Implementation of the uterine fibroids Option Grid patient decision aids across five organizational settings: a randomized stepped-wedge study protocol

The multi-component implementation strategy is guided by the CFIR and NPT. CFIR is a pragmatic, multilevel framework that guides the evaluation and design of implementation studies [21, 22]. It comprises five domains (intervention characteristics, outer setting, inner setting, characteristics of individuals, and process) and 39 related constructs. These domains interact with one another to influence the effectiveness of the implementation strategy. In the context of our study, CFIR will be used at a macro level to guide implementation by comprehensively addressing four of the five CFIR domains: intervention characteristics, inner setting, characteristics of individuals, and process. Table 1 provides details on the CFIR domains.

In addition to CFIR’s multilevel approach, NPT will be used at a micro level to focus on the dynamic process that leads to the successful implementation of an innovation in routine clinical practice. NPT is an explanatory theory that helps evaluators understand the process by which organizations embed interventions into their normal work [23, 24]. NPT has four theoretical tenets: (i) coherence, supports individual and collective consensus about an intervention and its purpose; (ii) collective action, the tasks allocated to the various members within the organization to build and sustain use; (iii) cognitive participation, the relational work that influences “implementation and legitimation”; and (iv) reflexive monitoring, the communal appraisal work that aids assessment of the intervention [25].

The primary outcome measure is the number of eligible patients who receive the uterine fibroid Option Grid (three formats available).

Eligible patients will be identified in advance by the project team at each site through their outpatient scheduling systems. This process will enable us to determine the total number of patients who are eligible to participate in the study (the denominator). To calculate the number of patients who receive an Option Grid (numerator), we will provide each site with a box filled with text and Picture Option Grid paper pads. Pads will be available in every consultation room within each sites’ obstetrics-gynecology clinic. To use a paper-based Option Grid, the clinician will tear it from the pad and a numbered stub will remain where they will write their name and indicate if they are an attending clinician or a resident (see Fig. 3 for a mockup of the paper pad stub). The research assistant at each site will refer to the numbered stubs to document the number of patients who received an Option Grid on a weekly basis. The research assistant will keep a log of the numerator and denominator in an Excel file and will report the data back to the Dartmouth project staff on a weekly basis via the Qualtrics database.

For sites with EHR access to the online uterine fibroid Option Grid, we will derive weblog data from EBSCO Health including the date and time the tool was generated; this data will be incorporated into measurement of the primary outcome.

In addition to recording each site’s chosen implementation strategy and if/how it changes during the implementation and sustainability phases, we will collect the following secondary outcome measures (as outlined in Table 2):

Throughout the pre- and active implementation phases, eligible patients at each site will be invited by clinic staff to take a survey on a tablet computer following their clinical encounter (see Additional file 2 to view the patient survey). Clinic staff may provide the patient with the tablet computer or refer them to a kiosk located in the clinic that will contain the tablet. At the start of the survey, patients will be given the opportunity to view the study information and provide consent to participate in the study. The survey will collect the following patient-reported outcomes, in addition to the patient’s email address which will be used to send patients a WebLink to the follow-up survey 3 months after their first survey is completed. The follow-up survey can be completed on a personal computer or mobile device. We anticipate that approximately 30% of eligible patients will complete the post-encounter survey. The following is a list of patient-reported outcomes that we will be assessing:

Dartmouth project staff met with clinicians at each site in January and February 2019 to introduce the project. Co-investigators and stakeholder partners at each site will help identify clinicians who are eligible to participate in the study prior to the start of their pre-implementation phase and ahead of their second site visit. Prior to the second site visit, the principal investigators at each site will receive an information sheet describing the aims and methodology of the study via email. They will disseminate the information sheets to eligible clinicians. Eligible clinicians who wish to participate can provide their written consent in-person during the project staff’s second site visit, email their consent forms to the project staff at their earliest convenience, or return them to project staff at each site. Clinicians can opt-in on the consent form to have a sample of 10 of their clinical encounters audio-recorded over the course of project. We will determine before the site visits if residents are eligible to have a sample of their encounters audio-recorded. Clinicians who opt to participate will be sent a secure web link to a survey via email, so they can provide their age, gender, and years of experience post-fellowship (or year(s) of residency).

The five participating sites were randomized by the project staff’s statistician, who was blinded to the identity of the sites, using R statistical software.

Prior to the start of the pre-implementation phase, stakeholders will complete the MORE survey. At the start of the pre-implementation phase, participating clinicians at each site will complete the ADOPT survey to indicate their attitude toward PDAs. We will audio-record five clinical encounters for each consenting participating clinician, and at the end of the study phase, we will conduct interviews with a convenience sample of clinicians and staff at each site to inform their completion of the NoMAD survey (see data collection section for details).

The Option Grid will not be used during the pre-implementation phase; however, patients will still complete the survey (see patient outcomes section for details) following their clinical encounters to provide baseline data for each site.

The study interventions will be carefully described to participating clinicians at each site during an initiation session that will take place within a 2-month phase between the pre- and active implementation periods. The principal investigators (GE and M-AD) will introduce the concept of shared decision making and Option Grid PDAs to stakeholders at each site using webinars, video-conferences, and face-to-face coaching based on the “three talk model of shared decision making” [39, 40]. The format of the initiation will be adapted for each site, based on MORE survey results. Each site will determine which Option Grid PDA formats to implement (text, picture, and online versions). During the initiation phase, the Dartmouth project team will also provide each site with its clinic-level collaboRATE top score percentage from the pre-implementation phase as insight into baseline patient-rated shared decision-making performance.

During each site’s active implementation phase, the uterine fibroid Option Grid PDA will be incorporated into the routine clinic workflow according to a tailored implementation strategy, as described in the “intervention” section above. Each site will keep a weekly record of the number of eligible patients who received the Option Grid (primary outcome).

As in the pre-implementation phase, patients will be asked to complete a survey (1) immediately following the clinical encounter and (2) 3 months after the clinical encounter.

Two months into the active implementation phase, research assistants at each site will begin audio-recording clinical encounters until each participating clinician (who consents to audio-recording) has had five of their encounters recorded. The audio-recordings will be sent to the project team at Dartmouth using a secure and HIPAA-compliant platform (Dartmouth Sharepoint). Two independent raters will assess (using the Observer Option-5 measure and fidelity checklist) the extent to which clinicians involve patients in the decision-making process using Option Grid tools. At the end of the active implementation phase, clinicians will complete the ADOPT survey, and we will conduct semi-structured interviews, guided by CFIR and NPT, with a convenience sample of participating clinicians to determine the tools’ utility, and the barriers and facilitators of their use in practice.

Prior to the sustainability phase, we will provide text and Picture Option Grid PDAs to each site. Project staff will limit communication with clinical sites throughout this phase to prevent unintended influence on continued PDA use. At the end of the sustainability phase, we will measure the number of eligible patients who receive the uterine fibroid Option Grid and clinician attitudes. We will also inquire about if and how study teams plan to use the text or Picture Option Grids after our study has ended, and we will assess the utility of the intervention from varied stakeholder perspectives at each site.

With the goal of offering guideline developers expertise on how to integrate shared decision-making processes into their documentation, we will contact organizations responsible for clinical practice guidelines regarding management of uterine fibroids. We will share the Option Grid and the evidence summaries created for the development of the existing tools, as well as indicating the new PCORI-generated evidence that we have considered in developing the PDA [41]. These steps will take place over the duration of the implementation study.

The primary analysis has an expected sample size of 2600 patients across 30 to 40 clinicians (an estimated 87 patients per clinician [n = 30]) at each of the five sites. Because the intervention varies within clinician and our statistical models include clinician random and site-fixed effects, the relevant clustering variable is clinician-project week. With 40 project weeks and at least 30 clinicians, clustering is likely to have minimal impact on the results. Therefore, for a binary dependent variable (e.g., whether the uterine fibroid Option Grid is used or whether the patient gives a top collaboRATE score) with a baseline proportion of 0.5, a symmetric two-sided 95% confidence interval of width of at least 0.048 will have an 80% coverage probability of containing the true effect of the intervention, i.e., attaining a confidence margin of 0.048. In the context of shared decision making, such a margin is small and below typical levels of precision obtained from randomized studies.

Patient-level data will be clustered by site and clinician. Patients will be assigned to a single phase based on the date of their first encounter with a clinician at the site during the project period. Instances in which a patient has a repeated visit in the pre- and active implementation periods will be recorded to enable sensitivity analyses accounting for repeated observations.

Regression analyses will compare quantitative outcome measures (including survey scores and Option Grid use) between the pre- and active implementation phases while adjusting for patient and clinician characteristics. The extent to which clinicians involve patients in decision-making for each recorded encounter will be assessed by analyzing the Observer OPTION-5 and Option Grid fidelity checklist data from each phase of the study. We will conduct a framework analysis, guided by NPT, of the clinician interview transcripts to determine the barriers and facilitators of Option Grid implementation. Completion of the NoMAD Normalization Process Theory survey, based on interviews with site stakeholders, will enable us to make across-site assessments of implementation sustainability.