## Wednesday, 7 December 2016

### How to spot a made up number and what to do with it

Paediatrics is absolutely full of made up numbers - we rely on them every day.  If you think about it, many of the numbers that we have been given to work with are too conveniently rounded to be believable.  I don’t want to spoil the movie for you, but you have been lied to an awful lot.

The question is then. what do we do with the made up numbers?  Knowing where they come from is the key to the answer to that question.

Lie number 1 – For a baby, the normal milk intake is 150ml/kg/day

This one goes way back.  It actually comes from a number of fluid ounces per pound per feed which was a rough guide produced back in the bad old days.  The thing is that no-one has ever done a robust study finding out how much the average breast fed baby takes in 24hrs because there has been no good way of checking.  Not only is the number 150 a bit conveniently round and decimal, no one can tell you what the 5th and 95th centiles are. These facts would be essential for you know if a baby is taking an abnormal amount of feed.  The reality is that the mean (which this may or may not be) is not what we need to know in the first place if we are trying to assess whether feeds are too much or two little.

Luckily, you don’t need to know.  The best way to tell if the feed is enough is by the effect it has on the baby.  If it is too little, the baby won’t grow.  If it is too much the baby might exhibit signs of reflux disease due to overfeeding.

This brings us back to the question of what to do with a made up number.

Just because a number has never been validated by research doesn’t mean it’s not at all valid.  This particular number has been in use for decades now and is a useful landmark even if we don’t know which side of the truth it sits on.  The fact that it is accepted practice tells you that it has value - if it was frankly misleading, someone would have noticed.  The 150ml/kg/day number does at least tell us whether a child might be having too much or too little feed.  Knowing that you’re nowhere close to that figure is also useful.  For example, if a baby has faltering growth and is found to be taking 50ml/kg/day of milk, then lack calories would be a reasonable diagnosis.

Lie Number 2 – A normal heart rate between the ages of 0 and 1 is 110-160

For many years there have been three main lists of ‘normal values’ for physiological values in children.  We rely heavily on these as warning signs when it comes to recognising ill children.  Those reference ranges were all different for a reason – they were decided upon by experts rather than being based on definitive evidence.

Two published papers (1,2) have shown that the numbers in pooled population data conflict in places with the normal values in these reference ranges.  Notably however, the authors of these papers acknowledge that (I paraphrase) there is no such thing as a normal value when it comes to heart rate or other parameters in children.  There are just too many variables each of which can have a huge effect: sleep, activity, fear, pain, fever...  You get the idea.  Do we need to have normal values for a three year old who has just eaten an ice cream but is upset because a sibling has taken a toy off them?  I would love to have an app that takes into account variables like tiredness and temperature.  In the meantime, what I have is a lot of numbers that give me some idea of what normal might be, if only there was such a thing as normal in a child.

In many ways, the studies referenced show that consensus or expert opinion can be pretty good at coming up with the answers.  The correlation between the study findings and the made up numbers is remarkable.

I find it liberating to know that a heart rate can’t be treated as a piece of information out of context.  After all, we should always be looking at the child.  If their numbers are lying to you, hopefully their smile will tell you the truth.

Lie number 3 – a baby should lose no more than 10% of their birth weight in the first week of life

By now, you should be able to spot these numbers for what they are:

• Still useful
• Only helpful in context

It’s that simple.

The lies go on and on.  I couldn’t even begin to list the drug doses that are either plucked out of nowhere or at best based on some research where a number plucked from nowhere was shown to work.  Adrenaline for cardiac arrest is in the former group amazingly.  Fortunately for most drugs, we are protected by a wide therapeutic margin of error.

I like numbers.  Unfortunately children are themselves hugely variable and rarely normal.  I certainly wasn’t.

Edward Snelson
Chaos conspiricist
@sailordoctor

Disclaimer - If you can't trust numbers, what can you trust?  Certainly not me.

References