Improving Clinical Pain Research: Past, Present, Future

Transcription

1 Improving Clinical Pain Research: Past, Present, Future Nathaniel Katz, MD, MS Analgesic Solutions, Natick, MA Tufts University, Boston, MA

2 Olson N, et al, J Clin Pharm, 2001

3 Hersh E, et al, Clin Ther, 2000

4 Relative Standard Effect Size SPID6 Ibuprofen liquigel 400mg vs. placebo: Hersh Delta 7.61 SD 4.85 SES 1.57 Sunshine Sunshine has 30% higher SES (Equivalent to reducing sample size from 100/arm to 60/arm)

5 Ray Houde ( ) Mitchell B. Max ( ) Louis Lasagna ( )

6 Case Study Assay Sensitivity (Inguinal Hernia Data) Research Site A (n = 126) All 24 Other Sites (n = 274) Primary efficacy endpoint: mean difference between active and placebo ( ) Pooled standard deviation (SD) Standardized effect size ( /SD) N needed for 80% power at alpha = 0.05 Subjects enrolled per site per month (mean) Overall Performance (time to 80% power) *utilizing one site at Lotus months* 36.6 months ** **utilizing 17 non-lotus sites in concert

7 Case Study Assay Sensitivity (Surgical Hemorrhoidectomy) Research Site A (n = 126) All 24 Other Sites (n = 274) Primary efficacy endpoint: mean difference between active and placebo ( ) Pooled standard deviation (SD) Standardized effect size ( /SD) N needed for 80% power at alpha = 0.05 Subjects enrolled per site per month (mean) Overall Performance (time to 80% power) *utilizing one site at Lotus months* 147 months** **utilizing 17 non-lotus sites in concert

8 What did they do? Single sites Experienced investigators & nurses Accurate diagnosis Empiric patient qualification Empiric nurse qualification Repeated use of good patients Talking to the patients

9 Reasons for Failure: Opioid Trials Trial structure Crossover and withdrawal better than parallel treatment Dosing Titration better than non-titration Flexible better than fixed Concomitant analgesics Prohibited better than allowed Rescue Prohibited better than allowed Primary endpoint AUC better than landmark Number of sites The fewer the better Katz N, et al, Neurology, 2005

10 Standardized effect size vs. number of sites, opioid trials

11 True vs. Actual Power Δ=1, sd=2.5, α=.05, power=.8

12 True vs. Actual Power

13 Low enrolling sites perform poorly Singla N et al, personal communication

14 Failure: Neuropathic Pain Trials Placebo response Positive Negative P-value Year (pub) PHN 16 4 Polyneuropathy Study design Katz J, Neurology, 2008

15 Failure: PDN Studies Correlation between design feature and SES Longer duration of PDN (r = 0.80) Shorter treatment duration (r = 0.65) Probably a proxy for earlier capture of primary endpoint Shorter titration (r = 060) Probably a proxy for shorter trials Smaller sample size (r = 0.55) Probably a proxy for fewer sites Fewer sites (r = 0.50) Fewer study visits per trial duration (r = 0.44) May be due to less nurturing nurse effect or potentially higher dropouts Two arms (SES = 0.60) vs. >2 arms (SES = 0.40) No rescue medication (SES = 0.60) vs. rescue medication (SES = 0.40) Dworkin R, Katz N, et al, unpublished

16 SES Clinical Condition Impacts SES PHN Clinical Condition DPN Dworkin RH, Pain, 2010

17 Psychophysical Assessment(Ф) I. Experimental Pain Rating Subjects rate 7 temperatures 7 times using VAS

18 VAS Rating VAS Rating Psychophysical Profile Samples Ф Evoked Thermal Pain Ratings Low variation reporter Low variation reporter (CoV=.42, ICC=.91, R 2 R=.72) Temperature ( o ( o C) C) Evoked Thermal Pain Ratings High variation reporter (CoV=.76, ICC=.58, R 2 =.47) Temperature ( o C)

19 Frequency Plots for Pain Reporting Skill N= 79 N Mean =.7363 Mean SD= SD N=79 Mean=.7826 SD= N=79 Mean=.5654 SD= CoV ICC R 2 Subjects demonstrated a large range of performance in pain reporting skill as indexed by CoV, ICC, and R 2.

20 Pain Matching Subjects adjust thermode temp until pain heat = pain OA (forced choice staircase procedure)

21 PM Score ( o C) VAS Score Delta Exercise Pain Results: Change in pain significantly different for PM not VAS SES Pre-exercise PM C p<.05 Post-exercise Pre-exercise Post-exercise VAS PM

22 Placebo reduction training program

23 SETS Positive Expectation Motivation to Participate Preliminary Results Positive expectations Motivation 2 No intervention PREP 7 No intervention PREP pre-mock post-mock 5 pre-mock enrollment post-mock

24 Bedside Sensory Testing Kit

25 Conclusions The old timers were probably doing something right We can quantify good research methods and standardize these approaches Doing so would accelerate the development of better analgesics