15250-m-cuypers

214 Chapter 11 Another development issue was the information density of the current version of the DA (chapter 4). According to Fuzzy trace theory, the delivery of this amount and type of precise information could be suboptimal to support decision-making. Fuzzy trace theory explains that people might rely on the gist of information, rather than its exact details. 20 . For future versions of the current DA, or for novel DAs, the amount of presented information should be reconsidered. Ongoing research in presentation of risk information provides opportunities to further optimize presentation of quantitative information, so that the risk for judgement biases and errors is minimalized, consisting of clear reference categories, the use of numbers over words, and additional visual formats, whereas in the current DA still large amounts of text were included 21-24 . Also,furtherdevelopmentsinthedisseminationofbigdatainhealthcare,withinthewider perspective of delivering value-based healthcare, could influence the future approach to DA development 25, 26 . Evaluating appropriate and effective care is increasingly relying on big data analyses. Insights that become available in this way, may also be useful for other purposes in the care process, such as decision support. For example, presenting the best available scientific evidence in a DA, can include presenting a broad margin of uncertainty (e.g. ‘between 40 and 80 out 100 patients will experience side effect X’). For a local hospital, individual doctors, and individual patients, other, more accurate, risk estimates may apply, and become insightful from big data analysis. In recent years, the number of studies into practical applications of big data to use in individual decisions has increased dramatically, and can enable personalized estimates of the effectiveness of different treatments 26 . Within Pca care, personalized risk estimates can already be calculated for Pca screening or suitability of patients for active surveillance, based on individual patient characteristics, 27-31 . Cancer in general, and cancer treatment choices specifically, are fields where data does not only come from scientific studies, but many (national) cancer registries routinely collect relevant data as well. With current developments in these fields, big data could be integrated into DAs and VCEs to present data in a personalized format, with risk estimation, and outcome predictions relevant to the individual patient. However, most of the data that are available for such analyses were not collected with the intention to support decision-making. It should therefore be critically evaluated it the available data is relevant to the individual patient. Methodological considerations PCPCC trial A main strength of the Prostate Cancer Patient Centered Care (PCPCC) trial (Chapters 5-10) entails the cluster randomized design. With cluster randomization, hospitals were randomized to the DA arm or control arm, instead of randomizing individual patients. As a result, all patients within the same hospital received the same care, and