Better thinking about performance improvement

  Better: A Surgeon’s Notes on Performance, Gawande, Atul

I’ve always been troubled by the phrase “best practices” thrown around loosely in business settings. In certain engineering and professional settings, the term can have an important legal meaning. Even then, “best practice” is always a moving target. Better, Atul Gawande’s most recent collection of essays nicely crystallizes my reservations and offers useful insight into how to think about performance and performance improvement in knowledge work environments.

Drawing on his experience as a surgeon, Gawande reflects on the connections between learning and practice; both as an individual practitioner and as a field. His essays provide fascinating insights into how the practice of medicine has evolved over time; ranging over such diverse topics as hand-washing, battlefield injuries, and obstetrics. For that alone, Better is well worth reading. But it offers broader lessons as well.

Rooted in science and medicine, one thread that Gawande examines is quality of evidence. The gold standard is that of the double-blind, controlled laboratory experiment. However, action in the world and the demands of day-to-day practice cannot always wait for that standard to be met. There’s a wonderful quote from Samuel Butler that captures this problem; “Life is the art of drawing sufficient conclusions from insufficient premises.” Many of Gawande’s stories shed light on the reality that we often must make decisions on the basis of imperfect information and knowledge. We may not be able always to meet a gold standard of evidence, but we still benefit from a methodological commitment to hypothesis, experiment, and measurement.

Gawande’s observations on measurement and performance evolution in obstetrics provides one good example. He starts with the development of the Apgar score; a simple, concrete, measure of a baby’s condition at one minute and five minutes after birth. I am particularly struck by the insight and cleverness represented by recording the score twice in such a short interval. That creates a connection between measurement and action that drives performance improvement; it creates a feedback loop well matched to the human system it is embedded in.

Moving up a level from an individual delivery to a hospital’s performance, the Apgar score also serves to drive performance improvement at a more systemic level. In addition to informed clinical judgments about performance, we now have some numbers we can compare against one another and over time. Because these numbers tie to clinical judgment and performance, they can be used to evaluate changes in practice. Changes that improve the scores stick; those that don’t are abandoned.

This logic sheds some interesting light on a tension between “evidence-based medicine” and performance improvement more broadly conceived. Careful, clinical studies of problematic deliveries showed that Caesarian-sections had no measurable advantage over forceps assisted deliveries. Yet, no obstetrician uses forceps anymore and C-sections are used more and more routinely to the point where some claim they are over-used.

Understanding why has important lessons for anyone interesting in improving the performance of knowledge work in organizations. The difference comes from whether you are looking at performance at the systems level or the individual practitioner level. Learning to use forceps is a complex skill; difficult to observe, difficult to learn and difficult to teach. A C-section, on the other hand, is straightforward as surgical procedures go, highly observable, and teachable to a wider range of competent OB/GYNs. If you are trying to improve the outcomes and reliability of the system as a whole, your payoff from pushing C-sections over forceps is much higher. This is a classic example of improving a system by reducing variability. It is also an important reminder to be clear about where you are trying to improve performance.