There are a number of methods to assess specific areas of behavior when evaluating changes in functioning, assessing the effects of an intervention, or measuring the effectiveness of a program of care. Some of these methods include: self-report, caregiver report, and family report of person-specific behaviors. However, self-reporting by dementia patients is often not reasonable given their cognitive decline and communication challenges. Ratings from other people can introduce measurement errors. This article provides information on the theoretical roots, administration methods, and psychometric properties of measures of direct observation of individual behavior.
This review looks at methods of direct observation of behaviors for use in long-term care settings, particularly with older adults who have dementia.
The authors conducted a literature review of direct observation measures of individual behavior for assessing people with dementia. They searched the literature from 1990 through 2007 in the AgeLine, PsycLIT, PubMed, and Sociological Abstracts databases using the keywords dementia and observation, behavior observation, behavior assessment, behavior evaluation, observational measurement, observational assessment, quality of life assessment, affect observation, pain observation, and pain assessment. Finally, the authors gathered relevant references to observational measures cited in the literature obtained from these searches.
This study consolidated and evaluated the reviewed literature into the following categories:
- Characteristics and Psychometric Properties of Direct Observation Measures of Individual Behavior
- Characteristics and Psychometric Properties of Direct Observation Measures of Individuals' Affect
- Characteristics and Psychometric Properties of Direct Observation Measures of Individuals' Activities of Daily Living
- Characteristics and Psychometric Properties of Direct Observation Measures of Individuals' Quality of Life
In using such methods, the authors concluded that a number of things should be considered, including:
- how often and when to observe,
- the behavior or range of behaviors to target, and
- the psychometric properties desired, depending on the assumptions held about the behavior observed (e.g.,stable trait, sensitive to environmental change).
Not applicable in this review, however, the authors discuss the limitations of this type of research method.