measures of quality of life are inferior to “hard,” clinical
indicators, such as histology and weight loss. However,
as the sophistication of quality of life instruments has
improved over time, so has the ability to provide informa-
tion of real clinical importance. In fact, patient’s ratings of
their quality of life before treatment have been found to be
predictive of length of survival—in some cases with even
more predictive power than traditional clinical indicators.
There are 22 studies in which patient’s ratings of their
quality of life before treatment began (at baseline) were
found to be predictive of survival.
These findings also were not limited to a single type of cancer, but were found
across 9 types (lung, breast, esophagus, head and neck,
bladder, cervical, pancreatic, colorectal, and prostate can-
cer), which provides greater confidence in the findings.
The quality of life scores that were predictive of survival
were global quality of life, symptoms (pain, fatigue,
anorexia, dyspnea), and functioning (physical, cognitive,
emotional, social, and role function). Satisfaction with life
was even predictive of survival in four studies. Because
the studies used different instruments to measure quality
of life, and patients received a wide variety of treatment
regimens, they increase our confidence that a relationship
between pretreatment quality of life and survival actually
exists.
In some cases, the quality of life scores were even more
powerful predictors than standard clinical indicators. For
example, in inoperable nonsmall cell lung cancer treated
by radiotherapy, Langendijk et al found that global quality
of life was the strongest prognostic factor of all; stronger
than node classification, weight loss, or performance status.
Additional clinical variables among the studies that were
found to be less predictive than quality of life scores were
tumor stage, disease measurability, age, gender, time since
diagnosis, tumor location, number of tumor sites, tumor
type, comorbidity, prior radiotherapy, hemoglobin level,
bone metastasis, estrogen and progesterone receptor status,
and disease-free interval.
On the basis of the findings such as these, Ganz et al6
recommended that patient-rated quality of life assessment
should be obtained as an integral part of cancer manage-
ment to serve as a guide to patient needs, as an outcome
measure, and as a prognostic variable for survival time.
Osoba has even suggested that pretreatment quality of life
ratings could be used as an eligibility criterion and stratifi-
cation variable in clinical trials.