The CONE-LCA framework is a promising framework that can be used to assess the impact of food items on human health and the environment. However, the nutrition-health assessment in the CONE-LCA framework quantifies the health impacts of food at the level of food items with the use of an indicator that is based on only few nutrients and otherwise broad food groups. Given that food items within food groups can have different impacts on health, disaggregating health impacts can be of added value and was explored here. An extended version of the Nutrients Rich Food index is used for both downscaling and cluster analysis. The correlation coefficient at the aggregated levels with the NRF index (-0.93) was substantially higher than the original health impact indicator of food items with the NRF index (0.36). These results illustrate that the use of a downscaled indicator improves the health impact assessment of food items. The downscaling analysis showed that downscaled health impacts can differ substantially between food items within food groups. Furthermore, the cluster analysis showed that some food groups (e.g. fruits, vegetables and red meat) are highly variable in nutrient density and therefore prioritizing these food groups for clustering can be of added value in future epidemiological studies. The improved nutrition health assessment model is not only relevant within the CONE-LCA framework but can be used on its own to assess health impacts of food items. The model may be very useful in policy applications as it enables comparison of food items on impact on human health and the environment, which can be the basis for sustainable dietary guidelines. Future research can focus on further development and validation of the model.