Data is crucial to running an effective health service, but it can be misleading.
The good thing about the NHS is that the data exists. The issue is how the data is processed and presented.
The problem with numbers is that they vary. Given a series of numbers, it is difficult to see whether there is a significant difference beyond natural variance.
Clinicians are taught to deal with variation to compare different systems. This is the impenetrable science of comparative statistics. But managers are being asked to improve a single system over time. Neither clinicians nor managers are taught the simple continuous statistical methods for doing this and so mistakes occur.
The first and most common mistake is to use only two numbers. Given the huge number of numbers in the NHS and our comparative culture, data is often processed to reduce the confusion of variation and only two numbers are presented: this year, last year; this week, last week etc. As the US statistician W Edwards Deming once said, given two numbers, there is a high probability that one will be bigger than the other.
So if you are asked to make a management decision based on the difference between two numbers, don't. Your diet has not been a success just because you are a pound lighter this week compared with last week. Your weight is varying all the time.
The trick is to demonstrate a change in the pattern of the variation over time, and the good news is that we don't need to resort to complex comparative statistical methods.
All you need to turn your data into information is at least 20 points in a time series (for example 20 consecutive patients, samples, minutes, days, months, years). Plot the dots with time on the Y-axis and the values on the X-axis and expose the pattern of the variation over time.
The second mistake is trend lines. Deming also pointed out that there is a low probability that a straight line drawn equidistant through a random set of numbers will be horizontal. Trend lines are therefore nonsense.
So to avoid being seduced by a trend line, plot the average as a line through the dots. This will keep your eye following the 'voice of the process' and not the 'voice of the target'.
If the dots are varying up and down but generally running horizontally across the page, the chances are that the process is stable and will carry on behaving like this forever unless we change it. Can you see the normal variation in the process? Similarly, are there obvious changes in the pattern of variation or obvious dots outside the range of the normal variation? Can you remember any significant events that might explain these special causes?
Needless to say, there is a scientific method to support the ocular method above. Statistical process control has been the mainstay of manufacturing improvements over the last 100 years. However, it has been ignored by healthcare professionals, who have yet to be weaned from the comparative culture in which we are reared.
Do yourself and the NHS a favour and get a copy of Understanding Variation: the key to managing chaos by Donald J Wheeler. It will certainly change your life, and might even save the NHS.
Dr Kate Silvester is national coach for the Osprey programme, which offers clinicians an opportunity to learn and apply manufacturing systems and engineering techniques to improve timeliness, cost, efficiency and quality.