Rehab success rates quietly drop the players who never came back

Counting outcomes only among those who finished removes precisely the cases that would show the programme failing.

A programme reports that ninety-four per cent of its athletes returned to their previous level. Ask who is in the denominator and the number starts to move.

The ones who transferred to another clinic six weeks in are usually gone from the count. So are the ones who retired during the rehab, the ones released by their club who stopped attending, and the ones whose insurance ran out. Every one of those departures is more likely to happen to somebody whose recovery was going badly. The exits are not random with respect to the outcome, which is the exact condition under which dropping them ruins the statistic.

Think about what a bad case looks like from inside. Three months in, still swollen, still limping, no confidence in the joint. That athlete is the most likely to seek a second opinion elsewhere and the least likely to appear in the final tally. The programme has not treated her badly. It has simply been allowed to stop counting her, and it will publish a figure describing only the people for whom things went well enough to keep showing up.

The same shape appears in return-to-play timelines. Median time to return is computed from those who returned. Anyone still not back when the audit closes is excluded, and they are excluded for the very reason that would have lengthened the median. Everything published about how long an injury takes is therefore biased short, sometimes badly, and it is that bias which most damages a player who is behind schedule and being told he is late against a standard built by excluding people like him.

What I want to see is the flow. Started with sixty. Forty-one completed. Nine left the sport, six moved away, four are still in the programme after two years. Then the outcome, expressed over the sixty. Uglier, and the one worth reading.

Hospitals report surgical numbers this way because they were made to. Sports rehabilitation, which sells itself on outcomes and answers to nobody in particular, mostly does not.

Any centre unwilling to publish the size of its starting group is telling you something, and it is not that the number would have been awkward to typeset.