OBJECTIVES: Cost-utility analysis assumes quality-adjusted life-years (QALYs) have equal value regardless of the characteristics of the patients gaining or losing them. However, this utilitarian view may not be in line with public opinion; eg. the public may prefer to fund therapies for patients with greater disease burden, or at the end of life. A number of studies have sought to measure the extent to which the public are willing to trade efficiency (the principle of QALY maximisation) in favour of prioritising patients, diseases or treatments with particular characteristics. We reviewed these studies with a view to developing a formula that could be used to adjust traditional QALYs to generate a value-weighted QALY. METHODS: We conducted a systematic review to identify primary research where members of the public quantified (directly or indirectly) the relative value of health gains depending on patient, disease or treatment characteristics. MEDLINE searches were supplemented by hand-searches of relevant congresses and reference lists of relevant reviews. Studies not presenting quantitative results were excluded. RESULTS: A number of studies using methods such as person trade-off and discrete choice experiments were identified. Overall, the evidence did not support the principle that all QALYs are of equal value. There was a preference for treating patients with more severe disease, but mixed opinions on whether patients with shorter or longer initial life expectancies should be favoured for treatment. Some studies examined broader factors such as patients’ socioeconomic status or culpability for their own disease. CONCLUSIONS: Published ratings of societal value weights for cost-value analysis (eg. Nord, E. 2015. Pharmacoeconomics; 33:89-95) are weighted for disease severity or proportional QALY shortfalls, but our review suggests that public preferences may support the consideration of a wide range of characteristics when determining priority setting for resource allocation. Further analyses are needed before quantitative weighting scales can be calculated.