Many patients with advanced cancer experience symptoms and problems, and some patients will be admitted to specialized palliative care (hospice or palliative care teams/units in hospitals) for symptom relief in the last part of their lives (often the last weeks to months). Previous studies of cancer patients at the start of specialized palliative care (SPC) have typically been small and in a single specialized palliative care unit. This PhD thesis studied symptoms, problems and overall QOL in cancer patients at the start of SPC at the population level, using nationwide data from all SPC units in Denmark aggregated in the Danish Palliative Care Database (DPD).
About 50% of the patients admitted to SPC filled in the EORTC QLQ-C15-PAL questionnaire at the start of SPC. The aim of Paper 1 was to test whether symptom/problem-scores differed between SPC units with the highest and lowest response rates, indicating a possible selection bias in SPC units with low response rates. Paper 2 investigated the levels of symptoms and problems in patients at the start of SPC and how gender, age and cancer diagnosis were associated with the levels of symptoms/problems. Paper 3 investigated whether the levels of symptoms/problems differed for patients referred by general practitioners and by physicians at hospitals. Lastly, Paper 4 studied if symptoms/problems were associated with survival time from start of SPC and whether taking symptoms/problems into account could improve one-week and one-month survival prediction.
Cancer patients admitted to SPC who died between 2010-2017 and who reported their symptoms and problems at the start of SPC using the EORTC QLQ-C15-PAL questionnaire.
In Paper 1, multiple linear regression analysis was used to compare symptom/problem-scores for the SPC units with the highest response rates to scores for SPC units with lower response rates. In Papers 2-3, ordinal logistic regression was used to study the associations between symptom/problem-scores and gender, age and cancer diagnoses, respectively (Paper 2), or referral sector (Paper 3). In Paper 4, Cox regression was used to study the associations between symptoms/problems and survival time; furthermore, logistic regression analyses with one-week and one-month survival as outcomes were used to test the predictive value of clinical variables and symptoms/problems.
In Paper 1, no clinically relevant differences were found in symptom/problem-scores between SPC units with the highest and lower response rates and therefore there was no indication of selection bias. Paper 2 found severe levels of symptoms/problems in cancer patients at the start of SPC. The most severe symptoms/problems were pain, appetite-loss, fatigue, poor physical function and poor QOL. The strongest associations between symptoms/problems and gender and age, respectively, were the increased risk of nausea in women as well as the increased risk of poor physical function and reduced risk of sleeplessness and pain with increasing age. Cancer diagnosis was significantly associated with all symptoms/problems. Paper 3 found that associations with referral sector were generally small and thus probably not clinically relevant. Paper 4 showed that physical function was the symptom/problem with the strongest association to survival time. Symptoms/problems (especially physical function) improved the overall accuracy for one-week- and one-month survival prediction.
Conclusions: There was no indication that the scores for symptoms and problems from SPC in units with low response were affected by selection bias. The cancer patients admitted to SPC were severely troubled by symptoms and problems. Gender, age and cancer diagnosis were significantly associated with most symptoms and problems, but the strength and direction of the association were dependent on the symptom/problem. Similar levels of symptoms/problems were found for patients referred by general practitioners and physicians at the hospitals. Most symptoms/problems were associated with survival time and inclusion of symptoms/problems improved prediction of one-week and one-month survival.
Read the thesis here, MB Hansen, Symptoms and Problems in Cancer Patients Admitted to Specialized Palliative Care MB_Hansen_PhD.pdf