Modeling Hospice Professionals' Understanding of Preparatory Grief: Development and Initial Validation

Background and purpose: Patients report a variety of terminal reactions at the end of life across affective, behavioral, and cognitive domains. Preparatory grief reactions are considered common and may result in a temporary decrease in patient quality of life (QoL). Pathological reactions, in contrast, are considered more dysfunctional and often result in a significant and prolonged decrease in patient QoL. Self-assessed measures of pain, QoL, and other terminal reactions are not always feasible in end of life care. Hospice professionals (e.g., physicians, social workers, nurses) must often complete these measures as proxies to guide comfort and care. Research has illustrated that patient and proxy-assessed ratings are often incongruent resulting in unmet patient needs. It is critical to assess how well hospice professionals are able to discriminate between preparatory grief and pathological terminal reactions in order to provide timely, targeted psychosocial interventions to maintain patient QoL. The primary objective of this study, therefore, is to develop a rapid-assessment instrument able to provide a measure of hospice professionals’ ability to accurately differentiate preparatory grief from pathological reactions in dying patients. The primary research question of this study is: Can hospice professionals’ identification and understanding of preparatory grief and pathological responses to terminal illness in the dying be usefully modeled and scaled? It is hypothesized that a three factor structure paralleling the affective, behavioral, and cognitive domains of terminal reactions will be found. Method: Initial item development for the Hospice Professionals’ Understanding of Preparatory Grief (HPPG) scale yielded 52 questions. Expert panel review (n=6) was utilized to examine content validity. Current hospice professionals (e.g., social workers, nurses) and PhD level social work faculty provided quantitative and qualitative feedback regarding initial items resulting in a reduced item pool (n=50). An online survey was then utilized to collect data from hospice professionals registered with two professional ListServs (Louisiana-Mississippi Hospice and Palliative Care Organization; Florida Hospice and Palliative Care Association). The data collection instrument included the HPPG, a measure of palliative care self-efficacy, and questions regarding professional and demographic characteristics. Cronbach’s alpha and stratified alphas were computed for sub- and global HPPG scales. Confirmatory factor analysis (CFA) was utilized to examine factor structure. Results: Initial item removal via examination of alpha-if-item-deleted values for HPPG subscales resulted in 33 items. These items were then utilized in the CFA to examine factor structure with final respecification yielding a χ²/df =2.22, CFI=.91, TLI=.90, RMSEA=.08, and SRMR=.14. Subsequent reliability analyses resulted in 24 items for the final instrument. The global HPPG scale demonstrated good internal consistency (ɑ=.812; α_stratified=.861) as did the subscale domains (affective: ɑ=.88; behavioral: ɑ=.74; cognitive: ɑ=.82). Conclusions and Implications: Initial evidence suggests adequate reliability and validity of the HPPG. Subsequent validations with larger, more representative samples should be considered with the final instrument. The HPPG may be utilized by hospice professionals and administrators to assess abilities in differentiating normal, preparatory grief reactions from pathological reactions in terminal patients at the end of life thereby impacting practice behaviors in hospice and palliative end of life care.