Twenty-one CDSMP classes (with parallel control groups)
were conducted in the period from October 2004 to January
2006 in 11 consecutive cycles, with at least one CDSMP class
in each cycle. In each cycle, after completion of baseline
measurements, patients were randomly allocated (per hospital)
to the intervention or control group using a computerized
allocation procedure in SPSS 12.0 (SPSS Inc., Chicago, IL,
USA) for Windows by an independent researcher blinded for
patients’ characteristics. Control patients received usual care,
consisting of regular checkups with the cardiologist and/or
the CHF nurse specialist at an outpatient clinic. Patients
allocated to the intervention group received the 6-week selfmanagement
group programme in addition to usual care.
Patients were not blinded to group allocation. For practical
reasons, an unequal randomization ratio was applied during
the recruitment period as in each cycle the time to recruit
patients for the study was limited. As the group process is
very important in the CDSMP, at least eight participants are
required to participate in the classes (Lorig et al. 2001b).
Therefore, based on the total number of recruited patients per
cycle, 59% were allocated to the intervention group and 41%
to the control group through the full recruitment period to
ensure sufficiently large CDSMP classes, taking possible
attrition into account (Dumville et al. 2006).