T. Kvist & C. Reit
Department of Endodontology and Oral Diagnosis, Faculty of Odontology, Göteborg University, Gothenburg, Sweden.

Introduction.
Several investigators have focused attention on the substantial variation amongst clinicians in the management of symptom-free periapical lesions in endodontically treated teeth (Smith et al. 1981, Reit & Gröndahl 1984, 1988, Petersson et al. 1989, Hülsmann 1994). Various aspects of the endodontic retreatment decision-making process have been explored (Reit et al. 1985, Petersson et al. 1989, Reit & Kvist 1998) and a theory explaining dentist behaviour has been suggested. In a ‘Praxis Concept theory’ (PC theory), Kvist et al. (1994) proposed that dentists perceive periapical lesions of varying sizes as different stages on a continuous health scale, based on their radiographic appearance. Inter-individual variations could then be regarded as the result of the choice of different cut-off points on the continuum for prescribing retreatment.

Assessment of clinicians inclination to re-treat.
On the basis of PC theory, Kvist et al. (1994) developed a technique for assessing and numerically expressing the inclination of a dentist to propose endodontic retreatment. It was assumed that their disposition could be summed up in a ‘Retreatment Preference Score’ (RPS). The RPS is based on the judgement of five periapical conditions in root-filled teeth (Fig. 1). These conditions are described in six cases presented as line drawings of simulated radiographs. For each case the individual’s cut-off point for retreatment is determined and scored between 0 and 1. The mean value is given as the RPS. Hence, the higher the value, the higher the inclination to re-treat. The RPS has been found to be subject to large interindividual variation amongst inexperienced decision makers (Kvist et al. 1994). The situation amongst experienced endodontists is not known.
The RPS may be influenced by several factors but the PC theory suggests that an individual’s placement of the cut-off point depends to a large extent on his/her personal values. It has been proposed that the value of endodontic retreatment lies in its contribution to a person’s well-being (Reit & Kvist 1998). Well-being has been defined as ‘the fulfilment of informed desire’ (Griffin 1986) and is concerned with what an individual regards as ‘good’ in life. In the present study it was suggested that, in endodontic retreatment decision making, these values are significantly related to the ‘goodness’ of an action, that is, the benefit a retreatment procedure might bring to a patient.

Figure 1. Assessment of the retreatment preference score (RPS). Five periapical conditions were positioned on a health continuum, with 'no lesion' and 'big lesion' at either end. For each case a cut-off point on the continuum was identified separating conditions requiring retreatment from those not requiring retreatment. The score of a case could vary between 0 and 1, with a higher RPS indicating a higher retreatment preference.


Figure 2. An individual may benefit from endodontic retreatment by moving from a state with an asymptomatic lesion (health state B) to a state where the lesion has healed (health state A). The numerical difference in assigned U-values was defined as the 'retreatment benefit' (RTB).

It is thought that there is a close connection between a person’s values and his value judgements (Österman 1995). The idea of systematic inclusion of value judgements in clinical decision making was introduced by Ledley & Lusted (1959) and discussed in detail by Lusted (1968) and Weinstein & Fineberg (1980). This approach is based on rational choice theory (for reviews see Hargreaves Heap et al. 1992, Bacharach & Hurley 1994). Within this theoretical framework ‘utility values’ ( U - values) are produced. A U -value represents the value a person assigns to a certain health state. This value can be numerically expressed using certain techniques (von Neumann & Morgenstern 1947, Torrance 1986) and placed on a scale ranging from ‘perfect’ health (given a value of 1) to ‘death’ (given a value of 0). This type of scale was transposed to an endodontic setting by Reit & Kvist (1998) (Fig. 2).

Assessment of retreatment benefit.
When a patient moves from a worse to a better health state he or she benefits, from a medical point of view. Thus, a patient potentially benefits from endodontic retreatment if he/she moves from a health state with a periapical inflammation to a postretreatment situation where the lesion has healed. In this model the distance on the scale between the two health states (the numerical difference in assigned U -values) will define the ‘retreatment benefit’ (RTB). RTB has been found to vary substantially amongst inexperienced decision makers (Reit & Kvist 1998) but the situation amongst endodontic specialists is not known. If U -values constitute a significant part of the personal values, the PC theory suggests that variations in RPS could be explained by variations in RTB.
In the present study the hypothesis that the PC theory describes the decision making strategy of endodontic experts was tested. The validity of the proposition that RPS and RTB are causally related was also investigated.

Materials and methods.

Endodontic experts.
Sixteen dentists (nine women and seven men) affiliated with the Departments of Endodontology at the dental schools in Gothenburg (7), Malmö (5) and Stockholm (4), Sweden, were enrolled to serve as decision makers. The mean age of the group was 47.

Assessment of Retreatment Preference Score (RPS).
Written forms presenting data on six cases of endodontically treated teeth were provided. Line drawings of simulated radiographs depicting the quality of root fillings and the presence or absence of posts and crowns were systematically varied. In each case five periapical conditions were judged (Fig. 1). An identical clinical history accompanied each case:

• The patient is 45 years old, in good general health and presents with a full set of teeth except for third molars. There are no symptoms from the teeth or oral tissues. The ‘radiographs’ were taken at a routine examination. The root-fillings are more than 4 years old. This is your first examination of the patient, who has no other dental problems, and no further dental treatment is being planned.

The examiner was asked to select one of the following five options for each case and periapical condition: no therapy, wait and see, non-surgical retreatment, surgical retreatment, or extraction. A ‘cut-off point’ was identified for each examiner and case to separate periapical conditions that were not selected for retreatment from those selected for retreatment. The cut-off point was numerically represented in the ‘retreatment preference score’ (RPS) on a scale from 0 to 1, with the higher values indicating a higher retreatment preference. The scores were: 0 = no retreatment selected, 0.2 = cut-off point at ‘big lesion’, 0.4 = cut-off point at ‘medium size lesion’, 0.6 = cut-off point at ‘small but clearly visible lesion’, 0.8 = cut-off point at ‘widened periodontal contour’, 1 = cut-off point at ‘no lesion’ (Fig. 1). The individual mean RPS of the six cases judged was calculated.

Assessment of U –values.
Two different dental health states associated with maxillary incisors were presented to the decision makers as radiographic images printed on paper. The prints were accompanied by the following texts:

• State A : ‘Imagine that this is a tooth in your own mouth. The tooth has a good-quality root filling. There are no clinical symptoms or any radiographically visible signs of periapical pathology. The situation is not expected to change for the rest of your life.’
• State B : ‘Identical to state A except that the radiograph shows a 5-mm large periapical radiolucency.’

The standard gamble technique (von Neumann & Morgenstern 1947, Torrance 1986) was used to give each subject a choice between two alternative courses of action. The options available were to continue living in the state of health described in the scenario, or to take a gamble that had two possible outcomes: a ‘best’ outcome or a ‘worst’ outcome.
In this study, the ‘best’ outcome was a hypothetical immediate restoration of perfect pulpal and periapical health and was given a utility of 1.0. The ‘worst’ outcome was an immediate loss of the tooth, which was given a utility of 0. For each health state (A and B) the probability of attaining the best outcome was systematically varied until the subject was indifferent to continuing to stay in that health state or taking the gamble. The probabilities were varied using a questionnaire where they were graphically illustrated with pie charts, and the utility of the health state was calculated as described previously (Reit & Kvist 1998). Negative utilities were accepted (Torrance 1984), but for calculations and representation in tables and figures such assessments were set to U = 0.

Calculation of retreatment benefit (RTB).
An individual’s potential retreatment benefit was defined as the numerical difference between the U -values of health states A and B (Fig. 2).

Judgement procedures.
Procedures to obtain RPS and U -values were carried out individually in the presence of one of the authors (TK). Procedures were repeated after 1 year; the experimental layout is shown in Fig. 3.

Figure 3. Experimental setup. The Retreatment Preference Score (RPS) and the retreatment benefit (RTB) were assessed by 16 endodontists on two occasions with a one-year interval.

Statistical methods.
Descriptive measures were given as means, standard deviations (SD), medians and ranges. Spearman’s rank correlation coefficient ( r s ) was used for correlation analyses. The intrarater reliability between assessments was expressed as standard deviations of the differences between first and second assessments and limits of agreement (Bland & Altman 1986). The Wilcoxon Signed Rank Test for matched pairs was used to test systematic changes between assessments. All hypothesis tests were two-tailed and conducted at the 0.05 level of significance.

The individual mean RPS range was 0.27–0.63. Elicited utilities for health state A and B ranged from 0.65 to 0.99 and 0–0.99. On five occasions negative utilities for health state B were obtained. Calculated RTB ranged from 0 to 0.99. Mean values, standard deviations (SD), medians and ranges are displayed in Table 1.
No statistically significant correlation between RPS and RTB was detected at the first ( P = 0.92; r s = 0.03) or second judgement ( P = 0.08; r s = −0.45).
Intrarater reliability over time is displayed graphically in Fig. 4 and the means, standard deviations (SD) and 95% limits of agreement are presented in Table 2.
No statistically significant systematic change between assessments was detected for RPS ( P = 0.43), health state A ( P = 0.45), health state B ( P = 0.66) or RTB ( P = 0.61).

Table 1. Results of RPS measurements, assessments of U-values for health states A and B and calculations of RTB.


Means, standard deviations (SD), medians and ranges are given.

Figure 4. Intraindividual stability of the Retreatment Preference Score (RPS) and retreatment benefit (RTB). Each square represents one decision maker.



Table 2. Results of intraindividual reliability assessments.


Means, standard deviations (SD) and 95% limits of agreement are given.

Discussion.
Intuitively, a considered medical procedure is regarded as meaningful only if it is thought to bring about some benefit to the patient. Accordingly, the consequences of treating or not treating the disease in question must be at the core of the clinical decision-making process. However, there is an element of subjectivity in what might be experienced as beneficial, and patients differ in their preferences regarding both medical and dental treatment. Studies have also demonstrated a marked difference in the value of treatment outcomes between patients and professionals (Fyffe & Kay 1992). Therefore, it must be noted that the present study was only concerned with dentists, and that other results would probably have been obtained if patients, rather than dentists, had been studied.
The traditional classification of endodontic treatment outcome into ‘success’ and ‘failure’ leaves little room for subjectivity or deliberation over consequences. For example, the idea underlying Strindberg’s (1956) system appears to be the concept that ‘normal’ biological conditions should always be restored. Consequences and preferences are neglected in the decision-making process, whilst the focus is on the ‘duty’ of the dentist to treat cases of persistent periapical lesions in root-filled teeth. This approach to case management is illustrated by one of the participants in the present study who, despite reporting no utility gain (RTB = 0) by moving from health state B to health state A, decided to re-treat every case where a lesion was present (RPS = 0.6).
The RPS data in the present study indicate that Swedish endodontic experts are not always inclined to re-treat cases with a persistent periapical lesion. Furthermore, there does not seem to be a consensus on a definite retreatment criterion. Similar observations amongst staff and instructors at 10 French-speaking European dental schools have recently been reported by Aryanpour et al. (2000). The interindividual variation in RPS is well explained by the PC theory and data support the hypothesis that PC is also valid amongst endodontic experts.
From a utilitarian ethical point of view, resources should be used to produce as much benefit, or utility, as possible. In the present study a methodology for numerical measurement of the benefit of endodontic retreatment (RTB) was proposed. Ideally, such a method may be used to make comparisons, direct capacity and set priorities. It was found that the assessment of RTB was subjected to substantial interindividual variation. This was mainly due to the experts’ deviations in their judgement of the U -value of the persistent periapical lesion (health state B).
The natural history of periapical lesions in root-filled teeth is not well known. Information from longitudinal epidemiological studies (Lavstedt et al. 1982, Eckerbom et al. 1989, Petersson et al. 1991), repetitive cross-sectional studies (Eriksen & Bjertness 1991) and case series (Strindberg 1956, Grahnén & Hansson 1961, Engström 1964, Kerekes & Tronstad 1979, Bergenholtz et al. 1979) is inconclusive. Experts differ greatly when they are asked to predict the risk of disease progression (Reit & Gröndahl 1987) or exacerbation of the process (Reit et al. 1985). Data from epidemiological studies indicate the risk of an acute exacerbation to be less than 5% per year for cases with chronic apical periodontitis (Eriksen 1998). Root canal infection as a systemic health hazard has been debated over the years, however, controlled clinical studies are rare and most authors judge the risk for the medically uncompromised individual as low (Debelian et al. 1994, Murray & Saunders 2000).
The inclusion of U -values in the decision-making process has been suggested in the ‘expected utility theory’ (EUT) concept (von Neumann & Morgenstern 1947, Weinstein & Fineberg 1980, Torrance 1986). However, the methodology used to elicit U -values is complex (Griffin 1986, Mulley 1989) and illogical results have been reported (Revicki 1992, Zug et al. 1995, Reit & Kvist 1998). In the present study the stability of U -values over time (1 year) was investigated. Values, particularly for health state B, were found to vary considerably. The intrarater reliability problem has met only scarce recognition. Torrance (1986) judged reliability to be ‘acceptable’ but to decrease with extended time intervals. Other investigators have found repeatability to be ‘poor’ (Groome et al. 1999). Based on the data from the present study, the use of U -values in direct EUT decision-making calculations cannot be recommended.
The RPS was found to be subject to substantial intraindividual variation over time. Apart from having methodological origins, the variation may reflect changes in attitude toward retreatment. Recent technical developments (surgical microscopes, ultrasonic retrotips, new root-end filling materials, nickel–titanium instruments, rotary systems) and reconsidered retreatment strategies (Molander et al. 1998, Waltimo et al. 1999) have changed the scope of retreatment and, consequently, may exert an influence on decision making. Decision making may also fluctuate over time for non-rational reasons. As has been suggested by Kahneman et al. (1982), and explored in endodontic retreatment situations (Reit et al. 1985), individuals faced with a complex decision-making situation often rely on a limited number of heuristic principles. The results of such decisions are expected to vary over time.
No statistically significant correlation between the retreatment prescriptions of the experts, as reflected in the RPS and retreatment benefit (RTB), was found. The personal values involved could not systematically be captured in the U -values or the RTB. A dentist’s clinical decisions are influenced by his moral values (Kay & Blinkhorn 1996). Morality may be defined as a system of beliefs and values concerning the rightness or wrongness of human acts (Björklund 2000). In the present study it was hypothesized that the endodontists above all acted in terms of possible consequences, and that the more potential utility that could be produced, the more the individual should tend to prescribe retreatment. However, it can be argued that although the decision makers acted from a position that considered the possible consequences, they might have used a decision strategy that was directed more toward minimizing the ‘losses’ than maximizing the ‘wins’. Instead of considering what could be gained by retreatment, they focused on ‘doing no harm’ and favoured a low-risk project. Such a strategy could not be captured by our experimental design. Furthermore, as mentioned above, individuals might approach clinical decision making not only from a consequentialist position, but might also regard the retreatment of periapical lesions as a moral ‘duty’.

References.

Aryanpour S, Van Niewenhuysen J-P, D'Hoore W (2000) Endodontic retreatment decisions. International Endodontic Journal 33 , 208-18.
Bacharach M, Hurley SL (1994) Foundations of Decision Theory: Issues and Advances. Cambridge, UK: Blackwell.
Bergenholtz G, Lekholm U, Milthon R, Heden G (1979) Retreatment of endodontic fillings. Scandinavian Journal of Dental Research 87 , 217-24.
Bj?rklund F (2000) Moral cognition: Individual differences, intuition and reasoning in moral judgement. PhD Thesis. Lund, Sweden: Lund University.
Bland JM, Altman DG (1986) Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1 , 307-10.
Debelian GJ, Olsen I, Tronstad L (1994) Systemic disease caused by oral microorganisms. Endodontics and Dental Traumatology 10 , 57-65.
Eckerbom M, Andersson J-E, Magnusson T (1989) A longitudinal study of changes in frequency and technical standard of endodontic treatment in a Swedish population. Endodontics and Dental Traumatology 5 , 27-31.
Engstrom B (1964) Bacteriologic cultures in root canal therapy. PhD Thesis. Umea, Sweden: University of Umea.
Eriksen H (1998) Epidemiology of apical periodontitis. In: Orstavik D, Pitt Ford TR, eds. Essential Endodontology . Oxford, UK: Blackwell Science, 179-91.
Eriksen H, Bjertness E (1991) Prevalence of apical periodontitis and results of endodontic treatment in middle-aged adults in Norway. Endodontics and Dental Traumatology 7 , 1-4.
Fyffe HE, Kay EJ (1992) Assessment of dental health state utilities. Community Dentistry and Oral Epidemiology 20 , 269-73.
Grahn?n H, Hansson L (1961) The prognosis of pulp and root canal therapy. Odontologisk Revy 12 , 146-65.
Griffin J (1986) Well-Being. Its Meaning, Measurement and Moral Importance. Oxford, UK: Clarendon Press.
Groome PA, Hutchinson TA, Tousignant P, Hanley JA (1999) The repeatability of three methods for measuring prospective patients' values in the context of treatment choice for end-stage renal disease. Journal of Clinical Epidemiology 52 , 849-60.
Hargreaves Heap S, Hollis M, Lyons B, Sugden R, Weale A (1992) The Theory of Choice. A Critical Guide. Oxford, UK: Blackwell.
Hulsmann M (1994) Retreatment decision making by a group of general dental practitioners in Germany. International Endodontic Journal 27 , 125-32.
Kahneman D, Slovic P, Tversky A (1982) Judgement Under Uncertainty: Heuristics and Biases. Cambridge, UK: Cambridge University Press.
Kay EJ, Blinkhorn AS (1996) A qualitative investigation of factors governing dentists' treatment philosophies. British Dental Journal 180 , 171-6.
Kerekes K, Tronstad L (1979) Long-term results of endodontic treatment performed with a standardized technique. Journal of Endodontics 5 , 83-90.
Kvist T, Reit C, Esposito M, Mileman P, Bianchi S, Pettersson K et al. (1994) Prescribing endodontic retreatment: towards a theory of dentist behaviour. International Endodontic Journal 27 , 285-90.
Ledley RS, Lusted LB (1959) Reasoning foundations of medical diagnosis. Science 130 , 9-21.
Lusted LB (1968) Introduction to Medical Decision Making. Springfield, IL, USA: Charles C. Thomas.
Molander A, Reit C, Dahl?n G, Kvist T (1998) Microbiological status of root filled teeth with apical periodontitis. International Endodontic Journal 31 , 1-7.
Mulley AG (1989) Assessing patients' utilities. Can the ends justify the means? Medical Care 27 (Suppl.), S269-81.
Murray CA, Saunders WP (2000) Root canal treatment and general health. A review of the literature. International Endodontic Journal 33 , 1-18.
von Neumann J, Morgenstern O (1947) Theory of Games and Economic Behaviour , 2nd edn. Princeton, NJ, USA: Princeton University Press.
?sterman B (1995) Value and Requirements. An Enquiry Concerning the Origin of Value . Aldershot: Avebury.
Petersson K, H?kansson R, H?kansson J, Olsson B, Wennberg A (1991) Follow-up study of endodontic status in an adult Swedish population. Endodontics and Dental Traumatology 7 , 221-5.
Petersson K, Lewin B, H?kansson J, Olsson B, Wennberg A (1989) Endodontic status and suggested treatment in a population requiring substantial dental care. Endodontics and Dental Traumatology 5 , 153-8.
Reit C, Gr?ndahl H-G (1984) Management of periapical lesions in endodontically treated teeth. A study on clinical decision making. Swedish Dental Journal 8 , 1-7.
Reit C, Gr?ndahl HG (1987) Endodontic decision-making under uncertainty: a decison analytic approach to management of periapical lesions in endodontically treated teeth. Endodontics & Dental Traumatology 3 , 15-20.
Reit C, Gr?ndahl H-G (1988) Endodontic retreatment decision making among a group of general practitioners. Scandinavian Journal of Dental Research 96 , 112-7.
Reit C, Gr?ndahl H-G, Engstr?m B (1985) Endodontic treatment decisions: a study of the clinical decision-making process. Endodontics and Dental Traumatology 1 , 102-7.
Reit C, Kvist T (1998) Endodontic retreatment behaviour: the influence of disease concepts and personal values. International Endodontic Journal 31 , 358-63.
Revicki DA (1992) Relationship between health utility and psychometric health status measures. Medical Care 30 (Suppl.), MS274-82.
Smith J, Crisp J, Torney D (1981) A survey: controversies in endodontic treatment and re-treatment. Journal of Endodontics 7 , 477-83.
Strindberg LZ (1956) The dependence of the results of pulp therapy on certain factors. Acta Odontologica Scandinavica 14 (Suppl. 21), 1-175.
Torrance GW (1984) Health states worse than death. In: van Eimeren W, Engelbrecht R, Flagle CD, eds. Third International Conference on System Science in Health Care . Berlin, Germany: Springer, 1085-9.
Torrance GW (1986) Measurement of health state utilities for economic appraisal. Journal of Health Economics 5 , 1-30.
Waltimo TM, ?rstavik D, Siren EK, Haapasalo MP (1999) In vitro susceptibility of Candida albicans to four disinfectants and their combinations. International Endodontic Journal 32 , 421-9.
Weinstein MC, Fineberg HV (1980) Clinical Decision Analysis . Philadelphia, PA, USA: W. B. Saunders.
Zug KA, Littenberg B, Baughman RD, Kneeland T, Nease RF, Sumner W et al. (1995) Assessing the preferences of patients with psoriasis. A quantitative, utility approach. Archives of Dermatology 131 , 561-8.