What does it take to have sustained use of decision aids? A programme evaluation for the Breast Cancer Initiative

Abstract

Background  The Breast Cancer Initiative (BCI) was started in 2002 to disseminate breast cancer decision aids (PtDAs) to providers.

Methods  We analysed BCI programme data for 195 sites and determined the proportion of sites involved in each of five stages of dissemination and implementation of PtDAs. We conducted cross‐sectional mail and telephone surveys of 79 sites with the most interest in implementation. We examined barriers associated with sustained use of the PtDAs.

Results  Since 2002 we attempted contact with 195 sites to join the BCI. The majority indicated interest in using PtDAs 172 of 195 (88%), 93 of 195 signed up for the BCI (48%), 57 of 195 reported distributing PtDAs to at least one patient (57%), and 46 of 195 reported sustained use (24%). We analysed data from interviews with 59 of 79 active sites (75% response rate). The majority of providers 49 of 59 (83%) had watched the PtDAs, and 46 of 59 (78%) distributed them to patients. The most common barriers were lack of a reliable way to identify patients before decisions are made (37%), a lack of time to distribute the PtDAs (22%) and having too many educational materials (15%). Sites that indicated a lack of clinician support as a barrier were significantly less likely to have sustained use compared to sites that didn’t (33% vs. 74%, P = 0.02).

Conclusions  Community breast cancer providers, both physicians and non‐physicians, express a high interest in using PtDAs with their patients. About a quarter of sites report sustained use of the PtDAs in routine care.

Background

The information needs of patients with breast cancer vary depending on where they are in the continuum of care, from diagnosis to treatment to survivorship. In the United States, national organizations such as the American Cancer Society, the National Cancer Institute, Breast Cancer Network of Strength and Susan G. Komen Foundation have a presence in many communities and provide educational materials and support to patients. On a local level, hospital and community‐based breast centres either develop their own materials or choose among books, pamphlets, websites and other resources available for patients with breast cancer. Given the abundance of educational materials available to patients about their choices for various breast cancer treatment options, it can be challenging for healthcare providers to determine which materials are of high quality and will have the biggest impact for their patient population.

Decision aids, a specific type of educational material, have been shown to be beneficial to patients who make decisions about medical treatment for chronic conditions such as breast cancer. Over 500 patient decision aids (PtDAs), including 15 in breast cancer, have been developed. ¹ , ² Decision aids are different from educational materials because they ‘guide patients through a process of deliberation so that a choice can be made that matches their informed preferences’. ³ PtDAs have been shown to increase patient knowledge, help undecided patients select a choice and reduce decisional conflict. ² A review of decision aids used in breast cancer found that the tools are acceptable to patients and surgeons and provide similar benefits in the cancer setting. ⁴ Patients also find the PtDAs easy to use and helpful in preparing them for the consultation with their provider. ⁵ , ⁶

Despite these positive findings, there is relatively little use of these tools in routine care. Previous research has examined some of the challenges for implementing PtDAs in the clinical setting. Holmes‐Rovner et al. ⁷ found that physician buy‐in was not sufficient for implementing two PtDAs (for heart disease and breast cancer) in three hospitals and that there was a lack of time for patient education. Graham et al. examined the acceptability of two decision aids, on hormone replacement therapy for post‐menopausal women and chronic obstructive pulmonary disease, to practitioners in family practice and gynaecology. They found that most physicians were willing to use decision aids with patients once they had the chance to review them. ⁸ Physicians’ willingness to use the programmes depended on their perceptions of the programme’s content, format, beliefs about their patients’ abilities to use the decision aid, and how much they thought the PtDAs would facilitate their work. ⁸

Gravel et al. reviewed previous studies on barriers and facilitators for implementing shared decision‐making (SDM) in clinical practice from the perspective of the healthcare provider. The top three reported barriers were ‘time constraints, lack of applicability because of patient characteristics and lack of applicability because of the clinical situation’ and the top three facilitators were ‘provider motivation, positive impact on the clinical process and positive impact on patient outcomes’. ⁹ Gravel et al. ⁹ concluded that aside from physicians, not much is known about the role of other health care providers and implementing SDM and suggested this be examined in future research.

To fill these gaps, Sepucha and colleagues started the Breast Cancer Initiative (BCI) in late 2002, with funding from the Foundation for Informed Medical Decision Making and support from Health Dialog, Inc. As part of the initiative, providers in the United States who consult patients with breast cancer about treatment options are able to receive up to ten free copies of five PtDAs. The videos or DVD‐based programmes and booklets are entitled Ductal Carcinoma In Situ: Choosing Your Treatment (DCIS), Early Stage Breast Cancer: Choosing Your Surgery (Surgery), Early Breast Cancer: Hormone Therapy and Chemotherapy – Are They Right For You? (Chemo), Breast Reconstruction: Is It Right For You? (Recon), and Living with Metastatic Breast Cancer: Making the Journey Your Own (Mets) (©Foundation for Informed Medical Decision Making and Health Dialog for all programmes). For our study, providers (such as physicians, nurses, and breast health coordinators) were located at breast cancer centres in hospitals and community health centres, at breast cancer support groups, and at private practices that diagnose and treat patients with breast cancer. Since 2002 the BCI has moved through three phases of dissemination activities.

In Phase 1, the BCI simply responded to incoming requests for programmes, mainly from collaborators and colleagues. We reported on the implementation by health care providers at nine Phase 1 sites and found high levels of usage (67%), but the small sample size and the predominance of academic medical centres made it hard to generalize those results. ¹⁰ In Phase 2, the BCI launched an outgoing recruitment campaign which targeted nurse coordinators and patient educators at 50 community sites in Massachusetts. Many of the providers at these sites were interested in using the programmes and about twenty percent of those sites reported implementing the PtDAs with their patients. Providers that were not using the programmes cited several barriers such as lack of physician buy‐in, lack of time, difficulty with integration into clinic process and lack of staff resources. ¹¹ Based on these results, we felt that it was important to increase the contact with doctors at the sites and to streamline the process.

The investigators refined the dissemination process and launched Phase 3 of the initiative targeting providers at sites across New England. Here, we report on the results of this more structured approach to disseminating breast cancer PtDAs to community breast cancer sites. A secondary objective is to examine the factors associated with successful, sustained implementation of PtDAs by the providers at these sites.

Methods

Dissemination framework

According to Roger’s theory of diffusion of innovations, diffusion is a process by which an innovation is communicated through certain channels over time among members of the social system. ¹² Four elements, the innovation, the communication channels, the time and social system are all objects of study for the adoption or rejection of the innovation. For this project, we focused on the innovation‐decision process, or how an individual passes from first knowledge of an innovation to sustained use or rejection. The innovation is the PtDAs, so we focused on the provider’s decision to adopt or reject using PtDAs with their patients. We used the five steps outlined by Rogers to develop the evaluation plan for the intervention: (i) awareness of the innovation (ii) interest in the innovation (iii) decision to use the innovation (iv) putting the innovation into practice and (v) sustained use of the innovation over time.

Research on diffusion has identified five characteristics of an innovation that have been consistently shown to impact the rate of adoption (i) relative advantage (whether it is perceived as better than status quo), (ii) compatibility (whether it is consistent with existing values, culture, needs, etc.), (iii) complexity (whether it is perceived as difficult to use) (iv) trialability (or the degree to which it can be experimented with or used on a limited basis) and (v) observability (or the degree to which the results are observable to others). ¹² In designing our written and telephone interview questions, we focused on getting providers’ perceptions of these issues for their sites.

Intervention

The structured methods for disseminating PtDAs in Phase 3 were as follows. We identified sites throughout New England through breast cancer resource directories and online searches. We did not include sites in Massachusetts that were in Phase 2. We contacted sites to get names and contact information for the medical director (usually a doctor) and person responsible for patient education (usually a nurse or breast health coordinator). First, we sent a letter (with and without a copy of the PtDAs) to up to two providers at each site to inform them of the availability of the programmes and to generate awareness of the innovation. Sites were able to fax or mail back a request form for a sample copy of each programme. We used this form as an indication of their interest in the innovation. We requested that sites watch the programmes, and then required them to sign a Participant Agreement (PA) form to get additional copies of the PtDAs to use with their patients. We used the PA form as an indication of their decision to use the programme with their patients. On the PA form, sites identify one contact person to receive correspondence for that site (e.g. notification of PtDA updates) and that person agreed to provide feedback on usage. We invited one provider at each site to participate in a short telephone interview 6‐12 months after the PA form was received to determine whether the PtDAs had been used and to examine their perceptions of the PtDAs. Finally, we mailed a survey annually after the phone interview to track continued use over time.

Control group

We did not have a concurrent control intervention. However, we did compare the results achieved in Phase 3 to a historical control group, comprised of sites approached during Phase 2. In that Phase, we identified 50 sites across Massachusetts using similar methods as Phase 3. We identified the person responsible for patient education at each site (again usually a nurse). Then, we sent a letter to that one provider, followed up with a phone call to assess interest and sent programmes to those providers who indicated interest verbally through the calls. We also communicated with providers via email at their request. We did not have any forms to sign, only verbal or email requests for programmes to indicate interest and adoption decisions.

Participants

A convenience sample of cancer centres, hospitals, private practices and resource centres that see patients with breast cancer who are facing decisions about treatment. Outgoing campaigns predominantly identified sites in New England (including Connecticut, Maine, Massachusetts, New Hampshire, Rhode Island, and Vermont) and one or two providers at each site were contacted. Incoming requests came were also handled from sites across the country.

All 79 active sites in the BCI were invited to participate in the mail and/or telephone survey. We defined ‘active sites’ as those who had completed a PA form and had indicated some interest in adopting the PtDAs. Of the 93 sites with a PA form, 14 sites indicated to BCI staff that they were not using the PtDAs and not interested in doing so, resulting in 79 active sites for the current sample. ¹⁰ , ¹¹ All sites were in the United States.

Study design

The programme evaluation had two main parts. First, we analysed the data tracked on all of the sites that had been contacted to participate in the BCI (e.g. sample set request forms and PA forms). Second, we conducted a cross‐sectional mail or telephone survey of active sites.

The provider contacts at all Phase 3 sites who signed a PA form were invited to participate in telephone interviews. Each provider was contacted at least twice and up to six times by email and/or by phone to set up an interview. The semi‐structured telephone interviews focused on the experiences using the PtDAs, the resources that enabled providers to successfully implement the PtDAs, and the barriers that prevented or limited their use of the programmes. Two study staff participated on each call, one conducted the interview and the other took detailed notes. The first 20 calls were recorded and audiotapes were reviewed to supplement the notes.

A mailed survey was sent to all active sites for whom we had already conducted a telephone interview in past years. A reminder call and/or email was sent a few weeks after the mailing and a reminder mailing was sent to all non‐respondents about 4 weeks after the initial mailing.

The study protocol was approved by the Institutional Review Board at Massachusetts General Hospital (protocol #001963). We obtained written informed consent from healthcare providers who agreed to participate in the telephone interview. Consent was implied by the voluntary completion of the mailed surveys.

Measures

The telephone interview guide contained four open‐ and 18 closed‐ended questions on site demographics such as size, kinds of services provided, patient population, the provider’s role, usage of the programmes, the provider’s evaluation of the programmes, and the barriers and resources for implementation. The telephone interviews each lasted for approximately 30 minutes. The mailed survey included the same 18 closed‐ended items.

Perceived barriers and facilitators: All respondents indicated whether the following six barriers prevented or limited their use of the programmes, (i) lack of a reliable way to identify patients before decisions are made; (ii) lack of time to distribute the PtDAs; (iii) too many other educational materials for patients; (iv) a lack of clinician support of the PtDAs; (v) lack of trust in the quality of the programmes; (vi) lack of staff to distribute the PtDAs. Respondents also indicated whether the following six facilitators would make it easier to use the programmes with their patients: (i) One‐on‐one training on implementation of PtDAs; (ii) CME/CEU credits for watching the PtDAs; (iii) web‐based versions of the PtDAs; (iv) having a brochure that can be placed in the waiting room to introduce the programmes to patients; (v) seeing published data on improvement of patient satisfaction, knowledge or other outcome measures; (vi) incentives for health care providers to watch the programmes. There was an option for respondents to verbalize (in the telephone survey) or write in other barriers and other facilitators that they experienced at their site.

Respondents who indicated that they had distributed the programmes were also asked about the following:

• 1 Physician support: respondents reported the number of physicians at their site who supported the use of the PtDAs on a 5 point scale (1 = none, 2 = a few, 3 = some, 4 = most, 5 = all). Respondents who were not in clinical centres with physicians (e.g. community resource centres) reported not applicable.
• 2 Perceived helpfulness: respondents rated the helpfulness for their patients of each PtDA that they had used on a 5 point scale (1 = very unhelpful, 2 = unhelpful, 3 = neither helpful nor unhelpful, 4 = helpful, and 5 = very helpful).
• 3 Perceived difficulty: respondents rated the difficulty of getting each of the PtDAs to patients before making a decision on an 11‐point scale [from extremely difficult (10) to somewhat difficult (5) to not at all difficult (0)].

Data analysis

The results of the dissemination activities were calculated for each of the five steps as follows:

• 1 Knowledge of PtDAs: Number of sites that received the information about the PtDAs/Number sites contacted;
• 2 Interest in PtDAs: Number of sample set requests/Number sites contacted;
• 3 Decision to adopt: Number of signed participant agreement forms/Number of sites contacted;
• 4 Implementation of PtDAs: Number of sites that reported using the PtDAs with at least one patient/Number of sites contacted;
• 5 Sustained use: Number of sites using the programmes regularly with patients at least 6 months after signing the participant agreement form/Number of sites contacted.

The denominator for each step was the total number of sites contacted, including incoming requests. Sites that did not respond to the request for telephone interviews were assumed to have not ever used the programmes. For sites for which we had data, we calculated the time it took, on average, for a site to move from step 1 to step 2 to step 3.

We report descriptive analyses of the telephone and mailed survey data. For physician support, we created a binary variable with all/most vs. some/a few/none. We averaged the responses across programmes to report on overall helpfulness and difficulty. We conducted bivariate analyses to examine whether certain site characteristics, aspects of the dissemination process and providers’ perceptions of helpfulness or barriers were associated with sustained use. Our hypotheses were that sites who perceived the programmes as helpful would be more likely to have sustained use. Conversely, sites who perceived the PtDAs as difficult to implement would be less likely to have sustained use. We conducted exploratory analyses to examine the association between specific barriers reported by the sites and sustained use, in an attempt to identify those barriers that may be most important to address. We used t‐tests for continuous variables, Fisher’s exact test for binary variables and chi square for categorical variables.

We conducted post hoc analyses comparing the Phase 3 dissemination results to those achieved in Phase 2 to see whether there were any improvements. We compared the proportion at each stage using Fisher’s exact test for two independent samples. All analyses were conducted using spss version 18.0 (PASW Statistics 18, Chicago, IL, USA).

Results

Dissemination of programmes

Since 2002 we were able to contact providers at 188 of a potential 195 sites. Most sites, 172 (88%), indicated interest in PtDAs by completing a sample set request and about half, 93 (48%), signed a PA form. Figure 1 shows the results of the dissemination steps for all sites across the three phases of the BCI. It also highlights the results achieved in Phases 2 and 3, and shows the results from miscellaneous sites that were not part of an outgoing campaign. The majority of the sites were recruited through outgoing solicitation, although 15% of sites in Phase 3 and 25% overall represent incoming requests, where sites contacted us after being referred by a colleague or otherwise learning about the BCI.

Figure 1.

 Results from all 195 sites contacted as part of the Breast Cancer Initiative.

For the Phase 3 sites, it took, on average, 7.1 months (SD 4.8) from the initial outgoing invitation letter, step 1, to receive a signed participant agreement form, step 3. Most of this time was to get from step 2 to step 3 which took, on average, 5.5 months (SD 4.1).

We completed interviews with 59 healthcare providers from 79 active sites (75% response rate) by telephone (n = 46/57) and by mail (n = 13/22), each representing one site. The majority of the providers we interviewed were women (97%). Thirty‐eight of the providers were based at a community centre (64%), 10 at a hospital (17%), 8 at a support group for breast cancer patients & their families (10%), and 3 in private practice (5%). About half of the providers interviewed were nurses (53%), 17% were physicians and 30% were other providers such as breast health coordinators, social workers, support group leaders and clinical administrators. The providers had been in practice an average of 19 years (SD 11 years). The patient volume varied across the sites. The median number of new breast cancer diagnoses was 60 (25th and 75th quartiles 20, 145) per year. The site characteristics are summarized in Table 1. Almost all of the sites that we interviewed had the PtDAs for at least 1 year. The surgery programmes were used by the most sites (DCIS 58% and Surgery 54%), followed by the Recon (41%), Chemo (39%), and Mets (31%) programmes.

Table 1.

 Site characteristics for responders

Characteristic	Responders (n = 59) N (%)
Campaign
Phase 2 – Massachusetts sites	10 (17)
Phase 3 – New England sites	31 (53)
Miscellaneous sites	18 (30)
Incoming/outgoing
Incoming	24 (41)
Type of site
Community centre	38 (64)
General hospital	10 (17)
Support group	8 (14)
Private practice	3 (5)
Median new early breast cancer diagnoses in the past year (25th, 75th quartiles)	60 (20,145)
Gender of contact
Female	57 (97)
Professional training
Nurse	31 (53)
Physician	10 (17)
Other	18 (30)
Mean years in practice (SD)	19 (SD 11)

Implementation of the programmes

The sites reported using a variety of methods for distributing the PtDAs, and the most common method was distribution by nurse or doctor during the patient visit (78%). Across the sites, the PtDAs were distributed by nurses (68%), breast health coordinators (14%), physicians (11%), or other staff such as support group leaders, dieticians or patient advocates (8%). Other distribution methods included placing the PtDAs in a library or resource room (29%), giving them to patients during support group meetings (11%), and mailing them to the patient’s home (4%). A few sites used a combination of these methods.

The majority of providers that we interviewed 50/59 (85%) reported that they watched the PtDAs. More providers who watched the PtDAs distributed them, compared to those who did not watch them, although the difference was not statistically significant (82% vs. 56%, Fisher’s exact test, two‐sided, P = 0.1).

Barriers and resources

On average, sites indicated 1.3 barriers (range 0–3) to using the programmes. The most commonly reported barrier was the lack of a reliable way to identify patients before decisions are made (37%). Frequency of other barriers was a lack of time to distribute the PtDAs (22%), too many educational materials (15%), a lack of clinician support (14%), lack of staff (10%) and lack of trust in the materials (2%). Respondents also wrote in additional barriers including hearing from some patients that they did not want the PtDAs (12%), literacy concerns with the PtDAs (10%), staff forgetting to hand out the PtDAs (8.5%), and the long length of the PtDAs (5%).

Providers also indicated a few facilitators that would make the programmes more attractive and easy to use. These included having web‐based versions of the PtDAs (25%), having a brochure that can be placed in the waiting room to introduce the programmes to patients (15%), seeing published data on improvement of patient satisfaction, knowledge or other outcome measures (10%), providing CME/CEU credits for providers who watch the programmes (10%), one‐on‐one training on implementation (10%), and incentives for staff or doctors to watch the programme (10%). Some respondents also wrote in an additional facilitator, modifying the programme itself (7%).

Factors associated with sustained use

The majority of respondents (71%) said they had support from all or most of the physicians at their sites. Sustained use of the PtDAs was not significantly associated with the level of physician support (85% vs. 73% for sites with support from all or most of physicians compared to support from none/few/some, P = 0.32).

Respondents rated the helpfulness for the PtDAs that they had used. The majority (81%) found them helpful or very helpful to their patients with percentages ranging from 67% for Metastatic and 88% for DCIS and Surgery. There was no difference in sustained use for sites who considered the programmes very helpful compared to other responses (80%, n = 24 vs. 79%, n = 20, P = n.s.). Figure 2 presents these results across the different programmes.

Figure 2.

 Helpfulness of PtDAs. The PtDAs are: Ductal Carcinoma In Situ: Choosing Your Treatment (DCIS), Early Stage Breast Cancer: Choosing Your Surgery (Surgery), Early Breast Cancer: Hormone Therapy and Chemotherapy – Are They Right For You? (Chemo), Breast Reconstruction: Is It Right For You? (Recon), and Living with Metastatic Breast Cancer: Making the Journey Your Own (Mets).

On average, providers reported that distributing PtDAs was somewhat difficult, with a rating of 3.9 (SD 3.1) out of 10. The sites who reported sustained use perceived less difficulty than those who did not have sustained use, but this was not statistically significant (mean difficulty rating 2.8 vs. 4.3, respectively, two‐tailed t‐test, P = 0.18).

We examined the relationship between specific barriers and sustained implementation of the programmes. Sites that reported lack of clinician support as a barrier were significantly less likely to have sustained use compared to those that didn’t (33% vs. 74%, Chi square=5.78, d.f.=1, P = 0.02). No other barriers were significantly related to sustained use.

Comparison of Phases 2 and 3

The more structured process employed in Phase 3 was associated with a significant increase in the response at the earlier steps compared to Phase 2 (step 1: 100% vs. 86%, P < 0.001, and step 2: 94% vs. 68%, P < 0.001). It did not result in significant differences for the latter steps (step 3: 41% vs. 32%, P = 0.3, step 4: 28% vs. 20%, P = 0.3, and step 5: 25% vs. 16%, P = 0.4).

Discussion

This study presents data on the evaluation of the dissemination of breast cancer patient decision aids to almost 200 sites. The highly structured dissemination process adopted in Phase 3, based on Rogers’ theory of diffusion was associated with higher rates of contact and more interest, but not with higher rates of adoption or sustained use. We believe the early increases are largely because of the fact that whenever possible, we had two contacts at each site, a medical director and the individual in charge of patient education (usually a nurse or breast health coordinator). It may also be due to the fact that it was easier for providers to indicate interest (via faxing in a request form).

A large portion of the sites dropped off between step 2 and step 3, where we required providers to watch one of the programmes before they could join the initiative (see step 3, Fig. 1). By removing this step from the process, it is likely that we would have had higher number of sites sign participant agreement forms. Although the difference was not significant, providers who reviewed at least one of the programmes were more likely to distribute them to patients (82% vs. 56%), which suggests that this step may play an important role in the decision to adopt the intervention. In future outgoing campaigns, we plan to relax this requirement and examine rates of implementation both in the short and long term. In addition, we plan to offer providers continuing education credits for watching the programmes to try to make this step less of a barrier.

A number of sites also dropped off after signing the PA form and never implemented the programmes. From the interviews, we found that interest in the programmes was not always enough to overcome barriers such as the difficulty in reliably identifying patients before decisions are made and a lack of time to distribute the PtDAs. These same barriers have been reported in other studies of decision aid implementation as well. ⁷ , ⁸ , ⁹ Brace et al. (2010) ¹³ surveyed cancer physicians in Ontario, Canada, and found that less than half were aware of PtDAs relevant to their clinical practice, although 70% were interested in using them with patients.

One quarter of the sites were able to overcome time and resource limitations, and reported sustained use of the PtDAs. Given the very limited support that sites were given to implement the programmes, we find this result very promising. The study provides evidence that decision aids can be implemented into routine breast cancer care. Unfortunately, we do not have much data on how many patients these sites are reaching and that data would be valuable to evaluate the quality of the implementation.

We do have data on the factors that sites felt would help facilitate implementation. The two most common requests were for web‐based versions of the PtDAs and for brochures that could be placed in waiting rooms to introduce the programmes to patients. These methods would make the PtDAs more visible and accessible to patients, therefore reducing the burden on providers to remember to offer the PtDAs to patients. It will be important in future research to examine the quality of the implementation, including how well the sites reach eligible patients and whether the PtDAs produce measurable improvements in the quality of decisions.

Contrary to our hypotheses, we did not find a significant association between sustained use and perceived helpfulness or with the perceived difficulty of implementing the PtDAs. This lack of association might be due to the fact that we only collected survey data from those sites who signed a PA form, in other words sites who were very interested in using the PtDAs. A better approach would be to ask about perceptions earlier in the process and then prospectively document how those beliefs were linked to the decision to adopt the innovation both for those who did and for those who did not sign a PA form. We plan to make these adjustments to the assessments in future BCI recruitment campaigns. Improving our understanding of the different factors that relate to sustained use of PtDAs will be an important area of future research.

There are several limitations to the study. First, the vast majority of the sites are located in New England and the invitation and mailings came from investigators at Massachusetts General Hospital, a well‐respected hospital in the region. Whether providers in other parts of the country would be as receptive to the initial contact is not clear. Second, we relied on reports of one provider at each site for data on implementation and perceptions of the PtDA. This contact reported on the usage of the programmes at the entire site, including usage by other providers. Some respondents did mention that they had talked with colleagues in anticipation of the interview to learn about their experiences. However, it would be preferable to survey all providers who are involved and even patients. Further, direct observation would also provide a rich data set to examine implementation at the sites.

Another limitation is that the interview and survey data were collected retrospectively, after the decision to adopt or reject the innovation had been made. We also did not gather survey data from those sites that did not indicate interest in or sign up for the initiative. It is a big challenge to collect data reliably from sites that do not participate, as the response rate for these sites is very low. We also do not have reliable estimates for utilization and are thus limited in estimating the actual reach of the PtDAs within each site.

Conclusions

There is high interest in using breast cancer PtDAs among community breast cancer providers in the northeastern United States. About a quarter of sites are able to integrate the PtDAs into their routine of clinical care without much outside support. Distribution methods varied by site. A more structured approach increased the ability to reach and engage more providers, but did not impact the sustained use of breast cancer PtDAs.

Conflict of interest

Dr Sepucha receives research and salary support from the Foundation for Informed Medical Decision Making (FIMDM). FIMDM produces the decision aids used in this project. The other authors declare that they have no competing interests.

Source of funding

Foundation for Informed Medical Decision Making, 40 Court Street, Suite 300, Boston, MA 02108, USA.