Inaccurate cancer news can have adverse effects on patients and families. One potential way to minimize this is through media literacy training-ideally, training tailored specifically to the evaluation of health-related media coverage. We test whether an abbreviated health-focused media literacy intervention improves accuracy discernment or sharing discernment for cancer news headlines and also examine how these outcomes compare to the effects of a generic media literacy intervention. We employ a survey experiment conducted using a nationally representative sample of Americans (N = 1,200). Respondents were assigned to either a health-focused media literacy intervention, a previously tested generic media literacy intervention, or the control. They were also randomly assigned to rate either perceived accuracy of headlines or sharing intentions. Intervention effects on accurate and inaccurate headline ratings were tested using OLS regressions at the item-response level, with standard errors clustered on the respondent and with headline fixed effects. We find that the health-focused media literacy intervention increased skepticism of both inaccurate (a 5.6% decrease in endorsement, 95% CI [0.1%, 10.7%]) and accurate (a 7.6% decrease, 95% CI [2.4%, 12.8%]) news headlines, and accordingly did not improve discernment between the two. The health-focused media literacy intervention also did not significantly improve sharing discernment. Meanwhile, the generic media literacy intervention had little effect on perceived accuracy outcomes, but did significantly improve sharing discernment. These results suggest further intervention development and refinement are needed before scaling up similarly targeted health information literacy tools, particularly focusing on building trust in legitimate sources and accurate content.