Evaluation of a systematic digitized training program on the effectivity of subgingival instrumentation with curettes and sonic scalers in vitro

All operators had the same setup and options to choose their instruments for root surface instrumentation; (1) Gracey normal-shape curettes nos. 5/6, 7/8, 11/12, and 13/14 (American Eagle Instruments, Missoula, MT, USA) and (2) sonic scaler (Synea, W&H, Bürmoos, Austria) with air pressure and water-cooling (30 ml/min) as recommended by the manufacturer on level two, which means “medium” amplitude, combined with a straight, right and left curved slimline tip (1AP, 2APr, 2APl, W&H, Bürmoos, Austria) with round cross-section.

New instruments were used only at baseline (T0), for following visits the curettes were randomized to each operator, whereby prior to each test day (T1–T3), the curettes were controlled by a trained dental hygienist regarding sharpness and if necessary sharpened with an Arkansas stone and grinding oil. Similarly, each tip for the sonic device was controlled for their length as advised by the manufacturer or any possible destruction.

All participants received the same training program including two lectures of theoretical application information according to our clinical and the manufacture’s guideline as well as practical sessions in vitro before baseline test (Fig. 1). Afterwards, both groups of instruments were trained equivalently 10 times for 45 min over a 14-week period by using a digitally modified learning program (for AIR/GRA 24/32 working steps) based on a program introduced in 1994 [12]. Concomitant to the digitized training program the group was supervised and monitored through the education staff (3 board certified specialist of periodontology). In a previously published study [15], we could demonstrate that training of root surface instrumentation should not only focus on efficacy, but also on ergonomics. The digitized visualization program is supposed to support the teaching of the necessary working steps and ergonomic aspects separately for each instrument, and concomitantly reduces time effort for supervision by the staff. Partly by animated GIF or short video sequences the whole set up, technique and sitting position of the operator and patient are explained. Additionally, all participants were during training sessions clinically calibrated regarding the application pressure (3–5 N for GRA and < 1 N for AIR), but no measuring of the root surface destruction or roughness were done at the test days.

Before the study, at the beginning of the workshop all students of the 7th semester (n = 31) were asked to participate in the study, and after their written consent (n = 30), they received a 2-h theoretical introduction for both instruments and to the training program, but without practical exercises. One student, refused to take part in the study.

Participants were evaluated at baseline (T0) and between the training weeks at week 6 (T1), week 11 (T2) and week 14 (T3) while instrumenting 12 comparable test teeth with GRA (tooth 11, 14, 16, 31, 37, 45) and with AIR (tooth 21, 25, 17, 34, 36, 43). The instrument to begin with was randomized (Microsoft Excel 16, Microsoft Corporation, One Microsoft Way Redmond, WA, USA) for each test day and participant. Participants were asked to instrument the test teeth until they achieved subjective maximal elimination of simulated hard and soft deposits. Time for treating the predetermined six teeth per instrument was recorded for each participant.

Both groups of instruments were performed in similar manikin heads, equipped with modified periodontitis models (Frasaco, Tettnang, Germany), exhibiting pronounced periodontitis with moderate to advanced horizontal bone loss, isolated and deep vertical pockets, resulting in teeth being differently difficult to instrument regarding both anatomy and accessibility. Mean (SD) pocket probing depth (PPD) was 5.8 (2.1) mm (range, 3–11 mm), separated for test teeth in GRA group (5.69 (2.01) mm (range, 3–9 mm)) versus AIR-group (5.67 (2.22) mm (range, 3–9 mm)). Gingival masks were fixed that no operator was able to lift the mask during the test trials.

Twelve teeth and seven single- and five multi-rooted teeth, were coated with a thin layer of transparent fluorescent varnish (Shiny White–Rival de Loop Young, Berlin, Germany) between the artificial cemento-enamel junction and the alveolar bone in order to simulate adhering plaque. Coating was performed by a reproducible and standardized procedure of dipping to provide similar thickness of the applied varnish [16]. The used varnish fluoresced with bright blue color when exposed to ultraviolet light. Gingival masks of persistently pliable silicon (Frasaco, Tettnang, Germany) were used to cover the models to ascertain instrumentation without visual control. Beside this simulation of biofilm, test-teeth were also randomly coated with commercial varnish (A-CK, Frasaco, Tettnang, Germany), modified for the ratio of varnish/thickener, to simulate adhering subgingival hard deposits. The hard deposits were located in spot size (diameter around 1.5 mm) at each tooth type at the same area by one investigator (P.F.) according to pictures taken during the first test. The varnish on the root surfaces was placed at a distance of at least 2 mm from the bottom of the pocket and 1 mm subgingival from the marginal gingiva, to simulate the pathologic circumstances as realistic as possible.

Artificial biofilm and hard deposits as well as teeth and gingival masks were renewed after each test day (T0–T3).

After each test day T0–T3, the effectivity of instrumentation was planimetrically assessed. Instrumented teeth were mounted on standardized prepared plastic units (Lego GmbH, Grasbrunn, Germany), individually fitting to each of the 12 teeth. Plastic units with mounted teeth were then fixed on the camera table in a reproducible position. Following, teeth were submitted to ultraviolet light (UV-A, 350–370 nm), and one image of each of the four tooth surfaces (mesial, distal, vestibular, oral) (in total 1728) was taken using a camera with a 100-mm macro-zoom (Canon EOS D30, Tokyo, Japan). The images were used to subsequently perform an evaluation of the cleaned surface area by digital image subtraction (Image J, NIH, Bethesda, USA) to calculate the relative cleaning efficacy of simulated biofilm removal subgingival (RCE-b in %) (Fig. 2). For multi-rooted teeth with furcation involvement (16, 37, 17, 36), no evaluation of the furcation area was done (please see Fig. 2c for details). The relative cleaning efficacy of hard deposits (RCE-d) was assessed through one investigator (P.F.) by counting the non-removed hard deposits on reproducible images of each tooth site.

As the primary outcome, the percentage of removed artificial biofilm and hard deposits were determined. Secondary outcomes were treatment time, difference between test days and instruments according to the cleaning effectiveness as well as operators’ experience reports.

At each test day (T0 -T3), operators answered a questionnaire to estimate their own effectiveness in removing simulated biofilm/hard deposits for each group of instruments as well as questions according gender, age and previous vocational training only at T0 (closed-ended question).

According to the sample size calculation with data of a comparable investigation [12] we found n = 175 teeth per group at each test day as sufficient to detect under five percent difference for removed hard deposits between the groups of instruments (power of 80%).

Randomization, means, percentages and standard deviations were calculated using Microsoft Excel (Microsoft Excel 16, Microsoft Corporation, One Microsoft Way Redmond, WA, USA). Values were mostly ordinal (yes/no) and with the use of Microsoft Excel, crosstabs have been drawn up. Data were entered into SPSS Statistics (SPSS Statistics 24, IBM, Chicago, IL, USA). Normal distribution of artificial biofilm, residual hard deposits and treatment time were tested by Kolmogorov-Smirnov, Shapiro-Wilk as well as Chi Quadrat Test. The artificial biofilm at the two test days (T2 and T3) was not normally distributed (p > 0.0001). Artificial biofilm removal, therefore, were tested with non-parametric tests. All tests were two-sided; statistical significance was assumed p ≤ 0.05 and adjusted with the Bonferroni method for multiple comparison (p = 0.05 / 6 = 0.0083). Associations between variables were analyzed using Pearson correlation and Spearman's rank correlation coefficient analysis.

All 30 operators were in the seventh semester at Kiel dental school (female/male 21/9) and in mean (SD) 25.50(4.11) years of age (range 21-42). Eleven had a previous education in dentistry e.g. dental assistants or technician (female/male 7/4).

In general, the learning curve of the yet-untrained operators begins with a nearly equal efficacy level for GRA of 63.18(25.74) % and for AIR of 62.25(26.69) % RCE-b at T0 (p = 0.172). In the GRA-group RCE-b decreases until T1 to 58.62(24.18) % (p < 0.001) and subsequently improves significantly at T2 (RCE-b 64.05(24.51) %, p < 0.001) and reaches highest efficacy at T3 (RCE-b 71.54(23.90) %, p < 0.001) (Table 1, Fig. 3). For AIR-group, we found a different performance with continually improvement of the RCE-b values after T0 until T3. At T1, the operators showed a 63.56(24.55) % efficacy without significant changes between T0 and T1 (p = 0.203), whereas at T2 (69.17(23.55) % and T3 (71.75(23.05) % efficacy significantly improved (T1–T2: p < 0.001 and T2–T3: p = 0.001) (Table 1, Fig. 3). Despite the different progression during the training phase, all operators removed simulated biofilm without significant difference between both groups of instruments at the final evaluation (T3: p = 0.864).

Considering hard deposits (RCE-d), the learning curves in the AIR-group are comparable with RCE-b. The comparison of RCE-d of both groups of instruments demonstrates significant superiority of the GRA-group at all time points (T0: GRA 85.48(12.32) % versus AIR 65.71(15.27) % efficacy; p < 0.001 and T3: GRA 84.68(16.84) % versus AIR 77.85(13.98) % efficacy; p < 0.001). RCE-d values for AIR at T0 and T1/T2 (65.48(14.90) %, p = 0.750/67.49 (16.11) %, p = 0.012 (adjusted to Bonferroni)) do not differ significantly. The improvement between T0 and T3 for RCE-d in the AIR-group is significant ((77.85(13.98) %, p < 0.001) (Table 1). The learning performance for GRA had a different pattern; after equal efficacy at T0 to T1 (84.52(13.58) %, p = 0.257) the lowest efficacy was found at T2 (79.26(15.71) %, p < 0.001) and increased again until T3 (84.68(16.84) %, p < 0.001). The difference between T0 and T3, however, was not significant (p = 0.417) (Table 1, Fig. 3).

Before the beginning of the training phase, for both groups of instrument no significant difference for removing simulated biofilm in the upper jaw versus the lower jaw were measured (GRA/AIR: p = 0.011 (adjusted to Bonferroni)/p = 0.367) (Table 2). At T3, both groups of instruments were able to reach a significant higher efficacy in the lower (GRA/AIR 68.88(23.36) %/70.05(23.18) %) versus upper jaw (GRA/AIR 74.23(24.18) %/AIR 73.39(22.38) %) (all groups T0–T3: p ≤ 0.001) without significant differences between the groups of instruments (upper jaw/lower jaw: p = 0.514/p = 0.638). For removal of hard deposits at T3 we found compared with RCE-b in the AIR-group, a higher efficacy in the upper jaw versus lower jaw (83.68(19.76) %/74.26(12.11) %, p < 0.001), but for GRA, results were nearly similar (95.95(11.52) %/(93.14(11.28) %, p = 0.311). Contrary to the training effect in simulated biofilm removal in both groups, an improvement in hard deposit removal was only observed for AIR.

On tooth-level, at T3 in the AIR group, most simulated biofilm was removed in premolars compared to molars and front teeth (p < 0.001), but there are no such differences for the GRA group in RCE-b (Table 2). Using GRA in molars lead to significant better RCE-b results (p = 0.001) compared with AIR, whereas in premolars efficacy was higher in the AIR group (81.28(19.0) %; GRA 71.35(22.09); p < 0.001).

With only one exception at T2 (p = 1.000), the treatment time for the six teeth per group was always shorter with GRA (Table 1, p < 0.001). At the final evaluation (T3), the calculated time per tooth was 2.78 (0.55) min per tooth for GRA (AIR 3.30(0.75) min). As illustrated in Fig. 3c, from baseline T0 until T3 the treatment time per tooth with curettes decreased around 2.0 min (AIR 2.1 min) (both: p < 0.001) and the majority of all thirty operators in both groups of instruments treated under 4 min (GRA: n = 29 and AIR: n = 23) (Fig. 4c). All operators reduced their own treatment time during the training phase with curettes, whereas with sonic scalers two operators failed to improve their own treatment time between T0 and T3. Regarding the performance as efficacy of RCE-b over treatment time, a continuous improvement from T0 to T3 can be observed for GRA and AIR (Fig. 3d).

At baseline T0, 60% of all operators in the GRA group were unable to reach an efficacy of RCE-b of 70% (73.33% of the AIR group). After the systematic training phase at T3 in both groups of instruments, this number of operators could be reduced to 33.33% (GRA) and 36.67% (AIR). For RCE-d, the learning curve was similar. Accordingly, at T0, no proband in the AIR group was able to remove all hard deposits (GRA 26.67%), but 10% at T3 (GRA 36.67%). Further changes in individual efficacy levels of RCE-b and RCE-d are shown in Fig. 4.

Furthermore, 18 operators showed higher efficacy to remove simulated biofilm with curettes at T3 versus T0 (AIR 24 operator) and 12 operators a lower efficacy (AIR 6 operators), respectively. For removal of hard deposits, 13 operators in the GRA group improved their own efficacy (AIR 22 operators), and 12 operators deteriorated (AIR 6 operators). Five operators showed similar results at T0 and T3 for RCE-d (AIR 2 operators).

At all test days, the operators self-estimated their efficacy of simulated biofilm removal (RCE-b) in both groups of instruments equally high with 60%-70% (Table 3).

Contrary to older position paper, in which powered devices were often described as more effective while, e.g., saving treatment time compared with hand instruments [17] which could not be supported by the results of our study. However, differences between a manikin model and a real clinical situation should be kept in mind. Untrained operators needed in the beginning of their systematic training phase independent of the instrument used nearly five Minutes per tooth improving to about 3 min at T3, favoring GRA at both time points (Table 1, p < 0.001). Therefore, we have to reject one part of our hypothesis that less-experienced operators after 6 months of training with sonic scaler will neither reach a higher efficacy nor save treatment time compared with curettes. We can only hypothesize on these differences. In the present study, the amplitude for the sonic scaler was fixed at medium level as advised by the manufacturer, but power adjustment of powered devices will influence the results significantly [18, 19]. However, these findings are in line with previous investigations [12, 16]. Our research underlines the recommendation for a comprehensively training of all powered devices, whose often times supposed perceived as simpler than that for hand instruments [20].

The operators of the current study could, however, estimate their own efficacy and treatment time more realistically compared with previous investigations for both instruments [21]. The training took the same amount of time, but our digitized training was supported by less supervisors (here with a ratio of 1:10). Nevertheless, the influence of motivation for the training seems to be enormous [21]. Motivated students (e.g., prizes for good performance, atmosphere of competition and enthusiasm during the whole training program) can achieve about 25% higher efficacy in removal of simulated biofilm compared with a group without motivational incentives [21]. Furthermore, motivation helped to estimate the efficacy more precisely and will greatly influence learning of effective root debridement [21]. Future studies could evaluate further developed digitized training programs with, e.g., personalized digital animation and immediate feedback function.

Furthermore, magnifying loupes might be of major importance already during training. They increase visual acuity and can therefore improve clinical outcomes in dentistry [22‐24]. Despite the potential advantages in wearing loupes, in the current investigation their use was allowed, but not obligatory. However, the benefit of loupes during periodontal treatments is controversially discussed in literature [23‐27], although a better view into the sulcus can be assumed.

For successful periodontal therapy, a major goal besides effective removal of biofilm and hard deposits is not to damage the root surface [28]. This outcome was however not assessed in the present study. According to our previous in vitro investigations focusing solely on simulated biofilm [15, 16] or on hard deposits [27], it might be helpful to consider the goal of the treatment for unexperienced operators. Thus, given the great variety of scope of interest and study designs, all comparisons between studies require caution [16, 19, 28‐34]. Further on, due to the character of the in vitro simulation, e.g., using varnish to simulate the subgingival biofilm and hard deposits on the root surface of plastic teeth does not allow transferring our findings to clinical settings in general [29]. Using varnish to simulate subgingival biofilm and hard deposits allowed reproducible and standardized conditions in vitro, but the location of artificial biofilm under the simulated hard deposits as done in the current investigation is not comparably with a clinical situation. In vivo the development of hard deposits is a dynamic process that starts with a non-mineralized biofilm, which eventually calcifies. Nevertheless, it can also be seen that it does not matter whether there is a biofilm under or above a hard deposit as both have to be removed. Complete removal of hard deposits inevitably correlates with simultaneous removal of biofilm. Further, we failed to sub-analyze results of removed simulate hard deposits on jaw/tooth level, as we calculated only the percentage of the number of removed deposits compared with the total number of deposits due to technical reasons. Also, we cannot quantify possible effects of the used gingival masks, with a possible limited penetration of curettes into the sulcus and also possible damping of oscillating instruments [12, 16, 29, 35]. Furthermore, the morphology of the jaw and the tooth type are likely to have influenced our results as mentioned before. The furcation area of the involved molars was not assessed in this study, which might explain, why AIR failed to show better cleaning efficacy in molars as we assumed [17], whereas another in vitro investigation showed good accessibility when the horizontal pocket depth was less than 2 mm for both type of instruments [36]. Furthermore, as the tests were performed within one periodontitis model, the test teeth were not exactly corresponding but much with regard to type of tooth (front teeth, premolar, molar) in the upper and lower jaw and the profile of the used periodontitis models for both groups of instruments (AIR all quadrants vs. GRA: right upper and both lower quadrants). Therefore, we failed to control for possible confounder of handedness [37]. The analyzed teeth were selected to be comparable in terms of position (tooth type) and simulated probing depth/loss of attachment. Although, there is a difference in the location of the teeth between the two groups of instruments for premolars (teeth 14, 25 in the upper jaw and 34, 45 in the lower jaw), the morphology of the roots of all premolars are comparable (no furcation involvement).

Additionally, the haptic of the plaster material could as well have influenced our results due to the difficulties of simulating realistic bone morphology and haptic. However, the presented in vitro analyses allow investigating defined and clinically not measurable parameters under reproducible conditions, thereby increasing the sensitivity of the comparisons. Last but not least, the authors want to point out that the present study focused on the evaluation of the modified training program in routine teaching procedure and not to the best possible effectiveness of removal of simulated biofilm or hard deposits of undergraduate students.

Future studies should investigate during routine students’ workshops the combination of sonic scalers and curettes versus ultrasonic and subgingival air polishing with non-abrasive powder under clinical setting [10]. By teaching all these treatment methods for periodontal non-surgical therapy in dental school routine, it can be presumed that the next generation of dentists will work with the highest effectiveness concomitant with the lowest presumed number of side-effects. However, this was not evaluated in the current study. Therefore, all efforts should be done to improve the training programs and lectures in periodontology.