Basic Data Analysis for Time Series with R

18 AN ANALYSIS OF SOME PROSTATE AND BREAST CANCER DATA

18.1 BACKGROUND

It has been observed that both breast cancer and prostate cancer are rarer near the equator and more prevalent toward the poles. This is a worldwide phenomenon. Some of the research done by Dr. St Hilaire, Rakesh Mandal, and others highlights this pattern and gives a plausible explanation (Mandal et al., 2009; St-Hilaire, 2010, 2011).

These biologists suspect that pollutants combined with meteorological conditions cause areas far from the tropics to collect pollutants. The nature of prostate cancer and breast cancer are such that both could be caused by certain pollutants known as endocrine disruptors.

Some of the work involves:

1. Showing that male prostate cancer and female breast cancer have an uncanny association (and part of this is showing cancers, in general, have little as- sociation).
2. Showing ER+ (estrogen response positive) and ER− (estrogen response negative) breast cancers have different prevalence patterns and that ER+ breast cancer seems to fit the “pollution-caused” model, while ER− breast cancer does not. More specifically, endocrine disruptors (and pollution in general) and ER+ breast cancer are both on the rise, while ER− breast cancer seems to be level over time.

Each of these data sets indicates a way in which small sample sizes can be handled in the context of time series. Generally, time series analysis is based on large data sets, but reality does not always furnish large data sets. ...