The tumour suppressor gene encoding p53 has been shown from experimental studies to have a crucial role in how cells respond to DNA damage. p53 has important functions in apoptosis, cell-cycle arrest and DNA repair, largely mediated by its activity on gene transcription. However, despite this wealth of in vitro data, its role in how tumours respond to DNA damage induced by chemotherapeutic drugs remains controversial. In this review, we highlight some of the problems surrounding design and analysis of studies of p53 as a prognostic marker of clinical outcome, using ovarian cancer as an example. We aim to build on the knowledge of the published literature in ovarian cancer to identify criteria for clinical studies that should give a more definitive estimate of the role of p53 in clinical drug resistance. A search of three public databases using keywords combined with Boolean operators identified 64 clinical publications investigating the relationship of p53 to clinical outcome following chemotherapy in ovarian cancer. Although 43% of 215 published analyses from the 64 papers reported a significant correlation between p53 status and a clinical endpoint relevant to chemoresistance, only six analyses fulfil minimum criteria and none of these finds a statistically significant correlation of p53 with chemotherapy-resistance endpoints. The results from published clinical studies suggest a more complex role of p53 mutation in the mechanism of resistance in ovarian cancer than is suggested by in vitro studies.