An method on dealing with field failure data that are heavily censored(failures recorded as 0.5 -1.5% for each month). The data consists of failures for the past 28 months. Need to estimate failures for the next 12 months.

However, there are a number of other survival analysis methods that are used in biomedical research and you could just as well use any of these: http://www.statsoft.com/textbook/stsurvan.html

plot the hazard function and see whether it looks like the hazard function for any theoretical distribution. if yes ( you are lucky) then that should guide your analysis. for instance if the hazard is weibull, then you assume weibull distribution and then go forward from there.