Later, a family member's life was prolonged by a treatment from M D Anderson Cancer hospital using a protocol for a previously resistant illness. The algorithm was a simple graphical presentation of what was the first line treatment, what were the alternative treatments should the first fail, what were the third line treatments, and so forth. Using this model, Jobson made a commitment to try to improve medication choice in psychiatry through what became known as the International Psychopharmacology Algorithm Project.

In 1985, Jobson contacted a group of colleagues to share algorithms. The Delphi method — each participant's treatment sequences would be submitted, then all shared among the group, minimizing the influence of "expert" opinions — was used. This "project" was well-received and informative. Eventually the group — which included faculty from Duke, Emory, Harvard, Stanford and Yale Universities; National Institutes of Mental Health and multiple international sites including the universities of Vienna and Stellenbosh (South Africa) and from Sendai, Japan — addressed, one by one, the major Axis I psychiatric illnesses in this way.

In 1992, speaking with longtime friend and colleague Bill Potter, then head of extra-mutual research at NIMH, Jobson learned that there was virtually no interest in funding research about medication choice sequencing (algorithms). Jobson explained that clinicians make those decisions daily and spoke about his informal algorithm project. They agreed that it would be worthwhile to have a national conference to create awareness of the need for psychopharmacology algorithms. So the project was formalized as the International Psychopharmacology Algorithm Project (IPAP) and the first educational conference was held in 1993 at the National Institutes of Mental Health.

Introducing O.R.


In 1992, Jobson enlisted Dean Hartley (then at the Oak Ridge National Laboratory) to provide operations research expertise to IPAP. Jobson had the idea that, as there was a better way to improve medication choices (algorithms), there must be better ways to design and communicate the algorithms. "Operations Research — the Science of Better" was to become an integral part of psychopharmacology algorithms.

The earliest O.R. part of the project was to examine the diagnostic intake reports of prominent U.S. and international psychiatrists. The question was, What parts were common to all and which of the differences would be valuable if made common?

This examination was educational for Hartley. The treatment of psychiatric disorders contains two broad traditions: therapy and pharmacology. Therapy, while popularly known from the theories of Freud and Jung, has more modern forms that have been medically proven to be effective — cognitive behavioral therapy (CBT) and interpersonal therapy [1]. Pharmacology involves the prescription of psychoactive drugs, including such early drugs as haloperidol (Haldol) and Lithium, to more recent drugs such as Prozac and selective serotonin reuptake inhibitors (SSRIs). In practice, these two traditions are often combined, and in the international community the methods of diagnosis, the standard therapies and the available drugs can vary widely from U.S. practice.

Designing Algorithms


The earliest "algorithms" were clinical guidelines, which are text-based products. Some have been organized with a chapter for each treatment modality (all for one psychiatric disorder), discussing the nature and supporting evidence for each modality. Others have been organized as unified examinations of the published articles relating to the disorder, with a more focused conclusion [2]. IPAP initially focused on the visual impact of flowcharts, with text amplification. Hartley introduced the concept of deepening the algorithm by adding Gantt chart, influence diagram and decision tree views of the same algorithm [3]. This concept has not yet been implemented, as it exposed a serious lack of basic data concerning the probabilities of success, partial success, failure and side effects at all stages of the algorithms, but especially at the levels following the failure of the first-line treatment.

As mentioned above, the first efforts by IPAP were efforts to elicit expert opinions. In 1997, IPAP experimented with using an interactive Web conference in the elicitation process [4]. This conference created two algorithms, one for schizophrenia and one for unipolar depression [5,6]. In addition to the creation of algorithms, the conference included an analysis of the conduct of the algorithm creation process. Ten professionals in the fields of O.R., informatics, information science, library science, pharmacoeconomics, technology and medical practice observed the interactions and decision-making and the impact of the technology on the proceedings [7]. A group of prominent psychiatrists acted as a synthesis committee to collect the results and make general recommendations [8].

In addition to publication in the scholarly literature, IPAP constructed a Web site on which it posts its algorithms and supplementary materials (www.ipap.org). IPAP has used the recommendations of the Web conference to improve its algorithm design process. In particular, the need for increased pre-conference work has borne fruit in successive algorithm designs. One of the failings noted by the observers was insufficient evidence connections for the recommendations. This has had a profound effect on the current set of algorithms, as discussed below.

During the period from 1998 through 2002, IPAP held conferences in Japan and China and participated in conferences in several other countries. These conferences resulted in the creation of the Japanese Psychopharmacology Algorithm Project (JPAP), the Chinese Psychopharmacology Algorithm Project (CPAP) and country specific algorithms for Japan and China.

Evidence-Based Algorithms


In 2003, IPAP made the shift from expert-based algorithms to evidence-based algorithms. The current model involves a separate international faculty of experts for each psychiatric disorder, together with consultants and a Webmaster/technologist. Disclosure statements are also included on the Web. Each algorithm is created from the best currently available medical evidence. The algorithm flowchart is displayed on the Web site, and a user may click on any node to obtain the detailed notes for the node, with the level of evidence given in the notes.

The flowchart for the first evidence-based algorithm, one for schizophrenia, is shown in Figure 1. The algorithm is dated December 2004; however, recent developments have caused the faculty to consider whether changes are warranted. At the very least, changes to the notes will be implemented to reflect the impact of recent research.

Figure 1: Flowchart of evidence-based algorithm for schizophrenia.

The flowchart for the second evidence-based algorithm is shown in Figure 2. This algorithm, dated June 2005, is for Post-Traumatic Stress Disorder (PTSD), which is especially timely in the U.S. civilian population, given the impact of the 9/11 attacks and Hurricane Katrina, and in the allied military population, given the impact of fighting in Afghanistan and Iraq. This algorithm is visibly more complex than the schizophrenia algorithm because of the chronic course of the disorder and its high co-morbidity (accompanied by other disorders).

Figure 2: Flowchart of evidence-based algorithm for post-traumatic stress disorder.

IPAP has most recently created its third evidence-based algorithm. This third algorithm is for Generalized Anxiety Disorder (GAD). The processes for the second and third algorithms differ from the first in using an executive committee-created starting algorithm, a technique selected to reduce the time required for algorithm design. IPAP is learning from its experiences.

As mentioned above, static algorithms lose value over time because the evidence base changes and new drugs and drug classes are created. At this time, the faculty model appears to be the correct model to support dynamic algorithms, as recent developments in schizophrenia have prompted the schizophrenia faculty to re-examine and reaffirm that algorithm. If this model proves to be inadequate over longer periods of time, IPAP will make changes to its operations.

The Future of Psychopharmacology Algorithms


The current IPAP Web site has been adequate technically, supporting interactive display of the algorithms, delivery of PDF versions of the algorithms and supplemental articles, and interactive searches of the supporting citations. However, the persuasive power of the Web site was unknown and the addition of a third algorithm will make the current design too clumsy. Therefore, IPAP sought out B. J. Fogg who created "captology," the study of computers as persuasive technology [9]. Fogg has been working with IPAP to analyze and improve the Web site's presentation and investigate other methods of getting the algorithms in the hands of users.

Jan Fawcett makes the point that "one of the common problems in yielding benefit from treatment algorithms" is in "getting clinicians to follow them and use them to increase the benefits of treatment" [10]. IPAP agrees with this point and has been investigating how to make the "information diffusion" of algorithms more effective. In conjunction with the University of Buffalo's School of Informatics, IPAP sponsored a conference in 2006 to address the problem of making psychopharmacology algorithms "more useful and utilized" [11]. Major algorithm project representatives from around the world, as well as experts in correlated fields, participated. Perhaps the algorithms should be part of hospitals' medical information systems. Perhaps the algorithms should be made available on notebook computers or hand-held devices.

Fawcett also makes the point that evidence is missing to support complete algorithms. "At this point, the clinician frequently finds the algorithm has run out of data-based recommendations for many patients," he says. He holds out hope that the algorithms will create enough pressure to obtain the data needed to improve them. Perhaps the future will allow the influence diagram and decision tree views of algorithms envisioned earlier.



Dean Hartley is the principal of Hartley Consulting (http://dshartley3.home.comcast.net). Previously he was a senior member of the research staff at the Department of Energy Oak Ridge Facilities (Oak Ridge National Laboratory, Y12 Site and East Tennessee Technology Park). A Senior Fellow with the George Mason University School of Public Policy, Hartley is a past vice president of INFORMS, a past director of the Military Operations Research Society (MORS), a past president of the Military Applications Society (MAS) and a member of the College on Simulation of INFORMS.

Dr. Kenneth Jobson has a clinical practice in psychiatry and psychopharmacology, Psychiatry and Psychopharmacology Services PC in Knoxville, Tenn. He is the founder and chairman of the board of the International Psychopharmacology Algorithm Project (www.ipap.org), on the clinical faculty at the University of Tennessee, Department of Psychiatry, and co-editor of a textbook, "Treatment Algorithms and Psychopharmacology." He has facilitated the establishment of algorithm projects in Europe and Asia.

References

1. Kruglinski, Susan, 2006, "The Discover Interview: Nobel Laureate Eric Kandel," Discover, April 2006, pp. 59-61.
2. Jobson, Kenneth O., Matthew J. Friedman, and Dean S. Hartley, 2005, "Psychopharmacology Algorithm Development," Psychiatric Annals, Vol. 35, No. 11, pp. 921-925.
3. Hartley, Dean S., III, 1999, "The Language of Algorithms," in textbook "Treatment Algorithms in Psychopharmacology," Jan Fawcett, Dan J. Stein, and Kenneth O. Jobson, editors, Wiley.
4. Jobson, Kenneth O. and Alan Schatzberg, 1998, "Introduction," Psychopharmacology Bulletin, Vol. 34, No. 3, p. 347.
5. Pearsall, Rowland, et al, 1998, "A New Algorithm for Treating Schizophrenia," Psychopharmacology Bulletin, Vol. 34, No. 3, pp. 349-353.
6. Trivedi, Madhukar H., 1998, "Developing Treatment Algorithms for Unipolar Depression in Cyberspace: International Psychopharmacology Algorithm Project (IPAP)," Psychopharmacology Bulletin, Vol. 34, No. 3, pp. 355-359.
7. Hartley, Dean S., III., 1998, "Observers' Report," Psychopharmacology Bulletin, Vol. 34, No. 3, pp. 361-368.
8. Jobson, Kenneth O., et al, 1998, "Synthesis Committee Report," Psychopharmacology Bulletin, Vol. 34, No. 3, pp. 369.
9. Fogg, B. J., 2003, "Persuasive Technology: Using Computers to Change What We Think and Do," Morgan Kaufman.
10. Fawcett, Jan, 2005, "Editorial: What We Need is 'Living Treatment Algorithms'," Psychiatric Annals, Vol. 35, No. 11, p. 873.
11. Adamson, Maureen, 2006, "Symposium on Diffusion, Adoption, and Maintenance of Psychiatric Treatment Algorithms, White Paper: Blueprint for Collaborations," IPAP.

Table of Contents

OR/MS Today Home Page


OR/MS Today copyright © 2007 by the Institute for Operations Research and the Management Sciences. All rights reserved.


Lionheart Publishing, Inc.
506 Roswell Rd., Suite 220, Marietta, GA 30060 USA
Phone: 770-431-0867 | Fax: 770-432-6969
E-mail: lpi@lionhrtpub.com
URL: http://www.lionhrtpub.com


Web Site © Copyright 2007 by Lionheart Publishing, Inc. All rights reserved.