As the rate of spread of invasive species increases, consumer-resource communities are often populated by a combination of exotic and native species at all trophic levels. In parasitoid-host communities, these novel associations may lead to disconnects between parasitoid preference and performance, and parasitoid oviposition may result in death of the parasitoid offspring, death of the host, or death of both. Despite their relevance for biological control risk and efficacy assessments, the direct and indirect population-level consequences of parasitoids attacking and killing their hosts without successfully reproducing (non-reproductive mortality) are poorly understood. Non-reproductive mortality induced by egg parasitoids (parasitoid-induced host egg abortion) may be particularly important for understanding the population dynamics of the invasive agricultural pest Halyomorpha halys (Hemiptera: Pentatomidae) and endemic stink bugs in North America, which are attacked by a suite of both native and introduced egg parasitoids. It is unclear, however, how various factors controlling parasitoid foraging and developmental success manifest at the population level. We constructed two related versions of a two-host-one-parasitoid model to evaluate the population-level consequences of non-reproductive host mortality. Egg abortion can result in strong negative or positive enemy-mediated indirect effects, taking the form of apparent competition, apparent parasitism, apparent amensalism, or apparent commensalism. For parasitoids limited in their reproductive output by the number of eggs they can produce, higher non-reproductive host mortality can reduce the strength of the positive indirect effect in cases of apparent parasitism, and it can reduce the negative indirect effect on the more suitable host in cases of apparent competition. For time-limited parasitoids, unsuitable hosts with high levels of non-reproductive parasitoid-induced mortality can be strongly suppressed in the presence of a suitable host, while the suitable host is only negligibly impacted (i.e., apparent amensalism). We evaluate these model-derived hypotheses within the context of H. halys and its native and nonnative parasitoids in North America, and discuss their application to risk assessment in biological control programs.