THE ACHIEVEMENT AND MAINTENANCE OF INTER-EXAMINER CONSISTENCY IN THE ASSESSMENT OF PLAQUE AND GINGIVITIS DURING A MULTICENTER STUDY BASED IN GENERAL DENTAL PRACTICES

This study set out to demonstrate that it was possible to train genera l dental pracitioners (gdps) to achieve and maintain high levels of in ter-examiner consistency in the use of simple periodontal indices over a 12-month period. The gdps were trained by one trainer in the use of the plaque index (P1I) and a modified version of the gingival index ( mGI) which assessed gingival bleeding only. All the gdps underwent int ensive training and employed a technique, when assessing inter-examine r consistency in the use of the P1I, such that the Ist examiner did no t disturb in situ plaque from 50% of the circumference of a tooth. The problem of variation due to repeat probing, when assessing inter-exam iner consistency for mGI, was overcome by one examiner lightly probing gingival margins and both examiners scoring the results. The problems relating to the multicentre nature of the study included: distance be tween the centres, the need for strict adherence to the study protocol , consistency in the use of forms and instruments, in the application of periodontal indices, and of inclusion and exclusion criteria. Techn iques for overcoming these problems included: the planning and applica tion of a coherent study design, which employed simple indices, a deta iled protocol, the recruitment of very well-motivated gdps of similar age and experience as examiners, the recruitment of an experienced tra iner who trained the gdps thoroughly and monitored their performance t hroughout the study, and repeat visits to the practices involved to ex plain the nature of the study to all their staff members. Initially, t he gdps achieved inter-examiner kappa scores of 0.78-0.85 (mean 0.81) for P1I and of 0.73-0.94 (mean 0.87) for mGI when assessing 168 sites for each variable. During the following 12 months, individual kappa sc ores, assessed every 3 months at 42 sites, ranged from 0.51-0.90 for P 1I and from 0.73-1.00 for mGI. Mean kappa for P1I scores achieved by t he five gdps fell during the study from 0.81 to 0.76, whereas that for mGI rose from 0.87 to 0.92. It is concluded that it is possible to tr ain gdps to achieve high levels of inter-examiner consistency in the u se of P1I and mGI and for these high levels to be maintained during a 12-month multicentre study.