Thursday, 31 July 2014

Understanding Median Predicted Survival

This year’s UK CF Registry Report shows that the median predicted survival for people with cystic fibrosis is a bit lower this year. To explain this change, Dr Janet Allen, Director of Research and Care, explains what we mean by median predicted survival and how “wobbles” happen.

You will have noticed in the annual report for the Registry that the median predicted survival is 36.6 years and that this is lower than the 43.5 years reported last year.

First of all, it is important to understand that median predicted survival is not the same as life expectancy. Median predicted survival is calculated from the registry data and represents the age beyond which half of the people in the UK CF Registry today would be likely to live. The annual values for median predicted survival tends to wobble around from year to year as the calculation is sensitive to the number of deaths recorded in any one year and there is a tendency to over-interpret the yearly wobble. This year the value appears to have dropped because the number of deaths recorded rose to 146 cases (of over 10,330 people on the Registry, representing around 1% of the UK CF population). A similar downward blip was observed in 2009 when the median predicted survival was 34.4 and the deaths recorded at 141. Apart from these dips, the trend in median predicted survival continues to rise and it may be better to report the overall trend rather than the year to year variation.

By way of an analogy, it may help to think of a country’s economy.

Over the past 50 years the UK economy has greatly, and largely consistently, increased, and yet there will be patches where, for a number of reasons, it contracts.
However when it does this, it never goes back to the start. It’s a case of always taking several steps forward, but every now and then one or two back.

The median predicted survival for people with cystic fibrosis has done the same and increased over the past 50 years, and yet, as happened in 2009, last year there was a noticeable but not statistically significant decline, even though the trend remains upwards.

Remember that when the Trust was founded in 1964, 90% of people with cystic fibrosis died before the age of 10.

Life expectancy answers a different question; for a child born today how long could they be expected to live?

These are complex calculations and are not calculated every year. The last report for the UK was over 20 years ago when it was calculated that life expectancy for a newborn child with cystic fibrosis in 1991 was in the order of 40 years. It is now estimated to be over 50 years for a baby born today, and the Trust is exploring having these calculations done again to show how far we have come since 1964.

You can find the UK CF Registry Report over on the Trust’s website both in Summary Form and as a Full Data Report.


  1. Thanks for this, Janet - very helpful.

    Could you please confirm that predicted life expectancy for children born today is based on currently available therapies, and does not account for the hoped-for improvements that may be delivered by small molecule therapies such as 809/661 and/or gene therapy (should these come to fruition)?

    1. The last time the life expectancy was published was in the early 1990s. At that time, the calculations indicated that for a baby born in the 1990s the life expectancy was in the order of 40 years. Clearly these are calculations and not everyone will reach that age. These calculations were done at a time when the new transformational treatments were not available. We now estimate that life expectancy for a baby born today is in the 1950s but we need to do the difficult maths to do the formal calculations. Because the new treatments are currently only available to a small set of people with cystic fibrosis (G551D) and has only been available for a short time, the calculations will not be unduly affected by these new treatments.