A few months ago I was stuck in traffic on my way to give a talk at the Royal College of General Practitioners. I thought of phoning the venue to warn them I’d be late, but decided they’d probably just tell me to call back at 8 a.m. the following morning.

When did that whole thing start? It’s now routine for GPs’ surgeries to make you phone first thing, and to book appointments for that day only. In my view, it overlooks the Law of Unintended Consequences. While it might help the surgery meet some bureaucratic ‘responsivity target’, it may also have the unfortunate psychological side effect of encouraging unnecessary bookings. Let me explain this issue by describing the trajectory of most illnesses.

Day 1. ‘Ooooh, I don’t feel very well. I think I’ll wrap up in a rug in front of the TV and have a Lemsip.’

Day 2. ‘Oh my God, I’m going to die.’ Start Googling symptoms of Ebola, before deciding it’s probably just dengue haemorrhagic fever or plague.

Day 3. ‘Feeling very slightly better. I might actually survive to the weekend.’ Downgrade Google diagnosis to stage three man flu.

Day 4. ‘Hmmm. This isn’t too bad actually. Why don’t I dig out a onesie and catch up on season two of something on Netflix.’

The vital inflexion point here is between day two and day three. The moment you start to feel the faint glimmer of recovery, your urge to visit a doctor disappears along with your paranoia. But the scarcity bias created around ‘You can only phone at 8 a.m.’ makes me phone the doctor a day too soon, before my immune system has a chance to do its work. If appointments were reservable a few days in advance, many people would end up cancelling once they began to feel better, freeing time for urgent cases and avoiding overmedication.

Even if I am wrong here, the NHS should test different systems, as the difference between the best and the worst will be immense. Tiny changes to choice design have enormous effects on behaviour: the brilliant idea of issuing postdated prescriptions for antibiotics (‘If you still feel ill on Friday, take this to the chemist’s’) means 65 per cent of the people handed such prescriptions do not use them.

Not only should we test different systems, we should continue to test them. Over time, people may start to game the system — or dodge it altogether by going to A&E. What works at first may not work indefinitely.

This seems to be the case with low-cost airlines and their practice of charging for checked-in luggage. In the beginning it worked quite well. People checked in large suitcases and carried small, squidgy holdalls on to the plane. In time, everyone learned to play the system by acquiring a rigid suitcase of the maximum allowable onboard size. These were comparatively rare when the fees were introduced; now, alas, every tosser owns one. Today the worst part of flying is the Gadarene rush for overhead locker space. It delays boarding and slows down disembarkation, since people are forced to stow bags a long way from where they are sitting. Some airlines have now started charging for the overhead lockers. Samoa Air has the ultimate solution — it charges by luggage weight plus body weight.

Despite what you may have heard, Einstein, I am happy to say, never said that the definition of insanity is ‘doing the same thing over and over again and expecting a different result’. It is not the definition of insanity — it is the definition of complexity theory. Lots of things work really well for a time, and then don’t. Once people understand this simple fact, political debate will become a good deal saner.