The Effect of Risk Information Communicated in Plain Language Combined with Natural Frequencies on Individual Risk Perception of Medication Side Effects

Objectives: The primary objective of this study was to determine whether describing the probability of experiencing a particular medication side effect in plain language and natural frequencies would result in a difference in the perceived risk of the medication, perceived harm of the side effects and intent to remain adherent as compared to describing side effect risks in plain language only. The secondary objective was to determine the correlation with perceived risk and harm of side effects and intent to remain adherent. Methods: The study design was an experimental two by three factorial survey. The effects of three independent variables (communication type [plain language vs. combined], severity of side effects [high vs. low], and frequency of side effects [high vs. low]) were measured across the dependent variables of perceived risk, perceived harm and intent to remain adherent. Four drug facts boxes were shown to participants, who were then asked to rate their perceived risk of experiencing the side effects, the perceived harm of the side effects and their intent to remain adherent to the medication on visual analog scales. Each drug facts box listed uses of the medication and one specific side effect. The probability of experiencing the side effect was described in either plain language or a combination of plain language and the natural frequency (i.e., 5 out of 100 people). Multiple analysis of variance (MANOVA) were conducted for the model and significant effects were analyzed with a poc-host Tukey test. Means were significant at p <0.05. Results: 100 students at the University of Houston College of Pharmacy completed the survey. MANOVA showed there were no significant effects for communication type (plain language or combination) alone (F3,390 = 1.97, p = 0.118). However, the model of communication type, severity and frequency was significant (F3,390 = 4.33, p = 0.005). The overall effects of communication type, severity and frequency had a significant effect on the dependent variable perceived risk (F7,392 = 190, p < 0.0001). The effect of communication type on perceived risk was significantly different (F1,392 =5.5, p = 0.0195). There was also a difference between severity (F1,392 = 17.87, p < 0.0001) and frequency (F1,392 = 1225, p < 0.0001). Tukey post-hoc analysis indicated that the combination communication type group had significantly higher risk perception scores (M = 46.6, 95% CI [44.2, 49.1]), compared to the risk perception scores for the plain language group (M = 42.5, 95% CI [40.1, 44.9]), p = 0.019). The overall effect of communication type, severity and frequency on intent to adhere to therapy was significant (F7,392 = 2.6, p = 0.01). There was no significant difference between communication type on the intent to remain adherent (F1,392 = 0.07, p = 0.786). There was a significant difference on intent to remain adherent for severity (F1,392 = 6.84, p = 0.009) and frequency (F1,392 = 8.2, p = 0.004). Mean scores for intent to adhere in the combined communication type group were slightly higher (M = 85.5, 95% CI [82.9, 88.0]) compared to the plain language group (M = 84.9, 95% CI [82.4, 87.5]), but were not statistically significant (p = 0.786). Mean scores in the high severity group were significantly lower (M = 82.8, 95% CI [80.3, 85.4], p < 0.009) than mean scores for the low severity group (M = 87.6, 95% CI [85.1, 90.2]). The mean scores for the high frequency group were also significantly lower (M = 82.6, 95% CI [80.0, 85.2], p = 0.004) compared to means for the low frequency group (M = 87.9, 95% CI [85.3, 90.4]). The overall model of communication type, severity and frequency and the effect on perceived harm was significant (F7,392= 91.46, p < 0.0001). There was no significant effect of communication type on perceived harm (F1,392 = 0.2, p = 0.659). There was a negative correlation between perceived risk and intent to remain adherent (r = -0.189, n = 400, p = < 0.001). A spearman product-moment correlation coefficient was computed to assess the relationship between the variables perceived harm and intent to remain adherent. There was a negative correlation between perceived harm and intent to remain adherent (r = -0.261, n = 400, p = < 0.001). Conclusions: Results of this study suggest that communication style does play a role in risk perception of medication side effects. However, as side effects become more severe and occur more often, this effect seems to diminish. There was no effect of communication type on participants intent to remain adherent. The participants were all students at various stages in pharmacy education and so are educated on the importance of adherence. It is striking to see that the known importance of adherence is shared amongst students in pharmacy school. Further studies in the general population may be warranted to test the effects of communication style on adherence.