Date Published: July 3, 2019
Publisher: Public Library of Science
Author(s): Dafina Petrova, Guiliana Mas, Gorka Navarrete, Tania Tello Rodriguez, Pedro J. Ortiz, Rocio Garcia-Retamero, Valerio Capraro.
We investigated what factors may foster or hinder physicians’ cancer screening risk literacy–specifically the ability to understand evidence regarding screening effectiveness and make evidence-based recommendations to patients. In an experiment, physicians in training (interns and residents) read statistical information about outcomes from screening for cancer, and had to decide whether to recommend it to a patient. We manipulated the effectiveness of the screening (effective vs. ineffective at reducing mortality) and the demand of the patient to get screened (demand vs. no demand). We assessed participants’ comprehension of the presented evidence and recommendation to the patient, as well as a-priori screening beliefs (e.g., that screening is always a good choice), numeracy, science literacy, knowledge of screening statistics, statistical education, and demographics. Stronger positive a-priori screening beliefs, lower knowledge of screening statistics, and lower numeracy were related to worse comprehension of the evidence. Physicians recommended against the ineffective screening but only if they showed good comprehension of the evidence. Physicians’ recommendations were further based on the perceived benefits from screening but not on perceived harms, nor the patient’s demands. The current study demonstrates that comprehension of cancer screening statistics and the ability to infer the potential benefits for patients are essential for evidence-based recommendations. However, strong beliefs in favor of screening fostered by promotion campaigns may influence how physicians evaluate evidence about specific screenings. Fostering physician numeracy skills could help counteract such biases and provide evidence-based recommendations to patients.
Many decisions about health involve the consideration of complex numerical information about risks and benefits. On such occasions, medical professionals are expected to be risk literate decision makers and advisors to their patients [1–3]. Risk literacy broadly refers to one’s practical ability to evaluate and understand risk in the context of informed decision making: for instance, to understand the benefits and harms of available treatments and to be able to make informed, value-consistent decisions based on the information at hand . Risk literacy is closely related to statistical literacy, which refers to physicians’ ability to understand the terminology and statistical aspects associated with the design, analysis, and conclusions of original research [1,5]. Thus, when physicians are required to make recommendations to patients based on research evidence, some statistical literacy would be required to understand the relevant evidence.
About half of the participants (N = 89, 51%) reported completing a methods and/or statistics course and only 12 (7%) reported having published a scientific article in an indexed journal. Table 2 shows that, on average, participants had strong positive beliefs about screening, answered about 3 out of 7 numeracy questions correctly, 2 out of 3 science literacy questions, and 2 out of 4 knowledge of screening statistics questions. The comprehension of evidence assessment achieved good discriminability following a normal distribution. Table 1 shows the percentages of correct responses to the individual items.
This study demonstrated the importance of understanding evidence about screening effectiveness for preventing biased and misleading physician recommendations. It also identified (a) a-priori positive beliefs about screening as markers of low cancer screening risk literacy, (b) physician numeracy as a specific skill that can foster comprehension and help counteract biases, and (c) knowledge of screening statistics as specific knowledge required for the correct evaluation of screening effectiveness.
Despite a rich literature on patient risk literacy, not many studies have addressed what influences physicians’ risk literacy. Given the multiple nuances and challenges of doctor-patient communication, research on physician risk literacy beyond artificial scenarios and in actual interaction with patients is needed. The current results demonstrate that in the context of cancer screening, a-priori positive beliefs about the goodness and desirability of screening, likely reinforced by multiple screening campaigns, and low physician numeracy can be important precursors of low physician risk literacy and biased, misleading recommendations.