Dentist Empathic Accuracy Is Associated With Patient-Reported Reassurance
A wide array of studies have established that dental anxiety is deeply rooted in previous negative dental experiences or, less frequently, in other types of distressing and traumatic life events.1, 2, 3, 4, 5, 6, 7, 8, 9 Risk factors for experiencing pathologic forms of dental anxiety (a spectrum ranging from elevated fear and dental phobia to posttraumatic stress disorder) include having undergone an invasive dental procedure (eg, injections, drillings, and extractions), extreme emotional responses during negative experiences (eg, felt helplessness, nausea, embarrassment, or sense of suffocation), and characteristics related to the dentist (eg, lack of understanding, impoliteness, rudeness, insufficient information provided during treatment, and criticism).1, 2, 3, 4, 5, 6,10, 11, 12 Creating a pleasant dental experience for the patient, and thereby preventing the development of a pathologic form of fear of the dental treatment, is most likely when the dentist is properly trained and willing to conduct behaviours that affect the internal state of the patient. In other words, it is imperative13, 14, 15, 16 that dentists have the empathic capabilities (we use the term “empathic accuracy”) to accurately estimate their patients’ level of anxiety and, moreover, to respond accordingly. Examples are providing emotional reassurance and a creating a sense of control by the patient.12,17,18
Regarding the accuracy with which dental professionals estimate their patients’ anxiety levels during treatment, only one cross-sectional study by Höglund et al. that included 1128 adult patients from several Swedish public dental offices has been conducted.19 The accuracy with which dental professionals estimated their patients’ anxiety levels was examined by obtaining dental anxiety scores before an annual oral checkup, whilst their dentist was asked to rate how anxious they thought the patients felt. Both ratings were found to be positively correlated (r = 0.45, P < .001), but because the authors found “no correlations between any of the clinicians’ and patients’ ratings of dental anxiety” (p. 458) within a small subgroup of highly anxious individuals, they concluded that “clinicians were unsuccessful in identifying a dentally anxious patient” (p. 455). An explanation for this finding might be that annual checkups are generally not invasive and do not rank amongst the most feared dental situations,10 and as such this may have led to relatively little variability in scores. However, empathic accuracy has not yet been studied during invasive dental procedures, and it is unknown to what extent empathic accuracy is related to the degree to which patients feel reassured during treatment.
Therefore, the first purpose of the present study was to determine dentists’ ability to accurately estimate patients’ anxiety level by conducting a cross-sectional study in 3 general dental clinics in the Netherlands, amongst patients undergoing dental treatments that have been shown to be perceived as invasive in previous studies.3,10,11 We assessed dentists’ accuracy in estimating their patients’ level of anxiety during treatment (ie, “empathic accuracy”) and the extent to which patients felt reassured during treatment. Second, based upon previous suggestions,10,12,17,18 we hypothesised that greater accuracy of dentists’ assessment of patients’ anxiety level (empathic accuracy) would be significantly associated with patients’ self-reported felt (perceived) reassurance during treatment. Third, because a previous study by Höglund et al.19 revealed that it proved more difficult to adequately estimate the anxiety level of highly anxious patients, we investigated whether this relationship could be replicated in our sample. Therefore, we also determined the association between dentists’ empathic accuracy and patients’ anxiety level. We hypothesised that dentists’ empathic accuracy would be significantly lower amongst patients reporting relatively high levels of anxiety than amongst patients reporting relatively low levels of anxiety. In the same vein, we determined the association between patients’ self-reported felt reassurance during treatment and patients’ anxiety levels. We hypothesised that felt reassurance would be significantly less amongst patients with relatively high anxiety levels vs patients with relatively low anxiety levels. We were also interested in patient-, dentist-, and treatment-related factors that may influence empathic accuracy. Specifically, we examined whether empathic accuracy would be associated with dentists’ years of experience, dentists’ and patients’ gender and age, and type of dental treatment.
Between December 2014 and April 2016, on 34 unscripted days selected based on the researchers’ availability, a total of 177 consecutive adult patients of 9 dentists in 3 general dental practices (adjacent to the residence of data collecting researchers) in The Netherlands (Amsterdam, the Hague, and Alphen aan den Rijn) who were scheduled for invasive dental treatments were included. At arrival in the waiting room for the treatment appointment, they were asked to participate in this study by a researcher (SAZ, MJR, or AL). After potential participants had given oral and written informed consent to the researcher, their dentists were informed about their participation. Patients were informed by the researcher that their dentist was blinded to their answers on study questionnaires. Demographic variables were subsequently recorded.
This study was designed as a cross-sectional observational study. The Institutional Review Board (IRB), “the Medical Ethics Committee of the Academic Medical Centre in Amsterdam, The Netherlands” decided that formal assessment by the IRB was not required for this observational study (decision W 15_084 # 15.0098). Participant data were anonymised and were stored on a protected server of the Academic Centre for Dentistry in Amsterdam, The Netherlands.
Materials and procedure
All participants underwent an invasive dental treatment (ie, dental extractions or procedures requiring both injections and drilling; root canal treatments, crown and bridge preparations, and fillings).3,10,11 After leaving the treatment room, they were asked by a researcher to fill out a questionnaire in the Dutch language containing five 100-point Visual Analogue Scale (VAS) items19, 20, 21 ranging from 0 (“not/none at all”) to 100 (“extremely”). The first 2 VAS items pertained to the experience of the treatment that day (“To what extent did you feel reassured by the dentist during treatment?” and “How much anxiety did you experience, on average, during treatment?”). The patients were also requested to respond to 3 different VAS items pertaining to their experience of the treatment.
Separately, after treatment, the dentist, blinded to the patients VAS scores, was instructed by the researcher to imagine being that particular patient and to estimate the patient’s level of anxiety during treatment (“How much anxiety did the patient express during treatment, according to you?”) using a 100-point VAS ranging from 0 (“none at all”) to 100 (“extremely much”). Finally, the dentist filled out a number of operative variables on the case report form (ie, duration of treatment, amount of local anaesthetic injected, and nature of the treatment).
Empathic accuracy was calculated as an absolute difference between patient-reported intraoperative anxiety (100-point VAS) and dentist estimation of the patients’ anxiety (100-point VAS). This means that an empathic accuracy of “0” represents perfect accuracy and that the maximum inaccuracy is “100” if the patient-reported anxiety is found to be 100 and the dentist estimates the anxiety to be 0, or vice versa. The direction of the difference is deliberately not addressed in these absolute values so that the group means are not averaged out by both positive and negative values and, as such, reflect the true (absolute) means. By means of this formula, we were able to quantify per data point (dentist estimated value vs patient-reported value) how accurately a dentist could infer the feelings from a patient and subsequently determine a correlation coefficient for these individual data points on a group level.
To avoid the Hawthorne effect22 (ie, when people behave differently because they know they are being watched), patients were informed that their dentists were not aware of the answers they provided on their case report form.
Sample size calculation
To ensure a sufficient sample size, a sample size calculation was conducted beforehand for the empathic accuracy (correlation coefficient) with G*Power 3.1 software.23 A power of 80% is seen as a standard in clinical studies.24 We calculated that, assuming a correlation of 0.45 between 2 measurements based on previous literature,19 with a power (1-β error probability) of 0.8 and an alpha (α error probability) set to 0.05, a total sample size of ≥36 participants would be required. However, because the aforementioned previous study was performed during annual oral checkups19 and found that higher patient anxiety scores were associated with lower empathic accuracy, we could not exclude the possibility that during the invasive dental procedures in our study, empathic accuracy would be found to be lower. Therefore, under the assumption of a small effect, an optimal sample size of ≥175 participants would be required.
Degree of conformity between dentists and patients in rating the patient’s anxiety was determined by calculating intraclass correlation coefficients (ICCs; measures of agreement), whereas the association between continuous variables was determined by calculating Pearson’s product-moment correlation coefficients (linear association). For interpreting effect sizes in the present study, we used the recommendation of recent insights in psychology, considering Pearson’s r correlations of 0.10, 0.20, 0.30, and 0.40 to be small, typical, large, and very large, respectively.25,26 If distributions of continuous variables were skewed (felt reassurance), variables were (natural log) transformed. In addition to the main analyses, we tested the effects of dentist-, patient-, and treatment-related factors (eg, gender and years of dentist’s experience) on patients’ ratings concerning anxiety and empathic accuracy. Paired t tests were used to compare the means of 2 dependent groups (dentists and patients). One-way analysis of variance was used to compare means of more than 2 unpaired groups (patients, grouped by treatments). Subgroup analyses were performed if explanatory variables contained outliers and, based on previous literature,19 a median split analysis27 was performed to explore empathic accuracy in a subgroup of highly anxious patients. All the statistical analyses were carried out using SPSS Statistics for Macintosh, Version 25.0 (IBM Corp.). The level of significance for all statistical analyses in this study was set at a = 0.05 (2-tailed). We did not adjust for multiple testing, because correcting for capitalisation on chance is based on the assumption that the null hypothesis is always true but a reference value for the main outcome (empathic accuracy) has been previously reported; hence, correcting for multiple testing in order to reduce the type I error is of less importance than reducing the chance of type II errors.28
A total of 177 consecutive patients were included. Initially, 189 individuals were asked to participate; 12 refused to take part in this study for personal reasons. The mean age was 41.4 years (range, 19-90). Ten dentists took part in this study, with a mean of 6.1 years of professional experience (range, 1-31). Descriptive statistics are displayed in Table 1, and the strengths of all associations are displayed in a Pearson correlation matrix (Table 2). The data that support the findings of this study are available from the corresponding author upon reasonable request.