We applied a decision tool for rationing choices, with a predetermined budget of about 11 US dollars per household per year, to identify priorities of poor people regarding health insurance benefits in India in late 2005. A total of 302 individuals, organized in 24 groups, participated from a number of villages and neighborhoods of towns in Karnataka and Maharashtra. Many individuals were illiterate, innumerate and without insurance experience. Involving clients in insurance package design is based on an implied assumption that people can make judicious rationing decisions. Judiciousness was assessed by examining the association between the frequency of choosing a package and its perceived effectiveness. Perceived effectiveness was evaluated by comparing respondents' choices to the costs registered in 2049 illness episodes among a comparable cohort, using three criteria: 'reimbursement' (reimbursement regardless of the absolute level of expenditure), 'fairness' (higher reimbursement rate for higher expenses) and 'catastrophic coverage' (insurance for catastrophic exposure). The most frequently chosen packages scored highly on all three criteria; thus, rationing choices were confirmed as judicious. Fully 88.4% of the respondents selected at least three of the following benefits: outpatient, inpatient, drugs and tests, with a clear preference to cover high aggregate costs regardless of their probability. The results show that involving prospective clients in benefit package design can be done without compromising the judiciousness of rationing choices, even with people who have low education, low-income and no previous experience in similar exercises.