What is B-SHADE
Sampling is a very important method used to understand and master the population. It has been widely applied in various domains such as natural resources, environmental pollution, and public health. After collecting sample data, some kind of estimation about the population (for example, mean and sum) is made with an appropriate model. Usually, a best and unbiased estimation is expected. However, if the samples were not carefully selected under the model's assumption, then the estimated result is biased from the population's real value. For example, when setting sentinels to estimate a disease's prevalence or incidence in a city, hospitals with good equipment and doctors are more likely to be selected by planners. Then the sentinels' average visitor number would be much higher than the real average visitor number of all hospitals. By considering the correlation between samples and their population, the B-SHADE (Biased Sentinel Hospital based Area Disease Estimation) model can generate an unbiased estimation of the population with biased samples. Although originally designed for biased sentinel hospitals' patient number/incidence estimation, it is a common method for biased samples’ population estimation.
How to get it
A package is developed for the B-SHADE model. It implements two important functions (a) statistical inference from samples and (b) optimal sample selection. The package installer can be downloaded here. It can be used and distributed freely.
System requirements to run the package:
(1) Microsoft Windows XP or later.
(2) Microsoft .Net Framework 2.0 or later.
How to use it
The package is concentrated, which means only the most necessary functions are included. Even non-experts easily use it. Here, a case is presented to show how to use the B-SHADE model package. (Manual, Case)
Reference
The B-SHADE model was originally published in PLoS ONE:
Wang J-F, Reis BY, Hu M-G, Christakos G, Yang W-Z, et al. (2011) Area Disease Estimation Based on Sentinel Hospital Records. PLoS ONE 6(8): e23428. doi:10.1371/journal.pone.0023428. (pdf)
Contact Us
Email: humg@lreis.ac.cn (Maogui Hu)