The use of proxy raters of patients' quality of life has been suggested as a means of facilitating the factoring of quality-of-life considerations explicitly into the medical decision-making process and of resolving the problem of missing data in longitudinal quality-of-life investigations. This review addresses two questions related to the potential role of such proxy raters in clinical research and practice: (1) to what extent are health care providers and lay individuals involved in the care of patients ("significant others") able to assess accurately the quality of life of patients with chronic disease? and (2) under what conditions, if any, is inclusion of such proxy ratings in quality-of-life investigations warranted? Although the extant literature yields few unequivocal findings, a number of clear trends can be identified: (i) health care providers and significant others tend, in general, to underestimate patients' quality of life; (ii) health care providers and significant others appear to evaluate patients' quality of life with a comparable degree of (in)accuracy; (iii) health care providers tend to underrate the pain intensity of their patients; (iv) proxy ratings appear to be more accurate when the information sought is concrete and observable; and (v) while significant others' ratings tend to be more accurate when they live in close proximity to the patient, they can also be biased by the caregiving function of the rater. There is need for more methodologically sound studies that: (a) incorporate head-to-head comparisons of health care providers and significant others as proxy raters; (b) employ well-validated quality-of-life measures; and (3) employ a longitudinal design in order to examine the effect of changes in patients' health status over time on the ability of proxies to provide valid quality-of-life assessments.