Introducing Decision Aids into Routine Prostate Cancer Care in The Netherlands: Implementation and Patient Evaluations from the Multi-regional JIPPA Initiative

Each of the three DAs involved was developed according to the International Patient Decision Aids Standards (IPDAS) and contained information about the disease, treatment options, and (dis) advantages of all options based on (inter) national guidelines and international literature [22]. Patients, urologists, and radiation oncologists were involved in the development and review process of the DAs [23‐25]. In each DA, the same choice options were presented: surgery, brachytherapy, and external beam radiotherapy, as well as the option of active surveillance. The DAs varied in their format and length. DA1 was a concise booklet (14 pages), DA2 was an even more concise (in diagram style with short explanations) booklet or online DA (by patient choice), and DA3 was an elaborate online DA with values clarification exercises (VCEs). The DA format coincided with the intended moment of use. DA1 and DA2 could be incorporated in clinical consultation, or used at home, while DA3 was, by design, supposed to be used outside of consultations. The characteristics of the DAs are presented in Table 1. Detailed descriptions of the separate trials investigating the DA effects have been published earlier [23‐26].

Thirty-three hospitals (out of a total of 90 hospitals in The Netherlands) implemented one of the three DAs in treatment counseling. Each DA was implemented in a specific region of The Netherlands (DA1 East, 8 hospitals; DA2 North-West, 16 hospitals; DA3 South, 9 hospitals). Per DA, hospitals were recruited to participate based on convenience (e.g., distance); the allocation of DAs to hospitals was not randomized. The DAs were handed out to patients newly diagnosed with localized Pca. For all 3 DAs, patients were eligible to participate if they had the possibility to choose between at least two treatments covered by the DA. The assessment of whether the DA was applicable (e.g., eligibility for at least two treatments covered by the DA) was done by the patient’s urologist. Actual distribution of the DA was done by either the urologist or a specialized nurse, depending on what best fitted with existing local care pathways. After the treatment decision was made, but before treatment started, patients received a questionnaire to evaluate receipt and usage of the DA. All data were collected between July 2013 and June 2016. Research protocols from each DA group were reviewed by their respective local institutional ethics committees, which each provided a waiver for further ethical assessment.

Our primary outcome measure was the implementation rate. This rate was calculated by the proportion of patients who received a DA compared with the estimated total number of eligible Pca patients per hospital during the period the DA was implemented. Since the total number of eligible patients was not prospectively registered in a structured manner in all participating hospitals, an estimation was based on hospital-specific registry data of the 6 years prior to the current project, retrieved from the Netherlands Cancer Registry.

After a treatment decision was made, a questionnaire was used to collect self-reported data about patient’s demographic variables (age, marital status, having children, and educational level). Evaluation measures consisted of DA distribution procedures (e.g., “Who presented the DA to you?”), DA user-friendliness (e.g., “Did it occur fonts were too small?”), and a 24-item list of barriers and facilitators for DA use (e.g., “I had insufficient trust in the DA”) based on literature [21] (items presented in Table 3). All three DA groups used the same questionnaires to evaluate DA use in order to enable combined data analyses.

Many DA initiatives struggle to get structurally embedded in clinical routine, despite ample evidence revealing the benefits of using DAs when making medical decisions [7, 12]. At the onset of a multi-regional implementation initiative of three new Pca treatment DAs in Dutch clinical practice, a consortium was formed to jointly measure implementation rates and patient evaluations (i.e., barriers and facilitators from the patients’ perspective) from these three DAs. Overall, 40% of eligible Pca patients received a DA. For all DAs alike, implementation was quite successful (implementation rate > 80%) in a limited number of hospitals, whereas uptake varied widely at other sites (2–80%). Overall, patient evaluations were supportive of implementation of each DA; however, the online DA3 was evaluated as having the least facilitators.

The format of the implemented DAs as well as their level of information density varied [23‐25]. DA1 and DA2 could be incorporated in clinical consultation, or used at home, while DA3 was, by design, supposed to be used outside of consultations. Despite the variation between DAs, implementation results showed the same variation between hospitals with each DA, and successful implementation (> 80%) was only achieved in a limited number of hospitals. Increasing the number of hospitals for implementation, as DA2 was implemented at 16 hospitals, compared with 8 and 9 hospitals for DA1 and DA3, did not result in more hospitals with successful implementation. This could suggest that for each DA, support was present in some hospitals prior to the start of implementation, and that for upscaling implementation, more structural encouragement and monitoring of implementation progress are needed in hospitals where the baseline support (in terms of care providers attitude or available resources) for DAs is lower.

When patient-perceived barriers were reported, most were related to DA characteristics (unpractical, unadjusted to needs) or expectations (no confidence, expected no benefits or reduction of uncertainty). Although overall report of barriers was low, barriers were reported most often for the online, elaborate DA3, and least for the very concise hybrid DA2. However, both DAs achieved similar implementation rates that were lower than the concise paper DA (DA1). This finding seems inconsistent with previous studies concluding that web-based DAs are the most promising modality for improving implementation [27, 28]. However, care providers have also shown hesitance towards online tools [29, 30]. Future research is needed to gain a deeper understanding of how the benefits of online tools, such as tailoring to patient information needs and enabling interactive VCEs, can be balanced with patients’ apparent preference for a more concise, paper format. One solution might be to provide concise, paper add-ons to online tools, which can be introduced during consultation and may enhance the user-friendliness of online tools.

The joint implementation efforts by the JIPPA consortium may have contributed to raising national awareness for SDM in both urology and oncology in The Netherlands. Many care providers have been introduced to the DA and to the principles of SDM, and during the course of the projects, consortium members contributed to national Pca treatment guidelines with a section on SDM and DAs (www.​oncoline.​nl). Therefore, the study in itself increased awareness for SDM and the existence of DAs and educated many teams in using DAs in clinical routine. However, it may also have caused a barrier, as clinical practice was unclear about which DA should be applied, and what the differences between the available DAs entailed. To the best of our knowledge, no earlier studies have reported (national) implementation rates for Pca DAs, and comparability to other DA implementations studies is difficult to interpret as they were aiming at different patient populations (e.g., women with breast cancer, or orthopedic patients) and settings (e.g., screening decisions often include the general practitioner) [10, 11, 31]. Further research is needed to determine if having different types of DA can help implementation since patients and care providers can select the DA they prefer most, or that the variety in available DAs hinders implementation since each DA has its specific characteristics and usability aspects that require training. Moreover, future research could study if specific DA characteristics have an effect on implementation rates, by randomizing distribution of different DA types across hospitals.

A strength of the current study was that we were able to investigate implementation of three DAs by using a similar questionnaire at a similar point in time. As a consequence of studying three different DAs, sample size and number of participating hospitals were higher than most previous Pca DA studies [9, 32]. Eventually, one in three Dutch hospitals was exposed to one of the three DAs. Hospitals from different levels (academic and non-academic) and from different regions were included in the study, increasing the generalizability of our findings.

A limitation of the current study is that the implementation rate was calculated based on actual receivers of a DA as proportion of an estimation of the total number of eligible patients. Since the number of patients eligible for study inclusion was not systematically registered by the hospitals included in our studies, we relied on the hospital-specific retrospective cohorts of PCa patients from the cancer registry. This ensured the sample was determined via the same method for every hospital. However, since the total number of patients eligible for DA receipt was estimated, this entailed that no information was available about patient characteristics from those patients who were possibly eligible but were not offered a DA. In particular, in hospitals with low implementation rates, a selection bias could have occurred if only patients were included who favored DA use. Another limitation is that the implementation period was not exactly simultaneous for all three DAs. Implementation of DA1 started almost a year ahead of DA2 and DA3. Moreover, a previous version of DA1 was studied in an earlier trial, which could have helped achieving the higher overall implementation of DA1 [25]. Furthermore, each participating hospital was linked to one of the three regions, and consequently implemented its respective DA. Possibly, some patients or care providers could have been more supportive of another DA and overall DA uptake would have been higher if all formats would be matched according to patient or care providers’ preferences. For example, one patient might benefit more from an elaborate DA, while for another patient, optimal understanding and satisfaction are reached with a concise DA [33‐36]. Finally, no information was available from patients who received, and possibly also used, a DA but did not consent to participate in the survey study.

Patient evaluations from the three DAs in the current study were all favorable towards implementation. To further understand the observed differences in implementation rates between hospitals, future steps towards sustained DA use should include further investigation into barriers at the level of care providers and organizational barriers.

Overall implementation rate of the DAs in clinical routine was 40%. A wide variation in uptake across hospitals was observed for each DA. Most patients were satisfied with the DA they received, and only few barriers of usage were perceived by patients. Offering an online-only DA led to less patient-reported facilitators compared with a paper-only or hybrid DA.

Patients appeared to be satisfied with each DA format. Sustained implementation of DAs in clinical routine requires further encouragement and attention, and could require a tailored distribution approach per hospital site [37].