In endodontic literature, the so-called success rate of conventional r
oot canal treatment is reported to range between 70% and 95%. This has
been calculated as the percentage of successfully treated teeth of al
l teeth followed up or included in the clinical trial. This approach,
however, does not allow for valid assertions on the prognosis of root
canal treatment as the individual observation times are not considered
. This article discusses some methodological and statistical aspects o
f how to design a prognostic study which focuses on the outcome of end
odontic therapy and of how to analyse the data appropriately. Methodol
ogically, the response variable should preferably be the individual ti
me required for the occurrence of an event, e.g., success or failure o
f endodontic therapy which should clearly be defined on the basis of w
idely accepted criteria in endodontology. Event times can appropriatel
y be analysed by the Kaplan-Meier method, which estimates the probabil
ity that the event will not occur within a fixed time. This probabilit
y, together with the approximate 95% confidence interval (CI), permits
an evaluation of the prognosis of a particular treatment. Two data se
ts were re-analysed to clarify the rationale behind the analysis of ev
ent times. Accordingly, the probability that an endodontically induced
lesion will completely heal, e.g., within the first 3 years after roo
t canal therapy ranges between 0.87 (CI: 0.74-1.00) and 0.89 (CI: 0.80
-0.98). In this situation, the simple calculation of success rates wou
ld overestimate the chance of complete periapical healing within the f
irst years after therapy but underestimate it over longer observation
periods. Another example was used to analyse the time to occurrence of
periapical pathosis associated with root canal treated teeth not dise
ased periapically. In this case the chance of recording a successful e
ndodontic treatment is initially underestimated by the percentage of s
uccessful cases of all teeth integrated in the study but is markedly o
verestimated for longer observation periods. Potential risk factors af
fecting the outcome of the endodontic therapy and thereby the event ti
mes can adequately be determined by applying the Cox's or Aalen's regr
ession model.