Giuseppe Schinaia

The Mover-Stayer Model for the Hiv/Aids Epidemic in Action

Short- and medium-term projections of the HIV/AIDS epidemic indicators are of great interest to those evaluating the needs for health care and prevention interventions. We developed a simulation procedure to obtain forecasts of the HIV/AIDS epidemic and used it to estimate the characteristic regional parameters of the epidemic in Italy. The simulation procedure is based on a hybrid compartmental model, in which the epidemic evolves via nonrandom mixing patterns. Because of its structure, the model is suitable for policy making; in particular, for evaluating prevention campaigns, alternate forms of health care for people with AIDS, and drug supply needs. The model will also be used to estimate the number of intravenous drug users in Italy and the number of AIDS cases not reported or reported with a delay to the Italian surveillance system.

Estimating the size of the HIV/AIDS epidemic : Complementary use of the empirical bayesian back-calculation and the mover-stayer model for gathering the largest amount of information

The HIV/AIDS epidemic has reached the end of its second decade, and estimations of its size and dynam ics are very important, particularly for developing and planning prevention and therapeutic interventions. Since the HIV infection is characterized by a long asymptomatic period and latency time, observational studies are not suitable to study the epidemic. Dy namic models and the Back-Calculation (BC) methods are two general methodologies which provide estimates and projections of HIV/AIDS incidence and prevalence, while focusing attention on different aspects of the epidemic. For example, the Mover-Stayer (MS) Model, a dynamic compartmental model developed to study the epidemics spread among a general population, is a useful tool to perform scenario analysis, while the Empirical Bayesian Back-Calculation (EBBC) method allows investigation of the dynamics of the epidemic by risk category. Both methods, with timely updating to reflect changes in the natural history of the HIV/ AIDS epidemic, were recently applied within the EU Concerted Action on Multinational AIDS Scenario Analysis and the Italian Consensus Conference. An overview of some dynamic models and BC methods, and an application of the MS Model and EBBC method to the Italian HIV/AIDS epidemic, are presented here.

Modeling the HIV/AIDS epidemic via survivor functions

An original approach to simulation modeling of the HIV/AIDS epidemic is proposed. This approach uses survivor functions estimated from cohort studies conducted with seropositive and AIDS-diagnosed individuals. The model can be considered an alternative to the usual Markov models and accounts for time-dependent HIV progression to AIDS, and AIDS progression to death. By using various forms of survivor functions, it can also easily be extended to accommodate natural history events, as well as long-term survivors and cofactor effects, when appropriate data are available.

EMPIRICAL DATA AND MATHEMATICAL STRUCTURES IN THE EPIDEMIC MODELING OF PARENTERAL HEPATITIS IN ITALY

Viral liver infections with parenteral transmission in Western countries are mostly caused by hepatitis B and hepatitis C viruses (HBV and HCV). This paper presents a mathematical model that describes the history of the spread of HBV and HCV infections in the general population in Italy. The analysis of the model and the results also provide some new insight into the mechanisms of the epidemics. The model structure is based on an underlying analysis of the various effects of the infection progression in the host, in order to incorporate into its parameters most of the information available from the literature. Moreover, incidence and prevalence curves of both HBV and HCV infections and of HBV/HCV co-infections are generated and qualitative aspects of the epidemic, such as possible endemic steady states and the basic reproduction number, are also analyzed.

EM Estimation of Diagnosis

The diagnosis/prognosis problem has already been introduced by the authors in previous papers as a classification problem for survival data. In this paper, the specific aspects of the estimation of the survival functions in diagnostic classes and the evaluation of the posterior probabilities of the diagnostic classes are addressed; a latent random variable Z is defined to denote the classification of censored and uncensored individuals, where early censored individuals cannot be immediately classified as Z is not observed. Parameter estimation of the mixture survival model thus derived is carried out using a proper version of the EM algorithm with given prior probabilities on Z and diagnostic/prognostic information provided by the observable covariates is also included into the model. Numerical examples using AIDS data and a simulation study are used to better outline the main features of the model and of the estimation methodology.

Low dose rate brachytherapy (LDR-BT) as monotherapy for early stage prostate cancer in Italy: practice and outcome analysis in a series of 2237 patients from 11 institutions

OBJECTIVE Low-dose-rate brachytherapy (LDR-BT) in localized prostate cancer is available since 15 years in Italy. We realized the first national multicentre and multidisciplinary data collection to evaluate LDR-BT practice, given as monotherapy, and outcome in terms of biochemical failure. METHODS Between May 1998 and December 2011, 2237 patients with early-stage prostate cancer from 11 Italian community and academic hospitals were treated with iodine-125 ((125)I) or palladium-103 LDR-BT as monotherapy and followed up for at least 2 years. (125)I seeds were implanted in 97.7% of the patients: the mean dose received by 90% of target volume was 145 Gy; the mean target volume receiving 100% of prescribed dose (V100) was 91.1%. Biochemical failure-free survival (BFFS), disease-specific survival (DSS) and overall survival (OS) were estimated using Kaplan-Meier method. Log-rank test and multivariable Cox regression were used to evaluate the relationship of covariates with outcomes. RESULTS Median follow-up time was 65 months. 5- and 7-year DSS, OS and BFFS were 99 and 98%, 94 and 89%, and 92 and 88%, respectively. At multivariate analysis, the National Comprehensive Cancer Network score (p < 0.0001) and V100 (p = 0.09) were correlated with BFFS, with V100 effect significantly different between patients at low risk and those at intermediate/high risk (p = 0.04). Short follow-up and lack of toxicity data represent the main limitations for a global evaluation of LDR-BT. CONCLUSION This first multicentre Italian report confirms LDR-BT as an excellent curative modality for low-/intermediate-risk prostate cancer. ADVANCES IN KNOWLEDGE Multidisciplinary teams may help to select adequately patients to be treated with brachytherapy, with a direct impact on the implant quality and, possibly, on outcome.