Date of Award


Document Type


Degree Name

Doctor of Philosophy (PhD)



Research Advisor

Michael A. Carter, DNSc, DNP


Patricia Cunningham, DNSc Karen Feldt, PhD Margaret Hartig, PhD Michael Huffman, PhD


Cancer, Dementia, End of Life, Nursing Home, Pain, Pain Assessment


Aims: We sought to identify differences in pain management between two groups; nursing home residents with malignant cancer and dementia with and without hospice services.

Methods: Decedent records from 2003-2009 were assessed for diagnosis of dementia and cause of death as cancer. Ten malignant cancer diagnoses were determined a priori from the CDC 2004 data on the top 10 malignant cancers for all races and genders. Fifty-five decedents from 10 nursing homes were included in the final sample. Four instruments were used: Minimum Data Set (MDS) a standardized assessment tool required of most U.S. nursing homes. A large comprehensive assessment is conducted yearly followed by smaller quarterly assessments. The MDS collects demographic and diagnostic variables, as well as clinical, functional, psychosocial, and cognitive assessments. Cognitive Performance Scale (CPS scored from 1 borderline-intact to 6 very severe impairment); Discomfort Behavior Scale (DBS scored from 0 no discomfort behavior identified to 102 maximum identifiable discomfort behavior); and Equivalent Dose Units (EDU’s) of opioid analgesic calculated and totaled over the last 2 weeks of life. We calculated the CPS score from the admission MDS because we believed cognitive levels were unlikely to improve over time. DBS scores were calculated from the last MDS prior to death in an effort to capture active cancer pain at the end-of-life. We realized the last MDS assessment might have been in the previous 90 days. The study received exempt status from the office of human protection.

Results: Total EDU’s were significantly greater among hospice enrollees (U 226.5, p <.05). There is a significantly greater likelihood of being prescribed a scheduled narcotic analgesic (OR 5.5; 95% CI 1.8-18.8) and a PRN narcotic analgesic (OR 3.6; 95% CI 1.2-11.3) when enrolled in hospice. Nursing home residents not enrolled in hospice had a significantly (U 195.0, p < .01) lower CPS scores than those enrolled in hospice. Decedents with lower cognitive levels were more likely (OR 4.9; 95% CI 1.6-15.6) to have a DBS score of zero. Forty percent of decedents with metastatic cancer and dementia received no opioid during the last 2 weeks of life.

Conclusions: We have demonstrated that pain among nursing home residents with dementia and cancer is a serious problem. Forty percent of our decedents received no opioid at the end of life. None (n = 12) of the subjects identified in the lowest CPS score (6) category were enrolled in hospice. One reason appears to be blunted pain related behaviors. Among the severely cognitively impaired whose pain behaviors become blunted, scheduled pain medications may be the best way to manage their pain. Pain behavioral tools are better suited for individuals with mild and moderate cognitive impairment but become less reliable in the severely cognitively impaired. New methods for assessing pain in this highly vulnerable population are needed.