Early orthodontic treatment for Class II malocclusion reduces the risk of trauma to the front teeth: How do we interpret this data; Odds, Risk and Numbers needed to treat?
In my last blog, I described the results of our Cochrane Systematic review into the effectiveness of treatment for Class II malocclusion (http://goo.gl/9aLWlU). In the review we concluded that the only benefit of providing treatment when a child was 8-10 years old was a reduction in trauma to the front teeth. I also pointed out my interpretation of the odds ratio for the data we presented was
“the odds of trauma for patients who were treated early was 41% less likely than for the patients treated in adolescence only”.
What does all this mean?
I have received several comments on this conclusion and this has led me to explore these findings further and attempt to come to a clear conclusion on what we should be doing for our young patients with prominent incisors. I have leant heavily on these three main sources.
- The module on dichotomous data from the Cochrane Collaboration opoen learning material for reviewers. http://goo.gl/t3gEcS
- The effect size by Nick Pandis http://goo.gl/i89M38
- Risk ratio vs odds ratio by Nick Pandis http://goo.gl/zzola6
You may also like to refer back to my previous post on reducing uncertainty in orthodontic treatment http://goo.gl/TuI55k
I have based all my calculations on Table 1.6 in the review and a shortened version is included here:
New trauma in adolescence
How do we quantify the risk?
I first calculated the risk of trauma for each group, this revealed that for the early treatment group it was 19.7% and for the adolescent group it was 29.7%. The difference in risk was, therefore, 9.6%. Meaning that if we treated early there was a reduction in trauma of 9.7% compared to if we treated in adolescence.
The next measure that I considered was relative risk or risk ratio. This is the risk in one group (early treatment) divided by the risk in the other (adolescent treatment). I calculated this to be (34/172)/(54/185)=0.67
This means that providing early treatment reduced the risk of trauma by 33% of what it was when the treatment was provided in adolescence. Again an interesting finding.
The last important measure is the numbers needed to treat(NNT) This gives us the number of children we would have to treat early to prevent one episode of trauma. This calculation is easy, is it 1/Risk difference. For this data 1/0.1. Therefore, the number needed to treat is 10 with a very wide 95% confidence interval of (6-214).
If you want to calculate similar values for other datasets this can be done easily by using this web calculator www.phsim.man.ac.uk/nnt/default.aspx
What can we conclude?
One of the important factors of which I have become more aware, is that is it is up to us to interpret the data and we should not all simply “go along” with the authors conclusions. So, the decision to treat early is up to you. You need to consider the following main points if we treat early
1 The chance of incisal trauma is reduced by 10% compared to a group of patients treated in adolescence
2 The odds of trauma will be 41% less likely than in the adolescent group
3 The risk of trauma will be reduced by 33% compared to the risk in the adolescent group
4 To prevent one episode of trauma, you will need to treat 10 patients.
How certain are you of this data? You need to remember that the confidence intervals are very wide. http://goo.gl/TuI55k
So what would I do?
I think that I am influenced mostly by the reduction in trauma of 10% and the numbers needed to treat of being high. I would not treat early. What would you do?
I would like to acknowledge the help of Badri Thiruvenkatachari and Nick Pandis with this one!
Badri Thiruvenkatachari,, Jayne E Harrison,, Helen V Worthington,, & Kevin D O’Brien (2013). Orthodontic treatment for prominent upper front teeth in children Cochrane Database of Systematic Reviews (11) DOI: 10.1002/14651858.CD003452.pub3