Assessing health-related quality of life in palliative care: Comparing patient and physician assessments Morten Aa. Petersen a, *, Henrik Larsen a , Lise Pedersen a , Nan Sonne a , Mogens Groenvold a,b a The Research Unit, Department of Palliative Medicine, Bispebjerg Hospital, Bispebjerg bakke 23, Opgang 20d, DK-2400 Copenhagen NV, Denmark b Institute of Public Health, University of Copenhagen, Copenhagen, Denmark ARTICLE INFO Article history: Received 16 December 2005 Received in revised form 4 January 2006 Accepted 6 January 2006 Available online 18 April 2006 Keywords: Agreement Questionnaires Palliative care Patients Physicians Proxy Quality of life ABSTRACT It is often difficult to recruit patients for palliative care studies and severe attrition must be expected resulting in biased findings. This may be avoided if equivalent information could be obtained from sources other than the patients. Therefore, we investigated whether phy- sician assessments can be used to evaluate the patients’ health-related quality of life (HRQOL). Patient and physician assessments of the patients’ HRQOL were obtained once a week for up to 13 weeks using EORTC QLQ-C30 items. The agreement between patients and physicians at first contact (N = 115) and for the following 13 weeks combined (total N = 263) was investigated. Significant differences between patient and physician assess- ments were observed for all HRQOL domains assessed. Physicians reported patients to have fewer problems/symptoms than patients did for all HRQOL domains except for physical and social functioning. The agreement between patients and physicians was poor. Using physician assessments may bias findings and cannot be recommended as a substitute for patient self-assessment in palliative care. Ó 2006 Elsevier Ltd. All rights reserved. 1. Introduction Palliative care aims at providing the best possible care to im- prove the quality of life of patients and their families. 1 To achieve this goal, it is necessary to evaluate the effect of the care. However, it is often difficult to recruit patients for palli- ative care studies and severe attrition due to deterioration must be expected. 2,3 This biases results and limits the broad applicability of findings. 4 If equivalent information could be obtained from other sources than the patients, this would be preferable in descriptive studies in palliative settings. This could potentially improve the generalizability and the range of descriptive studies ethically and practically feasible in pal- liative care. Such alternative sources or ‘proxies’ could be physicians, nurses, or significant others. There is general agreement that the patient is the most appropriate source of information regarding his/her quality of life. Therefore, proxy assessments should only be used if in agreement with the patients’ own ratings. Several studies have compared patient and proxy ratings in general. A review from 1992 concluded that the concordance between patient and proxy ratings was far from optimal. 5 The ‘up-date’ from 2002 was more positive, concluding that proxy ratings are reasonably accurate and substantial discrepancies occur in a minority of cases only. 6 However, it was noted that proxies tended to overestimate the patients’ problems and symptoms. 6 A number of studies have compared the re- sponses of terminally ill/palliative care patients with proxy responses 7–16 (see Ref. [17] for a review). Except for one, these studies were small (less than 50 participants 8,9,13,15 ) or of 0959-8049/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.ejca.2006.01.032 * Corresponding author: Tel.: +45 3531 2025; fax: +45 3531 2071. E-mail address: map01@bbh.hosp.dk (M.Aa. Petersen). EUROPEAN JOURNAL OF CANCER 42 (2006) 1159 – 1166 available at www.sciencedirect.com journal homepage: www.ejconline.com