Here is a recent paper on lionfish prey compiled at a regional scale. The Abstract is pretty straight forward, so I will just paste that below. I think the key sentence is the last. Since lionfish are such generalist predators, their trophic role will vary substantially among regions, islands, ecosystems and individual habitats. My opinion remains the dire predictions of their impacts are exaggerated, but it is unquestionable they are now integral components of Caribbean food webs.
Numerous location-based diet studies have been published describing different aspects of invasive lionfish (Pterois volitans and Pterois miles) feeding ecology, but there has been no synthesis of their diet composition and feeding patterns across regional gradients. 8125 lionfish stomachs collected from 10 locations. Our regional data indicate lionfish in the western Atlantic are opportunistic generalist carnivores that consume at least 167 vertebrate and invertebrate prey species across multiple trophic guilds, and carnivorous fish and shrimp prey that are not managed fishery species and not considered at risk of extinction by the International Union for Conservation of Nature disproportionately dominate their diet. Correlations between lionfish size and their diet composition indicate lionfish in the western Atlantic were analyzed to provide a generalized description of their feeding ecology at a regional scale and to compare their diet among transition from a shrimp-dominated diet to a fishdominated diet through ontogeny. Lionfish total length (TL) (mm) was found to predict mean prey mass per stomach (g) by the following equation mean prey mass =0.0002*TL1.6391, which can be used to estimate prey biomass consumption from lionfish lengthfrequency data. Our locational comparisons indicate lionfish diet varies considerably among locations, even at the group (e.g., crab) and trophic guild levels. The Modified Index of Relative Importance developed specifically for this study, calculated as the frequency of prey a 9 the number of prey a, can be used in other diet studies to assess prey importance when prey mass data are not available. Researchers and managers can use the diet data presented in this study to make inference about lionfish feeding ecology in areas where their diet has yet to be described. These data can be used to guide research and monitoring efforts, and can be used in modeling exercises to simulate the potential effects of lionfish on marine food webs. Given the large variability in lionfish diet composition among locations, this study highlights the importance of continued location-based diet assessments to better inform local management activities.