Richard Smith: Our need for clockware and swarmware

Richard Smith

Tackling the global pandemic problem of non-communicable disease (NCDs) is a complex problem that needs clockware and swarmware. I imagine that most BMJ readers have no idea what that sentence means, but read a few more paragraphs and you may learn something useful.

Problems, said Sian Williams, executive officer of the International Primary Care Research Group, at the group’s meeting in Edinburgh, can be classified into three types: simple, complicated, and complex. (her full talk, which is well worth reviewing, is available at:

An example of a simple problem is making a Victoria sponge. (As Williams said this, I had a vision of some of my collapsed or burnt cakes and thought that making a Victoria sponge was not perhaps the most simple of problems.) Making a Victoria sponge is a simple problem because there is a recipe that can be followed by (almost) anybody and will every time produce an edible cake.

Putting a man on the moon or building the first atomic bomb are examples of complicated problems. There is no simple recipe, but they are problems that we know when they are solved and that can be solved with care, planning, smart people, resources, and ingenuity. Reducing global deaths from infectious disease might, suggested Williams, be a complicated problem. With vaccines, drugs, and focused programmes smallpox was eradicated and polio might be, and deaths from childhood infections have reduced dramatically.

Tackling NCDs, in contrast, is a complex problem—like bringing up children. There is no simple formula. You never “solve” the problem, and having done it once does not guarantee that you can do it again and get the same result. You can plan and build programmes, but the course of the programme will always be uncertain, full of unexpected twists and turns. Tackling NCDs is a complex problem because it cannot be done simply by the health system with drugs and vaccines but needs—in the words of Margaret Chan, director general of WHO—”a whole of government and whole of society approach.” This means bringing together many government departments, professional groups, NGOs, academics, and the private sector, a task even more complex than guiding a child through adolescence.

I think that Bill Gates’s recognition that tackling NCDs is a complex problem that has kept him away from the problem—apart from tobacco control, which has arguably through evidence and the Framework Convention on Tobacco Control been reduced to a complicated problem.

Clockware and swarmware are two means to tackle complex problems. Clockware means gathering the best possible (but usually far from perfect) data on all aspects of the problem. You then use swarmware—a whole host of responses that are infused with both risk and hope—to try and make progress with the complex problem. The clockware will tell you which responses are working and you abandon or adapt those that produce no benefit and build and develop those that do.

Complex problems can respond to simple rules—the way that a flock of birds fly in formation being the classic example—and Williams ended her talk by suggesting simple rules for tackling ideas:

  • Prevent and treat tobacco addiction
  • Concentrate on “best buys” (those interventions that will produce the biggest benefit at the lowest cost)
  • Prioritise pragmatic research that includes information on costs
  • Empower and incentivise those on the ground
  • Secure the income of professionals who add the most value
  • Support self care and give self confidence to patients

Unfortunately using swarmware does not feel comfortable to conventional researchers. They are used to designing studies with clear protocols and designs, often randomized trials, and sticking with them until they get a result. Rapidly adapting the intervention feels “unscientific.” There is thus often a tension between advocates of continuous improvement and researchers, but I’m sure that there’s a need for both. I also fully support Williams’s proposal to prioritise pragmatic research.

Richard Smith was the editor of the BMJ until 2004 and is director of the United Health Group’s chronic disease initiative.

Competing interest: RS spoke at the meeting as well and had his expenses paid by IPCRG. He persuaded them to let him travel first class on the train, arguing that by buying the tickets in advance first class was much cheaper than an open standard class return. Plus he didn’t need a hotel as he imposed himself on friends.