Although emotional exhaustion scores decrease with respect
to time, the reason no significant difference was found between
the groups may be sample losses. Because the intention-to-treat
approach was based on the analyses, it was an anticipated result
that the effect of the intervention was lower (Gross & Fogg 2004).
In the study conducted by Marine et al. (2006), it was reported
that one of the biggest problems in studies on work stress was
losses in samples. Because the sample losses encountered in this
study posed an important problem, a qualitative study was conducted
to clarify these losses (Günüs¸en & Üstün 2009). Nurses
described burnout as resulting from work-related issues, and
their attendance was impeded by working conditions and lack of
support. Although we had support for our programme from the
senior nursing administration, this did not always translate into
support at the unit level, where nurse managers did not consistently
develop schedules that allowed attendance. Most participants
stated that they need work-directed rather than persondirected
interventions to reduce stress and burnout. Therefore,
their belief that person-directed interventions would be useless
had a negative impact on the level of participation. This suggests
that intervention programmes are more likely to be effective
if they include person-directed as well as work-directed
approaches (Günüs¸en & Üstün 2009). It is recommended to the
researchers who wish to conduct this type of studies that they
should take measures to minimize sample losses.