Data could have helped England succeed in Euro 2016!

And what are the lessons for healthcare?

Data analytics and wider Big Data technologies have led to great advancements in finance and (more recently) sports analytics. Incredible insights have been generated, leading to better performance and improving the standard of both teams and individuals.

The healthcare industry as a whole has been much slower to adopt data analytics and even slower in using it to deliver meaningful insight. There are several lessons that can be learnt from the way other sectors have been using data, and the Euro championship provides a convenient event to compare with.

Retrospective data analysis identifies trends and potentially explains rationale

In the Euro 2012 tournament, Joe Hart was beaten by the Andrea Pirlo “Panenka” penalty. Though Hart had studied the Italian players and their technique, the data (provided by OPTA) would have shown that Pirlo had attempted this penalty twice previously in high pressure games.

Analysis of Christiano Ronaldo’s penalty technique is quite interesting because he has a pattern of hitting the ball to the bottom left. Of the last 47 penalties taken, 20 have been in the bottom left with the rest spread out around the goal. This doesn’t necessarily mean that he’s going to hit the ball in that direction every single time, but if you’re a goalkeeper and you’re going to gamble, diving left is the safest option.

In fact, Ronaldo’s penalty in the 2011-12 Champions League semi-final shoot out for Real Madrid against Bayern Munich was saved by Bayern keeper Manuel Neuer, who guessed correctly by diving to the left!

Similarly, the NHS has a wealth of retrospective data, analyzing various trends such as in patient admissions, A&E attendance, outpatient admissions and various other things.

In fact Monitor undertook analysis on A&E performance for the 2014/2015 period (seen as the worst in the history of the NHS) and found that the main factor for NHS trusts failing to meet the 4 hour wait target was not down to inadequate levels of staffing, negligence or increased demand, but rather the real bottleneck occurred when it came to finding beds for patients being admitted from A&E . Patients were not being discharged fast enough, or once discharged they were not being transferred out of the hospital (known as Delayed Transfers of Care - DTOCs).

While this has been evidenced clearly, the national debate at the time centered on funding levels, doctor morale and a host of other things – yet to date, the main factor hasn’t yet entered the debate adequately. Data can uncover the trend and help identify the rationale and root cause so correct solutions can be implemented.

Data without contextual insight can lead to the wrong conclusion

Data needs to be taken in context and shouldn’t be divorced from expert opinion.

For example can a defensive dribble be measured with the same weight as the Cruyff Turn? Glen Johnson had the best dribble completion rate of any player in 2012 (in the premier league), yet most fans would hardly have noticed him.

OPTA recently analysed previous tournaments and some interesting facts have emerged, such as

• Portugal’s Cristiano Ronaldo has had 70 shots from outside the box at the Euros but never scored from distance!
• Italy have never scored more than two goals in a total of 33 European Championship games.
• Republic of Ireland recorded a goal difference of minus eight in 2012, the joint-worst in Euro group history.

The trouble with data and the reason why it sometimes draws criticism is that without contextual insight and wider knowledge of its implication, it can lead to some strange and frankly incorrect conclusions.

 For instance, the fact that Italy score fewer goals is explained by their national ethos, their emphasis on defending a lead and their confidence that they can score goals when needed.

 With the use of insight, deeper conclusion can be found. For instance, look at current champions Spain in the 1990s and their pass completion rate is not exceptional at an average of 78.02%. In fact, in their EURO 96 quarter final, England comprehensively out-passed Spain by 598 successful passes to 527!

Look at Spain in the last two European Championships, however, and the influence of the tiki-taka football of Barcelona is clear. Not so many shots, but significantly more passes and an impressive combined pass completion rate of 87.69%.

Football evolves in cycles. As one philosophy of playing takes hold, it leaves gaps and a new one exposes those gaps, until something else comes along and so on.

This factor is crucial to note as it explain why Greece managed to win Euro 2004, with the second worst pass completion rate in the whole tournament of 67.21%! Denmark in 1992, had a similar poor pass completion rate too. Greece simply defended as a team and let the opposition wear themselves out.

For healthcare providers context can help make better use of data and analyse trends crucial to care. For example, an analysis of COPD risk factors can help stratify patients at most risk of being admitted into some type of care and help pre-empt that by re-allocating funds into preventative or self managed care.

Predictive Analytics

Predictive analytics is the combination of data, analytics and statistical modeling to predict future outcomes.

In the opening game of France 2016, Olivier Giroud seemed to miss chances and appeared static. The question is – how should he have performed?

It is possible to predictively model based on where he was on the pitch along with a multitude of other factors how many goals he should have scored. If that expected figure was one, then Giroud did as expected. If as some suggest, it should have been three, then perhaps he should be replaced?

This type of analytics is still in early stages in sport, but already has a place in healthcare analytics. For instance, using a combination of population data, benchmarked inpatient, outpatient and A&E attendance, as well as regional factors and seasonal variations, it is possible to predict hospital performance, and primary care demand with relative accuracy.

In conclusion, the wealth of data that already exists, couple with the insights generated as well as Big Data technologies already allow for powerful analytics, genuine insight and ways to develop better solutions for healthcare.

As for England in 2016 – data insight can help them better prepare for the tournament, analyze performances of both individuals and teams and help predict better outcomes if used correctly. Data can’t however substitute for skill and technique on the pitch!

Jilani Gulam ( https://twitter.com/Jilani_HealthiQ ) is co-founder and CEO of Health iQ.