- Email: [email protected]
Response to letter from Lipsitz et al.
Journal of Psychiatric Research 57 (2014) 180e181
Contents lists available at ScienceDirect
Journal of Psychiatric Research journal homepage: www.elsevier.com/locate/psychires
Letter to the Editor
Response to letter from Lipsitz et al. Dr. Lipsitz and his colleagues make important points in their letter regarding our recent publication in the Journal of Psychiatric Research (Khan et al., 2014a). We agree with Dr. Lipsitz and his colleagues that it's important to assess change in drug as well as placebo. We were unable to include the drug arm because none of the investigational antidepressants were approved, and there were not enough public domain data available to adequately assess the changes with drug in this group of seven trials. However, with the approval of levomilnacipran by the US FDA and availability of additional data, we have recently published a manuscript showing the magnitude of response to both antidepressants and placebo among approved antidepressants when two different methods of ratings are used (Khan et al., 2014b). Like in the paper under discussion the two methods of ratings in this analysis consisted of either traditional free-flow interviews or interviews that closely followed SIGMA format, were audio-taped and had RAPS follow up. There were no differences in the antidepressant response between the two methods of rating. However, the magnitude of placebo response was more than twice as large in the two trials using the SIGMA audio-taped ratings and follow-up RAPS feedback. It is important to note that simply obtaining an increased reliability as a goal for using enhanced interview techniques does not address the ultimate question of whether such a change in assessment procedure improves accuracy and accomplishes the intended purpose of lowering the magnitude of placebo response and hence increasing antidepressant-placebo differences. This is important since we have detected a paradoxical effect albeit in data from two small samples at a single research site. Our data suggest that a potential increase in reliability of ratings may in fact decrease the assay sensitivity. Although there are no data as to how this could occur, it is worth considering possibilities that need verification. Specifically, asking the same structured interview questions in a repetitive and consistent manner may in fact foster a shift to the mean. This is a well- established phenomenon with many types of evaluations, not limited to psychopharmacology trials, regardless of treatment assignment, placebo or antidepressant. The possibility that structured interviews with taping may foster the ‘placebo drift’ is very real as depressed patients tend to respond to all sorts of “treatment interventions” including antidepressants, psychotherapies, placebo, exercise and acupuncture to name a few, suggesting considerable expectation effects (Khan et al., 2012). Thus, being taped may further increase the expectations of the patient/subjects and thus enhance placebo response.
DOI of original article: http://dx.doi.org/10.1016/j.jpsychires.2014.06.014. http://dx.doi.org/10.1016/j.jpsychires.2014.06.016 0022-3956/© 2014 Elsevier Ltd. All rights reserved.
Subtle differences in response to antidepressants versus placebo may be detected by clinicians who are specially trained to pick these signs and symptoms up, but may be beyond the quantification attempts of relatively simple rating scales such as MADRS and HAM-D. Thus, modifying the depression interviews may result in the antidepressant-placebo differences being lost. Not surprisingly, there are previous examples of situations in psychiatric diagnosis where increasing the reliability has not yielded desired results as illustrated by the classical Midtown Manhattan Study. Specifically, using highly reliable methods of questioning the population in Midtown Manhattan resulted in a frequency of diagnosis of mental illness that was far higher than that observed in clinical evaluations (Srole et al., 1962). We appreciate the clarification made by Dr. Lipsitz and his colleagues about the definitions of RAPS items pertaining to rapport. Lipsitz et al. (2004) make middle of the road recommendations for factors such as neutrality and rapport, however, when we have reviewed the actual RAPS feedback comments, they generally don't follow the spirit intended in the publication. For example, a clarification question by a rater may be interpreted as a leading question by the RAPS reviewer with no recourse about this sort of criticism. Thus, this may cow the raters into asking bland questions that are considered ‘neutral’ and this in turn may foster a shift to the mean. Lastly, the concept developed by Dr. Zimbroff that explicit conveying of empathy will affect placebo response has never been substantiated (Khan and Brown, 2001). In summary, it is our intention to advance a dialogue about the utility and effectiveness of antidepressant clinical trial designs and conduct. We hope that as a field we are able to come up with the best and most feasible model/s. As we have mentioned in both our publications, it is critical to have more of these data in peer reviewed journals so that a broad based review is possible. We hope Dr. Lipsitz and his colleagues will publish the data they have accumulated regarding antidepressant-placebo differences and magnitude of placebo response when various interview techniques have been used. Funding source All of the authors were salaried by their institutions during the writing of this letter. No specific funds or salary were set aside for contributions to this letter. Contributors All of the authors have read and approved the final version of this letter. All of the authors contributed to the development of this letter.
Letter to the Editor / Journal of Psychiatric Research 57 (2014) 180e181
Declaration of conflict of interest Arif Khan, MD, Principal Investigator of over 380 clinical trials sponsored by over 85 pharmaceutical companies and 30 CROs has done no compensated consulting or speaking on their behalf, nor does he own stock in any of these or other pharmaceutical companies. Dr. Khan is not compensated for his role as author. James Faucett, M.S. is an employee of the Northwest Clinical Research Center. Walter A. Brown has no conflict of interest to declare. Acknowledgement None. References Khan A, Brown W. The placebo enigma in antidepressant clinical trials. J Clin Psychopharmacol 2001;21:123e5. Khan A, Faucett J, Lichtenberg P, Kirsch I, Brown WA. A systematic review of comparative efficacy of treatments and controls for depression. PLoS ONE 2012;7(7):e41778. http://dx.doi.org/10.1371/journal.pone.0041778. Khan A, Faucett J, Brown WA. Magnitude of placebo response and response variance in antidepressant clinical trials using structured, taped and appraised rater interviews compared to traditional rating interviews. J Psychiatr Res 2014a;51: 88e92. Khan A, Faucett J, Brown WA. Magnitude of change with antidepressants and placebo in antidepressant clinical trials using structured, taped and appraised rater
181
interviews (SIGMA-RAPS) compared to trials using traditional semi-structured interviews. Psychopharmacology 2014. http://dx.doi.org/10.1007/s00213-0143584-4:1-7. Lipsitz J, Kobak K, Feiger A, Sikich D, Moroz G, Engelhardt N. The rater applied performance scale: development and reliability. Psychiatry Res 2004;127:147e55. Srole L, Langer TS, Michael ST, Opler M, Rennie T. Mental health in the metropolis: The Midtown Manhattan study. New York: McGraw-Hill; 1962.
Arif Khan* Northwest Clinical Research Center, Bellevue, WA, USA Duke University School of Medicine, Department of Psychiatry, Durham, NC, USA James Faucett Northwest Clinical Research Center, Bellevue, WA, USA Walter A. Brown Brown University Department of Psychiatry and Human Behavior, Providence, RI, USA * Corresponding author. Duke University School of Medicine, Department of Psychiatry, Durham, NC, USA. Tel.: þ1 425 453 0404. E-mail address: [email protected] (A. Khan).
10 June 2014
Contents lists available at ScienceDirect
Journal of Psychiatric Research journal homepage: www.elsevier.com/locate/psychires
Letter to the Editor
Response to letter from Lipsitz et al. Dr. Lipsitz and his colleagues make important points in their letter regarding our recent publication in the Journal of Psychiatric Research (Khan et al., 2014a). We agree with Dr. Lipsitz and his colleagues that it's important to assess change in drug as well as placebo. We were unable to include the drug arm because none of the investigational antidepressants were approved, and there were not enough public domain data available to adequately assess the changes with drug in this group of seven trials. However, with the approval of levomilnacipran by the US FDA and availability of additional data, we have recently published a manuscript showing the magnitude of response to both antidepressants and placebo among approved antidepressants when two different methods of ratings are used (Khan et al., 2014b). Like in the paper under discussion the two methods of ratings in this analysis consisted of either traditional free-flow interviews or interviews that closely followed SIGMA format, were audio-taped and had RAPS follow up. There were no differences in the antidepressant response between the two methods of rating. However, the magnitude of placebo response was more than twice as large in the two trials using the SIGMA audio-taped ratings and follow-up RAPS feedback. It is important to note that simply obtaining an increased reliability as a goal for using enhanced interview techniques does not address the ultimate question of whether such a change in assessment procedure improves accuracy and accomplishes the intended purpose of lowering the magnitude of placebo response and hence increasing antidepressant-placebo differences. This is important since we have detected a paradoxical effect albeit in data from two small samples at a single research site. Our data suggest that a potential increase in reliability of ratings may in fact decrease the assay sensitivity. Although there are no data as to how this could occur, it is worth considering possibilities that need verification. Specifically, asking the same structured interview questions in a repetitive and consistent manner may in fact foster a shift to the mean. This is a well- established phenomenon with many types of evaluations, not limited to psychopharmacology trials, regardless of treatment assignment, placebo or antidepressant. The possibility that structured interviews with taping may foster the ‘placebo drift’ is very real as depressed patients tend to respond to all sorts of “treatment interventions” including antidepressants, psychotherapies, placebo, exercise and acupuncture to name a few, suggesting considerable expectation effects (Khan et al., 2012). Thus, being taped may further increase the expectations of the patient/subjects and thus enhance placebo response.
DOI of original article: http://dx.doi.org/10.1016/j.jpsychires.2014.06.014. http://dx.doi.org/10.1016/j.jpsychires.2014.06.016 0022-3956/© 2014 Elsevier Ltd. All rights reserved.
Subtle differences in response to antidepressants versus placebo may be detected by clinicians who are specially trained to pick these signs and symptoms up, but may be beyond the quantification attempts of relatively simple rating scales such as MADRS and HAM-D. Thus, modifying the depression interviews may result in the antidepressant-placebo differences being lost. Not surprisingly, there are previous examples of situations in psychiatric diagnosis where increasing the reliability has not yielded desired results as illustrated by the classical Midtown Manhattan Study. Specifically, using highly reliable methods of questioning the population in Midtown Manhattan resulted in a frequency of diagnosis of mental illness that was far higher than that observed in clinical evaluations (Srole et al., 1962). We appreciate the clarification made by Dr. Lipsitz and his colleagues about the definitions of RAPS items pertaining to rapport. Lipsitz et al. (2004) make middle of the road recommendations for factors such as neutrality and rapport, however, when we have reviewed the actual RAPS feedback comments, they generally don't follow the spirit intended in the publication. For example, a clarification question by a rater may be interpreted as a leading question by the RAPS reviewer with no recourse about this sort of criticism. Thus, this may cow the raters into asking bland questions that are considered ‘neutral’ and this in turn may foster a shift to the mean. Lastly, the concept developed by Dr. Zimbroff that explicit conveying of empathy will affect placebo response has never been substantiated (Khan and Brown, 2001). In summary, it is our intention to advance a dialogue about the utility and effectiveness of antidepressant clinical trial designs and conduct. We hope that as a field we are able to come up with the best and most feasible model/s. As we have mentioned in both our publications, it is critical to have more of these data in peer reviewed journals so that a broad based review is possible. We hope Dr. Lipsitz and his colleagues will publish the data they have accumulated regarding antidepressant-placebo differences and magnitude of placebo response when various interview techniques have been used. Funding source All of the authors were salaried by their institutions during the writing of this letter. No specific funds or salary were set aside for contributions to this letter. Contributors All of the authors have read and approved the final version of this letter. All of the authors contributed to the development of this letter.
Letter to the Editor / Journal of Psychiatric Research 57 (2014) 180e181
Declaration of conflict of interest Arif Khan, MD, Principal Investigator of over 380 clinical trials sponsored by over 85 pharmaceutical companies and 30 CROs has done no compensated consulting or speaking on their behalf, nor does he own stock in any of these or other pharmaceutical companies. Dr. Khan is not compensated for his role as author. James Faucett, M.S. is an employee of the Northwest Clinical Research Center. Walter A. Brown has no conflict of interest to declare. Acknowledgement None. References Khan A, Brown W. The placebo enigma in antidepressant clinical trials. J Clin Psychopharmacol 2001;21:123e5. Khan A, Faucett J, Lichtenberg P, Kirsch I, Brown WA. A systematic review of comparative efficacy of treatments and controls for depression. PLoS ONE 2012;7(7):e41778. http://dx.doi.org/10.1371/journal.pone.0041778. Khan A, Faucett J, Brown WA. Magnitude of placebo response and response variance in antidepressant clinical trials using structured, taped and appraised rater interviews compared to traditional rating interviews. J Psychiatr Res 2014a;51: 88e92. Khan A, Faucett J, Brown WA. Magnitude of change with antidepressants and placebo in antidepressant clinical trials using structured, taped and appraised rater
181
interviews (SIGMA-RAPS) compared to trials using traditional semi-structured interviews. Psychopharmacology 2014. http://dx.doi.org/10.1007/s00213-0143584-4:1-7. Lipsitz J, Kobak K, Feiger A, Sikich D, Moroz G, Engelhardt N. The rater applied performance scale: development and reliability. Psychiatry Res 2004;127:147e55. Srole L, Langer TS, Michael ST, Opler M, Rennie T. Mental health in the metropolis: The Midtown Manhattan study. New York: McGraw-Hill; 1962.
Arif Khan* Northwest Clinical Research Center, Bellevue, WA, USA Duke University School of Medicine, Department of Psychiatry, Durham, NC, USA James Faucett Northwest Clinical Research Center, Bellevue, WA, USA Walter A. Brown Brown University Department of Psychiatry and Human Behavior, Providence, RI, USA * Corresponding author. Duke University School of Medicine, Department of Psychiatry, Durham, NC, USA. Tel.: þ1 425 453 0404. E-mail address: [email protected] (A. Khan).
10 June 2014