This is the first cluster randomised
controlled trial evaluating the effectiveness of a decision aid versus a leaflet versus usual practice for a childhood immunisation decision. Impact was explored on two outcomes relevant to childhood immunisation, informed decision-making and MMR uptake [32–34]
. Only the decision aid achieved both outcomes. Several study limitations are acknowledged. Nineteen per cent of GP practices agreed to take part however only 16% provided parents (21% of all parents invited). Possible deterrents to practice enrolment were concerns about associated workload, ambivalence towards decision aids and fears of a negative impact on vaccination uptake. As a consequence we did not meet our initial recruitment target. However, we consider the study was sufficiently powered to draw meaningful conclusions because the ICC and the number of parents per cluster were lower than estimates used in the sample size calculation, meaning that the sample size inflation factor, and therefore the required sample size, had been over-estimated 
. We do not know if parents who took part were different to non-responders, although their characteristics were consistent with parents who find making the MMR decision difficult [30,31]
. We are confident that the sample is broadly representative of the target audience for this decision aid, that is ‘hesitant’ and ‘late or selective’ vaccinators [8,30]
representing 20–30% and 2–27% of parents, respectively [33,35–38]
. Due to the study timeframe we recorded MMR uptake at 15 months of age instead of 24 months which is used for national COVER (Cover of Vaccination Evaluated Rapidly; 
) data. Uptake may have been higher (in the leaflet and control arms) by 24 months. Finally, complete case analysis for the primary outcome was undertaken on just 55% of the data. Although we found no evidence of selection bias this remains a possibility.
Our finding that parents receiving the decision aid had lower decisional conflict than those receiving usual practice is consistent with evaluations of decision aids in other contexts 
. The positive effect of the leaflet mirrors our previous research 
. The positive impact of the decision aid on both vaccination uptake (which was higher than uptake across the five participating PCTs at the time of the study 
) and decisional conflict is in keeping with previous quasi-experimental evaluations of childhood immunisation decision aids [15–17]
. What are the possible explanations for first, the differences between vaccination uptake in the leaflet and decision aid arms (in which parents had levels of decisional conflict associated with informed decision-making); and second for the high uptake in the control arm when parents experienced high levels of decisional conflict?
First, the decision aid framed MMR vaccination as a choice between two options, to have or to not have the vaccination, rather than as an opportunity to have a vaccination as presented in the leaflet. Importantly, it included an interactive values clarification exercise [41,42]
which prompted parents to deliberate about their child having the vaccine or catching the diseases. It is plausible that this deliberation process may have enabled parents to use the information effectively [13,14]
to make a decision consistent with their values 
, thus prompting action. Reading the information leaflet appears to have been sufficient for parents to feel
that they had made an informed decision but they may have based their decision on the issues they focused on at that time, which are often emotions such as anticipated regret [13,14]
, rather than the evidence. This is consistent with the findings of our previous study 
and other childhood immunisation research [13,14,43]
. It seems likely that the lower rate of uptake in the leaflet arm reflected parental inertia rather than vaccination rejection.
Second, uptake in the control group (99%) was very similar to uptake in the decision aid arm (100%). One could then question whether interventions to support decision-making are worthwhile as an uptake rate of ≥95% is a successful population health outcome 
. However, the two arms differed significantly in the extent to which decision-making was informed. Post-intervention, decisional conflict was reduced to a level associated with informed decision-making in parents receiving the decision aid while in the control arm was unchanged from baseline. So while parents exposed to the decision aid were making a deliberative informed
decision to have their child vaccinated, parents in the control arm may have been adopting a position of ‘unquestioning acceptor’ 
. Our earlier research 
found that whilst most parents took their children to be vaccinated, two thirds reported that they had not made an informed decision, with some experiencing decision regret.
Finally, it may be argued that our findings can be attributed to the mode of information delivery (internet versus paper). However, increasing evidence indicates that it is the components of the decision aid (rather than the mode of delivery) that enable more informed decisions to be acted upon