The quest for certainty: goodbye to index cards.

When I was in physical therapy school more than 25 years ago, it was quite easy for most physical therapist students to do a classic assignment: making annotations from periodicals (the concept of peer-reviewed literature had not yet been articulated) in the library. Although I am tempted to say that when I was a student, we had to walk 10 miles in the snow to get to the library (thus suggesting how much easier students have it today), the library was really only across the street and down the block flora the School of Allied Health (the concept of autonomous practitioner also had not yet been articulated).

Once ensconced in my seat in the "R's" (with the Library of Congress System, all rehabilitation periodicals were coded as "R"), I had my choice of Physical Therapy, American Journal of Physical Medicine, and Archives of Physical Medicine and Rehabilitation. In any given library of the time, this was likely to be the exhaustive collection of current literature that was relevant to physical therapy. American medical journals such as the New England Journal of Medicine and JAMA rarely, if ever, published a manuscript that was related to rehabilitation. Later in my career, with better access to the library at the School of Medicine, I expanded my search strategies to include the Scandinavian Journal of Rehabilitation and the Scandinavian Journal of Rheumatology, which seemed to have a peculiar penchant for publishing articles pertinent to rehabilitation. Eventually, I discovered that I could find interesting patient-related studies in Nursing Research, particularly if I wanted to learn more about patient education.

My searches were usually done backwards: If I could find 10 related articles on a single topic, I would create my "clinical question" and then answer it. Frankly, I am not stile that I knew the difference between a "topic" and a distinct clinical question. I believed that each article was generally true in its entirety, and that a single article on a particular point was conclusive. Sometimes the articles were strictly didactic summaries of common wisdom, as were many of the few textbooks that were available at the time. My annotations would include case reports and, more frequently, single-group pre- and post-intervention quasi-experimental studies on a convenience sample (most likely "normal" subjects who were the author's students or classmates). When asked how I "knew" something, it seemed quite sufficient to respond, "I have an article on it." (Please note that having the article was sufficient; comprehending and analyzing the article were not really expected behaviors for "entry-lever" physical therapist students).

How times have changed!

In October 2003, Hooked on Evidence, APTA's searchable database on the effectiveness of physical therapy intervention, reached its first anniversary as a database open to Association members for searching (subscriptions for nonmembers are in the planning stage). By the time you read this, the Hooked project will have catalogued almost 1,000 articles from the peer-reviewed literature articles that are pertinent to the questions asked daily by clinicians. Hooked is a source of potential answers to very real clinical questions.

Using APTA's Guide to Physical Therapist Practice as its template, Hooked on Evidence is a work in progress to catalog all that is currently reported in the literature. A conscious decision was made to include all peer-reviewed and indexed literature--from case reports to randomized controlled trials--because, in some areas, case reports are the only form of evidence that we have. So often, pronouncements about "best practice" have been foisted on the clinician. Hooked on Evidence, however, is determined to build the edifice of our evidence for clinical effectiveness from the ground up, in full view. In fact, the critical next step in validating the database will likely reveal that we have made a few mistakes.

With the extensive cooperation of APTA's Section on Research, members will be checking each and every entry in the database against the original publication. Typos, misreadings, and misinterpretations will be corrected in public. Some people might wonder whether we should have waited to open up the Hooked Web site until we were absolutely sure that the database was perfect. We made a deliberate decision to go forward on good faith that the members who would be working on Hooked many of them students--would do their utmost to be accurate. It's not much different, really, than any of us who lead a study and attempt to apply the information to a particular patient's problem. We believe that the inferences we have drawn about patient care ale based on an accurate and precise understanding of the study's findings, If we, as a profession, are reluctant to publicly admit our misreadings and misinterpretations in a database, how will individual physical therapists have the coinage to confront their own mistakes in managing patients when new research shows that previous beliefs were wrong? Whether few or many, the corrections that we make to the Hooked entries will be a sign of a mature appreciation of what it means "to know" something.

Over the coming months and years, we will begin to sift through the evidence we have gathered about practice in order to determine whether there is sufficient evidence to draw conclusions about what is preferred physical therapist practice based on the evidence. We already knot, that the literature is silent on a number of important clinical questions, and we should be quick to emphasize that the absence of evidence cannot automatically be interpreted as evidence against the use of an intervention. Where the literature has ripened, we should be able to draw conclusions about what interventions may be most appropriate for particular problems in particular populations to achieve certain outcomes.

Physical therapy science has not yet reached consensus on what outcome measurements should routinely be used in research and what outcomes are the most appropriate to study for certain conditions in particular populations. To assist in the synthesis of a diverse literature, the Hooked database has not, begun to include graphical representations of various statistics related to treatment effect, such as odds ratios, relative risk reductions and absolute risk reductions for dichotomous outcome variables, and effect sizes for continuous measurements of outcome, where appropriate. All of these statistical methods will help us determine whether the evidence of practice that we have gathered also may function as evidence for practice, guiding clinicians toward the best possible choices of interventions for their patients with greater certainty of the likely outcomes. In essence, we will "know" what to do because we are able to use online technology to manage the expansion of data and statistical "technology" to grapple with the meaning of those data.

One caution, however, is in order. As we add new knowledge to our core set of beliefs about treatment effectiveness, we will be tempted to view what we think we know as being absolutely certain. Philosopher Ludwig Wittgenstein posited that doing so was both quite natural and inconsequential, with one caveat: "The difference between the concept of 'knowing' and the concept of 'being certain' isn't of any great importance at all, except as 'I know' is meant to mean: I can't be wrong." (1(p3)

Unless we are willing to acknowledge that what we believe we know may not be true at all, physical therapy will never move forward as a science. And for those who may counter that physical therapy is an art as well as a science and therefore "what I know is true for me," I urge you to remember this: Although great art may uncover universal troths, those truths are different from the ones derived from science. Our patients aren't necessarily going to be interested in the personal knowledge derived from our artistry, unless this knowledge also holds true for their experience.

Following the collection and synthesis of evidence through such mechanisms as the Hooked database, the next step in the profession's development as a clinical science, and in the physical therapist's maturation as a scientific clinician, will be to proceed as if we were certain in our knowledge--but with the full awareness that we might be wrong. Then, and only then, will we appreciate the gravity of being accountable for our professional actions. Each of us may struggle a bit with this notion in a variety of ways, but collectively we are ready. We have much more than an article, and we have no more need for index cards.

References

(1) Wittgenstein L. On Certainty. Anscombe GEM, von Wright GH, eds. Paul D, Anscombe GEM, trans. New York, NY: Harper & Row; 1972.

Authors with articles published in Physical Therapy or other journals are encouraged to enter an extraction of their article into Hooked on Evidence (www.apta.org/ hookedonevidence/index. cfm). A study is eligible for inclusion if it investigates the outcomes of one or more physical therapy interventions in a group of subjects who have the condition of interest. Contact David Scalzitti, Associate Director, Research Services, at 800/999-2782, ext 8555, for instructions.

Andrew A Guccione, PT, DPT, PhD, FAPTA APTA Senior Vice President for Practice and Research