Objective: The aim was to study the reliability of the available data, stemming from the concordance between two different observers' measurement of the level of free residual chlorine.
Design: An observational, crossover study, using measurements of the level of free residual chlorine repeated on the same day by two different observers.
Setting: Health Area of Albacete.
Patients and other participants: 176 double measurements of chlorine, carried out during the study period in 69 networks distributing different waters.
Measurements and main results: Concordance between observers on the absence or presence of chlorination was low (I.Kappa = 0.261), showing a bias in favour of the presence of chlorination in the measurements made by local technicians (McNemar p < 0.01). Applying a logistic regression model, it was observed that the risk of disagreement was as much as 8.5 times greater when chlorine measurement required the local technician to travel.
Conclusions: We concluded that the data available was of questionable value and that it would be useful to put forward some new proposals to improve the supervision of the network.