Diffusion of Medical Innovation

A few years ago, I read an interesting report about a surgeon who performed 120 successive lithotomies--operating for bladder stones--with a mortality rate of only 2.5%. This might not seem so significant, given the low mortality rate for most surgeries today, except for one fact. The operations were performed in the late 18th century by a French surgeon named Pouteau. His surgical approach is discussed in a fascinating book on the history of surgery that was written by Owen Wangensteen, the former chairman of the department of surgery at the University of Minnesota, and his wife Sarah Wangensteen, a medical historian at the same institution.¹ The authors reported that the secret to Pouteau's success was his strong emphasis on cleanliness during surgery, his recognition that hospital infections were contagious, and his use of disposable paper bandages for dressing surgical wounds--remarkable innovations for his day.

The obvious question is, Why weren't these innovations more widely disseminated and adopted by other physicians of the time? Why did it take almost a century before similar procedures of antiseptic and aseptic surgery, proposed by Semmelweis and Lister, spread throughout medicine and became the universal standard of surgical care?

Diffusion of innovations

The diffusion of innovations has been extensively studied by Everett Rogers² and others. Rogers defines diffusion as "the process in which an innovation is communicated through certain channels over time among members of a social system." The process of adoption usually proceeds through several reasonably well-defined steps, beginning when individuals first learn of the innovation and progressing through stages when they become persuaded of its importance, decide to adopt it, begin implementation, and finally, seek confirmation that they made the correct decision. Various factors influence the rate of adoption of an innovation, including its perceived attributes; how information about it is communicated to members of the society; and the beliefs, customs, and attitudes prevalent in that society.

Rogers identified 5 perceived attributes of an innovation that seem to explain how quickly it will be adopted. These include the relative advantage of the innovation over prior ideas; its compatibility with the beliefs, values, needs, and experiences of the potential adopter; its complexity; how easily a potential adopter can try the innovation; and how visible the results of the innovation are to observers. Rogers refers to the last 2 attributes as the innovation's trialability and observability.

Using Rogers' criteria, we can speculate why Pouteau's discoveries in aseptic surgery did not immediately spread throughout medicine. The relative advantage of his technique is obvious to us today, but probably was not to his contemporaries. Few 18th century surgeons compiled outcome statistics for their patients. Moreover, Pouteau's ideas were not compatible with their medical view of the world. They had no theory of infection and no reason to believe that unsanitary conditions were a major cause of patient mortality.

One reason for Lister's subsequent success is that he postulated a credible theory of the cause of wound infection- that the small organisms Pasteur found in wine might also produce a septic fermentation in a patient's wound. His process of antiseptic surgery was, therefore, consistent with the leading scientific knowledge of his day. Lister was also living in an age that was beginning to accept the process of scientific discovery in medicine, and that had far more methods of communication than those available to Pouteau. Although Pouteau's techniques appeared simple, his success suggests that he must have been exceptionally compulsive about cleanliness during and after surgery. Perhaps the average surgeons of his day who knew about these techniques considered them unnecessarily complicated and difficult to emulate. In that case, it is unlikely that they would have observed the same level of success that he did.

Antibiotic innovation

A 1953 study by Coleman and his colleagues³ on the adoption of tetracycline by physicians in 4 small Illinois cities demonstrated some of the important elements of diffusion theory in the modern era. The research team interviewed 228 physicians, 64% of the practicing physicians in the 4 cities, about their use of the drug. They discovered that the prescription of tetracycline began with a few innovative or early adopter physicians, was tried by 50% of physicians within 4 months, and was almost universally adopted within 17 months.

They also discovered that the most important factor influencing physicians to adopt the new antibiotic was the personal clinical experience of their peers, who reported that tetracycline was more effective and better tolerated by patients than the standard antibiotics of the day. Articles about the drug in medical journals and contact with pharmaceutical company representatives were of lesser importance.

In addition to its increased effectiveness, tetracycline also possessed several of the other attributes that Rogers postulated increased the rate of adoption of an innovation. The use of the drug was compatible with the accepted clinical beliefs, practices, and experiences of the average physician; it was simple to use; and the results of treatment were quite visible.

It should be noted, however, that a few scholars have reanalyzed Coleman's original data and arrived at a slightly different conclusion than he and his colleagues did. The new analyses suggested that the reports in the medical literature and the influence of drug company representatives were as much or more of a factor in a physician's adoption of tetracycline than was the physician's peer network.² In other words, the social network, although crucially important, was only one of the factors that influenced physicians to adopt the new drug.

Innovation in psychiatry

In psychiatry, several studies have shown that computer-assisted psychotherapy, an innovative method of delivering cognitive-behavioral therapy using a computer program, may be as effective as treatment by a live therapist. Selmi and associates⁴ demonstrated that a small group of patients with a diagnosis of major depressive disorder (n = 12) responded as well to 6 sessions of computer-assisted cognitive-behavioral therapy as those who received the same number of treatments administered by a live therapist (n = 12). Both groups did better than controls who were placed on the waiting list and received no therapy.

Greist and coworkers⁵ reported that both computer-guided (n = 74) and therapist- administered (n = 69) treatments were effective for obsessive-compulsive disorder, but the latter was more effective than the former. Moreover, many patients who did not respond to computer-guided therapy did respond to subsequent clinician-administered therapy. The converse was not true.

Osgood-Hynes and colleagues⁶ used treatment booklets and a computer-aided interactive voice response telephone system to treat 41 American and British patients with mild to moderate depression in an open, uncontrolled study. All 41 subjects (intent-to-treat group) completed the initial assessment, and 28 subjects (completers group) successfully finished the full 12-week therapy program. All subjects, with the exception of those in the British completers group (n = 6), showed a significant improvement in depression rating scores.

Freedman,⁷ commenting on these and other studies, questioned why the innovation of computer-assisted psychotherapy was not more widely adopted by the profession. He suggests that psychiatrists have not chosen to adopt computerized methods of treatment because the technique is inconsistent with the values of the social system within which they function. He argues that the use of computers for the diagnosis and treatment of mental illness "might be alien to psychiatrists' selfimage of their role," which is based on the model of "two people in a room, talking."

Yet that self-image has not stopped psychiatrists from adopting other nontalking therapies, such as electroconvulsive therapy for the management of severe depression and medication for the management of depression, psychosis, and anxiety. One could, in fact, argue that many psychiatrists are only too entrepreneurial in adopting new nontalk treatments, such as the latest antiseizure medication for treatment of mood instability. In the best therapeutic circumstances, however, each of these models is augmented with talking therapy. There is no reason that the same thing cannot happen with computer-aided diagnosis and treatment.

It is likely that the limited adoption of computer-assisted psychotherapy has as much or more to do with other attributes that influence the rate of diffusion of an innovation than with the concern of psychiatrists about their professional self-image. Rogers² argued that the relative advantage, compatibility, complexity, trialability, and observability of an innovation affect its rate of diffusion. Although traditional cognitive-behavioral therapy is a wellaccepted mode of treatment, the relative advantage of computer-assisted treatment is unproved. The studies cited earlier suggest that a therapist is unlikely to observe a significant improvement in outcome between the computer-assisted and traditional methods. The former is theoretically compatible with the underlying concept of cognitive-behavioral therapy, but that fact would not necessarily increase its diffusion, since cognitive therapy itself is only practiced by a small proportion of psychiatrists.

Moreover, implementing computer-assisted treatment is certainly more complex than initiating other forms of psychiatric treatment, such as dynamic psychotherapy or psychopharmacology. Practitioners would have to purchase and learn how to run the necessary computer equipment and software. That is assuming, of course, that they already know how to administer the traditional form of therapy. Given this complexity, the system cannot easily be tried without a considerable investment in time and money.

None of these arguments disprove the potential usefulness of computer-assisted cognitive-behavioral therapy. They simply suggest that it is not an innovation that provides the same obvious benefits as aseptic surgery, CT, or lithium in the treatment of patients. What it does provide is the potential to treat large numbers of patients with a small number of therapists. That may be valuable to companies and health plans that want to offer cost-effective treatment to their employees and members. Its effectiveness in that realm, however, has yet to be proved.

References:

References 1. Wangensteen OH, Wangensteen SD. The Rise of Surgery From Empiric Craft to Scientific Discipline. Minneapolis: University of Minnesota Press; 1978.
2. Rogers EM. Diffusion of Innovations. 5th ed. New York: Free Press; 2003.
3. Coleman JS, Katz E, Menzel H, et al. Medical Innovation: A Diffusion Study. New York: Bobbs-Merrill; 1966.
4. Selmi PM, Klein MH, Greist JH, et al. Computeradministered cognitive-behavioral therapy for depression. Am J Psychiatry. 1990;147:51-56.
5. Greist JH, Marks IM, Baer L, et al. Behavior therapy for obsessive-compulsive disorder guided by a computer or by a clinician compared with relaxation as a control. J Clin Psychiatry. 2002;63:138-145.
6. Osgood-Hynes DJ, Greist JH, Marks IM, et al. Self-administered psychotherapy for depression using a telephone-accessed computer system plus booklets: an open U.S.-U.K. study. J Clin Psychiatry. 1998;59:358-365.
7. Freedman J. The diffusion of innovations into psychiatric practice. Psychiatr Serv. 2002;53:1539-1540.