The Behaviour of Dental Nurses Applying Fluoride Varnish to Young Children

A preliminary study has shown the following:

1. Ethical permission for this work is attainable from a Research Ethics Committee.
2. Parents are willing to give consent to their children participating in the study.
3. Dental nurses are willing to give consent for their behaviours to be video recorded within their work settings.
4. Video recordings are feasible with nursery school children, EDDNs and parents present at a nursery school setting.
5. The technical difficulties of collecting audio-visual materials from nursery school settings (e.g., camera placement, storage of video data in the field setting) and anonymising the video data using a specialist masking software can be overcome.
6. Importing video files into the Observer XT coding software is successful.
7. Development of a coding system for nurse-child interaction involving a clinical procedure has been achieved (see Appendix 1 for a brief version of the coding scheme).

We have collected video recordings of more than 50 varnish application sessions in the pilot study of separate children and 6 nurses in three nursery schools. The pilot study has therefore successfully demonstrated the feasibility of collecting high quality data suitable for intensive analysis to improve our understanding of some of the processes involved in the fluoride varnish application in nursery school settings. One of the most interesting findings from the pilot study was that the permission-seeking statements used by EDDNs were negatively correlated with the treatment outcome.

Summary of next phase of study:

• To confirm the suitability of the preliminary coding scheme and to potentially refine and improve the coding scheme using a larger set of observational data collected in nursery school settings across three NHS Health Board regions.
• Further investigation to assess the generalisability of pilot study findings by collecting a larger set of observations of children systematically across a wider set of nurse pairings (with various levels of experience) and children (of different age groups: 3yrs vs 4 yrs) and with or without previous receipt of fluoride varnish application in the Childsmile programme.
• To identify the key features of the nurse-child interaction that can be used to predict a full completion of the fluoride application and to test the hypothesis that frequent targeted reinforcement is effective.
• To explore effects of some additional factors in relation to the relevant participants (e.g., EDDN's level of experience, child's previous dental experience) and some situational factors (e.g., general purpose room or specialist accommodation) on the treatment outcome.
• To develop additional training sessions for EDDNs using the findings of the study to improve the Childsmile programme and to increase acceptability of the varnish application.

Methods

Ethics proposal preparation

We obtained ethics approval for the pilot study. We have successfully addressed and dealt with the ethical issues that arose from the study in the pilot stage. We have experienced no degree of difficulty in consenting staff or children for recruitment of the pilot. We believe that obtaining further ethical approval for an extended study should be achievable due to our previous experience. The difference between the pilot and this main study is an extension in the numbers of child observations. This will require a multisite ethics proposal to be entered into IRAS. In terms of ethical approval from NHS Research and Development (R \& D) offices, we are working with the NHS Research Scotland (NRS), who will assist in obtaining the approvals from different NHS R \& D offices.

We have been welcomed into the setting by the nursing staff and have had no adverse effects in our work in the schools collecting the 50+ child observations. In the main study, we will collect approximately 500 child observations across several NHS Health Board regions. We will continue a positive working relation with EDDNs and nursery school staff developed in the pilot stage. We will be very careful in the consenting and recruitment of nurse and child participants without causing pressure for participation. Anonymity of personal details and confidentiality of the information provided by participants will be dealt with carefully by assigning a unique code to each of the participants and only the research team members will have access to the codes. We do not foresee any additional new factors that will jeopardise approval of an application to REC.

Sample

All EDDNs who have been trained in the Childsmile programme and participated in the delivery of the varnish application at nursery school settings in the three NHS Health Board regions (Tayside, Fife and Forth Valley) are identified as potential staff participants. All newly trained EDDNs who will be delivering the varnish application to nursery school children at the three regions are also included. All nursery school children in the three regions mentioned above whose parents have consented to the Childsmile varnish application programme are identified as potential child participants. The parents of those children are potential parent participants.

Setting up the recording and coding system

The Observer system consists of a powerful suite of programmes that integrates the input of digital video files and sound to a platform on the computer screen for repeated playback and real-time coding, diagrammatic display, simple categorical analysis and more complex sequential analysis. Various options are available for collection of the recordings and these may vary according to setting. Variables to determine the optimum system include: obtrusiveness of the camera and microphone, quality of video image required, and ease of use in the 'field'. The pilot study has shown that the materials collected from the 'field' can be transferred onto a university computer hard disc to be ready for coding purpose. The picture quality and positioning of the camera can be varied to optimise the study of the nurse-child interaction during the varnish application.

Pilot observations

In order to check the feasibility of the new recording system to be used in the main study (a digital DVD camera instead of a web camera in the pilot stage), we will need to make a small number of pilot observations of the nurse-child interactions. This process is important to ensure that the recorded video data can be successfully digitalized and subsequently transferred into the Observer XT8.0 for coding and analysis. Although we have had some experience of dealing with this process in the pilot stage, taking additional pilot observations is still an important step as we are using a new recording system and an updated version of the Observer XT software in the main study. These pilot observations will be also used to double check that the coding scheme developed from the pilot data is functioning properly for a new set of data.

Preparation of coding scheme

The preparation of a coding scheme is a vital component of the study and will require considerable effort and thought to develop a suitable system that is not too cumbersome but also detailed enough to illustrate the various strategies that are adopted by the 'actors' to progress the intervention. Of particular interest are the cases where there is unexpected behaviour or events that require inclusion in the profiling of the interaction. For example the child may refuse certain aspects of the varnish procedure and require coaxing through reassurance or distraction. Interruptions of the procedure by outside events, e.g. parental or nursery nurse involvement, fire alarms can all be present in the scheme and included or ignored in subsequent analysis. The Observer XT system enables the accurate collection of codes from video. The advantage of the updated version of the Observer XT system is the modification of the coding scheme during the coding procedure itself. Previous versions of the software did not allow this 'luxury'. Other available ('freeware') coding systems do not have this capability. Therefore, the coding scheme developed from the pilot data can be tested for fit using new datasets collected in the main study and the coding scheme can also be potentially improved if certain behaviours need to be included and/or modified.

Collection of observational data

Observations of nurse-child interaction will be recorded in the field using a digital DVD camera. The video observation data, initially stored on a DVD, will be transferred in digital format and then stored on a computer internal hard drive for coding and analysis. The Observer XT system is more stable if digital video data is read directly from a computer than from an external hard disc (e.g., DVD), the process of digitalizing video files and storing in a computer is, therefore, an important step for successful behaviour coding and data analysis. Each of the initial raw video files stored on DVDs will be assigned a unique code for data protection purpose and will be locked in a secure place. All digitalized video files stored on a computer internal hard disc will be permanently removed from the computer once coding and analysis is complete. These digitalized video files will then be exported onto an external hard drive that will have all data encrypted and access to data files password protected. An audit record will be compiled to allow inspection of the progress of video data files from video DVD to computer hard disc for coding and then exporting to external media for safe long-term storage. Audit will be monitored by the School Ethics Committee on a 6-month basis.

Coding of video data The research assistants will be responsible for the coding of the video data. This task is labour- intensive and highly skilled. There will be approximately 500 video clips to be processed, which will require a period of about 9-10 months to complete. So far, no specific coding schemes have been developed for this type of nurse-child interactive behaviours. The new coding scheme (Appendix 1) that we have devised from the pilot study consists of about 40 codes and we have just recently conducted a reliability analysis of the scheme. There was 86% agreement using a tolerance window of 2 seconds for coded events. The kappa statistic was 0.84 (p<.001).

Reliability testing

The Observer XT has appropriate reliability (kappa) statistical routines built into the software to provide estimates of agreement controlled for random chance. See 'Coding of video data' above for reliability estimate for the coding scheme.

Analysis of data

Video data

The main software used for coding and analysing video data is the Observer XT 8.0. The rationale of the coding and analysis with The Observer system will be two-folded, a descriptive element. After all verbal and non-verbal behaviours of both nurse and child have been coded according to the coding scheme, the coded behaviours along with the time point when the behaviour took place (if the behaviour is coded as an event) or the duration of the time (if coded as a state behaviour) will be stored as a log file for one particular nurse-child interaction. Frequencies of certain behaviours can then be calculated, for example, nurse's use of praise or fantasy statements; and whether the use of these statements is linked with positive application outcome.

Second, there will be an explanatory element. Sequential analysis will be employed to explore which behaviours tend to facilitate or impede the occurrence of other behaviours. The purpose of this type of analysis is to understand more complex behavioural relationships in terms of how nurses use motivational strategies to encourage cooperative behaviours of children in receiving the varnish application.