We all know that physician-patient communication is important. Poor communication, not deficient clinical skill, is the primary driver of malpractice suits; conversely, good clinical communication alone can improve patient outcomes. For example, patients of physicians with superior communication skills have been shown to have superior blood pressure control. Fortunately, physicians generally recognize the significance of good communication, and often take the time to ensure that it happens. Yet there remains a major challenge that these efforts do not address — that of managing numeracy, the ability of patients to make sense of numbers.
When we ask patients to participate in decision-making, we’re relying on their ability to understand and manipulate numerical data. A patient might be called on to choose hospitals and physicians based on quality of care metrics, to evaluate the results of screening tests using an intuitive grasp of sensitivity and specificity, or evaluate probabilistic risks and benefits of different treatments. All of this presumes both clear communication of statistical data as well as a level of patient sophistication – two assumptions that are often in fact, untrue.
For example, the NIH recommends that women aged 40 to 49 discuss the need for mammography screening with their doctors. This implies that patients must necessarily be comfortable with quantitatively assessing the risks and benefits – and that physicians must have the time and expertise to both explain the data and evaluate patients’ understanding, stepping in and guiding their thinking as necessary. But in fact, many patients, particularly less numerate ones, were not able to accurately understand the risks and benefits of mammography when presented with a simple brochure.
So it’s not surprising that numeracy, even more so than general health literacy, is a strong predictor of comprehension and decision-making in making comparative health care decisions. According to the National Adult Literacy Survey, almost half the general population has difficulty with basic numeric tasks, such as determining the difference between a regular price and a sales price. Even among educated people, one study found that 16% incorrectly answered straightforward questions about risk magnitudes (one question read: “what represents the larger risk: 1%, 5%, or 10%”).
Given both limited training and limited time to spend in each patient encounter, physicians aren’t always able to compensate for these difficulties. When dealing with patients from lower socioeconomic classes, physicians tend towards a less participatory interaction style, and in particular tend to give significantly less information and engaging less in partnership-building. Unsurprisingly, such patients reported less satisfaction with the clarity and explanations of their condition and care. And this lack of understanding may contribute to the worse health outcomes experienced by these patients.
One promising solution is to convey quantitative information through intuitive graphical representations. One such example from the Mayo Clinic displays how many high-risk patients out of 100 will experience an MI with and without statins, visually demonstrating the effect of statins on reduction the risk of MI in this population. Such a figure can quickly convey a great deal of information that would otherwise be hard for innumerate patients to understand.
Visual aids exploit rapid, automatic visual perception skills to interpret data. A well-designed visual display can reduce the amount of cognition necessary to perceive the same statistical information, thus providing clearer information to patients with varying educational backgrounds without requiring lengthy and time-consuming explanations from providers.
It’s important to note some caveats. Poor framing of visualizations may cause patients to develop a biased interpretation, and educated patients still have an advantage in interpreting most graphs.
Nevertheless, much work has been done illustrating the beneficial effect of different visualizations, whether icon arrays, histograms, risk ladders, or pie charts, in communicating risk information. Furthermore, a few healthcare institutions have been spearheading the use of these visualizations in patient care. As these methods continue to be refined, and as physicians deploy them clinically, visual representations of quantitative data have great promise in both enhancing patient understanding and in prompting behavior change.
Peter Wei is a medical student.