Following a request from the European Commission, the EFSA Plant Health Panel analysed a dossier submitted by Costa Rica Authorities to reach a conclusion on the host status of Phoenix roebelenii for Xylella fastidiosa. The Panel wishes first to stress the difficulties faced in providing compelling evidence for the non-susceptibility status of any particular plant species. The Panel acknowledges that the listing of P. roebelenii as a host of X. fastidiosa rests on a single report from California. Because isolation of X. fastidiosa from some hosts can be difficult, the Panel considers that the failure to isolate X. fastidiosa from P. roebelenii cannot be used to totally discard the detection of X. fastidiosa by ELISA and PCR. The Panel concludes that the detection of X. fastidiosa by two independent techniques provides sufficient evidence, although not totally conclusive, for the listing of P. roebelenii as a X. fastidiosa host plant. Concerning the survey data provided in the Costa Rican dossier, the Panel wishes to stress that such surveys cannot demonstrate the non-host status but can only provide a probability bound, upper estimate of the proportion of infected plants in the field. In the present case, and assuming all survey parameters to be optimal, the 95% confidence incidence threshold obtained is 0.2%, leaving the possibility that close to 25,000 P. roebelenii plants could be infected but undetected in the country. Accepting a scenario of local, non-systemic infection of P. roebelenii by X. fastidiosa would further increase uncertainties. In addition, the absence of data on the vector infection pressure further affects the ability to derive meaningful information on the P. roebelenii host status from the survey data. Appropriately conducted mechanical and/or vector-mediated inoculation experiments are critical to reach a more solid conclusion on the X. fastidiosa host status of P. roebelenii.