عنوان

Nonparametric exploration of waning vaccine effects using survival data

شماره

TLpq304391005

زبان متن نوشتاري يا گفتاري و مانند آن

انگلیسی

عنوان اصلي

Nonparametric exploration of waning vaccine effects using survival data

نام عام مواد

[Thesis]

نام نخستين پديدآور

L. K. Durham

نام ساير پديدآوران

I. M. Longini, Jr.

نام ناشر، پخش کننده و غيره

Emory University

تاریخ نشرو بخش و غیره

1997

نام خاص و کميت اثر

91

جزئيات پايان نامه و نوع درجه آن

Ph.D.

کسي که مدرک را اعطا کرده

Emory University

امتياز متن

1997

متن يادداشت

When evaluating the protective effect of a vaccine against infection, or vaccine efficacy (VE), it is important to consider whether the protection conferred on vaccinated individuals wanes with time. We consider the case that data are available on time to infection for both vaccinated and unvaccinated individuals, thus survival analysis techniques are used to estimate the relative risk (RR) for vaccinated compared to unvaccinated subjects. To estimate possible waning of efficacy, we use the generalization usd\rm VE(t)=1-RR(t)usd, where time the relative risk and thus the vaccine efficacy are functions of time t. In the general context of estimating a time-dependent treatment effect using the Cox model (1972), we consider two methods for obtaining nonparametric estimates of RR(t) that examine the time trend in the data without imposing parametric restrictions on the functional form. One is a Poisson model for survival data with time-varying effects that is fit in the Generalized Additive Models framework (Hastie and Tibshirani 1990). The second approach is that taken by Grambsch and Therneau (1994). They use Schoenfeld residuals obtained in the counting process formulation of the Cox model to produce smooth plots that represent the functional form of a time-varying effect. We compare these two methods in terms of their graphical RR(t) functions, estimation techniques, and power to detect time-dependent treatment effects. Due to its flexibility for use as either a diagnostic or an estimation technique, we prefer the smoothed scaled Schoenfeld residuals method for diagnosing and modeling time-dependent RR(t) functions. Using smoothed scaled Schoenfeld residuals, we obtain nonparametric estimates of VE(t) from smooth RR(t) estimates for a five year study of a whole cell and a B-subunit whole cell oral cholera vaccine in Bangladesh. There are also two possible biotypes of cholera in the population, classical and el tor. In general, the B-subunit whole cell vaccine appears to confer better initial protection than the whole cell vaccine, and both vaccines appear to be more effective against the classical biotype than against the el tor biotype. Neither vaccine appears to offer protection for a period longer than about 3 years. We also explore the use of the underlying prevalence function in computing the baseline hazard for infectious disease survival models.

موضوع مستند نشده

Bangladesh

موضوع مستند نشده

Biological sciences

موضوع مستند نشده

Biostatistics

موضوع مستند نشده

cholera

موضوع مستند نشده

epidemiology

موضوع مستند نشده

Health and environmental sciences

مستند نام اشخاص تاييد نشده

I. M. Longini, Jr.

مستند نام اشخاص تاييد نشده

L. K. Durham

نام الکترونيکي

فرمت انتشار

p

نوع ماده

[Thesis]

کد کاربرگه

276903

سطح دسترسي

a

تكميل شده

Y