Nolan [1] argues the importance, if not necessity, of understanding medical systems for improving health care. However, he fails to mention a key characteristic of a successful system, that of learning and adapting to environmental change through innovation. These attributes are essential if we are to achieve our purposes (his principle 1) of reducing waste, error, and unneeded variation and improving outcomes of care. This is at the core of complex adaptive systems based on chaos and complexity theory [2, 3]. I like viewing our system of medical care as a biological system rather than as a machine, the industrial model for organizations and systems. Because biological systems become more complex through the ebb and flow of differentiation and integration, we can see fragmentation and piecemeal care as an unintended consequence (Nolan's principle 3) of specialization (differentiation) for which we have yet to establish effective integrating activities. The emergence (a central attribute of complex adaptive systems) of family practice and emergency medicine in the 1970s came about not because new biomedical knowledge needed to be put to use but because these two specialties better integrated care for two populations of people.

Physicians like to organize themselves (self-organization being another characteristic of complex adaptive systems); we don't like to be organized by others, as evidenced by our placing a high value on autonomy. In a demonstration of paradox, another attribute of complex adaptive systems, most of us don't realize that our colleagues in well-organized multispecialty groups see themselves as having more clinical autonomy than those in solo or small group practice [4].

Nolan's two scenarios speak to better ways of integrating the myriad things we now do: improving primary care physician-specialist interactions through a common set of referral criteria, monitoring and providing feedback when inappropriate referrals occur, and using specialists to help primary care physicians expand their diagnostic and therapeutic skills to optimize referrals. These, however, are processes or interactions, not structures. Nolan's emphasis on structures in principles 2 and 4 is misleading. The human brain, the major complex adaptive system in our body responsible for much of our learning and adaptation, is much better evaluated by neuropsychological testing than by structural examination.

To develop our leadership capabilities and better organize ourselves, we need a solid conceptual model of systems that resonates with our experience. I believe that complex adaptive systems provide this framework much better than does general systems theory.