Worldwide dietary data for nutrition monitoring and surveillance are commonly derived from food balance sheets (FBS) and household budget surveys (HBS). We have compared food supply from FBS and food availability data from HBS among eighteen European countries and have estimated the extent to which they correlate, focusing on food groups which are comparably captured by FBS and HBS and for which there is epidemiological evidence that they can have a noticeable impact on population mortality. Spearman's correlation coefficient was +0·78 (P < 10− 3) for vegetables (including legumes),+0·76 (P < 10− 3) for fruits, +0·69 (P < 10− 3) for fish and seafood and +0·93 (P < 10− 3) for olive oil. With respect to meat and meat products, the coefficient was lower at +0·39 (P = 0·08). Moreover, we have examined whether the supply (FBS) or the availability (HBS) of food groups known or presumed to have beneficial effect on the occurrence of CHD and total cancer can predict overall, coronary and cancer mortality in ecological analyses. After controlling for purchasing power parity-adjusted gross domestic product and tobacco smoking we found that for vegetables, fruits, fish and seafood, as well as for olive oil, both the FBS and the HBS estimates were inversely associated with all three indicators of mortality, although the number of countries with complete information on all study variables hindered formal statistical documentation (P>0·05 in some instances). FBS and HBS have their own strengths and weaknesses, but they may complement each other in dietary assessments at the population level.