Endodontic retreatment decisions: no consensus

S. Aryanpour1, J.-P. Van Nieuwenhuysen1 & W. D'Hoore2

1

Department of Operative Dentistry, School of Dental Medicine and Stomatology; and 2Department of Hospital Administration, Universite Catholique de Louvain, Belgium

Abstract
Aryanpour S, Van Niewenhuysen J-P, D'Hoore W.
Endodontic retreatment decisions: no consensus. International Endodontic Journal, 33, 208218, 2000.

making a decision, and the technical complexity of the retreatment procedure. Results The results indicate wide inter- and also intra-school disagreements in the clinical management of root canal treated teeth. Analysis of variance showed that the main source of variation was the `school effect', explaining 1.8% (NS) to 18.6% (P < 0.0001) of the treatment variations. No other factor explained as much variance. Decision difficulty was moderately correlated to technical complexity (Pearsons' r ranging from 0.19 to 0.35, P < 0.0001). Conclusions No clear consensus occurred amongst and within dental schools concerning the clinical management of the 14 cases. The lack of consensus amongst schools seems to be due mainly to chance or uncertainty, but can be partly explained by the `school effect'. Keywords: behavioural science, decision making, endodontic retreatment. Received 14 October 1998; accepted 17 June 1999

Aim The objectives of the present study were to: (i) evaluate the consensus, if any, amongst dental schools, students and their instructors managing the same clinical cases, all of which involved endodontically treated teeth; and (ii) determine the predominant proposed treatment option. Methodology Final year students, endodontic staff members and instructors of 10 European dental schools were surveyed as decision makers. Fourteen different radiographic cases of root canal treated teeth accompanied by a short clinical history were presented to them in a uniform format. For each case the decision makers were requested to: (i) choose only one out of nine treatment alternatives proposed, from `no treatment' to `extraction' via `retreatment' and `surgery' (ii) assess on two 5-point scales: the difficulty of

Introduction
There are substantial differences in endodontic treatment outcomes between controlled studies conducted by specialists or supervised trainees and epidemiological studies based on various population groups treated in general dental practice. Controlled studies have reported success rates as high as 96% (Kerekes & Tronstad 1979, Sjo gren et al. 1990), whereas longitudinal studies of endodontic treatment in the general dental services have shown large
Correspondence: Dr S. Aryanpour, Universite Catholique de Louvain, Department of Dental Medicine and Stomatology, Avenue Hippocrate 15, 1200 Brussels, Belgium (fax: 32 (0)2 7645727).

numbers of inadequate root fillings associated with periapical disease (Petersson et al. 1986, Eckerbom et al. 1987, Eriksen et al. 1988, Petersson et al. 1989, Imfeld 1991, de Cleen et al. 1993, Buckley & Spa ngberg 1995, Saunders et al. 1997). These results suggest that poor technical standard affects the outcome of treatment in general practice and results in a growing demand for retreatment. To improve and audit the quality of endodontic treatment, guidelines have been provided to simplify endodontic retreatment decision making (British Endodontic Society 1983, American Association of Endodontists 1987, European Society of Endodontology 1994). The presence or absence of a periapical radiolu-

208

q 2000 Blackwell Science Ltd

International Endodontic Journal, 33, 208218, 2000

Aryanpour et al.

Endodontic retreatment decisions

cency and clinical signs and/or symptoms along with the radiographic quality of a root filling can be used as a means of assessing the outcome of root canal treatment. Retreatment is clearly indicated when a periapical lesion, clinical signs and/or symptoms are present (Friedman & Stabholz 1986, Lewis & Block 1988, Stabholz & Friedman 1988). Several studies have shown wide inter-individual disagreements in practitioners' management of periapical lesions associated with previously root treated teeth (Smith et al. 1981, Reit & Gro ndahl 1984, 1988, Hu lsmann 1994, Kvist et al. 1994). These variations cannot be attributed solely to differences in radiographic diagnosis of a periapical lesion or to the presence or absence of clinical signs and/or symptoms (Reit & Gro ndahl 1984, 1988, Hu lsmann 1994). The complexity of the operative procedures and the variety of treatment alternatives introduce variation into the choices of therapy (Smith et al. 1981, Kvist et al. 1994). In addition to the diagnostic and technical problems, other characteristics of practitioners (e.g. their age, clinical experience, confidence, specialty training) may influence their decision making (Reit et al. 1985, Reit & Gro ndahl 1987). The large number of the relevant factors in each study and the complexity of the decision making process itself, make data analysis and interpretation difficult (Weinstein & Fineberg 1980, Grembowski et al. 1991, Gerrity et al. 1992). There has been no research that includes both student clinicians and their instructors and makes comparison of their attitudes toward root treated teeth. Therefore, the objectives of the present study were to: (i) evaluate the consensus, if any, amongst dental schools, students and their instructors in management of the same clinical cases, all of which involve endodontically treated teeth; (ii) determine the predominant proposed treatment option and the influence, if any, of the `school effect' on decision variations amongst dental schools, taking into consideration the level of training and the country; and (iii) analyse the relationship between the technical complexity of the retreatment and the difficulty in choosing a treatment option.

schools were surveyed as decision makers during 1995 and 1996. These universities and dental schools were: University of Geneva, University of Lille, University of Lyon, University of Nancy, University of Nice, University of Reims, University of Rennes, University of Lie ge, University of Brussels and the Catholic University of Louvain. Clinical cases and radiographs Radiographs of 14 different endodontically treated teeth were carefully selected from the archives of the Department of Operative Dentistry and Endodontics from the Catholic University of Louvain, Brussels, Belgium. The cases selected represented a wide range of situations including endodontically treated teeth with: . . . . . . Subjective signs (cases 4, 7, 8, 14) Objective clinical signs (cases 2, 4, 8, 10, 14) Periapical radiolucency (cases 28, 10, 12, 14) Underfilling (cases 13, 5, 8, 1114) Overfilling (cases 46, 10) Silver point or fractured instrument (cases 2, 4, 5, 9, 13) . Post (cases 7, 8, 14) . Need of coronal restoration (cases 1, 2, 5, 9, 11, 12) Radiographs were photographed as slides and complemented by a different clinical history concerning each patient and tooth including: age, gender, complaints, clinical symptoms and restorative and/or periodontal treatment if planned. This was intended to simulate the patient's first visit to a new dentist. Procedure Decision makers from each university were gathered in the same room and, in order to standardize the terminology used, participants had a briefing concerning the treatment alternatives proposed. Individual data relating to each participant were recorded including: . Clinical experience of decision maker: student, endodontic staff member, teacher or instructor . Dental degree or diploma received: graduation year and institution . Age and gender . University or dental school . Country

Materials and methods

Decision makers Final-year students (within 1 month of graduation), endodontic faculty members and instructors (for the theoretical or/and clinical teaching of endodontics) of 10 European French-speaking universities and dental

q 2000 Blackwell Science Ltd

International Endodontic Journal, 33, 208218, 2000

209

Endodontic retreatment decisions Aryanpour et al.

Each case was presented to decision makers in a uniform format with both a radiograph and a clinical history projected onto a screen in a darkened room. Fourteen cases were presented successively in approximately 1 h and participants' decisions were recorded. For each case, decision makers were requested to: (i) Choose one out of the following nine proposed treatment alternatives: . No treatment necessary . Wait 6 to 12 months and re-examine (watchful waiting) . Selective retreatment of one or two canals . Complete retreatment of all canals . Apicectomy . Retreatment apicectomy . Root resection . Retreatment root resection . Extraction (ii) Assess the difficulty of making a decision by using a rating scale from 1 to 5 (1 easy to make a decision, 5 difficult to make a decision). (iii) Evaluate the technical complexity of the retreatment procedure using the hypothesis that they were obliged to manage the case by a nonsurgical retreatment, using a scale from 1 to 5 (1 procedure will be easy, 5 procedure will be technically difficult). Data analysis Each case was first individually analysed. To facilitate reporting of results, cases were divided into four categories according to presence of radiographic and clinical or/and subjective signs as follows: 1 2 Cases without clinical or radiographic signs (cases 1, 9, 11, 13). Cases with radiographically perceptible periapical lesion but without clinical symptoms or signs (cases 3, 5, 6, 12). Cases with a radiographically perceptible periapical lesion and only one clinical symptom or sign (cases 2, 7, 10). Cases with a radiographically perceptible periapical lesion and two clinical symptoms and/or signs (cases 4, 8, 14).

Table 1

Distribution of decision makers in the different

Students 10 37 50 37 23 10 50 12 29 30 288 Staff members and instructors 13 11 5 10 2 6 7 11 13 12 90 Total 23 48 55 47 25 16 57 23 42 42 378

schools
University Gene ve Lille Lyon Nancy Nice Reims Rennes Lie ge Brussels (ULB) Louvain (UCL) Total

The responses were coded, verified and entered into microcomputer files. The frequency of each kind of decision was computed for all subjects. One way analysis of variance was used to assess the influence of the country, university, and specialization on the four kinds of decisions. Nested models (university nested within country, and specialization nested within university) were then used to assess the relative contribution of factors. Computations were made with the Statistical Analysis System software 6.09 release (SAS Institute Inc. Cary, NC, USA).

Results
A total of 378 complete questionnaires were collected. Fifty-six per cent of participants were men and 44% were women. The age range was 2163 years with a mean of 27.7 years of age. The distribution of decision makers in the different schools is displayed in Table 1. Decision makers' treatment choices To determine the predominant proposed treatment amongst participants, we analysed the frequency of treatment choices. Table 2 presents the frequencies of participants' treatment option selections for each category of cases. As expected, the results indicate large inter-individual disagreements in participants' management of endodontically treated teeth. On average, no retreatment (52.18%) and nonsurgical retreatment (46.8%) were chosen most frequently as the appropriate alternatives for asymptomatic cases. For other cases, the presence of radiographic and/or clinical signs resulted in a more aggressive attitude, such as retreatment, surgery and

Treatment alternatives were pooled into four groups: no (re)treatment proposed, nonsurgical retreatment, surgical retreatment, extraction.

210

International Endodontic Journal, 33, 208218, 2000

q 2000 Blackwell Science Ltd

Aryanpour et al.

Endodontic retreatment decisions

Table 2

Mean frequencies of participants' treatment option selections for each category of cases
Category 1 Frequency Category 2 Frequency 0.62 (15.64) 0.97 (24.53) 1.33 (33.40) 1.05 (26.33) 4 (100) Category 3 Frequency 0.44 (15.04) 1.52 (50.96) 0.66 (22.38) 0.35 (11.85) 3 (100) Category 4 Frequency 0.02 (0.79) 0.84 (28.31) 0.73 (24.68) 1.38 (46.27) 3 (100)

No (re)treatment Nonsurgical retreatment Surgical retreatment Extraction Total

a b

2.08 (52.18) 1.86 (46.8) 0.03 (0.99) (0.06) 4 (100)

Figures in parentheses are percentages. Case categories: category 1, cases without clinical or radiographic signs; category 2, cases with radiographically perceptible periapical lesion but without other clinical symptoms or signs; category 3, cases with a radiographically perceptible periapical lesion and only one clinical symptom or sign; category 4, cases with a radiographically perceptible periapical lesion and two clinical symptoms and/or signs.

extraction, but no relationship was noted between the nature of the symptoms and selected treatment options. Contributing factors Overall, the variation in participants' treatment choice was wide. We therefore analysed whether disagreements amongst participants' treatment strategies were related to their: (a) gender; (b) years of experience; (c) hierarchical status (student, endodontic staff member, teacher, instructor); (d) difficulty of making a decision; (e) estimation of the technical complexity if a retreatment had to be performed; (f ) university or dental school; or (g) country. Correlation analysis There was no significant correlation between the gender and the decision makers' choice. Table 3 presents correlation results for four other contributing factors and proposed treatment strategies. None of the tested variables was statistically significant across all four categories of cases: Asymptomatic cases. For asymptomatic cases, decision makers with more experience chose nonsurgical retreatment more often; the age and the hierarchical status were inversely related to no (re)treatment option, that is, no therapy or watchful waiting. Technical complexity of retreatment had a negative effect on no (re)treatment rate. Participants who considered these cases as technically complex proposed more nonsurgical retreatment (conservative retreatment of one, two or all canals) than those who found these cases easy to retreat. Cases with radiographically perceptible periapical lesion but without clinical symptoms or signs. For these cases,

hierarchical status had a highly significant, positive effect on nonsurgical retreatment choice and a negative effect on surgical retreatment. Similarly, age also had a positive effect on nonsurgical retreatment rate and was inversely related to the choice of the surgical retreatment option, that is, apicectomy or root resection with or without a previous conventional retreatment. Difficulty of making a decision had a positive effect on surgical retreatment rate and an inconsistent negative effect on extraction choice. Cases with a radiographically perceptible periapical lesion and only one clinical symptom or sign. For these cases, the age and the hierarchical status had a positive effect on the extraction choice rate but a negative effect on the surgical retreatment option. Difficulty of making a decision had a positive effect on no (re)treatment and surgical options rates and a negative effect on nonsurgical retreatment choice. Cases with a radiographically perceptible periapical lesion and two clinical symptoms and/or signs. For these cases, the hierarchical status had a highly significant, positive effect on nonsurgical retreatment choice and a negative effect on surgical retreatment and extraction proposition rate. Age of the practitioner was not significant in treatment decisions. For these cases, difficulty of making a decision was positively associated with nonsurgical retreatment and inversely related to extraction option. Technical complexity of retreatment had an inconsistent positive effect on nonsurgical retreatment rate and a negative one on the surgical option. In general, decision makers with more experience and higher hierarchical status chose nonsurgical retreatment more often (except for cases with a radiographically perceptible periapical lesion and only one

q 2000 Blackwell Science Ltd

International Endodontic Journal, 33, 208218, 2000

211

Endodontic retreatment decisions Aryanpour et al.

Case categories: category 1, cases without clinical or radiographic signs; category 2, cases with radiographically perceptible periapical lesion but without other clinical symptoms or signs; category 3, cases with a radiographically perceptible periapical lesion and only one clinical symptom or sign; category 4, cases with a radiographically perceptible periapical lesion and two clinical symptoms and/or signs. Treatment alternatives: N, no (re)treatment; R, nonsurgical retreatment; S, surgical retreatment; E, extraction. 8P < 0:05; *P < 0:01; **P < 0:001; ***P < 0:0001, and NS as nonsignificant (P > 0:05).

Pearson correlation coefficients for variables: age, hierarchical status, decision difficulty, technical complexity and treatment alternatives in each of the four case categories

clinical symptom or sign). For asymptomatic cases the age and the hierarchical status were inversely related to the no (re)treatment option. Analysis of variance To assess the influence of the country, university, and specialization (hierarchical status) on the four kinds of decisions one way analysis of variance was used. Analysis of variance showed that the `school effect', could explain 1.8% (NS) to 18.6% (P < 0.0001) of the treatment variations (Table 4). This means that therapeutic choices vary significantly amongst universities. Two exceptions `no (re)treatment' in category 4 and extraction in category 1 are due to a very small proportion of such atypical decisions. Nested (or hierarchical) models were used to assess the effects of country and university (Table 5), and hierarchical status and university (Table 6). In both analyses, we considered that a local factor was nested within a more general factor. Table 5 shows that the principal source of variation was the `university'. Except in cases with radiographically perceptible periapical lesion but without other clinical symptoms or signs; the variable `country' did not explain variation in treatment decisions. In Table 6, analyses show again that `university' was the main source of variation, except in asymptomatic cases and cases with radiographically perceptible periapical lesion but without other clinical symptoms or signs, where hierarchical status explained a significant amount of variance. Within `universities' experienced and less experienced subjects may not make the same decisions, especially in cases where fewer signs were present. Universities' treatment choices Figure 1 presents an overview of the frequencies of different universities' treatment choices for each group of cases. As can be seen in Fig. 1. (a-d), the variations in schools' retreatment policy were large. For symptomatic cases the predominant proposed strategy was intervention (nonsurgical retreatment, surgical retreatment, or extraction). Consensus in terms of type of treatment, however, was low. Technical complexity and uncertainty Participants evaluated the technical complexity of a hypothetical retreatment case and the difficulty of

Category 4

Category 3

Category 2

Category 1

212

International Endodontic Journal, 33, 208218, 2000

Age Hierarchical status Decision difficulty Technical complexity

Table 3

0.25*** 0.30*** NS 0.16*

0.25*** 0.31*** NS 0.16*

NS NS NS NS

NS NS NS NS

NS 0.13* NS NS

0.14* 0.20*** NS NS

0.14* 0.19** 0.16* NS

NS NS 0.138 NS

NS NS 0.14* NS

NS NS 0.19** 0.138

0.17** 0.16** 0.128 NS

0.15* 0.128 NS NS

NS 0.14* NS NS

NS 0.19*** 0.22*** 0.118

NS 0.118 NS 0.118

NS 0.128 0.22*** NS

q 2000 Blackwell Science Ltd

Aryanpour et al.

Endodontic retreatment decisions

Table 4

Importance of the 'school effect` on treatment choice: percentage of variance explained by university affiliation of participants
No (re)treatment % Nonsurgical retreatment % 8.3*** 10.6*** 9.0*** 18.5*** Surgical retreatment % 5.8* 16.5*** 10.3*** 10.4*** Extraction % 1.8 (NS) 15.3*** 11.7*** 18.6***

Category Category Category Category

1 2 3 4

9.9*** 10.8*** 9.2*** 4.68

Case categories: category 1, cases without clinical or radiographic signs; category 2, cases with radiographically perceptible periapical lesion but without other clinical symptoms or signs; category 3, cases with a radiographically perceptible periapical lesion and only one clinical symptom or sign; category 4, cases with a radiographically perceptible periapical lesion and two clinical symptoms and/or signs. 8P < 0.05, * Significant at P < 0.01,*** Significant at P < 0.0001, and NS as nonsignificant P > 0.05.

choosing an option for all 14 cases. Technical complexity and decision difficulty were moderately correlated (Pearsons' r ranging from 0.19 to 0.35, P < 0.0001) (Table 7). Analysis was made of the influence of decision makers' characteristics (gender, years of experience, and hierarchical status) on their ratings of difficulty on making a decision and the technical complexity in each of the four case categories (Table 8). None of the decision makers' characteristics were statistically significant across all situations. Technical complexity was rated higher by women than by men. Similarly, years of experience and hierarchical status
Table 5

were positively correlated to ratings of technical complexity. Female participants showed higher rates of decision difficulties than males (except for asymptomatic cases). Participants with high hierarchical status reported less difficulties when making a decision than others (except for cases with a radiographically perceptible periapical lesion and two clinical symptoms and/or signs).

Discussion
Considerable interindividual variations in clinical management of endodontically treated teeth were

Effects of country and university on treatment strategy variations (university nested within country)
Effects on variations R (%)
2

Country NS NS NS NS

University *** ** * NS *** 8 *** *** *** ** *** *** NS *** *** ***

Cases without clinical or radiographic signs No (re)treatment Nonsurgical retreatment Surgical retreatment Extraction

8.2** 8.3** 5.9* NS

Cases with radiographically perceptible periapical lesion but without other clinical symptoms or signs No (re)treatment 10.8*** NS Nonsurgical retreatment 10.6*** *** Surgical retreatment 16.5*** NS Extraction 15.4*** ** Cases with a radiographically perceptible periapical lesion and only one clinical symptom or sign No (re)treatment 9.3*** 8 Nonsurgical retreatment 9.1*** 8 Surgical retreatment 10.3*** NS Extraction 11.7*** NS Cases with a radiographically perceptible periapical lesion and two clinical symptoms and/or signs No (re)treatment 4.78 8 Nonsurgical retreatment 18.6*** ** Surgical retreatment 10.5*** 8 Extraction 18.7*** NS

8P < 0.05, * Significant at P < 0.01,*** Significant at P < 0.0001, and NS as nonsignificant P > 0.05.

q 2000 Blackwell Science Ltd

International Endodontic Journal, 33, 208218, 2000

213

Endodontic retreatment decisions Aryanpour et al.

Table 6

Effects of hierarchical status and university on treatment strategy variations (hierarchical status nested within university)
Effects on variations R2 (%) University ** ** 8 NS Hierarchical status *** *** NS NS ** ** *** *** NS NS NS * *** NS NS NS

Cases without clinical or radiographic signs No (re)treatment Nonsurgical retreatment Surgical retreatment Extraction

20.5*** 20.3*** 8.48 NS

Cases with radiographically perceptible periapical lesion but without other clinical symptoms or signs No (re)treatment 17.9*** ** Nonsurgical retreatment 18.3*** ** Surgical retreatment 26.5*** *** Extraction 21.6*** *** Cases with a radiographically perceptible periapical lesion and only one clinical symptom or sign No (re)treatment 12.2** ** Nonsurgical retreatment 11.9** ** Surgical retreatment 13.6*** * Extraction 18.2*** *** Cases with a radiographically perceptible periapical lesion and two clinical symptoms and/or signs No (re)treatment 14.8*** *** Nonsurgical retreatment 22.4*** *** Surgical retreatment 13.3*** *** Extraction 22.5*** ***

8P < 0.05, * Significant at P < 0.01,*** Significant at P < 0.0001, and NS as nonsignificant P > 0.05.

shown in the present study. Similar results have been reported by other authors (Smith et al. 1981, Reit & Gro ndahl 1984, 1988, Hu lsmann 1994, Kvist et al. 1994). As expected, the presence of radiographic and/or clinical symptoms resulted in a more aggressive attitude, such as retreatment, surgery and extraction, but no relationship was noted between the nature of the symptoms and the selected treatment options. These results confirm the findings of other studies, which showed that some participants chose the `nonintervention' alternative not only for cases with radiographic evidence of periapical disease (Reit & Gro ndahl 1988), but also for cases with clinical symptoms (Hu lsmann 1994). Obviously other clinical considerations, such as the size of the periapical radiolucency (Reit & Gro ndahl 1984, Reit et al. 1985), the quality of the previous root canal filling, time since the placement of root filling and the presence of a post (Kvist et al. 1994), will also play a role in treatment selection. The symptom-based categorization used in the present study did not take these factors into consideration. Practitioners' characteristics may influence treatment decisions (Young 1987, Gerrity et al. 1992).

Studies in medicine suggested that practitioners' personal variables, such as gender and age, do not have a significant effect on their clinical decision making (Goldman et al. 1990, Gerrity et al. 1992). Our data confirmed that there was no significant correlation between the gender and the decision makers' choices. The factors `years of experience' (age) and `hierarchical status' (student, endodontic staff member, teacher or instructor) were linked in the present study, therefore age alone was not considered just as a personal variable. Since in this study older practitioners were endodontic staff members, teachers and instructors with a high level of expertise in endodontics compared with students, we examined the influence of their `specialization' (years of experience and hierarchical status) on the treatment strategy. In previous studies (Smith et al. 1981, Reit et al. 1985), general dental practitioners and endodontists' attitudes to treatment of asymptomatic periapical lesions in endodontically treated teeth were compared. They demonstrated the `specialization' influence on treatment decisions and also as an explanation for variations in assessement of the probabilities of disease and future complications. It was also noted that the

214

International Endodontic Journal, 33, 208218, 2000

q 2000 Blackwell Science Ltd

Aryanpour et al.

Endodontic retreatment decisions

100 90 80 70 60 50 40 30 20 10 0 No retreatment Nonsurgical retreatment Surgical retreatment Extraction

100 90 80 70 60 50 40 30 20 10 0 No retreatment Nonsurgical retreatment Surgical retreatment Extraction

Frequencies (%)

Frequencies (%)

10

10

(a)
100 90 80 70 60 50 40 30 20 10 0

Universities

(c)
100 90 No retreatment Nonsurgical retreatment Surgical retreatment Extraction 80 70 60 50 40 30 20 10 0

Universities

10

Frequencies (%)

Frequencies (%)

No retreatment Nonsurgical retreatment Surgical retreatment Extraction

10

(b)

Universities

(d)

Universities

Figure 1 Frequencies of universities' treatment choices in percentage for each group of cases: (a) cases without clinical or radiographic signs; (b) cases with radiographically perceptible periapical lesion but without other clinical symptoms or signs; (c) cases with a radiographically perceptible periapical lesion and only one clinical symptom or sign; (d) cases with a radiographically perceptible periapical lesion and two clinical symptoms and/or signs.

endodontists did subject smaller and medium-sized lesions to therapeutic measures more often than did the general practitioners. In the present study, for cases without clinical or radiographic signs, it was noted that years of experience and hierarchical status were inversely related to the no (re)treatment option and also had a positive effect on the nonsurgical retreatment rate. Overall it was found that decision makers with more experience and higher hierarchical status chose nonsurgical retreatment more often except for cases with a radiographically perceptible periapical lesion and only one clinical symptom. For these cases, years of experience and hierarchical status were related to the surgical retreatment option. These correlations have relatively small magnitudes and demonstrate that more experience and higher hierarchical status do not

necessarily lead to better consensus regarding the proposed treatment. Difficulty in making a decision and technical complexity were not statistically significantly correlated across all case categories. A possible explanation for the absence of constant and statistically significant effects of the tested variables might be that the symptom-based categorization of cases used in this study did not take into consideration some clinical considerations contributing to the practitioner's treatment choice. Pathology and case specifics, together with practitioners' characteristics, interact and are confounded with other factors, which make the clinical decision making process so complex. Analysis of variance showed that the main source of variation was the `school effect', explaining 1.8% (NS) to 18.6% (P < 0.0001) of the treatment variations, no other factor explaining as much variance. This study

q 2000 Blackwell Science Ltd

International Endodontic Journal, 33, 208218, 2000

215

Endodontic retreatment decisions Aryanpour et al.

Table 7

Correlation between technical complexity and decision difficulty for each of four case categories
Pearson correlation coefficients (r)* 0.32 0.19 0.19 0.35 Number of observations 376 378 377 368

Case categories Cases without clinical or radiographic signs Cases with radiographically perceptible periapical lesion but without other clinical symptoms or signs Cases with a radiographically perceptible periapical lesion and only one clinical symptom or sign; category Cases with a radiographically perceptible periapical lesion and two clinical symptoms and/or signs

*P < 0:0001 Table 8

Pearson correlation coefficients of three participants' variables (gender, age, hierarchical status) and technical complexity and decision difficulty in each of the four case categories
Category 1 Decision making Technical complexity NS NS NS Category 2 Decision making NS 0.18** 0.17** Technical complexity NS NS NS Category 3 Decision making 0.108 NS 0.15* Technical complexity NS NS 0.17** Category 4 Decision making 0.128 NS 0.128 Technical complexity NS 0.118 0.22***

Gender Age Hierarchical status

0.118 NS NS

Case categories: category 1, cases without clinical or radiographic signs; category 2, cases with radiographically perceptible periapical lesion but without other clinical symptoms or signs; category 3, cases with a radiographically perceptible periapical lesion and only one clinical symptom or sign; category 4, cases with a radiographically perceptible periapical lesion and two clinical symptoms and/or signs. Gender: male 1, female 0. 8P < 0.05; *P < 0.01; **P < 0.001; ***P < 0.0001, and NS as nonsignificant (P > 0:05).

showed no clear consensus recorded amongst dental schools concerning the clinical management of the 14 proposed cases, but there was a `school effect', i.e. some consensus within dental schools. An explanation for the `school effect' was that students from the same university have the same undergraduate curriculum on endodontics. This curriculum is not uniform across countries and dental schools. Final year students with little clinical experience manage their uncertainty of diagnostic and treatment decision making by: deciding not to decide, requesting more tests, adopting their teacher's or instructor's recommendations or `school of thought'. Clinical training by different instructors may be a possible explanation for disagreements within dental schools. This could explain the lack of accepted criteria for retreatment decisions within dental schools. As expected, participants' ratings of technical complexity of retreatment procedures and difficulty of making a decision were moderately correlated (Pearsons' r ranging from 0.19 to 0.35, P < 0.0001). None of the tested decision makers' characteristics (gender, age, hierarchical status) was statistically significant across all situations. The 14 cases used in the study were carefully selected to represent a wide range of clinical situations,

including endodontically treated teeth with or without radiographic evidence of periapical lesions and/or clinical symptoms and also with varying quality of root canal filling and coronal restoration. There is no reason to think that another selection could have led to different results. In order to make it possible to analyse the effects of radiographic and clinical symptoms on participants' decisions, it was decided to restrict the case classification to the nature and the number of symptoms: asymptomatic cases (without clinical or radiographic signs); those with radiographically perceptible periapical lesion but without clinical symptoms; those with a radiographically perceptible periapical lesion and only one clinical symptom; and those with a radiographically perceptible periapical lesion and two clinical symptoms and/or signs. Similar results were observed with analysis of individual cases. The use of radiograph-based clinical cases to evaluate practitioners' attitudes toward root canal treated teeth is not new. In contrast, there has been no research that included both student clinicians and their teachers and instructors managing the same clinical cases, all involving endodontically treated teeth.

216

International Endodontic Journal, 33, 208218, 2000

q 2000 Blackwell Science Ltd

Aryanpour et al.

Endodontic retreatment decisions

Such a radiograph-based clinical case study does not reflect the practitioner's decision making process in real life for two reasons (Jones et al. 1990). First, there is a considerable difference between what is learned from dental textbooks, what is expressed in a survey and what is actually done in clinical practice. Incorporation of clinical elements in theoretical concepts is a complex process and may explain these variations. Secondly, clinical decision making is a multifactorial problem. Practitioners consider several contributing factors when prescribing treatments and a radiographbased clinical case does not provide all the required elements. Multidisciplinary research is required if the effects of these contributing factors on the decision making process is to be understood. This study is therefore inadequate in identifying all of the causes of decision variations, but it provides evidence regarding the consensus (`school effect') and the disagreement rate amongst and within dental schools. The quantitative analysis of the `school effect' can be used in dental education as a tool in estimating the impact of teachers' opinions and teaching of endodontics on students' treatment decisions. In addition, further comparative research is needed to examine whether this `school effect' remains stable over time. The lack of consensus amongst and within dental schools, as shown by the results of the present study, indicates that there is an important need for uniformity and improvement in the undergraduate curriculum on endodontics. More emphasis should be given to endodontic retreatment, and established guidelines for root canal treatment must be incorporated into the undergraduate curriculum.

health care professionals to determine the most appropriate treatment strategy for a given clinical problem. It is hoped that the incorporation of these evidence-based methods and established guidelines for root canal treatment into the undergraduate curriculum will reduce the decision variation margins within and between dental schools.

Acknowledgements
We acknowledge the helpful assistance and cooperation of the endodontic staff members of the participating universities, particularly those of Professors B. Ciucchi (Geneva), A. Claisse (Lille), P. Farge (Lyon), M. Panighi (Nancy), J.-P. Rocca (Nice), J.-P. Camus (Reims), J.-M. Vulcain (Rennes), J. Charpentier (Lie ge), and T. Charles (Brussels).

References
American Association of Endodontists (1987) Quality assurance guidelines. Meeting of the American Association of Endodontists. Chicago, IL, USA: American Association of Endodontists, 127. British Endodontic Society (1983) Guidelines for root canal treatment. International Endodontic Journal 16, 1925. Buckley M, Spa ngberg LSW (1995) The prevalence and technical quality of endodontic treatment in an American subpopulation. Oral Surgery, Oral Medicine and Oral Pathology 79, 92100. de Cleen MJH, Schuurs AHB, Wesselink PR, Wu M-K (1993) Periapical status and prevalence of endodontic treatment in an adult Dutch population. International Endodontic Journal 26, 1129. Eckerbom M, Andersson J-E, Magnusson T (1987) Frequency and technical standard of endodontic treatment in a Swedish population. Endodontics and Dental Traumatology 3, 2458. Eriksen HM, Bjertness E, Orstavik D (1988) Prevalence and quality of endodontic treatment in an urban adult population in Norway. Endodontics and Dental Traumatology 4, 1226. European Society of Endodontology (1994) Consensus report of the European Society of Endodontology on quality guidelines for endodontic treatment. International Endodontic Journal 27, 11524. Friedman S, Stabholz A (1986) Endodontic retreatment. Case selection and technique. Part 1: criteria for case selection. Journal of Endodontics 12, 2833. Gerrity MS, Earp JAL, Devellis RF, Light DW (1992) Uncertainty and professional work: perceptions of physicians in clinical practice. American Journal of Sociology 97, 102251.

Conclusion
In conclusion, no clear consensus was recorded between dental schools concerning the clinical management of the 14 proposed cases, but there was a `school effect', i.e. some consensus within dental schools. The lack of consensus between schools appears to be due mainly to chance or uncertainty. Technical complexity explained little interindividual variation in the treatment choice and was related to decision difficulty. The multiplicity and the complexity of all the contributing factors which come into play when a clinical decision is made show how difficult it is to reduce the interindividual variation. Recently, decisional methods based on research of the best clinical evidence have been developed and used by

q 2000 Blackwell Science Ltd

International Endodontic Journal, 33, 208218, 2000

217

Endodontic retreatment decisions Aryanpour et al.

Goldman G, Pineault R, Bilodeau H, Blais R (1990) Effects of patients, physician, and hospital characteristics on the likelihood of vaginal birth after previous Caesarean section in Quebec. Canadian Medical Association Journal 143, 1017 24. Grembowski D, Milgrom P, Fiset L (1991) Dental decision making and variation in dentist service rates. Social Science and Medicine 32, 28794. Hu lsmann M (1994) Retreatment decision making by a group of general dental practitioners in Germany. International Endodontic Journal 27, 12532. Imfeld TN (1991) Prevalence and quality of endodontic treatment in an elderly urban population of Switzerland. Journal of Endodontics 17, 6047. Jones TV, Gerrity MS, Earp JA (1990) Written case simulations: do they predict physicians' behavior? Journal of Clinical Epidemiology 43, 80515. Kerekes K, Tronstad L (1979) Long-term results of endodontic treatment performed with a standardized technique. Journal of Endodontics 5, 8390. Kvist T, Reit C, Esposito M et al. (1994) Prescribing endodontic retreatment: towards a theory of dentist behaviour. International Endodontic Journal 27, 28590. Lewis RD, Block RM (1988) Management of endodontic failures. Oral Surgery, Oral Medicine and Oral Pathology 66, 71121. Petersson K, Lewin B, Hakansson J, Olsson B, Wennberg A (1989) Endodontic status and suggested treatment in a population requiring substantial dental care. Endodontics and Dental Traumatology 5, 1538. Petersson K, Petersson A, Olsson B, Hakansson J, Wennberg A (1986) Technical quality of root fillings in an adult

Swedish population. Endodontics and Dental Traumatology 2, 99102. Reit C, Gro ndahl HG (1984) Management of periapical lesions in endodontically treated teeth. Swedish Dental Journal 8, 1 7. Reit C, Gro ndahl HG (1988) Endodontic retreatment decision making among a group of general practitioners. Scandinavian Journal of Dental Research 96, 1127. Reit C, Gro ndahl HG (1987) Endodontic decision-making under uncertainty: a decision analytic approach to management of periapical lesions in endodontically treated teeth. Endodontics and Dental Traumatology 3, 1520. Reit C, Gro ndahl HG, Engstro m B (1985) Endodontic treatment decisions: a study of the clinical decision-making process. Endodontics and Dental Traumatology 1, 1027. Saunders WP, Saunders EM, Sadiq J, Cruickshank E (1997) Technical standard of root canal treatment in an adult Scottish sub-population. British Dental Journal 182, 3826. Sjo gren U, Ha gglund B, Sundqvist G, Wing K (1990) Factors affecting the long-term results of endodontic treatment. Journal of Endodontics 16, 498504. Smith JW, Crisp JP, Torney DL (1981) A survey: controversies in endodontic treatment and re-treatment. Journal of Endodontics 7, 47783. Stabholz A, Friedman S (1988) Endodontic retreatment. Case selection and technique. Part 2: treatment planning for retreatment. Journal of Endodontics 14, 60714. Weinstein MC, Fineberg HV (1980) Clinical Decision Analysis, 1st edn. Philadelphia, PA, USA: W.B. Saunders Company. Young EH (1987) Relationship of residents' emotional problems, coping behaviour and gender. Journal of Medical Education 62, 64250.

218

International Endodontic Journal, 33, 208218, 2000

q 2000 Blackwell Science Ltd