and adjusted odds ratio were used to analyse association between different variables and HIV status. Logistic regression was used to examine variables associated with HIV infection and risk behaviors (injecting drug use, sharing needles, and no use of condom).

RESULTS

Three hundred and eighty three opiate users were interviewed. It was possible to know HIV status of 94% of the subjects. Of them, 61% were positive (219). The best adjusted logistic model to predict associated variables with HIV included being female, primary school level, sickness absence, to attend because of organic pathology, and more years of parenteral use. Seventy five percent of the sample injected drugs during the past 30 days, and among them 30% shared syringes. The variables associated with a higher probability of having injected heroin or cocaine in the last 30 days were to have completed at least primary education, to be unemployed or reliant on illegal activities, not to be in drug treatment, and a larger number of drugs used in the last 30 days. A higher probability of sharing syringes was associated with a lower educational level, not to be in drug treatment, to live alone and a higher number of drugs used during last 30 days. Twenty one percent of the subjects who were sexually active always or nearly always used condom with regular partners and 56% with casual partners. Women were more likely to use condom than men with casual partners. Subjects who shared syringes during last 30 days were nearly three times more likely not to use condom with casual partners.

CONCLUSIONS

Although a high prevalence of HIV-1 was estimated among opioid users seen in an emergency room, it is not higher than estimates for intravenous drug users recruited from treatment centres, prison or needles exchange programs. A high frequency of risk behavior was also observed indicating a need to develop specific prevention programs for drug users.

Infecção pelo HIV entre usuários de drogas injetáveis: análise dos fatores de risco no Município do Rio de Janeiro, Brasil

O trabalho visa a determinar os fatores de risco para a infecção pelo HIV (Human Immunodeficiency Virus - vírus da aids) entre usuários de drogas injetáveis (UDI), no Município do Rio de Janeiro. Os entrevistados foram recrutados em centros de tratamento para usuários de drogas e nas "cenas de uso" ("ruas"). Foram utilizados três questionários: um destinado a colher informações mais gerais acerca dos UDI, adaptado de estudo anterior da Organização Mundial da Saúde (OMS); um segundo voltado para a pesquisa de informações suplementares sobre conhecimentos, comportamentos e atitudes em relação à aids; além de um terceiro instrumento destinado a avaliar níveis de disfunção psicológica. O estudo consistiu de 110 entrevistados que voluntariamente realizaram testes laboratoriais para a presença do HIV, além de outros agentes infecciosos de transmissão sangüínea (HBV, HCV, HDV, HGV/GBV-C, HTLV-I/II). A soroprevalência para o HIV-1 foi de 28,7%. Utilizando-se um modelo de regressão logística múltipla (Estatística de Hosmer-Lemeshow, c² = 1,89; p=0,98), foram identificados como fatores de risco para a infecção pelo HIV: "local de moradia de baixa renda" (OR=5,57; IC 95%: [1,39 - 22,27]), "fonte de renda irregular" (OR=3,26; IC 95%: [1,01 - 10,51]), "menor idade de início quando do consumo de drogas injetáveis" (OR=2,50; IC: 90%: [0,99 - 6,28]), "local de recrutamento nos centros de tratamento para uso de drogas" (vs. recrutamento nas "ruas") (OR=7,91; IC 95%: [2,03 - 30,87]) e "maior freqüência de compartilhamento do material de injeção nos 6 meses anteriores à entrevista" (OR=4,41; IC 95%: [1,33 - 14,64]).

Temporal trends in HIV risk behaviors of out-of-treatment injection drug users and injection drug users who smoke crack

This study compared the baseline sociodemographic characteristics and HIV risk behaviors of two groups of out-of-treatment injection drug users (IDUs): 366 who concurrently smoked crack (smoking IDUs) and 212 who did not smoke crack (IDUs) in the past 30 days. Temporal trends in recent risk behaviors were also assessed for each drug user group over an 18-month period, January 1992 through June 1994. Baseline data were collected in South Philadelphia before the implementation of a multisite HIV intervention research project funded by the National Institute on Drug Abuse. For the temporal trend analysis, the sample was grouped into four intake periods based on the date of the baseline interview. The results indicated that although both groups were economically disadvantaged and at high risk of HIV infection and transmission, smoking injectors had fewer economic resources and were at a moderately greater risk because of higher levels of sexual risk behaviors. Analysis of temporal trends revealed few reductions in drug risk behaviors and none in sexual risk behaviors. This study points to the need for examining differences between types of drug users, developing appropriate multidrug treatment programs and assessing the characteristics of communities so that theory-based interventions can be tailored for maximum effectiveness.

Deliveries, abortion and HIV-1 infection in Rome, 1989-1994. The Lazio AIDS Collaborative Group

The prevalence of HIV-1 among women of reproductive age is currently estimated at the time they give birth. We assessed HIV-1 prevalence at the end of pregnancy, whether they delivered or had an induced or spontaneous abortion. Women admitted at the end of pregnancy to hospitals in the Lazio Region, Italy, were tested for antibodies to HIV-1. Consent for testing was granted by 97.1% of 218,357 subjects; women who did not consent were tested anonymously. The prevalences of infection were 0.34% in 1989, 0.38% in 1990, 0.28% in 1991, 0.23% in 1992, 0.28% in 1993, and 0.24% in 1994. Significantly higher prevalences of infection were associated with induced abortion (0.49%) than with delivery (0.18%; OR: 2.72; 95% CI: 2.29-3.22) and among women who refused (0.85%) than among those who consented to testing (0.27%; OR: 3.14; 95% CI: 2.35-4.19). A significant temporal reduction in prevalence was observed only among women who delivered (0.15% in 1993 and 1994; 0.26% in 1989 and 1990). The prevalence of HIV-1 infection is thus higher among women undergoing induced abortions than among those who deliver and higher among women who refuse testing than among those who consent. Studies confined to neonatal testing or to voluntary testing of pregnant women would thus underestimate the prevalence of HIV-1 among women of reproductive age.

A persistent rise in mortality among injection drug users in Rome, 1980 through 1992

OBJECTIVES

The purpose of the study was to analyze overall and cause-specific mortality among injection drug users in Rome.

METHODS

A cohort of 4200 injection drug users was enrolled in drug treatment centers from 1980 through 1988 and followed up until December 1992.

RESULTS

The age-adjusted mortality rate from all causes increased from 7.8/1000 person-years in 1985/86 to 27.7/1000 in 1991/92. The rise was mainly attributable to acquired immunodeficiency syndrome (AIDS), but mortality from overdose and other causes increased as well. The cumulative risk of death by the age of 40 was 29.3%.

CONCLUSIONS

The impact of AIDS deaths appears to be additional to a persistent increase of mortality for all other causes.

Close Relationships and Elevated HIV Risk Behavior: Evidence and Possible Underlying Psychological Processes

A review of research on AIDS preventive behavior indicates that minority and nonminority heterosexual adolescents and adults, gay men, injection drug users, and commercial sex workers are all less likely to practice safer sex with close relationship partners, compared with partners they perceive to be “casual” sexual partners. Because many individuals in close relationships have engaged in HIV risk behavior over extended periods of time and are unaware of their actual HIV status, practicing unprotected sexual intercourse with a committed relationship partner who is not tested for HIV appears to be a major and unrecognized source of HIV risk. This article reviews the evidence for higher levels of HIV risk behavior in close relationships and then presents relevant conceptual and empirical work to explore the psychological processes that may underlie risky sexual behavior in close relationships, using as a framework the information–motivation–behavioral skills model of preventive behavior.

The HIV/AIDS epidemics among drug injectors: a study of contact structure through a mathematical model

Numerical results of a model with variable infectivity for the dynamics of HIV/AIDS (human immunodeficiency virus/acquired immune deficiency syndrome) have been compared with the data of AIDS cases among intravenous drug users in Italy, especially in the Latium region. We examined several hypotheses about the dynamics of the epidemics; for each we obtained, mainly through a least-square approach but also minimizing a different quantity, a best-fit estimate of the parameters. In the simplest model, the population is assumed to be homogeneous, and we estimate contact rate and year of start of the epidemics, obtaining a good fit up to 1989, less so after. A substantial increase in fit is obtained in assuming either a decrease of the contact rate over time or a heterogeneous population with a smaller active group. We have also compared models with different infectivity curves during the incubation period: the assumption of constant infectivity is untenable, whereas the often suggested hypothesis of a peak in infectivity shortly after infection seems to be in agreement with data.