When monitoring the quality of cancer care, the severity of symptoms is an important factor to consider. But assessments often rely on what the symptoms are like at one time point, such as the baseline interview at diagnosis or hospice admission, even though symptom severity changes rapidly in cancer patients, especially as the patient approaches death, say researchers from University of Toronto in Canada. They suggest that using repeated assessments at varying time points could both help improve care and predict the risk of death more closely.
They designed a longitudinal study of 66,112 outpatients, with > 310,000 assessments of symptoms. Using patient symptom reports from the Edmonton Symptom Assessment System (ESAS), the researchers compared a model using a time-varying covariate for each symptom with one that used only a time-fixed (baseline) covariate for each symptom. The median follow-up period was about 19 months, with a median of 3 assessments per patient and a median of 1.16 months between each assessment. Each assessment covered 9 symptom scores: fatigue, appetite, well-being, drowsiness, pain, shortness of breath, anxiety, depression, and nausea.
The repeated assessments improved predictions for risk of death. The strongest predictors were increased pain, fatigue, and reduced appetite. As symptoms became more severe, the hazard of death increased significantly. For example, at any given time, a patient with a severe pain score had about 2 times the risk of death compared with a patient with no pain. Similarly, among patients with breast cancer, a woman with the worst appetite score had 3.66 times the risk of death compared with a woman with the best appetite score.
Though shortness of breath was a predictor among women with breast cancer, the researchers found no associations between shortness of breath and risk of death in patients with lung cancer. (This could be because shortness of breath is well managed in these patients, the researchers note.) They also found no significant associations between depression and death, overall, or by any of the cancer types.
The researchers encourage recording measurements on symptom scores over time, at each visit, using paper-based visual scales or electronic surveys. The information about changing symptoms could prompt interventions such as exercise for fatigue and nutritional support for anorexia. On the other hand, worsening pain, fatigue, and appetite may be unavoidable, the researchers say, and a “flag for impending death”—in which case, a more timely assessment might sooner identify patients in need of palliative care.
Sutradhar R, Atzema C, Seow H, Earle C, Porter J, Barbera L. J Pain Symptom Manage. 2014;48(6):1041-1049.