In a combined analysis of two randomized controlled trials promoting SDM for prostate cancer screening, the intervention had mixed effects on the key components of SDM: it increased men’s perception that screening is a decision and men’s knowledge about prostate cancer screening, but had no effect on men’s preferred or actual participation in shared decisions. Men who were exposed to the intervention were significantly less likely to plan to get screened in the next 12 months and actually get screened according to chart review at 9 months.
Our findings suggest the ability of SDM interventions to increase men’s knowledge and alter their intent for prostate cancer screening, but additionally highlight the complexities of promoting SDM: namely helping men participate equally with their clinician, or, at least, participate at a level at which they desire. We designed our intervention to prepare both men and doctors for SDM. We encouraged men to consider the facts, decide what they wanted, and talk with their doctor and, then, provided advice on overcoming barriers to talking with their doctor. Additionally, we provided doctors with one session of education on the rationale and recommendations for SDM. With these efforts, we increased the proportion of men who perceived prostate cancer screening to be a decision and had adequate knowledge and decreased intent for screening. We did not, however, increase the proportion of men who shared a decision with their doctor or participated at the level they desired. There are four possible explanations for these findings: men either a) talked with their doctor and agreed that a primarily independent decision was appropriate, b) didn’t see the value in equal participation when they had already received the relevant facts, clarified their values, and considered what they wanted, c) didn’t (despite our intervention) know how to engage the doctor for equal participation when they wanted to, or d) had a doctor who (despite our educational session) didn’t engage in shared decisions.
An important question given our results is whether there is an added benefit of sharing the decision with a doctor beyond getting the facts and considering one’s own values. The theoretical benefit of sharing the decision is that it allows doctors to clarify men’s understanding of key facts and relevant values, highlight the unique circumstances that might alter the decision for any individual, and add a considered perspective on the decision. Several of these functions may not be necessary when decision aids are available, particularly if men are known to be health literate and the decision aid allows for deliberation on values and preferences. Instead, independent decision making (with the doctor in a supporting role) may sometimes be appropriate and improve men’s self-efficacy for following through with a decision . Whether a shared decision adds benefit over an informed decision is an empirical question that should be addressed in future research. An observational analysis of the outcomes that result when decisions are either shared, made at a level which men desire and feel comfortable, or made without regard to men’s preferences, would be helpful in resolving questions about SDM (especially if the analysis is focused on outcomes such as value concordance and adherence to decisions). Additionally, an analysis of the frequency of decisions in which clinicians’ input might substantially alter the decision making process would be helpful.
Future research should also attempt to measure the relative contribution of each component included in SDM interventions. We are unclear which components of our intervention had the most impact on decision making and screening outcomes or of the independent value of our novel coaching tool. Testing the relative effect of various components will aid construction of new interventions and refinement of existing ones. Future work might also explore other intervention components or component content that might make SDM interventions more effective in promoting shared decisions. For instance, researchers might explore alternate messages about how to share a decision (i.e. men should clearly state their preference for decision making; or men should ask their doctor whether their unique circumstances should alter the decision making process). Researchers might also explore stronger interventions for physicians (i.e. to help them identify patients’ preferences and encourage question asking). A few simple changes might refocus men and their physicians on the value of sharing decisions and facilitate the process, thereby promoting such decisions.
In considering future directions, researchers should also consider how to improve on the methods we used. First, our study sample included a convenience sample of men, who may have exhibited differences from the source population from which they were sampled. Future work should consider random sampling from the source population. Second, despite randomization, the small size of our study resulted in differential distribution of confounders among study groups. We controlled for this in multivariate analysis, but recognize the potential that unmeasured confounders may have affected our results. Future work should employ larger sample sizes to ensure the success of randomization. Third, because we randomized at the patient level, physicians saw patients in both the intervention and control groups, creating the possibility for contamination. Future work with greater resources should consider randomization of physicians or practices. Fourth, other factors may have biased our effect size. Providers may have altered their behavior merely because they knew they were being watched. Similarly, patients or providers may have altered their survey responses because based on their study assignment (or assessment thereof). Fifth, our measures of the key components of SDM (including our measures of knowledge and patient participation) haven’t been formally validated. Future work should consider a full assessment of the validity and reliability of our measures or alternate methods of measuring knowledge  and patient participation  to ensure the validity of conclusions. More extensive knowledge and participation measures and/or newer measures of informed and shared decision making might draw different conclusions [31–33]. Finally, the generalizability of our sample may be limited. We enrolled a convenience sample of men from 4 conveniently located academic and community practices in two cities in North Carolina. Approximately 45% of men had engaged in discussions with their physician in the last 12 months, suggesting they may be a more educated and activated group than most. Sampling from more diverse practices in diverse locations would improve the generalizability of data. Further, we addressed only 1 clinical decision (e.g. prostate cancer screening). To the extent that preferred and actual involvement in decision making differs across clinical decisions, the benefits of shared decision making interventions similar to ours (i.e. including a decision aid, coaching session, and physician education session) may differ across decisions and should be tested.