Please use this identifier to cite or link to this item:
Dynamic epidemiology of HIV/AIDS in Taiwan
Human immunodeficiency virus,Acquired immunodeficiency syndrome,Left censoring,Markov model,Survival analysis,HAART,
|Publication Year :||2008|
HIV/AIDS(human immunodeficiency virus/ human immunodeficiency syndrome)傳染病監測的困難，在於無法正確記錄到可感受者受傳染病感染的時間點，往往受感染者被觀察記錄到的時間，已是有症狀的發病期；如果沒有主動的傳染病篩檢系統，此種左設限(left censoring)的問題，將造成不易從現有的傳染病的通報資料來檢視傳染病之自然病程，亦造成傳染病流行病學監測的錯估。因此利用統計模式來預估HIV感染率及HIV到AIDS的轉移機率，即可克服左設限造成的困境。目前台灣少有研究是以–傳染病動力學來長期的預測HIV/AIDS的感染發生。HAART（Highly active anti-retroviral therapy）的介入更會使HIV感染及AIDS病例的預測更加複雜化。
利用HAART介入前所以HIV感染者及AIDS發病者的資料。馬可夫三階段模式(Markov three-state model)可以用來推測及模擬HIV及AIDS的自然病程動態。邏輯式成長模式(logistic growth model)可推估HIV在群族人口的成長速率。HAART對HIV感染者的死亡影響，採用時間相依的proportional hazard Cox regression model. 馬可夫三階段模式及邏輯式成長模式可以進一步用來預估在HAART實行後的HIV及AIDS變化。
Background It is a challenging task to estimate natural course of HIV and AIDS using empirical data because subjects infected with HIV are often asymptomatic and have remained occult for a long time (left-censored problem) without active surveillance system. It is imperative to apply a statistical model to estimate infection rate and the rate of conversion from HIV to AIDS making allowance for left-censored problem. In the view of infectious epidemiology, few studies in Taiwan considered using a dynamic concept to estimate and predict HIV infection and occurrence of AIDS with long-term follow-up. Furthermore, the administration of HAART further complicates the projection of HIV infection and AIDS cases.
Objects The present study aimed to investigate descriptive epidemiology of HIV and AIDS cases, to estimate annual infection rate and conversion rate from HIV to AIDS, to investigate the effect of HARRT on total death and death due to AIDS.
Methods Data used for the current study were derived from registered system of HIV and AIDS before the administration of HAART. A three-state Markov model was used to estimate the parameters pertaining to the natural history of HIV and AIDS. Logistic growth model was also used to estimate the growth rate of HIV. The effect of HARRT on reducing total death was investigated by the application of time-dependent proportional hazards Cox regression model. Both three-state-state Markov model and logistic growth model are further used to project HIV and AIDS after the HAART study.
Results Annual infection rate for HIV infection was 0.0000006519 per person year. Annual conversion rate form HIV to AIDS for the overall group was 0.139 per year, namely 7.58 years of the average dwelling time staying in HIV state. By looking at different transmission modes, annual infection rate for homosexual and heterosexual modes was approximately two-fold compared with intravenous drug users and hemophiliacs. However, the opposite results were noted for annual conversion rates in terms of transmission modes. Using logistic growth model, we estimated 50% reduction of HIV and AIDS after the administration of HAART. By using survival data, we found 67% reduction in total death among HIV cohorts. The numbers of projection of HIV and AIDS in the absence of HAART intervention based on our three-state model were 9540 and 4164.
Conclusion Stochastic model and logistic growth model together with data before the administration of HAART were used to elucidate the natural history of HIV and AIDS in the absence of HAART intervention. We also found HAART can reduce 67% of total death and 50% of HIV or AIDS cases.
|Appears in Collections:||流行病學與預防醫學研究所|
Files in This Item:
|994.99 kB||Adobe PDF|
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.