Statistical analysis
SAS version 9.1.3 will be used for data analysis (SAS
Institute Inc., Cary, NC). Descriptive statistics of participant
baseline characteristics will be presented for each
treatment group to assess their comparability as well as
the generalizability of the sample. We will use an intention-
to-treat approach for all analyses. The analyses for
the primary and secondary aims are described below.
We will not impute any missing data for analysis, but
will report the amount of missing data for each variable
and the reason it is missing.
Data Analysis for Primary Aim
For each of the 3 primary response variables, the
immediate changes in the pre to post treatment measurements
at BL2 and the changes in the pre-treatment
variables from BL2 to week 2 will be compared across
the 3 treatment groups using analysis of covariance
(ANCOVA) adjusting for the minimization variables.
The following 2 preplanned contrasts for the 2 degrees
of freedom for treatment will be tested at 0.05: HVLASM
vs. sham control and LVVA-SM vs. sham control
on the adjusted means based on the ANCOVA model.
Residual plots will be used to assess the adequacy of the
model assumptions. Data transformations will be
explored when model assumptions are violated. Further
adjustments for the following baseline characteristics
will be explored to increase the precision of the estimated
effects: RMDQ, NRS, Quebec Task Force classification,
Beck Depression Inventory and chiropractic care
(yes/no). Adjusted group means and mean differences
between SM and control groups will be reported with
95% confidence intervals determined under the final
ANCOVA model.