ORIGINAL RESEARCH PAPERSP Cleaton-Jones, N Daya, JA Hargreaves, D C o rtes,V Hargreaves, LP Fatti Examiner performance with visual, probing and FOTI caries diagnosis in the primary dentition Keywords: dental epidemiology, dental caries diagnosis fpllllu A . . v V ' S U M M A R Y To compare clinical reproducibility of dental caries diagnosis in the primary dentition under field conditions, a conve­ nience sample of 5-year-old children in a nursery school in Germiston, was examined for dental caries by four dentists using visual (mirror), visual plus tactile (mirror plus probe) and fibre-optic transillumination (FOTI) methods. Seventeen children were examined on day one and 11 re-examined on day two. Inter-examiner agreement was high, above 90%. Visual examination on its own is comparable with the tradi­ tional visual plus tactile method and to FOTI under field conditions. New caries data collected by visual diagnosis alone may, reasonably, be compared with historical data diagnosed with visual + tactile examination. S Afr Dent J 2001; 56: 182-185 O P S O M M IN G Vier tandartse het die kliniese herhaalbaarheid van tand- kariesdiagnose in die primere tandstelsel onder veldtoestande in 'n geriefsmonster van 5-jariges in kleuterskole in Germiston bepaal. Visuele (spieel), visuele plus sondering, en veseloptiese transilluminasie (FOTI) metodes is gebruik. Op die eerste dag is 17 kinders ondersoek van wie 11 op die tweede dag weer ondersoek is. lnterondersoeker ooreenkoms was hoog, bo 90%. Afsonderlik is visuele ondersoek verge- lykbaar met die tradisionele visuele plus sonderingmetode en FOTI in die veld. Nuwe karies data wat slegs deur visuele diagnose verkiy is, mag redelik met die historiese visuele plus sonderingdata vergelyk word. It is not clear when dental caries was first diagnosed with a mirror and sharp probe but the method was used in 1910 when Friel published the results o f the first field study o f the disease in South Africa. Until recently, P Cleaton-Jones, BDS, MB BCh, PhD, DSc (Dent) N Daya, BDS Dental Research Institute, Medical Research Council and University of the Witwatersrand, Johannesburg, South Africa JA Hargreaves, MChD, MA, MRC D (C) Faculty of Medicine and Dentistry, University of Alberta, Edmonton, Canada D Cortes, DDS, PhD Institute of Odontology, Gama Filho University, Rio de Janiero, Brazil V Hargreaves Victoria, British Columbia, Canada LP Fatti, PhD Department of Statistics and Actuarial Science, University of the Witwatersrand, Johannesburg, South Africa Address for correspondence: Professor P Cleaton-Jones, Dental Research Institute, PIBag 3,Wits 2050, Johannesburg, South Africa.Tel: (O il) 717- 2229, fax ( 0 1 1) 7 1 7 -2 1 2 1, e-mail 078cleat@chiron.wits.ac.za South African researchers as well as their North American colleagues continued to use the method (Roberts et a l, 1993; Disney ef a l, 1992). There has been pressure for some time, however, from British and European dental epidemiologists to use visual diagnosis alone, possibly with fibre-optic transillumination (FOTI), because of a potential to damage tooth structure or to transfer cariogenic bacterial strains with a sharp probe (Pitts and Evans, 1996). To change from visual plus tactile examination to visual only is easy but if this is done, may new field data be compared with the large amount of accumulated information in South Africa since 1910? The objective of the current study was to compare den­ tal caries diagnoses made under typical survey condi­ tions by four clinicians using three diagnostic methods. Materials and methods Population The study sample was all the 5-year-old children pre­ sent on two study days in a nurseiy school, one of 182 A P R I L 2001 V O L . 56 N O . 4 S A D J mailto:078cleat@chiron.wits.ac.za 15 such schools in Germiston that have participated in a caries surveillance study since 1981, so staff have become accustomed to dental surveys (Cleaton-Jones and Williams, 1995). Ethical clearance was obtained from the University of the Witwatersrand and all the parents gave informed consent before the children were examined. Each child gave verbal consent to be examined. primary or permanent dentition. Each tooth surface was classified as sound (S) or unsound (b) to produce patterns o f agreement (SSSS, UUUU) or disagreement (SSSU, SSUU, SUUU) between all four examiners. Kappa (Fleiss ef a i, 1979) and the Modified Percentage Reproducibility [MPR] (Shaw and Murray, 1975) were calculated. The chi-square and NcNemar tests were used with statistical significance set at P<0.05. Examiners Three of the examiners had been calibrated for visual plus tactile examination by examining some 200 extracted permanent teeth embedded in groups o f five in plaster blocks on two occasions to kappa scores >0.8 (Cleaton-Jones ef a/., 1989). One examiner was highly experienced in caries diagnosis with FOT1, having participated in international clinical trials using the method. With F0T1, dental caries is diagnosed when there is a clear dentinal shadow within a tooth. All four examiners had completed an extensive comparison of the same three diagnostic methods, plus radiographs, on extracted permanent teeth mounted in training mannikins two days before the current study. Examinations On day one the primary teeth of all 17 5-year-olds in the school were examined indoors in good, mixed nat­ ural and artificial light. Two examiners did visual diag­ nosis only with a plane mirror (VI, V2), one did visual plus tactile with a plane mirror and sharp probe (VP), and one used F0T1 with a 0.5 mm diameter tip. No radiographs were used for ethical reasons. Each child had four examinations, one by each examiner each day. On the second day 11 of the original 17 children, were re-examined in the same way. Earlier W HO dental caries diagnostic criteria that specified the use of a probe (WHO, 1987) were used for VP and current W HO crite­ ria which recommend a mirror only (WHO, 1997) were used for V I, V2. Both of these criteria specify that obvious cavitation is the minimum requirement for diagnosis; ‘white spots’ or ‘sticky fissures’ were not diagnosed as dental caries. F0T1 used dentinal shadows - the F0T1 tip was placed on all erupted tooth surfaces and a plane mirror was used to look for dentinal shad­ ows within a tooth (Cortes ef a l, 1994). M issing teeth were not included in the scoring. Statistical analysis A possible total of 11 918 surface observations were available for analysis with SAS (1990), variations in numbers o f surfaces examined by each examiner was due to a child having to go home with a parent before completion o f the examination, or a disagreement between examiners on whether a tooth was from the Results In the results replicate 1 refers to the results o f all 17 children seen on day one, replicate 2 results are for the 11 children seen on day two and replicate 12 indicates the observations on day one for the 11 children who were also seen on day two. The numbers and percent­ ages o f tooth surfaces diagnosed as carious are shown in Table 1. The percentage of carious surfaces ranged between 2.0 and 6.4; the lowest rates were for visual diagnosis and the highest was with F0T1. Except for V2, between replicates 12 and 2, there were no statistically significant differences between the replicates. Similarly, there were no statistically significant differences between V I, V2 and VP for replicates 1 and 12 but these differed significantly in replicate 2 (PcO.OOl) due to under-diagnosis by V2 in replicate 2. V I and V P did not differ significantly from each other in any replicate. When FOT1 was included in the analyses this differed significantly from the other methods in all three replicates (PcO.OOl). Table I. Dental caries surface prevalence rates by examiner method V I V2 V P F O T I Replicate 1 (n = 17) Total surfaces 1 564 1 565 1 618 1 476 Carious surfaces 53 53 72 93 °/o 34 3.4 4.4 6.3 Replicate 12 (n= 11) Total surfaces 1 388 1 392 1 442 1 044 Carious surfaces 46 48 49 67 o/o 3.3 3.4 3.4 6.4 Replicate 2 (n=11) Total surfaces 1 388 1 391 1 442 1 044 Carious surfaces 46 28 49 61 O/o 3.3 2.0 3.4 5.8 •Variation in total surface number is due to surfaces excluded from examination at a replicate. The inter-examiner agreements and disagreements for tooth surfaces are shown in Table 11. Agreement fre­ quency was high in all replicates. Kappa scores were high for three of the four examiners, and the M PR results were above 9 0 % (Table 111) but intra-examiner S A D J A P R I L 2001 V O L .56 N O .4 183 comparison showed that V2 had a statistically signifi­ cant difference for surface diagnoses between replicates 12 and 2 (P<0.05). No other intra-examiner compar­ isons showed statistically significant differences. Table II. Inter-examiner agreements and disagreements for tooth surfaces Replicate I Replicate 12 Replicate 2 Children N 17 11 11 Tooth surfaces N 1437 917 956 Agreements N o/o N % N o/o SSSS 1292 89.9 816 89.0 859 89.9 uuuu 14 1.0 10 1.1 10 1.0 Total 1306 90.9 826 90.1 869 90.9 Disagreements SSSU 35 2.4 22 2.4 28 2.9 SSUU 21 1.5 15 1.6 20 2.1 SUUU 75 5.2 54 5.9 39 4.1 Total 131 9.1 91 9.9 87 9.1 Table III. Intra-examiner reproducibility scores for tooth surfaces by examiner method V I V2 Vp F O T I Kappa 0.88 0.52 0.75 0.74 MPR (%) 99.2 97.8 91.1 97.1 McNemar test ( d 2) 0.36 4.12* 2.56 0.75 * = P < 0.05, MPR = modified percentage reproducibility Discussion The study was trouble free, all children enjoyed the experience. The sample size was limited by logistics. Firstly, the study was done at the end of a school year so attendance was down as families prepared for the summer vacation, secondly the F0T1 examiner with her equipment was in South Africa for only a short period. Fortunately, the low number of children was offset by the large number of surfaces examined. The study design needs some clarification. To fully test the effect o f four examiners and three methods each child would have had to be examined 24 times, 12 times each day. It was not ethically acceptable to us to subject the children to so many examinations, hence the current study design which limited examinations to four per day. F0T1 was used by a dentist who had been calibrated by international experts, VP was used by an epidemiologist experienced in the technique. Visual only diagnosis was used by two examiners, one (V1) a new graduate, the other (V2) an experienced epidemiologist, since this is the current method recommended (WHO, 1997). Analysis therefore included a combined effect of examiner and method. For credibility in a study such as this the caries rates of the study sample should be representative of those in the area. The caries prevalence of the 17 children was 59 .5% and the mean dfs was 3.1 (upper 9 5 % confi­ dence limit 4.9). Comparable values for 200 5-year-olds in 15 nursery schools in the surveillance study one year earlier were prevalence 51.4% and mean dfs 2.4 (upper 9 5 % confidence limit 4.7), which confirms the represen­ tativeness o f the sample. The study does not truly differentiate between diagnos­ tic methods - for that each examiner would need to use each method on the same subject. Rather, it is an inter- and intra-examiner comparison of examiners who used various diagnostic methods. There is no generally accepted ‘gold standard’ for diagnostic reproducibility but two methods are mentioned by W HO (1997) name­ ly kappa and percentage reproducibility. For kappa > 0.8 indicates good agreement (VI in this study), 0.6 - 0.8 substantial agreement (VP and F0T1) and 0.4 - 0.6 moderate agreement (V2) (WHO, 1997). For M PR the value that has been recommended by W HO is 8 5 -9 5 % (WHO, 1987). All examiners were above this. The reproducibility values in the current study were higher than a comparable study in the permanent den­ tition (Cortes et al., 1994). In that study, as in the cur­ rent investigation, decisions have to be made that sur­ faces are sound or carious, and the proportion of cari­ ous surfaces present will therefore influence the mea­ sured reproducibility. In the study sample the propor­ tion of carious surfaces was low but was representative of the caries rate in the area. The significant difference found for examiner V2 with the McNemar test indicat­ ed a slight asymmetry through under-diagnosis of caries on day two. Of the two methods to measure reproducibility kappa is strict and the Modified Percentage Reproducibility is lenient. Current W HO oral health survey methods (1997) recommend the kappa score. The high agreements in diagnosis between the examin­ ers indicate that if only visual diagnosis is used in future in South Africa, comparison to the mass o f accu­ mulated caries data diagnosed by the visual and tactile method will be reasonable. Conclusion Regarding choice of diagnostic method for field studies in the primary dentition, the levels of agreement between examiners indicate that all are acceptable. Regarding carious surface rate, F0T1 showed the highest rates and visual diagnosis was the lowest. 184 A P R I L 2001 V O L . 56 N O . 4 S A D ] Acknowledgements We gratefully record the generous assistance of Unilever Dental Research that provided the FOT1 equipment plus researcher (DFC). This paper is dedicated to the memory of John Anthony Hargreaves, Emeritus Professor at the University of Alberta, Canada and Visiting Professor at the University o f the Witwatersrand, and his wife Vera who both passed away before completion of the manuscript. R E F E R E N C E S Cleaton-Jones P, Hargreaves JA, Fatti LP, Chandler HG & Grossman FS (1989). Dental caries diagnosis calibration for clinical field surveys. Caries Research; 23: 195-199. Cleaton-Jones P & Williams S (1995). Caries trends in a South African com­ munity of 5-year-olds. Journal o f Dental Research; 74: 472 1ADR abstract. Cortes DF, Palma RG, Navarro MFL, Freitas SFT & Capelozza ALA (1994). Comparison of five caries-diagnostic methods for posterior approximal sur­ faces. Caries Research; 28: 192 ORCA abstract. Disney JA, Graves RC, Stamm JW, Bohannan HM, Abernathy JR & Zack DD (1992). The University of North Carolina Caries Risk Assessment study: further developments in caries risk prediction. Community Dentistry and Oral Epidemiology; 20: 64-75. Fleiss JL, Fischman SL, Chilton NW & Park MH (1979). Reliability of discrete measurements in caries trials. Caries Research; 13: 23-31. Friel G (1910). Teeth of African natives in Johannesburg. British Dental Journal; 15: 729-731. Pitts NB & Evans DJ (1996). The dental caries experience of 14-year-old children in the United Kingdom. Surveys conducted by the British Association for the Study of Community Dentistry in 1994/95. Community Dental Health; 13: 51-58. Roberts GJ, Cleaton-Jones PE, Fatti LP, Richardson BD, Sinwel RE, Hargreaves JA & Williams S (1993). Patterns of breast and bottle feeding and their association with dental caries in 1 - to 4-year-old South African children. 1. Dental caries prevalence and experience. Community Dental Health; 10: 405-413. SAS Institute Inc (1990). SAS Procedures Guide, version 6, 3rd ed, Caiy.NC: SAS Institute Inc. Shaw L & Murray JJ (1975). Inter-examiner and intra-examiner repro­ ducibility in clinical and radiographic diagnosis. International Dental Journal; 25: 280-288. WHO (1987). Oral Health Surveys, 3rd ed. Geneva: WHO. WHO (1997). Oral Health Surveys, 4th ed. Geneva: WHO. A BSTRA CT FROM OTHER JO U RN A LS Shear bond strength, microleakage and confocal studies of 4 amalgam alloy bonding agents Objective: This study was under­ taken to determine the relative shear bond strengths and microleakage o f 4 bonding agents to dentin and amalgam and to investigate the bonding to dentin through confocal laser scanning microscopy. M ethods and materials: Sixty non-carious molars were restored with 1 o f 4 different systems and the shear bond strengths were determined. For the microleakage study, Class V amalgam restorations were placed in 60 non-carious teeth. The specimens were thermo- cycled, the teeth were sectioned and dye penetration was assessed. For confocal examinations, the first component of the bonding adhe­ sives was labelled with rhodamine B. each of the adhesives was applied to 3 dentin specimens, which were examined under a con­ focal laser scanning microscope. Results: Of the 4 restorative sys­ tems tested, AmalgamBond Plus + HPA and Prime & Bond 2.1/base- catalyst showed significantly higher shear bond strengths. Prime & Bond 2.1/base-catalyst had the lowest microleakage value, which was significantly lower than that of AmalgamBond Plus + HPA and AmalgamBond Plus. AmalgamBond Plus + H PA had the highest varia­ tion in both the bond strength and microleakage values. Confocal laser scanning microscopy revealed tag formation, penetration of the bonding agents deep into the tubules, and hybrid layer formation for all 4 bonding systems. Conclusion: The 2 best systems, Prime & Bond 2.1/base-catalyst and, to a lesser extent, AmalgamBond Plus + HPA, utilised a supplemental bonding agent. Grob ler SR, O b e rh o lze rT G , Rossouw RJ, Groblei^Rabie A & Van W y k KotzeTJ Quintessence Int 2000; 3 I: 501 -508 S A D J A P R IL 2001 V O L .56 N O .4 185