Clinical Research Oriented Workshop (CROW) Meeting: Sept 9, 2021

 

Present:   Levi Bonnell, Justine Dee, Nancy Gell, Juvena Hitt, Rocky Kelley, Ben Littenberg, Adam Sprouse-Blum, Connie van Eeghen (8)

 1.                   Warm Up: Justine did not have to work on her manuscript over the holiday weekend 😊  Friday seminar starts tomorrow, with food, and maybe outdoors. 

2.                   Justine’s RCT: Looking for general feedback on clarity/wording, and help with the discussion. She also needs to cut out 671 words.

a.       To be submitted to PT and Rehab Journal

b.       As a RCT: reference the CONSORT 2010 checklist on the web

c.       How to manage the wealth of secondary outcomes for which there are no significant results

                                                   i.      Para: primary results, with CI and P values

                                                 ii.      Para: secondary results, just as a table, with brief narrative, e.g. no significant findings

                                               iii.      Results can be very short; save the pattern analysis for the Discussion section

                                               iv.      Start with: all subjects improved over time (see table).  However, there was no difference between groups. 

1.       In discussion, consider causes: regression to mean, natural history…

                                                 v.      Combine Table 2 and Figure 3

d.       Results section:

                                                   i.      Keep consort diagram description in first para

                                                 ii.      The role of pain: on the causal pathway to function, not the end result

1.       Pain got better regardless of central sensitization technique

2.       Function did not get better

a.       Therefore: we made pain better but we didn’t measure the reason why

b.       Why function not better

                                                                                                                           i.      Poor measure

                                                                                                                         ii.      Poor theory about function and pain

                                                                                                                       iii.      Not long enough treatment

                                                                                                                       iv.      Not long enough follow up

                                                                                                                         v.      Other…

3.       Improvement of pain not explained; we need a different strategy to explain how to improve pain.  Central sensitization may be a red herring.

                                               iii.      Are interactions mentioned in the Methods?  Talk about exploratory, subgroup analysis.

                                               iv.      OK to eliminate tables or narrative with no interesting results (not both)

                                                 v.      Evaluate the study population (those that did not drop out that you have data for)

1.       Then analyze the group that dropped group and compare; are they different?  Don’t include in Table 1. Should include the 98 as the analyzed group.  Include para about the 19 for whom the outcomes data don’t exist.  Put it in a separate table to think about it.

e.       Discussion

                                                   i.      Needs to include model (as a figure?): Trauma, CSI, pain, function

3.                   Next week:  TBD