Response
http://www.100md.com
《美国整骨期刊》
Philadelphia College of Osteopathic Medicine Philadelphia, Pennsylvania
MARVIN J. BERMAN, DO
Nazareth Hospital Philadelphia, Pennsylvania
MARILYN LERARIO, BSC, RPH
Philadelphia Veterans Administration Medical Center Philadelphia, Pennsylvania
As always, we appreciate comments by others on our clinical research. In response to those points raised by Roberto Cardarelli, DO, MPH, regarding our June 2005 original contribution (2005;105:273–279), we offer the following.
Although we are in basic agreement with Dr Cardarelli that "we, as a profession, continue to produce...large numbers of pilot studies on the effectiveness of OMT [osteopathic manipulative treatment]," our June 2005 study rendered statistically significant results in spite of the small number of patients. Furthermore, as clinicians and researchers, we chose not to enroll more subjects than necessary.
On the topic of sample size, Dr Cardarelli also noted that our study had 39 participants placed in four study groups, adding that "No discussion on the study's power or sample size calculations was provided for readers...to make accurate conclusions about the study results."
Accurate conclusions are typically not derived from proof-of-concept research that is appropriately classified as a "pilot." In addition, power analyses are typically performed if the planning estimates fail to hold and a nonsignificant difference of the primary end point is observed. In our case, power is not an issue given that a significant difference (P=.02) was detected relative to the dose of morphine sulfate.
As noted in our original contribution, our study protocol was approved by the institutional review boards at the City Avenue Hospital (Philadelphia, Pa), which reviewed our sample size estimates and carefully weighed all bioethical issues ascribed to this study, as well as the American Osteopathic Association's Bureau of Research (now the Council on Research), which conducted a similar review and subsequently provided financial support for our clinical investigation.
As previously mentioned, we felt it was not appropriate to randomize a greater number of patients in our study without the opportunity of gathering more data than we were already able to produce with 39 subjects. We did consider an adaptive study design that used Bayesian monitoring. However, in the end, we agreed that the clarity yielded with the design described was the most efficient option to address this research question.
Regarding the significance of our statistical findings, Dr Cardarelli wrote: "In addition, though the authors noted a statistically significant finding between Groups 3 and 4 with regard to reduced morphine use in the first 24 hours after elective total abdominal hysterectomy (P=.02), overlapping confidence intervals cast some doubt on the stability of the data and whether these findings were truly beyond chance alone."
It is our hope that readers understand that the objective of this level of research (ie, a pilot study) is to derive a signal from the randomized treatment groups. It is important to remember that probability values are only intended to provide insight relative to the likelihood of finding a difference of a certain magnitude. In this particular case, the probability that a difference of this magnitude would occur by chance is approximately 2%, which is an impressive finding.
As Dr Cardarelli indicated, subjective pain scores are certainly important for treating patients in general. However, subjective measures are not necessarily useful parameters for a scientifically based clinical investigation on pain management. In fact, we clearly stated in our paper that similar findings were also reported by Miaskowski et al1 and Adams et al.2 In addition, differences in pain scores were not necessarily expected knowing that morphine doses were patient controlled, ensuring an adequate level of analgesia. Thus, our study design was sanctioned, in part, by a methodology that was able to assume that all patients would receive the full complement of medication required for appropriate pain control.
Finally, Dr Cardarelli noted that our study did not include groups that received "saline or morphine alone (ie, without OMT or sham manipulative treatment)." He added, "according to extensive discussions at the Sixth Annual American Association of Colleges of Osteopathic Medicine Osteopathic Collaborative Clinical Trials Initiatives Conference, this remains a controversial methodology issue for OMT research." Although it was, admittedly, not clearly noted in our June 2005 JAOA article, our research was completed more than a year before the April 2005 conference to which Dr Cardarelli referred. Therefore, we feel it is not entirely appropriate to attempt to draw a connection between this conference and our study. Even if we grant Dr Cardarelli's point momentarily for the sake of argument, however, we would like to note that he does not state that there were definitive conclusions regarding this study methodology at this conference, only "discussions," which though certainly de rigeur at all conferences do somewhat temper his "connective" comment.
In any research involving OMT, there will always be a question of whether or not human touch alone will bring about beneficial changes in patients. It is clear that such benefits can, and do, occur.3–5 Therefore, our study design intentionally sought to separate "touch" from actual osteopathic manipulation. We wanted to—and succeeded in—developing a clinical research model that demonstrated that touching alone will not "do it"; osteopathic manipulation must occur for patients to receive clinical benefits. The objective of our study was, as stated, to determine whether there was a clinical effect of adding postoperative OMT to preemptive morphine sulfate in patients undergoing elective total abdominal hysterectomy. Adding study groups who received only preemptive saline or only morphine would address a different research question from the one we sought to answer with our study design.
References
1. Miaskowski C, Mack KA, Dodd M, West C, Paul SM, Tripathy D, et al. Oncology outpatients with pain from bone metastasis require more than around-the-clock dosing of analgesics to achieve adequate pain control. J Pain. February2002; 3:12 –20.
2. Adams HA, Saatweber P, Schmitz CS, Hecker H. Postoperative pain management in orthopaedic patients: no differences in pain score, but improved stress control by epidural anaesthesia. Eur J Anaesthesiol. 2002;19:658 –665.
3. Seskevich JE, Crater SW, Lane JD, Krucof MW. Beneficial effects of noetic therapies on mood before percutaneous intervention for unstable coronary syndromes. Nurs Res.2004; 53:116 –121.
4. McCaffrey R, Taylor N. Effective anxiety treatment prior to diagnostic cardiac catheterization [review]. Holist Nurs Pract. 2005;19:70 –73.
5. Burr JP. Jayne's story: healing touch as a complementary treatment for trauma recovery [review]. Holist Nurs Pract.2005; 19:211 –216.(FREDERICK J. GOLDSTEIN, P)
MARVIN J. BERMAN, DO
Nazareth Hospital Philadelphia, Pennsylvania
MARILYN LERARIO, BSC, RPH
Philadelphia Veterans Administration Medical Center Philadelphia, Pennsylvania
As always, we appreciate comments by others on our clinical research. In response to those points raised by Roberto Cardarelli, DO, MPH, regarding our June 2005 original contribution (2005;105:273–279), we offer the following.
Although we are in basic agreement with Dr Cardarelli that "we, as a profession, continue to produce...large numbers of pilot studies on the effectiveness of OMT [osteopathic manipulative treatment]," our June 2005 study rendered statistically significant results in spite of the small number of patients. Furthermore, as clinicians and researchers, we chose not to enroll more subjects than necessary.
On the topic of sample size, Dr Cardarelli also noted that our study had 39 participants placed in four study groups, adding that "No discussion on the study's power or sample size calculations was provided for readers...to make accurate conclusions about the study results."
Accurate conclusions are typically not derived from proof-of-concept research that is appropriately classified as a "pilot." In addition, power analyses are typically performed if the planning estimates fail to hold and a nonsignificant difference of the primary end point is observed. In our case, power is not an issue given that a significant difference (P=.02) was detected relative to the dose of morphine sulfate.
As noted in our original contribution, our study protocol was approved by the institutional review boards at the City Avenue Hospital (Philadelphia, Pa), which reviewed our sample size estimates and carefully weighed all bioethical issues ascribed to this study, as well as the American Osteopathic Association's Bureau of Research (now the Council on Research), which conducted a similar review and subsequently provided financial support for our clinical investigation.
As previously mentioned, we felt it was not appropriate to randomize a greater number of patients in our study without the opportunity of gathering more data than we were already able to produce with 39 subjects. We did consider an adaptive study design that used Bayesian monitoring. However, in the end, we agreed that the clarity yielded with the design described was the most efficient option to address this research question.
Regarding the significance of our statistical findings, Dr Cardarelli wrote: "In addition, though the authors noted a statistically significant finding between Groups 3 and 4 with regard to reduced morphine use in the first 24 hours after elective total abdominal hysterectomy (P=.02), overlapping confidence intervals cast some doubt on the stability of the data and whether these findings were truly beyond chance alone."
It is our hope that readers understand that the objective of this level of research (ie, a pilot study) is to derive a signal from the randomized treatment groups. It is important to remember that probability values are only intended to provide insight relative to the likelihood of finding a difference of a certain magnitude. In this particular case, the probability that a difference of this magnitude would occur by chance is approximately 2%, which is an impressive finding.
As Dr Cardarelli indicated, subjective pain scores are certainly important for treating patients in general. However, subjective measures are not necessarily useful parameters for a scientifically based clinical investigation on pain management. In fact, we clearly stated in our paper that similar findings were also reported by Miaskowski et al1 and Adams et al.2 In addition, differences in pain scores were not necessarily expected knowing that morphine doses were patient controlled, ensuring an adequate level of analgesia. Thus, our study design was sanctioned, in part, by a methodology that was able to assume that all patients would receive the full complement of medication required for appropriate pain control.
Finally, Dr Cardarelli noted that our study did not include groups that received "saline or morphine alone (ie, without OMT or sham manipulative treatment)." He added, "according to extensive discussions at the Sixth Annual American Association of Colleges of Osteopathic Medicine Osteopathic Collaborative Clinical Trials Initiatives Conference, this remains a controversial methodology issue for OMT research." Although it was, admittedly, not clearly noted in our June 2005 JAOA article, our research was completed more than a year before the April 2005 conference to which Dr Cardarelli referred. Therefore, we feel it is not entirely appropriate to attempt to draw a connection between this conference and our study. Even if we grant Dr Cardarelli's point momentarily for the sake of argument, however, we would like to note that he does not state that there were definitive conclusions regarding this study methodology at this conference, only "discussions," which though certainly de rigeur at all conferences do somewhat temper his "connective" comment.
In any research involving OMT, there will always be a question of whether or not human touch alone will bring about beneficial changes in patients. It is clear that such benefits can, and do, occur.3–5 Therefore, our study design intentionally sought to separate "touch" from actual osteopathic manipulation. We wanted to—and succeeded in—developing a clinical research model that demonstrated that touching alone will not "do it"; osteopathic manipulation must occur for patients to receive clinical benefits. The objective of our study was, as stated, to determine whether there was a clinical effect of adding postoperative OMT to preemptive morphine sulfate in patients undergoing elective total abdominal hysterectomy. Adding study groups who received only preemptive saline or only morphine would address a different research question from the one we sought to answer with our study design.
References
1. Miaskowski C, Mack KA, Dodd M, West C, Paul SM, Tripathy D, et al. Oncology outpatients with pain from bone metastasis require more than around-the-clock dosing of analgesics to achieve adequate pain control. J Pain. February2002; 3:12 –20.
2. Adams HA, Saatweber P, Schmitz CS, Hecker H. Postoperative pain management in orthopaedic patients: no differences in pain score, but improved stress control by epidural anaesthesia. Eur J Anaesthesiol. 2002;19:658 –665.
3. Seskevich JE, Crater SW, Lane JD, Krucof MW. Beneficial effects of noetic therapies on mood before percutaneous intervention for unstable coronary syndromes. Nurs Res.2004; 53:116 –121.
4. McCaffrey R, Taylor N. Effective anxiety treatment prior to diagnostic cardiac catheterization [review]. Holist Nurs Pract. 2005;19:70 –73.
5. Burr JP. Jayne's story: healing touch as a complementary treatment for trauma recovery [review]. Holist Nurs Pract.2005; 19:211 –216.(FREDERICK J. GOLDSTEIN, P)