The best magazine
Discussion
In this study, Marteau's model was used to determine whether the DA increased the number of informed choices about uptake of tamoxifene and raloxifene. In fact, our data revealed that women given a DA were more likely to have made an informed choice about uptake of tamoxifen and raloxifene than women who did not view the DA. Also, fewer women in the intervention group remained undecided about uptake of the drugs. As expected, three months after baseline the rate of informed decision making about actual uptake was still marginally higher in women who had received the DA, particularly because of declining knowledge levels. However, overall, levels of informed decision making are strikingly low.
Since most respondents had negative attitudes toward uptake of tamoxifen and raloxifene and most decided not to take chemoprevention drugs, uninformed choices were mainly related to insufficient knowledge. One potential reason for relatively low rates of knowledge in the group receiving the DA is that each drug had five risks and two benefits presented. The amount of information conveyed may have been overwhelming and reduced people's recall, even though we asked about gist-level information.
The rate of informed choices in the intervention group decreased from 54% (post-test) to 18% (three-month follow-up), but attitudes did not change from post-test to follow-up. The observed decline in informed choices is mainly due to decreased knowledge scores over time and measuring latent decision-making appears to be useful. This raises an interesting question - how long are decision-makers supposed to remember gist-level information once they have come to a decision?
Another issue is the important influence of the definition of what exactly constitutes 'sufficient decision-relevant knowledge' on the resulting proportion of informed choice. Consensus is needed on the levels of knowledge and specific content necessary to make informed decisions. A drawback of the complicated nature of the information being provided, and thus a limitation of this study, is that there were not enough participants with more than 50% knowledge retention to examine differences between groups.
One might expect that women who consider taking chemoprevention drugs are more likely to remember information because it is relevant to their decision, whereas women who have already decided not to take the drugs may be less motivated to remember the information. However, we did not find this effect.
Traditionally two kinds of informed choices are distinguished: sufficient decision-relevant knowledge combined with either positive attitudes and acceptance of the preventive measure or negative attitudes and a decline of the preventive measure. We identified a third type of informed choice, which is the combination of sufficient knowledge, neutral attitudes and being undecided about uptake of the preventive measure. The question is whether this third combination indeed reflects an informed choice rather than ambivalent attitudes, since -- as shown by earlier research -- ambivalence may lead to postponing decision making. We recommend further study into this topic.
It has often been shown that DAs lead to increased levels of knowledge in those who have used them. While many studies of DAs examine changes in knowledge, attitudes, intentions and behavior, few examine the relationship between knowledge and the concordance between attitudes, intentions, and behavior. Even for well-known population-based screening programs, such as breast or cervical cancer screening, interventions specifically aimed at improving informed decision-making are limited. This study, therefore, represents a valuable contribution to the literature. Our study has several limitations related to generalizability of the findings. First, despite oversampling eligible African American women, participants were mostly White. Other chemoprevention studies have also encountered this problem. Second, we used an online DA, which can ultimately increase intervention reach, but lack of Internet access or discomfort with using the Internet likely narrowed the pool of potential participants and possibly directed access to a more educated sample. However, findings of previous research comparing Internet samples to other samples are inconsistent in this regard. Third, we assessed attitudes with one item. Although this single item enabled the registration of negative attitudes considering own participation in the intervention - which is often not the case when attitudes are being assessed - we acknowledge that using one item is a bit limited.
While the above limitations decrease external validity of our study, including two control groups significantly strengthens internal validity. The standard control group allows us to test the effects of the DA compared to not receiving any information about tamoxifen and raloxifene, while the three-month control group allows us a clean control group to evaluate the impact of the DA over time. Our results may generalize to other insured patients, as we recruited participants from two geographically diverse health plan populations.
Practice Implications
Women given a DA describing risks and benefits of tamoxifen and raloxifene were significantly more likely to make informed decisions about undergoing chemoprevention for breast cancer. The intervention materials impacted not only knowledge but also alignment of attitudes with decisions. The online DA format makes this a potentially valuable adjunct in clinical practice, since women could utilize it prior to clinic visits. Finally, this study suggests a valuable way to evaluate the effectiveness of DAs.
Source: ...