Johnston et al developed a model for determining if symptoms occurred by chance or is associated with gastroesophageal reflux disease (GERD). This uses an expression for the binomial distribution to describe the events. The authors are from the Royal Victoria Hospital and Ulster Hospitals in Belfast, Ireland.


A key definition is that any symptom occurring during a reflux episode or within 2 minutes of its ending is considered reflux-related.


Patient testing: esophageal pH monitoring



(1) number of minutes recorded

(2) number of minutes pH < 4

(3) number of reflux episodes

(4) number of reflux-related symptoms

(5) total number of symptoms during monitoring


p =

= ((total number of minutes pH < 4) + (2 * (number of reflux episodes))) / (total number of minutes recorded)


The binomial distribution was used to determine the probability of the symptoms being a chance occurrence. If the probability is < 0.05 then it is not considered a chance occurrence.


probability of "x" symptoms occurring with acidity by chance =

= FACTORIAL(total number of symptoms) / (FACTORIAL(x) * FACTORIAL((total number) - (x))) * (p^(x)) * ((1 - (p))^((total number) - (x)))


probability of "x" or more symptoms occurring with acidity by chance =

= (probability of "x") + (probability of "x+1") + …. + (probability of (total number of symptoms))



• In Appendix B the data uses the results for "x" or more symptoms. However, since only "x" reflux symptoms occurred I am not sure why the probabilities for additional symptoms are included.


To read more or access our algorithms and calculators, please log in or register.