Developing a population data linkage cohort to investigate the impact on child oral health outcomes following the roll-out of the Childsmile programme in Scotland

Kidd, Jamie Brian Rutherford (2019) Developing a population data linkage cohort to investigate the impact on child oral health outcomes following the roll-out of the Childsmile programme in Scotland. PhD thesis, University of Glasgow.

Abstract

Background:
Good oral health is essential for eating, smiling and talking, yet dental decay (caries) is amongst one of the most common diseases worldwide, with untreated caries in deciduous teeth affecting 9% of the child population at a global level. In high income countries, dental care has remained focused on treating oral diseases, rather than preventing them in the first place. Oral diseases predominantly affect the most socioeconomically deprived members of society and have strong links to the social determinants of health. In 2002, the Scottish Executive’s consultation document ‘Towards Better Oral Health in Children’ reported that by the age of three, 60% of children living in the highest areas of deprivation were suffering from caries and that more than half of five-year-olds across Scotland were also burdened by this disease.

In 2005, the Scottish Government published the national oral health and dental service strategy ‘An Action Plan for Improving Oral Health and Modernising NHS Dental Services in Scotland’. In response to the Action Plan, Childsmile, the national oral health improvement programme for Scotland was initially launched as demonstration projects in the West and East of Scotland and then rolled-out nationally during 2010 and 2011. Both the delivery and the evaluation of Childsmile was developed using theory- and evidence-based approaches. This combination led to the development of a multi-agency and multi-service programme that included the involvement of health visitors, specially trained community-based lay workers (Dental Health Support Workers), nurseries and schools (including supervised toothbrushing), as well as dental services (and the wider dental team).

Aim:
The overarching aim of this thesis is to assess the impact of the measurable input at the individual child level of the Childsmile programme on the oral health outcome of ‘obvious dental caries experience’ of five-year-olds, taking into account socioeconomic deprivation. This aim will be met via answering the following research questions. (1) Is the Childsmile programme and its universal and targeted components being delivered as envisaged and does this differ by socioeconomic status of the child population? (2) What is the association between obvious dental caries experience and sociodemographic characteristics (age, sex, and area-based deprivation) of the five-year-old study cohort? (3) What is the impact of each of the individual components of the Childsmile programme on obvious dental caries experience, and is there variation of the impact by sociodemographic characteristics? (4) What is the independent effect of each of the Childsmile components over and above the other interventions and the relative contributions of each of the components of Childsmile on obvious caries experience, within both the whole child population and for children living in the areas of highest deprivation?

Methods:
The first challenge to be tackled in order to evaluate Childsmile in this way was to establish a series of linked and anonymised child level source datasets from multiple sources via a process of data management, quality and completeness checks, and then via data linkage, create a cohort that could capture the exposure to the components of Childsmile and be able to assess the impact on the oral health outcomes. There were two phases to this work due to the sheer complexity of the difficulties that presented themselves. Phase One was considered to be a pilot phase, and indeed was one of the national pilot projects that provided an early opportunity for the infrastructure of the National Safe Haven secure remote data linkage service in Scotland to be tested. The processes and learning not only informed the Childsmile evaluation, but also the wider methods of data linkage systems in NHS Scotland.

The development of the first phase towards creating the ultimate cohort included: i) successful navigation of the initial ethical and information governance approval processes for accessing and linking the source datasets that were to be used in the study; ii) identification of the appropriate datasets that were to be used in the evaluation of the outcomes of the Childsmile programme; iii) installation and pilot use of the National Health Service (NHS) National Service Scotland (NSS) electronic Data Research and Innovation Service (eDRIS) remote National Safe Haven infrastructure software; and iv) extraction and uploading of these datasets (including the initial linkage process) into the Safe Haven; and v) primary analysis of the datasets to measure and validate the data quality and completeness.

Phase Two consisted of: i) gaining updated consent to access and link multiple individual child level datasets to facilitate the outcome analysis of Childsmile; ii) using updated data linkage processes for the sharing and uploading of the refreshed source datasets into the Safe Haven; iii) further primary analysis of the datasets to measure and validate the data quality and completeness; and iv) initial primary analysis of the datasets to validate the linkage process, this step included developing comprehensive data dictionaries.

While the Childsmile Data Linkage project resource created included 24 datasets in total, the datasets that contributed to the final analysis cohort were: i) ‘The 2009 Scottish Index of Multiple Deprivation’ (SIMD) – which is an area-based socioeconomic measure; ii) ‘Child Health Systems Programme Pre-School 6-8 Week Review’ – which consists of proxy birth and population data; iii) ‘National Dental Inspection Programme’ (NDIP) – which provides data on child oral health outcomes; iv) ‘Management Information and Dental Accounting System Treatments’ (MIDAS) – primary dental care appointment data; v) Dental Health Support Worker (DHSW) contact datasets; vi) ‘Toothbrushing Consent’ – supervised nursery and school toothbrushing participation; and vii) ‘Fluoride Varnish Visits’ – nursery and school fluoride varnish applications. The datasets in the Safe Haven were validated using many different quality and completeness methods, including comparisons to published reports.

It was decided that in Phase Two, the year group with a 2014/2015 Primary One ‘P1’ (five-year-old) Basic NDIP dental inspection would be analysed. This was the most current NDIP year of inspection available at the time of this work and it was deemed to be appropriate as it was the first year group that had been born into the nationally ‘rolled-out’ Childsmile programme. The primary outcome for this work was the presence of ‘obvious caries experience’ in the Basic NDIP, which will be known in shorthand as ‘Caries Experience’. There were 57,410 P1 NDIP individual child records in the 2014/2015 cohort year. After various exclusion criteria were applied, 50,379 children (88%) remained in the final study cohort, which remained representative of both the population and the oral health outcomes of the published 2014/2015 NDIP report. Overall, the quality and completeness of the datasets to be used in the study were high with no concerns highlighted when the completeness of the variables were checked (although this was not the case in Phase One). Of the 50,379 in the cohort, 30% (n = 15,032) had Caries Experience. The four Childsmile components that were evaluated were: Dental Health Support Worker Contacts (‘DHSW Contacts’); Childsmile Contacts at a Dental Practice; Time Consented to Toothbrushing in the supervised nursery and school programme (‘Time Toothbrushing’); and Nursery and School Fluoride Varnish Applications.

Extensive statistical analyses were carried out. These included assessing whether or not the Childsmile programme was being delivered as envisaged, analysing the association between the potential confounders of age, sex, and SIMD (quintiles) with Caries Experience; the association between the Childsmile components and Caries Experience; the interactions between the Childsmile Components and the potential confounders; and the associations of the Childsmile components on Caries Experience after being individually adjusted for the potential confounders (Model One), and again after being adjusted for the potential confounders and the other three components (Model Two).

Results:
The delivery of the programme was being delivered mostly as envisaged in terms of the targeted and universal components. However, there remains room to improve the reach of the components. The delivery of the universal supervised toothbrushing programme overall is high, with children from more deprived areas having slightly better participation.

The variable ‘DHSW Contacts’ has four categories: ‘Not Targeted’, ‘0 contacts’, ‘1 contact’, and ‘2 plus contacts’. The children who were targeted for a DHSW contact who did not receive a contact (‘0 contacts’) were the referent category for comparisons. The Model One results are provided as odds-ratios that are adjusted by age, sex, and SIMD (AOR). Those who were targeted and who received only one contact had 37% lower odds of Caries Experience than those who were targeted and not reached, AOR = 0.63; 95% CI (0.55 to 0.72), whereas those who had received two or more contacts did not have significantly lower odds than those targeted and not reached, AOR = 0.91; 95% CI (0.76 to 1.10). The effect of DHSW contacts on Caries Experience after the Model Two adjustment attenuated slightly but did not change the overall results.

The variable ‘Childsmile Dental Practice Contacts’ has eleven categories; ‘0 Childsmile contacts’, ‘1 contact’ … ‘9 contacts’ and ‘10 plus contacts’. Children with ‘0 Childsmile contacts’ at a dental practice were the referent category. Those attending ten or more times (two or more visits per year) experienced a 67% reduced odds of Caries Experience, AOR = 0.33, 95% CI (0.18 to 0.60), compared to those who never attended. There was no change in the effect of the Childsmile contacts at a dental practice after the Model Two adjustment.

The Childsmile component ‘Time Toothbrushing’ has four categories; ‘0 (no consent)’, ‘Up to 1 year’, ‘1 to 2 years’ and ‘2 plus years’. Each category represents a single year i.e. ‘Up to 1 year’ is one day to one year of toothbrushing consent prior to the NDIP inspection date. Children that were not participating in the supervised toothbrushing component, ‘0 (no consent)’, were the referent category. Compared to those who were not consented to toothbrushing, those who had participated in the toothbrushing component for two or more years had lower odds of Caries Experience, AOR = 0.81; 95% CI (0.76 to 0.87). There was a minimal strengthening of the effect of toothbrushing after the Model Two adjustment.

The variable ‘Nursery and School Fluoride Varnish Applications’ has seven categories: ‘Not Targeted’, ‘0 applications’ (the referent category), ‘1 application’, ‘2 applications’, ‘3 applications’, ‘4 applications’ and ‘5 plus applications’. For children that were targeted, the odds of Caries Experience decreased with an increasing number of fluoride varnish applications but was only significant with four or more applications (‘4 applications’ AOR = 0.89; 95% CI [0.82 to 0.96]). After the Model Two adjustment, there was no reduction in the odds of Caries Experience for this component (‘5 plus applications’ AOR = 0.99; 95% CI [0.91 to 1.08]).

There was a strong interaction (effect modification) between SIMD and the association between Time Toothbrushing and Caries Experience. For children living in the 20% most deprived areas of Scotland (SIMD 1) there was a reduction in the odds of Caries Experience for those children that had only been toothbrushing for one year (SIMD 1 AOR = 0.77; 95% CI = [0.64 to 0.93]) compared to those that had never been consented to toothbrushing, and with each additional year of Time Toothbrushing, the odds of Caries Experience reduced further. A similar but not so marked effect (AOR = 0.89; 95% CI = [0.72 to 1.09]) was observed for children living in SIMD 2 areas. After the Model Two adjustment, the strengthening of the effect of toothbrushing was strongest among children from SIMD 1. A weaker interaction between the number of Nursery and School Fluoride Varnish Applications and Caries Experience by Area Based-Deprivation (SIMD) was observed. Children living in SIMD 2 and 3 initially had a reduction in the odds of Caries Experience after five and four varnishes respectively, but the effect of this component was attenuated after the Model Two adjustment, and a reduction in the odds of Caries Experience was only observed for children living in SIMD 2 after receiving five or more varnishes (Model Two AOR = 0.80; 95% CI = [0.67 to 0.95]).

Conclusions:
This thesis has shown that it was possible to create a study cohort via data linkage of routine administrative datasets and to undertake an initial evaluation of the impact of the components of the Childsmile Programme – which is a complex multifaceted national public health intervention - on the oral health of five-year-old-children.

The four main components of the Childsmile programme examined are largely being delivered as envisaged with respect to their differing targeted and universal aims, however, there remains room to improve the reach of aspects of the programme.

There was evidence to suggest that DHSW contacts were associated with a reduction in the odds of obvious caries experience when the child was contacted only once. This reduced risk disappeared if the child received additional contacts, which suggested that although there has been some success in DHSWs identifying children at a higher risk of obvious caries experience, the delivery of this component in terms of reducing the risk of caries for children at a higher need was less clear.

Attendance at a Childsmile dental practice was associated with a reduction in odds of obvious caries experience, with a clear dose response observed as the number of contacts increased and with no variation observed across the deprivation categories. It could be surmised that regular attendance at a dental practice may be a proxy for a positive approach to oral health within the family, and that the parents attending these contacts are already motivated or enabled towards caring for their child’s oral health. Moreover, no additional benefit of fluoride varnish application, over and above regular attendance at a dental practice, was seen in reducing the odds of caries experience.

Supervised toothbrushing was most effective at reducing the odds of dental caries when a child was living in an area of high deprivation, with the effect increasing the longer these children had been consented into the programme. For the children from the least deprived areas, there was no effect on the odds of caries experience observed regardless of the length of time that they had been participating in supervised toothbrushing.

The results of the analysis of the nursery and school fluoride varnish component show an initial independent effect, however, when those not contacted are taken into account, there is no overall effect, and reduced odds of developing caries only emerges among a very small number of children from SIMD 2 areas who received five or more fluoride varnish applications.

The findings of the thesis show differing risk associations between the various components of the programme and caries experience across area-based deprivation categories. This work has developed a resource, utilising nation-wide population routine administrative datasets, which can be used for further Childsmile evaluation. The findings can inform the future direction of the Childsmile programme and child oral health policy in Scotland.

• Item Type: Thesis (PhD)
• Qualification Level: Doctoral
• Keywords: Childsmile, data linkage, complex evaluation, oral health improvement, community dental, routine administrative data, dental caries, dental public health.
• Subjects: R Medicine > RA Public aspects of medicine > RA0421 Public health. Hygiene. Preventive Medicine; R Medicine > RJ Pediatrics > RJ101 Child Health. Child health services; R Medicine > RK Dentistry
• Colleges/Schools: College of Medical Veterinary and Life Sciences > School of Medicine, Dentistry & Nursing > Dental School
• Supervisor's Name: Conway, Professor David I. and Alex, Doctor McMahon
• Date of Award: 2019
• Depositing User: Dr Jamie B.R. Kidd
• Unique ID: glathesis:2019-76718
• Copyright: Copyright of this thesis is held by the author.
• Date Deposited: 23 Dec 2019 11:35
• Last Modified: 01 Aug 2022 08:08
• Thesis DOI: 10.5525/gla.thesis.76718
• URI: https://theses.gla.ac.uk/id/eprint/76718