Table 4

ANCOVA estimation of pre-specified primary, post-specified secondary, and ancillary outcomes

Outcome Variables

Treatment effect estimate

Standard error

P-value


Pre-specified primary outcome


Brief Fatigue Inventory (BFI)

-02.38

01.35

0.0776

Secondary outcomes


Perceived cognitive impairments, QOL impact(FACT-COG)

-00.44

00.56

0.4286

Perceived cognitive impairments (FACT-COG)

-00.16

00.30

0.6024

Perceived cognitive abilities (FACT-COG)

-00.02

00.45

0.9609


ANCOVA is fit with Huber's M estimation, a robust regression methodology.44 Rregression equations model score at follow-up as a function of an intercept, pre-treatment score and treatment (an indicator variable).

Johnston et al. BMC Complementary and Alternative Medicine 2011 11:49   doi:10.1186/1472-6882-11-49

Open Data