Background/Purpose: The Food and Drug Administration (FDA) requires that patients prescribed a bisphosphonate to treat osteoporosis receive an FDA-approved Medication Guide that warns of the risk of unusual thigh fracture. These Medication Guides state: “Some people have developed unusual fractures in their thigh bone. Symptoms of a fracture may include new or unusual pain in your hip, groin, or thigh”. However, no information is given concerning the probability of experiencing an unusual thigh fracture. We hypothesized that providing quantitative information concerning the probability of fracture would: (1) increase willingness to use the medication, (2) increase perceptions of medication safety, and (3) decrease perceptions of the likelihood of experiencing an unusual thigh fracture.

Methods: Participants (N=258) were recruited using Amazon Mechanical Turk. Using an internet-based survey design, each participant was randomly assigned to 1 of 3 groups that viewed a medication fact sheet with general information about a fictional bisphosphonate called, Bonemax. The fact sheet included information about medication benefits and the risk of atypical femoral fractures. For atypical femoral fractures, one format presented only non-numeric risk information, modeled after Medication Guides, while the other 2 formats presented numeric risk information for differing durations of therapy (0-2 years and 2+ years). The risk of atypical femur fracture was reported as 2 in 10,000 in the 0-2 year duration group and 8 in 10,000 in the 2+ year duration group. All 3 formats presented the same benefit information. The primary outcome variables were: 1) Likelihood of taking the medication if recommended by one’s physician; 2) Medication safety; and 3) Likelihood of causing an atypical femoral fracture. All responses were recorded on 7-point rating scales and statistical significance was set at alpha=0.05. Statistical analyses were performed using SAS^®9.4.

Results: All three hypotheses were supported. Participants in the non-numeric group reported being less likely to use the medication (Mean=3.15) compared to those in either of the other two groups (Means=4.14 and 3.96 in the 0-2 year and 2+ year duration groups respectively, both p’s < 0.01). Participants in the non-numeric group rated the medication as less safe (M=2.79) compared to those in either of the other two groups (Means=3.89 and 3.74 in the 0-2 year and 2+ year duration groups, both p’s < 0.001). Participants in the non-numeric group believed that the medication was more likely to cause an atypical femur fracture (Mean=3.20) compared to those in either of the other two groups (Means=1.71 and 1.56, respectively, both p’s < 0.001). None of the variables significantly differed between the numeric risk information groups.

Conclusion: Study findings suggest that the addition of numeric risk information to Medication Guides may facilitate patient decision making.

« Back to 2016 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/effect-of-quantitative-information-concerning-medication-side-effects-on-risk-perception/