Patients Prefer Pictures to Numbers to Express Cardiovascular Benefit From Treatment

INTRODUCTION

We are drowning in information while starving for wisdom.

—E.O. Wilson

Shared decision making is the centerpiece of the current model of patient-centered care. To realize this goal, patients need to be given, to understand, and to interpret information on their own personal risk¹ and on the benefits and harms of treatment. Risks and benefits can be expressed in numerous ways, such as odds ratios, number needed to treat (NNT), absolute or relative risk, and natural frequency statements; can be presented as either percentages or proportions; can be positively or negatively framed; and can be displayed in numeric or pictorial formats.

The way information is presented influences patients’ decisions.² Although some studies have examined the effects of manipulating risk and benefit information in clinical settings, given the range of ways the data can be presented and the range of relevant clinical settings, a literature review has demonstrated a relative paucity of evidence in this field.³ One study has indicated that patients are more likely to consider taking medication to prevent a heart attack if information on benefit is expressed in terms of the relative risk reduction compared with other methods of risk communication.⁴ A qualitative study of patients with familial hypercholesterolemia found that perceived risk may shift over time in response to social contexts and may be influenced by factors such as cardiac events or deaths in the family, illness experiences, or becoming a parent.⁵ Another qualitative study found that patients participating in focus groups did not think that knowing their risk numbers would necessarily motivate them to change their behavior.⁶

Physicians are also influenced by the way information is presented. One study found that physicians were more likely to prescribe a drug that reduces cardiac events if the trial results were expressed in terms of relative risk rather than absolute risk reduction, percentages of disease-free survival, or NNT⁷; however, presenting only relative risk information is misleading and potentially coercive.⁸ There may be tension, furthermore, between what health care professionals or payers see as best for the population and what individuals see as best for themselves. Although informing patients of their individual risk may increase their participation in screening programs, evidence is lacking regarding how to improve communication of risk to patients⁹ or patients’ informed decision making.¹⁰

In risk communication, it is therefore necessary to establish not only the format of expression most likely to influence patients’ behavior, but also the format they prefer in making informed decisions on their own management, either to adopt an intervention or to avoid unnecessary interventions.

The aim of this study was to determine which methods of expressing benefit would encourage patients with known cardiovascular disease to decide to take preventive medication, and which methods they prefer to help them understand the information.

METHODS

Our study sample was drawn from patients attending 4 family practices in Auckland, New Zealand, located in 1 low-income, 1 middle-income, and 2 high-income areas. We identified potential participants by searching the practices’ electronic medical records according to disease classification and drug prescriptions. Patients were eligible if they had heart disease (myocardial infarction, angina, or both) and were taking a statin. The latter requirement ensured that patients had experience with decision making regarding medications and with taking medication, and it provided a backdrop against which we could propose to them a hypothetical new medication.

To recruit patients, practices mailed the identified patients a participant information sheet, a copy of the study questionnaire, a consent form, and a postage-paid return envelope for the consent form. Patients who had not responded after 10 days were contacted by telephone and invited to participate.

Patients were not eligible to participate if they had any of the following exclusion criteria: (1) prior stroke, because we believed such patients would be more likely to take medication to prevent another event than patients who had only cardiovascular disease; (2) inability to communicate in English; or (3) failure to be contacted after 3 calls made at different times of the day and week.

One of the researchers (L.C.) collected data by telephone interviews, which lasted about 20 minutes. During the interviews, patients were asked questions from the questionnaire that had already been mailed to them, including demographic data (sex, age, ethnicity, formal education), their perceived risk of heart attack over the next 5 years, and their willingness to take a protective heart medication continuously for the rest of their lives, plus 2 additional questions to assess basic numeracy. This telephone approach was more efficient than requiring face-to-face meeting of the researcher with the participant.

The study questionnaire was developed using information from the literature with input from the researchers, and was modified after piloting in 1 clinic. The questionnaire informed patients of a hypothetical new medication that had few adverse effects and would be given to people who have had angina or a heart attack to prevent a future heart attack. For the sake of the study, the risk of a heart attack over the next 5 years was assumed to be 23% without the medication and 16% with the medication. These risk estimates were derived from a randomized trial evaluating the effect of cholesterol lowering with simvastatin on mortality and morbidity in patients with a history of myocardial infarction or angina.¹¹

Patients were presented with this same medication benefit in 8 different ways (Table 1⇓), consisting of 7 numerical formats (relative risk, absolute risk with positive framing, absolute risk with negative framing, NNT, odds ratio, detailed natural frequencies, and simplified natural frequencies) and 1 graphical format (Figure 1⇓). They were then asked (1) whether information presented by each of these methods would encourage them to take the medication daily, (2) which method they preferred for expressing the medication’s benefit, and (3) whether they preferred positively or negatively framed information. Patients were also asked whether they preferred that physicians give them numbers and risks or give them their own opinion.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Graphic expression of the risk of heart attack for 100 people over 5 years without and with new medication.

We entered the raw data into Microsoft Excel (Microsoft Corp, Redmond, Washington). Statistical tests were undertaken in Stata 9.0 (Stata Corp, College Station, Texas) using the χ² test, analysis of variance (ANOVA), and logistic regression analysis. The study was approved by the Auckland Ethics Committee.

RESULTS

We identified 272 patients as potential participants, of whom 84 were ineligible. Of the 188 eligible, invited patients, 100 participated (53% response rate); 60 were male and 40 were female. Seventy-seven were of New Zealand European ethnicity, 16 were of Maori ethnicity, and 7 were of other ethnicities. One-third had a tertiary education. Their ages ranged from 39 to 87 years, with an average age of 66 years.

Some 32 patients were from the practice in the low-income area, 33 were from the practice in the middle-income area, and 15 and 20 were from the 2 practices in high-income areas. Forty-seven percent of patients demonstrated basic numeracy skills. Patient ethnicity, education, and numeracy differed significantly across practices; however, when these 3 factors were entered into logistic regression models that estimated perceived risk of heart attack or preferred format for communicating benefit, adding practice to the model did not significantly alter the results. We therefore analyzed patients from all practices together. Eighty-one percent of patients were willing to take medication daily for the rest of their lives to protect their hearts.

Table 1⇑ shows that no matter which method was used to express the benefit of the new medication, the great majority of patients (67%–89%) indicated that they would be encouraged to take it.

Sixty-eight participants could decide which of the 8 methods they considered to be the best for expressing benefit (the others could not make this decision). Of these 68, 39 (57%) preferred the graph, which was significantly more (P <.001) than the 13 (19%) who chose the next most preferred option, relative risk. Only 9 (13%) preferred absolute risk (either positively or negatively framed), and 6 (9%) preferred natural frequencies (either simplified or detailed). Only 1 patient preferred the odds ratio (1%), and none preferred NNT.

Of those 69 patients who responded to the question on framing risk, 62 (90%) preferred positive framing, consistent with other research.^12,13

When asked whether they preferred physicians to give them numbers and risks or to give their opinion regarding treatment, about two-thirds of patients preferred the physician’s opinion, compared with about one-fifth who preferred being presented with numbers and risks. The remaining roughly one-fifth of patients had no preference or were undecided. Sex, ethnicity, age, tertiary education, and willingness to take protective heart medication continuously for the next 5 years had no significant effect on whether patients preferred numbers and risks or a physician’s opinion.

DISCUSSION