Factors influencing prosthodontic decision-making

among general dentists

Ridwaan Omar,* BSc, BDS, LDS RCS, MSc, FRACDS, FDS RCSEd
Lamees Abu Nassif,** BDS, May Al-Kokani,** BDS, Nazeer B. Khan,* Phd
* College of Dentistry, King Saud University, Riyadh, KSA
** General Dental Practice, Jeddah, KSA

Variations in clinical decision-making may, in part, be due to dentist-related factors.  The aim of this study was to examine the role of dentist-related factors in the process of decision-making in a prosthodontic context, by evaluating the importance that general dentists attach to a series of listed criteria in typical clinical settings. A questionnaire was distributed to 228 dentists, of whom 137 completed it, giving a response rate of 60%. Four clinical categories, or Paper Patient Cases (PPC), representing choices between treatment options, were defined: (1) crown or plastic restoration, (2) fixed or removable partial denture, (3) fixed partial denture or single tooth implant restoration, (4) the choice to replace or not to replace a single missing posterior tooth. A visual analogue scale (VAS) was used to grade responses. There were generally large variations in the individual responses to various items, but smaller variations among genders and working sectors. Greatest importance was given to "prognosis" and "aesthetics of final result" across all situations, and less importance to "time for treatment" and "number of visits". In factor analysis, items in each PPC could be reduced to between 3 and 5 principal component factors, which explained between 59% and 70% of the variances. The reduced factors broadly encompassed 'clinical dental status', 'time', 'patient influence', 'outcome', 'general health', and 'non-dental treatment barriers'.  While some common explanations could be seen across some of the PPCs, there were also factors which were unique to a given PPC, suggesting that decision-making in prosthodontics could be regarded as a multi-dimensional process involving a wider range of factors than those included here.

It is widely acknowledged that, when faced with a clinical situation requiring intervention, dentists differ in their choice of treatment.1 Variations in diagnoses and treatment decisions occur equally when dentists are asked to examine extracted teeth or radiographs.2-4 The reasons for these variations are not well understood. This has led to a perception that clinical decisions tend to be made in an implicit, intuitive way, and that dentists do not share a common decision-making process.5 Even though dental education is based on scientific knowledge and acquired clinical experience, it is possible that variations in clinical decision-making are also influenced by individual preferences and styles.6 For this reason, it has been suggested that decision-making might not be easily amenable to structured characterization, and may be best described as an art rather than a science.7 

In many clinical situations, almost any one of a number of possible treatments could be suitable according to well-established clinical principles and practice. Therefore, the way in which the dentist evaluates the available information in a given situation plays an important role in the decision-making process. Although the process is undoubtedly very complex, involving many influential factors, which may be both patient- and dentist-related, the mechanisms involved in the selection, evaluation and application of information remain unclear. If there is to be less of the apparent subjectivity affecting the process, then the mechanisms underlying diagnostic thinking should be better understood.8 

Investigations into decision-making have focussed on empirical theories, which seek to understand the continuous process culminating in actual decisions, or on normative theories, which seek to define how decisions should be made by prescribing norms. Normative theories have been presented in many reports.6,9,10 Until recently, there have been few studies on empirical decision-making,11,12 which reflects the inherent difficulties of carrying out such research. 

In view of its ever-broadening scope, brought about by many new and improved materials and techniques, prosthodontic decision-making is becoming increasingly challenging. The role of more informed, and more litigious patients at the individual level, as well as greater demands for cost-effectiveness of dental health care by funding authorities at the societal level, add to the complexities.11 Recent Swedish studies13-16 found large individual variations among dentists regarding their ranking of patient-related criteria when deciding treatment in hypothetical clinical scenarios, so-called Paper Patient Cases (PPC). 

Studies have shown that variations in treatment decisions,4 and treatment patterns,17,18 occur between, as well as within countries. Recently, following a questionnaire survey among general dentists in Riyadh, associations between the reported time spent on prosthodontics, and a number of dentist-related factors were reported,19 and confirmed the aforementioned variations in treatment patterns. The aim of the present study was to expand our previous investigation19 to specific clinical situations. Using factor analysis, we investigated the relative importance that dentists assigned to various patient-related items that might be associated with treatment decisions for four clinical situations in which a choice between two feasible treatments were offered.

The selection and composition of the study population of general dentists, has been previously described.19 Convenience sampling was applied in a broad attempt to represent relevant sub-groups of the total dental workforce, including those from the different working sectors, viz. government (GS) and private sectors (PS), and secondly from across all the different districts of the city.   

A total of 228 dentists were surveyed, with 137 returning completed questionnaires, yielding a response rate of 60%. The study population of 137 general dentists was fairly evenly divided with regard to gender (47% males, 53% females), nationality (46% Saudis, 54% non-Saudis), and working sector (49% PS, 51% GS). The mean age of the group was 33 years (range 22 to 63

The construction, validation, distribution and collection of the questionnaire have been previously described.19 In summary, it was based on one used in a previous survey of prosthodontic decision-making among Swedish general dentists,13 but adapted to the local situation in some respects. The questionnaire comprised three parts, the results of the first two parts having been previously reported.19 The third part used a PPC technique, which described four typical prosthodontic situations involving treatment choices, as follows:

-        PPC1 - The choice to treat a lesion such as a cusp fracture on a maxillary premolar with a crown or an amalgam/composite restoration.

-        PPC2 - The choice to provide a patient with a fixed partial denture (FPD) or a removable partial denture (RPD), both options being technically possible.

-        PPC3 - The choice to replace a single missing tooth in the maxillary anterior region with a conventional FPD or a single tooth implant restoration.

-        PPC4 - The choice to replace or not to replace a single missing mandibular molar.           

A series of items relating to clinical decision-making for each of the cases was given (Tables 1-4). A visual analogue scale (VAS) response technique was used, with alternatives coded in 8 equidistant steps, ranging from "unimportant" (1) to "decisively important" (8). Clarification of certain key aspects of the methods used, for example the VAS, was given. In structuring the questions for each of the PPCs, it was important to prevent a feeling among respondents that their clinical judgement was being questioned or their responses "directed". Thus, preceding the questions, there was an orientation to the exercise, and reassurance that there were no "right" or "wrong" answers, the only objective being to record dentists' spontaneous reactions. This could be interpreted as reflecting dentists' own views, and not a view that they felt was the prevailing, correct one.  

Data were first analysed in contingency tables. Means and standard deviations of the VAS scores were calculated and differences evaluated with the t test for gender, nationality, and working sector. The responses to the PPC questions were analysed using principal components analysis (PCA), a standard psychometric method to assess common variation between attitude questions.20 Thus, a large number of variables, in our case each set of PPC items, were reduced to fewer factors which showed internal homogeneity. For each PPC, the number of factors was determined after inspection of scree plots and by the Kaiser criteria. Items with communality of < 0.4 were excluded in the final factor analysis. The factors were rotated by the Varimax method to maximize the total variance explanation retaining the dimensionality of factors. All statistical analyses were performed on an IBM Personal Computer using SPSS 10.

Relative frequencies of VAS responses to the four sets of PPC questions appear in Tables 1-4. The tables also list means and standard deviations, indicating wide individual variations in responses to most questions. High importance, with low variations were recorded for "prognosis" in all PPC cases (SD, 0.9-1.1), and for "aesthetics of final result" (SD, 1.0-1.1) in the first 3 PPCs. Items consistently regarded as less important were "time required for treatment" and "number of visits required", but with wide individual variations (SD, 2.2-2.3). In specific situations, great importance, and with relatively small individual variations, were observed for " remaining tooth structure" in PPC1, "abutment condition" in PPC2, "oral hygiene" in PPC3, and "presence of opposing tooth" and "occlusal stability" in PPC4. 

Bivariate analysis revealed some significant differences in responses related to gender, nationality and working sector (Tables 1-4). In all PPCs, the most significant differences were observed in "time required for treatment" and "number of visits required", for which Saudi and GS dentists consistently scored the items higher (P<0.01 or P<0.001).  

In the factor analysis, none of the items in any of the PPCs had a communality of < 0.4, and thus all were included in the final factor analyses (Tables 5-8). From the analyses, the items judged by dentists as important in their treatment choices, could be condensed into between 3 and 5 dimensions, or factors, which captured most of the variation. Variance explanations of between 59% and 70% were obtained. In interpreting the data, it should be kept in mind that the order of the factors reflects only their variance explanation, not their substantive interpretation. 

There was no common explanatory pattern across all of the PPCs, even though some factors showed remarkable cross-case similarity in dimensionality. For all four PPCs, a 'time' dimension was fairly consistently identifiable (Tables 5-8), except that in PPC3 and PPC4 some other general items were also included in the factor. Other factors, identified as 'outcome', 'general health', 'clinical dental status', 'patient influence', and 'non-dental treatment barriers', showed less cross-case consistency.

Notwithstanding the limitations of the convenience sampling, the purpose here was to explore the mechanisms underlying diagnostic thinking and treatment planning. The approach used was to see whether patterns could be observed in the way dentists say they evaluate a list of issues in defined clinical consultation contexts. It was not the intention here to define what clinicians ought to be doing,21 nor whether what they say conforms to a given norm.22

The PPC technique has been applied in studies on decision-making in general dentistry in the US,11,12 and in prosthodontics (amongst general dentists) in Sweden,13-16 with the conclusion that it is a useful research instrument for such purposes. By replicating the clinical situations used in the latter studies, but with the addition of a fourth clinical situation, direct comparisons of the present findings with the Swedish results could be made. In general terms, the present results support the view that dentists from different countries differ in terms of their stated criteria for making treatment choices.4 

This can be clarified by a further discussion of the results. Although there were similarly wide individual variations amongst Riyadh and Swedish dentists in their evaluations of the importance of various patient-related factors, there were some differences between the two groups. In Riyadh, "prognosis" and "aesthetics of final result"

 were highly ranked in each of the first three PPCs. This is perhaps a reflection of the increasing emphasis on the aesthetic outcome of dental treatment. For the same clinical situations, Swedes scored "patient's wish" and "prognosis"

consistently high.13-16 The Riyadh group considered "patient's wish" relatively less important, but not unimportant. 

There were a number of case-specific items, mainly technical in nature, which were also considered important by Riyadh dentists. This was so in the cases of the FPD vs RPD (PPC2) and the crown vs restoration (PPC1) choices. In the case of FPD vs single tooth implant (PPC3), dentists considered "adjacent teeth with restorations" as important. In the case of replacing vs not replacing a single missing molar (PPC4), the "presence of an opposing tooth" was considered most important by our group.  This is a commonly held clinical view, but one which lacks strong scientific support.23 Other traditional concerns such as "functional needs" and "arch integrity" ranked lower, and could be seen as evidence for a growing awareness among dentists that not all teeth distal to the premolars must always be replaced.24,25 

"Time required for treatment" and "number of visits needed" ranked lowest, or close to lowest, in all situations. Individual variations within these items were also wide, which may not be surprising given the influence of practice profile on such a factor.26 Indeed, GS dentists attached significantly more importance to this factor than PS dentists. It is not clear why GS dentists should be more time-bound, although it could be speculated that the work-load is higher on account of a free GS service. Time-related factors were also ranked significantly higher amongst the Saudi subset than the non-Saudis. If, as seems likely, Saudis make up most of the GS workforce and non-Saudis the PS, then the two findings would not be contradictory. Gender differences were hardly evident, except weakly for a few ite

Clearly, an important consideration in any treatment decision is the patient's treatment need. The focus of the study being an evaluation of dentist-related factors, the patient need factor was intentionally not included in the questionnaire, except indirectly through some of the items relating to the case.  The large individual variations in dentists' responses to given items should not automatically be interpreted as being indicative of equivalent variations in treatment, even though such variations are known to be large.4  The scoring of an item as important by one clinician, and unimportant by another, does not mean that two opposite treatment decisions will result. In the totality of the clinical situation, the same treatment decision could well arise, due to a complex interaction of many decision-making factors. Explanations for such questions are, of course, outside the scope of questionnaire techniques.        

Nevertheless, the PPC method appeared to work well for the present purpose of detecting variations in the factors considered by dentists when deciding treatment. Although the factor analysis differed in its ability to condense dentists' judgements into consistent dimensions for the different PPCs, the solutions for the first 2 cases were quite consistent, with 5 and 4 factors capturing 70% and 69% of the variation, respectively (Tables 5 and 6). In a Swedish study, data reduction by principal components analysis was not possible in the case of crown vs restoration, which was ascribed by the authors to the complexity of prosthodontic decision-making,14 even though an earlier pilot study by the same authors had produced a stable model in factor analysis.13 This, as with our result, supports the concept that only a few basic factors might influence decisions.27 On the other hand, even though the FPD vs implant and replacement vs non-replacement items were similarly reducible into factors (Tables 7 and 8), with high variance explanations, their interpretation as specific characteristics was not consistent. This presents an apparent dichotomy, since factors other than those represented in the PPCs may be influential in the choices made.  For example, in PPC3, the decision may have been largely theoretical since in practice most dentists would not have been confronted with this decision. Thus it may be that the decision-making approach is easier to resolve into a few factors when not subject to actual "hands-on" clinical experience in the treatment option. And in PPC4, with patient factors deliberately unspecified (to avoid "leading" respondents to a particular answer), many different perceptions of the situation could arise, and with it a theoretically "correct" response.

From the range of findings in this study, it appears that prosthodontic decision-making is a multi-dimensional process, much of which remains only poorly understood.

making items

Address reprint requests to:

Prof. Ridwaan Omar
Department of Prosthetic Dental Sciences
College of Dentistry, King Saud University
P.O. Box 60169, Riyadh 11545, KSA
Tel./Fax: +9661 488 5241
email: This e-mail address is being protected from spambots. You need JavaScript enabled to view it