| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Research Reports |
VJ Robertson, PT, PhD, is Professor, University of Newcastle, Teaching and Research Unit, Gosford Hospital, Gosford, New South Wales 2250, Australia
Address all correspondence to Professor Robertson at: Val.Robertson{at}newcastle.edu.au
White swans had an important role in European philosophy. Different versions of the following sentences were used to illustrate logical arguments and refutations:
Premise 1: All swans are white.
Premise 2: This bird is a swan.
Conclusion: Therefore, this bird is white.
The finding of black swans in Australia in the late 18th century put this to rest and added a new set of variants to these sentences.
How do swans and the present article1 relate? There are 2 connections: the first concerns an obvious epistemological flaw, and the second concerns evidence of knowledge changing over time. Prior to exploring these connections, I commend the authors on analyzing patient records from a 2-year period. Mining existing patient records can be invaluable and should be attempted more often.
Back to the first connection with swans: the present article implies that knowledge of the outcomes for more patients would increase our conviction of the value of this therapy. The danger of this is obvious: with white swans, centuries of repeated instances of them always being white convinced observers that the first premise was true. One black swan changed this! Likewise, with studies of ultrasound, more observations should not necessarily increase our certainty.
Staying with certainty, the more times we see the sun rise, the more likely we are to believe it will do so again tomorrow. With the sun, the underpinning science does justify our high level of certainty about its rising again tomorrow. This is not so for the effects of many patient interventions. At best, sophisticated statistical analyses tell us the probability of particular outcomes following an intervention under specified conditions.
Research methods can help increase our certainty, particularly if patients are randomly allocated to either a treatment group or a nontreatment group. All treatments then should be provided in a double-blinded context: patients not knowing whether they are in an active or placebo treatment arm, administering therapists not knowing either, and assessors blinded to the group assignment of participating patients. If a treated group has a particularly different outcome, we can ascribe this to the intervention if the procedures were implemented as planned, appropriate assessment methods used, and so on. We still prefer more extensive testing in multiple centers with different researchers to be convinced that the outcomes were not due to contiguous events.
Moving on from Epistemology 101: a cursory review of the present article raises important questions that should affect our confidence in its findings. The critical reason, common in retrospective studies, concerns the lack of procedural detail and rationale for treatments. Starting at the beginning, no rationale was provided for the exclusion criteria. Why were patients’ data excluded if they were treated less than twice a week? Why not 3 times a week? Was this a convenience, related to the data set, or a decision based on an unspecified rationale or institutional treatment protocols? The description of the "conventional wound care" used includes "other biophysical technologies." What were they, and were they used routinely or occasionally or only on wounds associated with particular medical conditions or in certain locations? Why were nonhealing wounds defined as those with less than 15% closure in the prior 2 weeks? While no doubt justifiable, it could be seen as opportunistic in the absence of an examinable justification.
Another concern has to be with the lack of detail regarding the treatment intervention. This is a long-standing issue for many of us interested in electrophysical agents (EPAs). If we do not know how an EPA was applied, how can we systematically replicate the procedure and better establish the effectiveness or not of a treatment or identify the conditions under which it is effective? We need dosage details. What distance was the applicator from the wound? What was the application time per unit area treated? What parameter options can be controlled by the operator? If any can be, what parameters were used? Did the therapists comply and accurately record dosages? Was the output of the equipment technically verified at adequate time periods? Were all patients treated the same number of days per week and minutes per area of wound? Clearly, the times per week varied considerably, but I am puzzled as to how the range was from 1.5 to 4 times. What does a frequency of 1.5 times per week mean given the inclusion criterion was twice a week minimum and the number of treatment weeks varied considerably and for unidentified reasons? What were the criteria for starting wound care with ultrasound and for terminating it? Saying they "included changes in exudation and amount of devitalized tissue" begs the question. Were these changes objectively ascertained or was a reliable system of grading used? In summary, many essential treatment details appear to be missing. The implications of such omissions have previously been identified and discussed.2 That the usual focus is on megahertz-frequency ultrasound does not change the implications if newer uses of kilohertz frequency are being reported.
We also need to consider the outcome measures used. Is the method used to measure change in wound size intertester reliable, valid, and sensitive? Were measurers tested for reliability? Again, I am sure these questions can easily be addressed by the researchers, but, as they have not been, we should remain uncertain about the findings.
Assuming acceptable responses to the questions I have raised to date, I would have hoped for some more interesting analyses. For example, the duration of wound chronicity was very large. Did chronicity affect the outcome? What was the effect for smokers versus nonsmokers with comparable wounds (size, chronicity, location)? Of the 13 wounds that closed completely during the study period, what distinguished them from the majority of wounds that did not close? What was the week-to-week closure rate of wounds in different body regions? Or the size change per 4 or 5 applications of ultrasound to wounds? Without these types of analysis, it is very difficult to make sense of the subgroup analysis provided, as treatment conditions appear to have been very different. How similar was the rate of healing to previously reported studies using other treatment methods? Did some people have more than one wound being treated? The clinical significance of a drop in pain of 1.8 points across 26 patients is uninterpretable without knowing what scale was used. Furthermore, statistical significance is not necessarily a major consideration in a clinical phenomenon like pain.
Why do these omissions and lack of analyses matter? One of the references cited in the present article answers this question.3 That study reported the outcomes of a multicenter, randomized controlled trial to examine the effects of 40-kHz ultrasound on wound healing. The researchers carefully specified essential experimental factors, including the procedure; how they managed differences between the sham and active ultrasound equipment; and how they measured changes in the wounds.
Despite the care they clearly took, Ennis et al3 subsequently identified important differences in the protocol implementation in different centers. This finding appeared to be serendipitous, a consequence of unexpectedly improved rates of healing of some sham-treated wounds. The application distance from the wound was found to have been inverted for the sham and active methods in 5 centers. Data for 42 patients were deleted from final analyses, although an intention-to-treat analysis was provided. What would have happened had the findings confirmed the expected results? Presumably, the methodological differences would not have been identified and the delivery of 40-kHz ultrasound via a continuous irrigation stream would have been assumed to have limited effectiveness.
The significance of this for the present study should be obvious. Very careful specification and checking of procedures is essential. The present article has not provided sufficient details for us to be sure what was done and how procedural equivalence was ensured. Regarding the unexpected benefits of an inadvertently altered sham protocol: the findings of Ennis et al3 appear to confirm the importance of ultrasound for debridement. They also raise questions about the relative contributions to wound healing of 40-kHz ultrasound and of its mode of delivery, an irrigation system. I hope future research will address this, if it has not yet been done.
After reading the present article, I would not know when to consider using kilohertz-frequency ultrasound for treating wounds. I remain reluctant to rely on material provided by manufacturers and sales representatives. See, for example, the unfortunate conflation of radio-frequency and ultrasound diathermy by a manufacturer of deep brain stimulators.4 Such errors tend to be repeated by clinicians.5
In summary, different ways exist to test our certainty about the effectiveness of treatment interventions. Finding more confirming examples is not ideal. The authors could have strengthened their article considerably by providing adequate procedural details and discussing the inherent weaknesses in their method. A reference on their reference list provided an excellent example of what was required. For the record, I note both studies of 40-kHz frequency ultrasound discussed in this commentary1,3 reported receiving support from the manufacturer. This was carefully outlined in both instances and seemed appropriate.
Returning to the present study: the second connection among this article, this commentary, and swans is to a series of items Professor Alex Ward and I published about 45-kHz ultrasound, also known as "long-wave ultrasound."6–8 We were concerned with what we perceived as uncritical acceptance of a study reporting the treatment of acute and chronic musculoskeletal problems with 45-kHz ultrasound. We examined the theoretical basis of the claims made for it and reviewed existing empirical studies.6
Reading the present article, therefore, reminds me of the dangers of being certain all swans are white. Like ugly fluffy cygnets yet to become beautiful black or white swans, our knowledge of uses of low-frequency ultrasound is still developing. However, convincing studies are required—methodologically adequate studies that probe the effectiveness of 40-kHz frequency ultrasound for wound healing and compare its relative contribution to that of the means of delivering the irrigating fluid.
 | References |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  A. L Bell and J. Cavorsi Author Response Physical Therapy, December 1, 2008; 88(12): 1526 - 1528. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
