Most physicians spend the overwhelming majority of their professional lives in their own individual practice, operating independently of oversight, delivering care, and following a standard set of flows that have been laid out for them. At the end of the day, each of these physicians is essentially running their own retail business, seeing patients according to a tightly time-boxed regime that defines what they will be reimbursed for. The success or failure of their businesses is thus determined by how effective they are at recovering reimbursements from insurance, and how successful they are at avoiding lawsuits. And a belief in personal responsibility is deeply engrained in their professional culture, and is responsible for their fierce attachment to individual autonomy.

In many cultures, systematic improvements are seen as external pressures that are idealized, 'academic' efforts which are disconnected for the pressures of daily practice. They also often see such efforts as taking away their own discretion or ability to contribute, and reduce their job to a scripted, mindless role, and may even be perceived as questioning the value which they produce. It is not uncommon in such situations for changes, even when effective, to quickly regress to prior patterns, especially under times of high stress. Simply put, it is basic human nature for individuals tend to rely on behaviors that have rewarded them in the past, rather than the promise of rewards in the future. This doesn't mean that change is not possible, but it does mean that constraints on change must be removed. Consider the following excerpt from the book Better:

Each year, according to the U.S. Centers for Disease Control, two million Americans acquire an infection while they are in the hospital. Ninety thousand die of that infection. The hardest part of the infection-control team’s job, Yokoe says, is not coping with the variety of contagions they encounter or the panic that sometimes occurs among patients and staff. Instead, their greatest difficulty is getting clinicians like me to do the one thing that consistently halts the spread of infections: wash our hands.

We always hope for the easy fix: the one simple change that will erase a problem in a stroke. But few things in life work this way. Instead, success requires making a hundred small steps go right—one after the other, no slipups, no goofs, everyone pitching in. We are used to thinking of doctoring as a solitary, intellectual task. But making medicine go right is less often like making a difficult diagnosis than like making sure everyone washes their hands....

According to the Center for Disease Control, Control, two million Americans acquire an infection while they are in the hospital. Ninety thousand die of that infection. The hardest part of the infection-control team’s job, Yokoe says, is not coping with the variety of contagions they encounter or the panic that sometimes occurs among patients and staff. Instead, their greatest difficulty is getting clinicians like me to do the one thing that consistently halts the spread of infections: wash our hands. There isn’t much they haven’t tried. Walking about the surgical floors where I admit my patients, Yokoe and Marino showed me the admonishing signs they have posted, the sinks they have repositioned, the new ones they have installed. They have made some sinks automated. They have bought special five-thousand-dollar "precaution carts" that store everything for washing up, gloving, and gowning in one ergonomic, portable, and aesthetically pleasing package. They have given away free movie tickets to the hospital units with the best compliance. They have issued hygiene report cards. Yet still, we have not mended our ways. Our hospital’s statistics show what studies everywhere else have shown—that we doctors and nurses wash our hands one-third to one-half as often as we are supposed to. Having shaken hands with a sniffling patient, pulled a sticky dressing off someone’s wound, pressed a stethoscope against a sweating chest, most of us do little more than wipe our hands on our white coats and move on—to see the next patient, to scribble a note in the chart, to grab some lunch. This is, embarassingly, nothing new. In 1847, at the age of twenty-eight, the Viennese obstetrician Ignac Semmelweis famously deduced that, by not washing their hands consistently or well enough, doctors were themselves to blame for childbed fever. Childbed fever, also known as puerperal fever, was the leading cause of maternal death in childbirth in the era before antibiotics (and before the recognition that germs are the agents of infectious disease). It is a bacterial infection—most commonly caused by Streptococcus, the same bacteria that causes strep throat—that ascends through the vagina to the uterus after childbirth. Out of three thousand mothers who delivered babies at the hospital where Semmelweis worked, six hundred or more died of the disease each year—a horrifying 20 percent maternal death rate. Of mothers delivering at home, only 1 percent died. Semmelweis concluded that right. Yet elsewhere, doctors’ practices did not change. Some colleagues were even offended by his claims; it was impossible to them that doctors could be killing their patients. Far from being hailed, Semmelweis was ultimately dismissed from his job. 

Lloyd was bitten by the positive deviance idea—the idea of building on capabilities people already had rather than telling them how they had to change. By March 2005, he and Perreiah persuaded the veterans hospital leadership in Pittsburgh to try the positive deviance approach with hospital infections. Lloyd even convinced the Sternins to join in. Together they held a series of thirty-minute, small group discussions with health care workers at every level: food service workers, janitors, nurses, doctors, patients themselves. The team began each meeting saying, in essence, "We’re here because of the hospital infection problem and we want to know what you know about how to solve it." There were no directives, no charts with what the experts thought should be done.

Ideas came pouring out. People told of places where hand-gel dispensers were missing, ways to keep gowns and gloves from running out of supply, nurses who always seemed able to wash their hands and even taught patients to wash their hands, too. Many people said it was the first time anyone had ever asked them what to do. The norms began to shift. When forty new hand-gel dispensers arrived, staff members took charge of putting them up in the right places. Nurses who would never speak up when a doctor failed to wash his or her hands began to do so after learning of other nurses who did. Eight therapists who thought wearing gloves with patients was silly were persuaded by two of their colleagues that it was no big deal. The ideas were not terribly new. "After the eighth group, we began to hear the same things over and over," Sternin says. "But we kept going even if it was group number thirty-three for us, because it was the first time those people had been heard, the first time they had a chance to innovate for themselves." The team made sure to publicize the ideas and the small victories on the hospital Web site and in newsletters. The team also carried out detailed surveillance—taking nasal cultures from every hospital patient upon admission and upon discharge. They posted the monthly results unit by unit. One year into the experiment—and after years without widespread progress—the entire hospital saw its MRSA wound infection rates drop to zero.

Ultimately, change leaders must thus overcome this cultural resistance to self-discipline by engaging the practioners directly, appealing to their values, and addressing the constraints that typically stand in the way in implementing improvements. They can do this by instilling a foundation that will help reveal systemic insights, by providing adequate resources to support improvements, by establishing effective implementation strategies, and by reinforcing progess through accurate and timely communications of information. The AMA material supporting IHI's change efforts puts in this way: “It is better to build on what is working than to obsess about what is not working. It is easier to evolve the culture than to change it.”

Watts Humphrey recognized the importance of this early in his work to define process maturity models. In an Executive Brief article, he describes the lack of stickiness of new methods, and the importance of historical context, this way:

The support issue refers to the long time required for adoption of any new method or technology and the need for a continuing support and enhancement effort during this time. One example of this problem is illustrated by the Fagan inspection process at IBM. With Fagan’s help, development managers were convinced that inspections would sharply cut test time, shorten development schedules, and improve product quality. Many inspection courses were given and the inspection process was implemented in all of the development laboratories.

Some years later, I found that several laboratories had stopped doing inspections. On investigation, we found that after inspections had been in place for several years and many of the managers and developers had changed jobs, people started to complain about the time that the inspections were taking. The inspection advocates explained that this method was important in holding down testing costs. However, since testing was not then seen as a problem, management decided to make inspections optional. Quality then rapidly deteriorated and testing times grew accordingly. With no defined process or measures, however, the history was lost and nobody recognized what had happened.

The implementation issue concerns the skill and competence of the groups using CMMI methods. This issue relates to the support issue and concerns the problems caused by addressing symptoms instead of causes. While the evidence for poor software performance might include lack of documented plans, ill-defined processes, or incomplete data, these are only symptoms. Fixing the symptoms might get a suitable maturity rating but it does not mean that the plans were properly made and used, the processes actually followed, or the data analyzed and used. The problem is that when the DoD started requiring CMMI maturity levels for contracts, some groups found that it was easier to produce artifacts than to change engineering behavior. However, with processes, the connection between symptom and performance is behavior, and unless behavior changes, performance will not change. Without an understanding of history, it is hard to change behavior.