When reviewing studies in the field of endodontic epidemiology, it is apparent that the most common individual studied is the one who seek dental care at a dental school and have been examined with full-mouth radiographs and/or panoramic x-rays. This is a convenient approach since researchers do not have to make an effort to invite individuals to the examination. Also, if individuals examined are exposed to radiation in the context of seeking dental care, there will be a lesser ethical dilemma as to whether the radiological examination was justified or not. However, it is reasonable to assume that individuals seeking dental care are not representative for the whole population. This assumption may be especially true for patients seeking dental care at a dental school. These patients may have more extensive treatment needs and may have smaller financial resources than the population as a whole. Thus, in order to render samples representative for the population, researchers should consider other approaches such as studying a randomized sample of individuals. If large enough, the randomized sample may be regarded as representative for the population from which it was sampled. Studies using randomized samples are in minority in endodontic epidemiology. Despite the methodological considerations discussed, the number of studies using convenience sampling is still in majority.

Does the periapical destruction always represent apical periodontitis ? In a clinical context non-root-filled teeth with a periapical destruction may be examined with regard to pulpal sensitivity to confirm the diagnosis. In an epidemiological study exclusively based on radiographs, this information is lacking but it is usually considered uncontroversial since the most probable diagnosis is apical periodontitis [3]. Older studies investigating root-filled teeth and apical periodontitis, more often than today, used apical radiographs. During recent decades panoramic radiographs have emerged as a simpler and more economical technique. Also, and more important, it exposes the individual with a lower radiation dose compared to a full mouth examination using apical radiographs. In the context of epidemiological studies, the panoramic radiograph has been demonstrated as reasonably effective as the apical radiograph when apical periodontitis is studied [4]. It performs worse than the apical radiograph when root filling quality is studied. Cone beam CT (CBCT) is a rather new technique which has shown to be promising as an adjunct in endodontic diagnostics. However, it has not been used in endodontic epidemiology other than in a few studies [5, 6]. It may be viewed as doubtful if it is justifiable to expose healthy individuals with a much larger radiation dose (compared to a full mouth examination) when studying the prevalence and frequency of apical periodontitis. Apical periodontitis is a prevalent condition and not life-threatening other than for selected patients. It may thus be argued that CBCT is not suited for screening. Others claim that it is justifiable and advocate the use of CBCT in epidemiological studies, highlighting the drawbacks with two-dimensional techniques [5].

It is acknowledged that when studying apical periodontitis or root filling quality, researchers have to consider a variation both between observers and also within observers over time [7]. This is often referred to as inter- and intraobserver variation, respectively. In order to cope with the problem, two different strategies have been developed to reduce observer variation when studying apical periodontitis.

Reit and Gröndahl [8] suggested that observer variation may be reduced if the number of false-positive findings is kept to a minimum by instructing the observers to only register a periapical destruction when certain, introducing a five-scale index where score 1 = “periapical destruction of bone definitely not present” and 5 = “periapical destruction of bone definitely present.” This index is also called the PRI-index (probability index). Reit [9] found that calibration of observers had only limited benefits in reducing observer variation.

Örstavik et al. [10] presented the periapical index (PAI) . In contrast to Reit [9] it is proposed that observers should be calibrated, and in contrast to Reit and Gröndahl [8] there is no overall strategy to reduce false-positive findings. PAI works as follows: observers are presented with a five-graded scale with radiographs with different periapical expressions ranging from periapical health (score 1) to an aggravating periapical condition (score 2–5) (Fig. 2.1). The periapical expressions on the radiographs have been validated with the histological periapical expression in a previous study using biopsies from an autopsy material [3]. For the purpose of calibration of observers to PAI, observers are instructed to use the scale when observing 100 radiographs. When in doubt, observers are instructed to assign a higher score. The reason for this is findings from Brynolf [3] where the histological periapical expression always was more severe than the radiological periapical expression. The registrations are then compared to a “golden standard” constructed by a panel of observers who have assigned “true” scores to all 100 teeth. If the observer variation is low enough compared to the “golden standard,” observers may use PAI in their study.

An alternative to PAI, when defining a healthy and diseased periapex, modified Strindberg criteria may be used [7, 11]. There is only a text reference to describe the periapical expression (Table 2.2).