Assessment of Hopelessness in Older Adults: The Development and Initial Validation of the Hopelessness Inventory for Later Life (HILL)

Abstract

Assessment of Hopelessness in Older Adults: The Development and Initial Validation of the Hopelessness Inventory for Later Life (HILL) Jeffrey J. GreggHopelessness has been shown to be a strong, independent risk factor for physical illness, suicidal behavior, and mortality. It is an especially important construct for assessment in later life, as older adults are at an elevated risk of suicide worldwide. Studies have generally supported the use of the Beck Hopelessness Scale (BHS;Beck et al, 1974) with older adults; though, it was not designed with this population in mind. On the other hand, the Geriatric Hopelessness Scale (GHS; Fry, 1984) was constructed specifically for use with older adults, but has demonstrated unstable psychometric properties. In addition, many items on the GHS lack face validity or may be too culturally specific. Both the BHS and the GHS contain items that may be conflated with realistic appraisals of remaining time or physical illness. Based on these characteristics, the creation of a new measure for geriatric hopelessness is warranted. The current study sought to construct and validate a self-report scale for late-life hopelessness that contains future-oriented thematic content relevant to older adults (e.g., legacy, social support, pain) and excludes items that may be inherently biased against older adults or individuals with life-limiting illnesses. Items were drafted based on a literature review and sent to experts in the field of geriatric depression/suicide for revision and to establish content validity. After review, the preliminary version of the Hopelessness Inventory for Later Life (HILL) contained 30 items and utilized a 4point Likert-type response format. A sample of 265 older adults completed a survey battery containing the preliminary HILL, as well as other relevant measures in order to further revise the item content and examine its psychometric properties. The sample included older adults recruited in-person (i.e., from senior centers, a family medicine clinic, churches), through a mailing list, and online via Amazon mTurk (M age=71.1 SD=6.7). An examination of itemresponse characteristics (e.g., skewness) revealed one item as a candidate for deletion. An unrotated principal components analysis was then used to examine unidimensionality and to identify additional items for deletion. Based on these analyses, two viable versions of the scale, the HILL and the HILL-Shortened (HILL-S) were proposed. Both exhibited strong itemresponse characteristics, as well as preliminary evidence of unidimensionality, internal consistency (α=.96 and α=.89, respectively), and construct validity. Regarding construct validity, the HILL and the HILL-S were strongly associated with measures of hopelessness (BHS), geriatric depression, perceived burdensomeness, and social support. Moderate correlations were found with suicide risk, anxiety, self-rated health, physical health, and pain interference. The HILL and the HILL-S demonstrated small associations with social desirability and pain frequency. ...