Background

Dental caries in pre-school children remains a problem in Hong Kong. In a recent population-based oral health survey, the caries experience of pre-school children was over 50% in Hong Kong, with an extensive proportion of untreated carious teeth (over 90%) [1]. One in 20 of the children had a dental abscess associated with an extensively decayed tooth. The prevalence of dental caries among pre-school children showed limited improvement over the past decade and the extent and severity have increased. Among pre-school children, the number of decayed, missing and filled teeth was 2.3 in 2001 and 2.5 in 2011 [1, 2]. Children from disadvantaged and socially marginalised populations have even higher caries experience and are at higher risk for caries.

The high degree of untreated dental caries indicates that there is inadequate access to quality oral healthcare for pre-school children in Hong Kong. Management of pre-school children’s oral health requires commitment from professionals, patients and their families. Dentists’ attitudes can contribute to the problem, resulting in a barrier to provide adequate and quality dental care [3]. Dentists in Hong Kong do see the value of dental treatment for pre-school children, but most general dental practitioners consider children’s coping skills to dental treatment as a problem and have a negative attitude towards the provision of dental care for children [4]. A wide variation in caries management approaches among dental practitioners has also been identified [5], reflecting the uncertainty within the profession about the most suitable and effective dental care approaches for pre-school children with dental caries. Thus, these factors highlight the need for clinical guidelines relating to dental caries for pre-school children in Hong Kong.

Clinical practice guidelines are systematically developed statements that can assist practitioners and patients in making decisions about appropriate healthcare for specific clinical circumstances [6, 7]. It is the major available tool to assist in the translating of research evidence into practice. With the best available evidence and opinion in guidelines, healthcare professionals can be assisted in making appropriate clinical decisions, minimising variation in practice, and promote effective and safe patient outcomes.

Expectations and requirements have been proposed internationally for the development of quality clinical guidelines [8, 9]. The development and implementation of guidelines that meet international standards require substantial time, expertise and resources, especially during systematic identification and critical analysis of evidence. Guideline adaptation is a systematic approach for considering the endorsement or modification of guidelines produced in one setting for application and implementation in another as an alternative to de novo guideline development or as a first step in the process of implementation, while preserving evidence-based principles [10]. The ADAPTE process is a comprehensive framework for guideline adaptation. ADAPTE can reduce duplication of effort and enhance efficiency in guideline development. The manual and toolkit for the ADAPTE process are available at: http://www.g-i-n.net/gin [11]. The ADAPTE process consists of three main phases: set up, adaptation and finalisation.

The Delphi technique is a formal iterative structured process that aims to gather consensus of opinion, judgement or choice among a panel of experts [12–15]. The attitudes, needs and priorities of the panel can be explored during the interactive process. Delphi is suitable for reaching consensus on topics with dissenting opinions, uncertainty or absence of evidence. Guidelines developed with consensus of stakeholders are more likely to be accepted by the dental profession.

There are several quality guidelines for managing dental caries for pre-school children worldwide. Nevertheless, guidelines ascribed to other countries or areas can encounter barriers to their practical application in a context with a different culture and healthcare system [16]. Localised or adapted clinical guidelines on the basis of available evidence in the local context are advocated, with the consideration of cost-effectiveness and suitability for the local population. This project aims to develop consensus evidence-based clinical guidelines to inform the practice regarding the determination of caries risk and developing preventive and management strategies for pre-school children in Hong Kong using the ADAPTE process and a web-based Delphi consensus technique.

Methods

Set up phase

The project was led by a team of researchers based at the Faculty of Dentistry, the University of Hong Kong. The team comprised of specialists in the field of Dental Public Health and Paediatric Dentistry. Members of the team reported no conflict of interests. The scope and key questions for the guidelines were established. The key stakeholder, the Hong Kong Society of Paediatric Dentistry (HKSPD), was identified and invited to be the collaborator on developing the guidelines. An introduction to guideline development and the ADAPTE process was presented to the members of the Society.

Adaptation phase

Search and screen guidelines

A search for relevant guidelines, protocols or policy statements in guideline repository websites, websites of various key guideline development organisations, and dental professional organisations, plus MEDLINE (Ovid), PubMed, Google and several other databases was conducted. The subject heading ‘dental caries’ was used for the search in combination with several other keyword terms: ‘tooth decay’ or ‘diagnosis’ or ‘management’ or ‘prevention’ or ‘approach’ or ‘caries risk’ or ‘caries risk assessment’ or ‘caries assessment’ or ‘caries prediction’ or ‘children’. An appropriate search strategy for each database was employed.

Twenty non-duplicate guidelines, protocols or policy statements were identified in the search. No language restrictions were applied at this point. The retrieved guidelines were screened according to list of a priori inclusion/exclusion criteria: (1) dental caries, regarding its diagnosis, prevention, management or risk assessment was the main topic of the guidelines; (2) the target population of the guidelines was pre-school children with primary dentition; (3) the guidelines were for local or national use; (4) publication date should be from 2000 onwards to date of the search (December 2011); (5) indication that a literature search, ideally using a systematic method, had been conducted; (6) references were included; and (7) explicit link between the recommendations and the supporting evidence was indicated.

Access guidelines

Eleven guidelines/protocols were retained for further assessment after initial screening based on the inclusion/exclusion criteria. A panel group with four clinical teaching staff from the Department of Paediatric Dentistry of the University of Hong Kong was set up. The panel members were briefed about the guideline development and evaluation process for the retrieved guidelines by the guideline development group. They were invited to review the 11 retrieved guidelines and appraise their quality using the AGREE II (Appraisal of Guidelines for Research and Evaluation II) instrument individually [17]. The AGREE II instrument is a generic instrument to assess the process of guideline development and reporting of this process in the guidelines. The instrument consists of 23 items on a 7-point response scale comprising six quality-related domains: scope and purpose, stakeholder involvement, rigour of development, clarity of presentation, applicability, and editorial independence. The instrument’s overall assessment includes the rating of the overall quality of the guideline and whether the guideline would be recommended for use in practice. The scores for each of the six AGREE II domains of each of the retrieved guidelines were calculated according to the rubrics. Pooled domain scores and overall assessment scores for the 11 guidelines are presented in Table 1.

Table 1 The six AGREE II domain scores and overall assessment scores of the 11 retrieved guidelines
Full size table

Three ‘Source’ guidelines for adaptation in the ADAPTE process were selected [18–20]. The selection was based on the AGREE II scores on Domain 3 (rigour of development) and overall assessment scores. The currency of the guidelines, applicability and extent of the key questions for the guidelines were also taken into account. Although the Scottish Dental Effectiveness Programme 2010 [21] had higher scores than American Academy of Pediatric Dentistry 2011 [20], the guideline was not selected because the content and extent of key questions covered was very similar to the Scottish Intercollegiate Guidelines Network (SIGN) 2005 [18]. Therefore, the next guideline on the list, that of American Academy of Pediatric Dentistry 2011 [20], was selected. The three ‘source’ guidelines provided different perspectives on caries prevention and management strategies and on caries risk assessment, and together addressed most of the key questions for the developing guidelines. With regard to the clarity of presentation of the guidelines, the other three retrieved guidelines [21–23] were kept as ‘satellite’ for reference.

Drafting of guidelines

The recommendations from the three source guidelines were compiled in a recommendation matrix and compared. The grading and levels of evidence associated with each of the recommendations provided by the source guidelines were placed within each cell and was critically appraised using the SIGN methodology checklist [24].

The consistency between the evidence presented in the source guidelines, their interpretation of the evidence, and the resulting recommendations were evaluated. The original evidence supporting the interpretations and recommendations in the source guidelines were reviewed. The currency of the source guidelines for the adaptation process was also assessed to ascertain whether the most current evidence had been included.

A literature search of databases and websites for systematic reviews and randomised clinical trials, quasi-randomised trials or longitudinal cohort studies since the source guidelines published was conducted. The relevant publications were critically appraised and graded. The publications searched were between 2004 and June 2012. Year 2004 was the end date of search in the oldest source guideline by SIGN [18]. This limited search period was established to identify new publications that were not included in the three ‘source’ guidelines. The search was later updated to the period between 2004 and December 2012 following external review for the preparation of the final guidelines.

Draft-adapted/adopted recommendations and summaries of the evidence were assessed by the guideline development group for (1) consistency between the evidence presented, (2) its interpretation and the resulting adapted/adopted recommendations, and (3) the effect of any new evidence on the recommendations. For each of the draft-adapted/adopted recommendations, their acceptability and applicability in the local Hong Kong context, taking culture, organisation and expertise of dental services and population characteristic, beliefs and values into account, were discussed and considered in the meetings among the guideline development group. Using informal consensus, the recommendations were either adapted, updated or rejected based on the latest evidence identified. A preliminary list of 91 draft recommendations was developed.

Finalisation phase

A consultation draft of the adapted guidelines was produced with the preliminary list of 91 draft recommendations. The draft was circulated among the guideline development group members for comment and was revised where necessary.

External review process

All 68 ordinary and associate members (excluding guideline development group members and panel members) of the HKSPD (key stakeholder) were invited to be external reviewers by email. Ordinary HKSPD members were either registered paediatric dentists or practitioners who had completed structured training in the specialty and were experts in Paediatric Dentistry in Hong Kong. Associate members of HKSPD were general medical or dental practitioners who were interested in the practice of Paediatric Dentistry. All of them would be the potential users of the developing guidelines.

Electronic copies of the consultation draft of the adapted guidelines were circulated to the external reviewers. The external reviewers were invited to contribute any comments on the consultation draft of the developing guidelines. An Internet-based structured questionnaire on a 5-point Likert scale was used to ask the external reviewers about their view on the strengths and weaknesses of the consultation draft, any need for modification, whether they would approve the draft guidelines and use them in their practice, and how it would impact or change their current practice. The 22 items about the external reviewers’ general view on the guidelines were added up to create a general view score ranging from 0 to 110, with a higher score indicating a more positive attitude towards the guidelines. The external reviewers consented to participate in the guideline external review process and for the data to be published.

Ranking of recommendations

A two-round Internet-based password-protected Delphi consensus process was conducted among the external reviewers. The Delphi process was carried out to collect opinions or comments from external reviewers so as to establish consensus recommendations for the guidelines. Before each round, external reviewers were reminded of the upcoming surveys by at least four emails. The draft recommendations were presented to the external reviewers. They rated their agreement with each of the 91 preliminary adapted/adopted recommendations using a 9-point Likert scale (1 = strongly disagree, 5 = no opinion or not enough information or experience to judge, 9 = strongly agree) and to contribute any comments on the recommendations. They were given the option of not commenting on a recommendation when the topic was outside their respective areas of expertise. This process was repeated during the second round, in which reviewers’ ratings, dissenting opinions or uncertainty, or absence of evidence were combined and provided anonymously to all of them. They were allowed to revise their first round responses in view of further responses in the second round. After the first round, the guideline development group met to modify some recommendations for clarity and presentation for the second round.

Classification of recommendations

For the first round of Delphi, recommendations rated by < 20% of the external reviewers were discarded. Recommendations rated by 80% of the external reviewers were discarded if the overall median score fell within the bottom tertile (1–3, disagree) and deferred until the second round if the overall median score was 4–6 (neutral). Recommendations with overall median scores within the top tertile (7–9) were classified as follows: those rated by < 80% of the external reviewers were deferred until the second round, and those rated by ≥ 80% of the external reviewers were classified as ‘high agreement’ if they had an overall median score of 9 or as ‘agreement’ they had an overall median score of 7 or 8.

Similarly, in the second round of Delphi, recommendations rated by < 20% of the external reviewers were discarded. Recommendations rated by 80% of the external reviewers were discarded if the overall median score fell between 1 and 6. Recommendations with an overall median score within the top tertile (7–9) were classified as follows: those rated by < 80% of the external reviewers were discarded, and those rated by ≥ 80% of the external reviewers were classified as for the first round. The classification of recommendations according to the overall median score and percentage of ratings by external reviewers in the two-round Delphi process is illustrated in Fig. 1.

Fig. 1
figure 1

Classification of recommendations according to the overall median score and percentage of ratings by external reviewers in the first and second round of Delphi surveys

Full size image

Data analysis

Data were analysed using IBM® SPSS Statistics 19 (SPSS Inc., Chicago, IL, USA). Recommendation-specific medians were estimated for each round. The inter-rater agreement between participants in each round was evaluated using intraclass correlation coefficient with 95% confidence intervals (CI). Microsoft Excel for Mac 2011 was used for preparing tabular and graphic presentation.

Production of the final guideline

All the feedback from the external reviewers were documented and presented in the guideline development group meeting for discussion. The guidelines and recommendations were further revised to incorporate feedback from the external reviewers where necessary.

Results

The flow chart in Fig. 2 describes the number of guideline recommendations and participation of external reviewers in the external review process. Forty-one out of 68 invited members agreed to participate (response rate: 41/68 = 60.3%). In the Delphi process, there were 41 external reviewers in the first round and 36 for the second round (drop-out rate: 5/41 = 12.2%; overall response rate: 36/68 = 52.9%). The intraclass correlation coefficients were 0.96 (95% CI, 0.95–0.98) and 0.92 (95% CI, 0.87–0.95) for the first and second rounds, respectively.

Fig. 2
figure 2

Recommendations and external reviewers’ participation flow chart

Full size image

All the participated external reviewers rated all the items. Of the original 91 recommendations, 70 were retained and classified as follows: five items reached consensus (‘high agreement’), 65 items reached near consensus (‘agreement’), 21 items were discarded. Results of the recommendations in the two-round Internet-based Delphi consensus surveys are listed in Table 2. Only those recommendations (N = 70) that had gained consensus from the external reviewers in the Delphi process were included in the final list of recommendations of the guidelines.

Table 2 Results of the recommendations of the adapted guidelines in the two-round Delphi survey
Full size table

The ratings of the external reviewers about their general views on the strengths and weaknesses of the guidelines are listed in Table 3. The majority of the external reviewers (82.9%) agreed that there was a need for the guidelines on caries prevention and management by caries risk assessment for pre-school children in Hong Kong (statement 2). Over half of them (>60%) agreed to practice the guidelines when they are implemented (statement 21–22) and agreed that the guidelines were properly developed (statement 1, 3, 4). They were also positive about the relevance and applicability of guideline recommendations in practice and on patients (statement 5–13, 19). However, their views regarding the possible effects and outcomes of the guidelines were neutral (statement 14–18). Over 60% of the external reviewers approved the guidelines as clinical practice guidelines. The mean general view score of the guideline was 75.66 (SD = 7.80; out of a maximum possible score of 110).

Table 3 General view of the external reviewers on the strengths and weaknesses of the guideline
Full size table

Discussion

Clinical practice guidelines are the major means to summarise and translate rapidly changing research evidence into practice and assist clinical decision making [25, 26]. Clinicians require knowledge of the current expert recommendations and guidelines to manage specific clinical problems and improve their patients’ outcomes. Developing evidence-based guidelines on caries prevention and management by caries risk assessment for pre-school children in Hong Kong is important. The guidelines have great potential in improving the current oral health and dental care of pre-school children in Hong Kong if implemented.

Developing clinical practice guidelines using the ADAPTE process and Delphi consensus has not been reported in the literature. This approach allows developing guidelines with consensus evidence-based recommendations. The literature suggests that guidelines developed without active participation and involvement of the intended users and which were partially without credible scientific basis were not fully supported by the profession and were poorly accepted by the practitioners [27, 28]. ADAPTE is a comprehensive systematic framework for adapting guidelines produced in one setting to be used in a different cultural context. The framework was developed to meet the current challenges of guideline development and implementation that require lots of time, expertise and resources [29]. The process is designed for a range of stakeholders, such as guideline developers, healthcare providers, consumers and policymakers at the local, national and international levels and in different healthcare sectors [10]. The process also allows users to customise the process to their own context, resource constraints and expertise. In this project, in addition to the external review process set up in the ADAPTE framework, the Delphi consensus technique was employed to gain reviewers’ comments and consensus on the guideline recommendations. The developed guidelines are evidence-based with up-to-date knowledge and have gained consensus among the profession. Concerns in the acceptability and barriers in implementation are reduced.

A consensus approach using the Delphi technique was employed during the external review process of the guidelines. Delphi strengthened the external review process of this guideline development. In addition to external reviewers’ comments and views on the draft-adapted guidelines, their consensus on the guidelines’ recommendations was achieved. External reviewers were members of the Hong Kong Society of Paediatric Dentistry, who were either specialists of paediatric dentistry or dental/medical practitioners with active participation in the field. This heterogeneous sample ensured that a wide spectrum of opinions of the guideline could be collected. The group consisted of 41 members in round 1 and 36 members in round 2, thus reducing the potential of selection bias of ‘experts’ for the Delphi process [13, 14, 30]. Moreover, the invited external reviewers are likely to be the major users of the guidelines when implemented. Involving them in the external review process using Delphi promotes ownership and acceptance of the guidelines, ultimately reducing the barrier of guideline adherence in practice in the future [31]. Involvement of the general public in the external review process could be considered in the future. They are one of the key stakeholders of the guidelines. Their input could help develop more comprehensive evidence-based guidelines with minimum barriers during implementation.

The response rate in the second round of Delphi (52.9%, 36/68) dropped when compared to the first round (60.3%, 41/68). Recruitment of dental practitioners in a research study is a challenge and it is also difficult to retain a high response rate in each round of Delphi [13, 14]. Instead of employing the classical four-round Delphi, a two-round web-based Delphi was used [14]; the use of Internet avoids the need for face-to-face meetings among the external reviewers. The external reviewers can review the guidelines at different geographic locations and time.

The consensus level of the Delphi technique in the literature ranges from 51% to 80%; a bench mark of 70% is used in many cases [30]. We used 80% as the level of consensus in this guideline development project. The intraclass correlation and CIs were good. Consensus was reached for a large proportion of recommendations with high inter-rater agreement between rounds.

The online implementation of the Delphi technique for external review was well received. External reviewers completed the two-round web-based Delphi external review of the guideline over a period of 5 months. They contributed extensive and supportive comments on the guidelines and the recommendations.

Not all evidence-based recommendations were accepted by the external reviewers. Although the recommendation, ‘All children should be encouraged to brush their teeth under adult supervision with fluoride toothpaste containing at least 1000 ppm F’ was based on strong evidence [32], it was not able to gain agreement from the external reviewers. A number of external reviewers were concerned about dental fluorosis in relation to the use of high concentration fluoride toothpaste in a water fluoridated area like Hong Kong. This reflected that the strength of the evidence, applicability to real-life conditions and associated socioeconomic costs influenced the acceptance of recommendations. Moreover, this showed that the guideline development process was effective in identifying potential local barriers in applying research evidence to clinical practice and recommendations allowed their modification to fit local circumstances in response to reviewers’ opinions.

Conclusions

Using the ADAPTE process and Delphi consensus, evidence-based clinical guidelines for caries prevention and management by caries risk assessment for pre-school children in Hong Kong were developed. These guidelines can facilitate the translation of evidence into practice. The framework for developing the guidelines could be employed elsewhere.

Abbreviations

AGREE II, Appraisal of Guidelines for Research and Evaluation II instrument; HKSPD, Hong Kong Society of Paediatric Dentistry; SIGN, Scottish Intercollegiate Guideline Network (SIGN)