DILIGENCIAS DE INVESTIGACIÓN REALIZADAS POR LA AUTORIDAD ELECTORAL
3. De igual forma y a efecto de allegarse de mayores elementos, para el esclarecimiento de los hechos denunciados, y determinar lo que en derecho
8.3.1.1 A multivariable analysis of factors independently influencing a
referral for DHSW intervention through all pathways: across Scotland
A consequence of the underuse of the Childsmile referral pathway is that looking at the factors associated with health visitor referrals through this pathway does not accurately inform us about the factors associated with whether the right children were being targeted by any pathway, including local forms and birth books. This problem could only have been identified after the data had been linked during the project and, although surprising, highlighted a major issue with the referral pathway, that would not have been picked up without the linkage. We now address this issue in this section.
Section 8.2 has shown that some health visitor referrals through the Childsmile referral pathway did not result in a response from a DHSW and section 8.3 has shown that some ‘incomplete’ records resulted in a DHSW response. In order to assess if the right children were being referred, it was therefore necessary to look at DHSW response (attempted intervention) rather than health visitor
referral. ‘Attempted DHSW intervention’ is an indicator that a child has been referred through an available pathway.
The numbers and percentages of children referred for DHSW intervention by all pathways is broken down by geographical and family variables in the cohort table in Table 6.1 and 6.2 under the heading ‘Children who received an attempted DHSW intervention 1st Sept 2010-31st Dec 2012’.
Univariable and multivariable logistic regression models were produced to establish the key variables associated with a referral by any pathway. The results table can be viewed in Appendix 21. Multivariable results are reported and any deviation from the univariable models is identified in the following descriptions.
8.3.1.2 Health board
The models show a greatly reduced likelihood for children in some health boards to receive an attempted intervention when compared to NHS Ayrshire and Arran. It should be considered that this is, in many cases, due to the DHSW role being established at a later date in some health boards than in NHS Ayrshire and Arran, which was one of the health boards in which the role was first introduced.
8.3.1.3 Area-based deprivation
In both the univariable and multivariable models there is a gradient in
association between SIMD and attempted DHSW intervention, as there was for health visitor referral through the Childsmile referral pathway. Those children living in increasingly deprived areas are increasingly more likely to receive an attempted DHSW intervention.
8.3.1.4 Urban-rural classification
In a univariable regression, children living in large urban areas are more likely to receive an attempted DHSW intervention than those in remote rural areas. This association is no longer significant in a multivariable model. Children living in all other types of urban and rural area are more likely to receive an attempted intervention than those living in a remote rural area, when accounting for all
other factors (‘other urban area’ AOR=1.14 95% CI [1.01, 1.29]; ‘accessible small town’ AOR=1.79 95% CI [1.56, 2.09]; ‘remote small town’ AOR=2.67 95% CI [2.33, 3.06]; ‘accessible rural’ AOR=1.26 95% CI [1.10, 1.44]). Unlike the association with health visitor referral through the Childsmile referral pathway, those living in large urban areas were not significantly more likely to receive an attempted intervention that those living in remote rural areas.
8.3.1.5 Feeding type
The association between feeding type and health visitor referral through the Childsmile referral pathway is also seen for attempted DHSW intervention. Children who are bottle-fed are slightly more likely to receive an attempted DHSW intervention than those who are breastfed (AOR=1.06 95% CI [1.00, 1.13], p=0.04), independent of all other factors.
8.3.1.6 Smoking
Smoking was not found to be independently associated with attempted DHSW intervention (AOR=0.95 95% CI [0.89, 1.02], p=0.19) across Scotland.
8.3.1.7 Health Plan Indicator
In a univaraible model, those with an ‘additional’ care plan are shown to be less likely to receive an attempted DHSW intervention than those on a ‘core’ plan (OR=0.59 95% CI [0.57, 0.61], p=<0.001). However, when adjusted for all other variables in a multivariable model, children who were assigned an ‘additional; care plan and those assigned an ‘intensive’ care plan were more likely to be referred by any means than those on a ‘core’ plan (‘Additional’= AOR=1.26 95% CI [1.20, 1.32], p=<0.001; ‘Intensive’= AOR=1.52 95% CI [1.32, 1.75], p= <0.001).
8.3.1.8 Level of risk
In a univariable regression level of risk had a gradient in association with attempted DHSW intervention, with attempted intervention being increasingly more likely as the number of risk factors increased. There was a slightly stronger gradient of association for this analysis of all referral pathways than for the Childsmile referral pathway only. Level of risk was not entered into the
were included. The model for Scotland and each health board can be viewed in Appendix 22.
8.3.1.9 A multivariable analysis of factors independently influencing referral for DHSW intervention through all pathways: within health boards
A multivariable regression for NHS Forth Valley, NHS Grampian, NHS Lothian and the Island boards could not be performed due to low numbers of children
receiving attempted interventions in these health boards at this early stage of implementation. There was substantial variation in the factors influencing referral within health boards, with referrals in some health boards being much less influenced by geographic or family variables than others. Where the same factors emerged from the model, the size of effects differed.
The association between attempted DHSW intervention and area-based deprivation held a gradient for the majority of health boards (NHS Ayrshire & Arran, NHS Fife, NHS Greater Glasgow & Clyde, and NHS Tayside) and those in the most deprived areas were also more likely to receive an attempted DHSW intervention in NHS Highland, and NHS Dumfries & Galloway. Results can be viewed in Appendix 23. In NHS Lanarkshire, those in the three most deprived groups were more likely than the least deprived to receive an attempted DHSW intervention but there was not a clear gradient and the likelihood of referral was similar between categories (‘SIMD Q1’ AOR=1.21, 95% CI [1.08, 1.37]; ‘SIMD Q2’ AOR=1.30, 95% CI [1.15, 1.47]; ‘SIMD Q3’ AOR=1.14, 95% CI [1.01, 1.29]).
Where feeding type was included in the final regression models, children who were bottle-fed were more likely to receive an attempted intervention. Where smoking was included, children who were exposed to smoking were more likely to receive an attempted intervention. There was a gradient in association for HPI in NHS Ayrshire & Aran, NHS Fife, NHS Greater Glasgow & Clyde and NHS Dumfries & Galloway. In NHS Borders, NHS Highland and NHS Lanarkshire, the ‘additional’ care plan was associated with increased likelihood of attempted DHSW intervention.
A univariable regression for level of risk was carried out for each health board (Appendix 22). Within health boards the likelihood of an attempted intervention increased as the number of risk factors increased in all but NHS Lanarkshire (NHS
Grampian, NHS Lothian, NHS Forth Valley and the Island boards could not be included due to low numbers in each risk category). As with health visitor referral through the Childsmile referral pathway (see Chapter 7), in NHS Lanarkshire the percentages of children in each risk category were similar
(48.65-55.34%) and there was no significant association between level of risk and referral by any means.