Quality-adjusted survival (QALYs and Q-TWiST)
Quality-adjusted survival counts time alive but discounts the months spent with severe side effects or progressing disease, so a treatment that adds three months of misery scores lower than one that adds three good months.
Overview
Overall survival treats every month alive as equal. Quality-adjusted measures weight each period of time by how good it was. The quality-adjusted life-year, used by health technology assessors such as NICE, multiplies time by a utility between 0 (dead) and 1 (full health), usually derived from a questionnaire such as the EQ-5D completed during the trial; a year with a utility of 0.7 is 0.7 of a QALY. Q-TWiST, developed for cancer trials, partitions each patient's survival into time with treatment toxicity, time without symptoms or toxicity, and time after relapse or progression, then weights the first and third states and adds them to the second. Both approaches let a trial show whether a survival gain was made up of good time or bad, and both feed directly into cost-effectiveness, where the price of a treatment is expressed per QALY gained.
QUARTZ is the corpus example of a trial that used quality-adjusted survival as its primary endpoint. It randomised 538 patients with lung cancer and brain metastases unsuitable for surgery or radiosurgery to supportive care with or without whole-brain radiotherapy and found no meaningful difference in quality-adjusted life-years (46.4 against 41.7 days) or in survival, a result that restrained routine use of the treatment. TheraP showed the other use of the idea: the radioligand gave better quality of life and fewer side effects than cabazitaxel with similar survival, a benefit invisible to a survival-only comparison. MARS 2 found that a major mesothelioma operation gave worse survival and poorer quality of life, so the quality-adjusted verdict was doubly negative. CheckMate 9ER and the CE.6 glioblastoma trial reported quality of life preserved or similar alongside survival gains, which is the pattern a good treatment should show.
The weights are the weak point. Utilities differ by who is asked (the public tends to rate health states lower than patients living in them), by questionnaire and by country, and small changes in the weights can flip a cost-effectiveness verdict. Q-TWiST thresholds for a clinically important gain are conventions, not laws. Quality-adjusted results should sit next to, not replace, the raw survival and the patient-reported outcomes they were built from.
Similar pages
not linked directly; found by shared links- TermReading a hazard ratio
Shares Median survival, Absolute versus relative benefit (number needed to treat), Endpoint, Overall survival (OS).
- TermKaplan-Meier curve, censoring and proportional hazards
Shares Median survival, Landmark and milestone survival (5-year survival, median follow-up), Absolute versus relative benefit (number needed to treat), Phase 1, 2 and 3 trials.
- TermConfidence interval
Shares Median survival, Absolute versus relative benefit (number needed to treat), Endpoint.
- TermNICE technology appraisal methods (2022)
Shares Financial toxicity, Health technology assessment (HTA), reimbursement and QALYs, Quality of life and patient-reported outcomes (QoL, PRO).
- TermHazard ratio (HR)
Shares Median survival, Landmark and milestone survival (5-year survival, median follow-up), Absolute versus relative benefit (number needed to treat), Endpoint.
- TermTime to progression (TTP) and time to next treatment (TTNT)
Shares Quality of life and patient-reported outcomes (QoL, PRO), Overall survival (OS).
- TermData maturity (immature vs mature survival data)
Shares Landmark and milestone survival (5-year survival, median follow-up), Overall survival (OS).
- TermState biomarker testing coverage laws
Shares Financial toxicity, Health technology assessment (HTA), reimbursement and QALYs.