If you care about health care, go read this latest piece from Atul Gawande. It's almost unquotably good. He compares our efforts to improve the American health care system with the agricultural reformation that began in the early 20th century, driven by pilot programs and experimentation.

What seemed like a hodgepodge eventually cohered into a whole. The government never took over agriculture, but the government didn’t leave it alone, either. It shaped a feedback loop of experiment and learning and encouragement for farmers across the country. The results were beyond what anyone could have imagined. Productivity went way up, outpacing that of other Western countries. Prices fell by half. By 1930, food absorbed just twenty-four per cent of family spending and twenty per cent of the workforce. Today, food accounts for just eight per cent of household income and two per cent of the labor force. It is produced on no more land than was devoted to it a century ago, and with far greater variety and abundance than ever before in history...

There are, in human affairs, two kinds of problems: those which are amenable to a technical solution and those which are not. Universal health-care coverage belongs to the first category: you can pick one of several possible solutions, pass a bill, and (allowing for some tinkering around the edges) it will happen. Problems of the second kind, by contrast, are never solved, exactly; they are managed. Reforming the agricultural system so that it serves the country’s needs has been a process, involving millions of farmers pursuing their individual interests. This could not happen by fiat. There was no one-time fix. The same goes for reforming the health-care system so that it serves the country’s needs. No nation has escaped the cost problem: the expenditure curves have outpaced inflation around the world. Nobody has found a master switch that you can flip to make the problem go away. If we want to start solving it, we first need to recognize that there is no technical solution.

Our most challenging long-term health care goals - reduction in costs and improvements in quality of care - aren't solely technical challenges that can be solved with a single plan. The path to "fixing" health care, then, is just that: a process rather than operation. It requires investigation, experimentation, and measurement across a disparate array of social and regulatory contexts. We can effect institutional change across a landscape this fragmented by prototyping extensively and seeing what works best; we need a portfolio of innovation. This is exactly the approach the current Senate bill takes, as Gawande notes: "it contains a test of almost every approach that leading health-care experts have suggested...None of this is as satisfying as a master plan. But there can’t be a master plan."

This is how we build new companies and products, and it's refreshing to hear it being applied by government. It's also awesome to see another analysis of this caliber by Gawande. He's established himself as likely the best health care journalist we have.