Prospective application of implementation science theories and frameworks to inform use of PROMs in routine clinical care within an integrated pain network

Chronic pain (CP) is a complex condition that affects the individual’s daily activities and quality of life, and social and family environments [1]. Management of CP patients is fragmented, with patients seeking a wide variety of primary and specialty care services [2‐5] resulting in poorer outcomes and higher healthcare costs as patients often receive unneeded diagnostic and medical procedures [3, 6]. Pain has the potential to impact a range of domains of a person’s health and quality of life including physical, mental, emotional, and cognitive function. Patient-Reported Outcome Measures (PROMs) enable clinicians to incorporate the patient voice in evaluating the impact of their health condition on their function and health-related quality of life (HRQL), [7‐9] and in optimizing treatment planning. There is evidence to support that PROMs are necessary to capture outcomes that are meaningful to individuals with chronic pain and to support a patient-centred care approach [10, 11].

To increase their widespread adoption and to optimize their usefulness to a range of stakeholders, PROMs must be perceived as relevant, meaningful, and actionable to the patients and the clinicians using them and the organizations implementing them. Using implementation research can help identify strategies that address the contextual barriers and facilitators to using PROMs in a specific clinical context. There are over 60 frameworks to guide the implementation of healthcare innovations [12]. Using frameworks to guide the application of PROMs in clinical practice is important as they can facilitate identification of determinants of implementation, guide the selection of implementation strategies, and evaluate if implementation was effective. Frameworks can also guide the selection of domains that influence sustainability of the use of PROMs, and their impact on quality of care and health outcomes, and provide a common language to synthesize results across studies [13].

As part of a 5-year strategic plan, the McGill RUISSS Center of Expertise in Chronic Pain (CECP) developed processes to implement and evaluate the impact of PROMs across an integrated care network. Taking a grass roots approach, the CECP leadership asked chronic pain clinics about their interest in implementing PROMs. The clinics that expressed interest worked with the research team, the CECP and patients, to develop their electronic PROM (ePROM) implementation plan. Two complementary implementation science frameworks were selected to guide the process and to identify the determinants of implementation of ePROMs at the level of patients, clinicians, and the organization. The objective of this study was to present the knowledge translation (KT) [14] and implementation science approaches that we used to inform the implementation of ePROMs across an integrated chronic pain network that includes primary, rehabilitation, and hospital-based care. Specifically, we present how we applied two implementation science frameworks to evaluate determinants of implementation, to guide the selection of strategies to implement ePROMs in clinical practice, and to select implementation outcomes.

To address the gaps in chronic pain management, the CECP developed a strategic plan to provide integrated care for individuals that included early referral to the right service based on individuals’ needs. In line with evidence-based practices, interdisciplinary teams were integrated in primary care clinics in four health regions, with referral to secondary and third lines of care for more complex chronic pain cases.

In order to successfully implement ePROMs to screen for patient service needs, develop individualized intervention plans, and monitor changes, results from two implementation research studies were integrated, one in primary care [15], and a one in tertiary and rehabilitation care [2, 11]. A steering committee reviewed the results of each study during four 1.5-h working meetings and developed and iteratively reviewed the implementation plan. Committee members consisted of a decision maker and clinician from each site, the director of the CECP, two research team members, and two patient partners met during the studies and at the end.

The choice of the theoretical implementation framework used in each study or setting was mainly driven by the specific clinical context, and the understanding of our team at the time about implementation research and its application to ePROMs (Fig. 1). In primary care, the Theoretical Domains Framework (TDF) was selected as it allowed for identifying determinants of implementation of ePROMs more at an individual clinician level and for developing theory-based behavioural change strategies to address barriers and train clinicians in the use and application of PROMs for patient care. In these settings, the main focus was to train individual team members to interpret and apply ePROM scores to treatment planning within a well-defined clinical process endorsed and supported by the healthcare organization. In tertiary and rehabilitation care, where there is more variability in care delivery processes, we needed to consider determinants related more to organizational change in addition to determinants of clinician and patient behaviour change. In these settings we used the Consolidated Framework of Implementation Research (CFIR) to evaluate clinician, individuals with chronic pain, and organizational factors that influence implementation of ePROMs.

In this paper, we present an overview of the use of each framework (Table 1) and how the use of these frameworks resulted in a final PROM implementation plan. The steering committee integrated results across both studies to inform PROM implementation and evaluation. The Proctor framework was used to guide the development of an evaluation plan for the implementation of PROMs. The details of the methodology and results for each study are presented elsewhere [2, 15, 16]. Here, we focus on the use of the frameworks and the resulting implementation plan to feedback PROM scores to patients and clinicians. Finally, we contrast the use of the TDF and CFIR frameworks, assess whether these frameworks resulted in identifying similar barriers, and present the final PROM implementation plan, in addition to the evaluation plan that will be used across all CECP settings.

The steering committee reviewed all information regarding the determinants of implementation of ePROMs identified across both studies and reviewed the objectives of the ePROM system implementation plan across the integrated pain network. The program objectives encompassed two broad categories. The first was adoption of PROMs, mainly, increasing feasibility and clinician and patient acceptability of ePROMs, and enhancing clinician and patient knowledge about ePROM score interpretation. The second category were objectives regarding quality of care and included increasing patients’ engagement and satisfaction with care, improving both patient-provider and healthcare team communication, informing shared decision making and treatment planning between patients and providers, facilitating goal setting and problem-solving, and increasing the focus of services on patient-centred outcomes.

The steering committee agreed to pursue the following steps for implementation at an organizational level. First, develop a business plan to incorporate the costs related to the ePROM system by directly integrating resources needed into the sites’ budget. This will be accomplished by examining potential cost savings associated with using ePROM scores to triage patients more efficiently and administrative time needed to complete paper-based forms for standardized and non-standardized patient information. The second organizational change recommended was to link ePROM scores to the clinic’s electronic medical record to facilitate integration of ePROM scores into the clinical workflow.

The steering committee defined the following process for facilitating implementation at the level of the clinical unit: administrative staff will send an email to patients containing the link to the ePROM questionnaire 2–4 weeks prior to the first visit or follow-up appointment. The patient will be invited to complete ePROMs at home 1 week prior to first visit or follow-up appointment. If PROMs were not completed at home patients will be encouraged by email and/or telephone to arrive before his/her appointment to complete ePROMS at the clinic.

The ePROM results will be shared with clinicians and patients through an online feedback report before the clinic visit. The feedback report includes patients’ scores over time for each measured domain, and the interpretation of each PROM score using graphs and text to indicate the current scores and the changes over time in each domain. Prior to the first visit or follow-up appointment, clinicians will be prompted to review the results in preparation to reviewing the scores with patients during the clinical encounter.

Evaluation of the determinants played a critical role in deciding on the implementation strategies that will be included in the development of a KT intervention that will increase the use of ePROMs. The process of deciding on which strategies to include in the KT intervention are presented elsewhere [2, 15, 16]. The main selected strategies were an interactive training program. The program will include educational/ instructional material on the selection, application, and interpretation of PROM scores and the clinicians’ and patients’ roles in using PROMs to plan treatment and monitor changes in outcomes in collaboration with patients. The program will also include a half-day training workshop on the use of PROMs in clinical practice that will include interpretation of feedback reports of individuals’ PROM scores.; The second strategy will be designating an opinion leader to provide clinicians with one-on-one coaching when needed to support the implementation of the KT intervention components and to provide coaching to clinicians on the use of PROMs. Table 4 presents the final interactive training program that will be adopted by the McGill RUISSS CECP.

Guided by the Proctor framework of evaluation, [13] the steering committee agreed on a set of outcomes that will be used to evaluate implementation and impact of using PROMs. The PROMs that will be implemented as part of patient care are those that are part of a CORE set prioritised by patients and clinicians [29] using PROMIS measures [30]. The framework “positions implementation outcomes as preceding both service outcomes and client outcomes, with the latter sets of outcomes being impacted by the implementation outcomes” [13]. The selection of outcomes was guided by the CECP objectives for using ePROMs, and the interrelationship of outcomes we expected to be impacted by adoption of the ePROM intervention. For example, we would expect to see the strongest impact on patient outcomes (e.g. general health, pain interference, work ability) as the ePROM uptake and sharing and discussion of scores with patients increases in a clinical setting. The Proctor framework derives service outcomes from the six quality improvement aims set out in the reports on crossing the quality chasm: the extent to which services are safe, effective, patient-centred, timely, efficient, and equitable [31, 32]. Table 5 presents the implementation and impact outcomes that will be used to evaluate the ePROM intervention across the McGill RUISSS CECP integrated network.

In this review of two studies, we demonstrated the use of the TDF and CFIR as guiding frameworks to evaluate determinants of implementation of ePROMs. As recommended by Birken et al. (2017), we explicitly specified the rationale for using the TDF or the CFIR to contribute to our understanding of each framework, its scope, and its limitations of using both frameworks to implement PROMs. Using theoretical frameworks facilitates identifying barriers and enablers to implement ePROMs in clinical settings and guides the development of theory-based interventions aimed at closing the gaps between knowledge and practice. The results of this study provide researchers and clinicians that are planning to monitor implementation of ePROMs an understanding of how the TDF and CFIR can be applied and highlight possible barriers and enablers that will likely need to be addressed.

The contrast and integration of results from both studies provided us with rich insight into the application of both approaches to the implementation of PROMs in clinical care. Similar to previous studies, our rationale for selecting the TDF or CFIR was driven by the purpose at the time for evaluating the determinants of implementation. The TDF was used in the context of an interdisciplinary team perspective, where each team member had to consider use of PROM scores for creating an intervention plan related to their discipline, and to link the resulting treatment plan to the overall program delivered by all team members in collaboration with the patient. The TDF targeted well the individual-level determinants of implementation that needed to be considered. Also, the clinical process of the LBP program was well defined and endorsed by the healthcare organizations where these programs were implemented as it was part of the provincial CECP Action Plan. Using ePROMs in these programs was endorsed and was an organizational expectation from the start. In contrast, we used the CFIR in rehabilitation and tertiary care where organizational-level factors were felt to play an important role in the adoption of ePROMs. The CFIR operationalized the individual characteristics of clinicians and patients (e.g. knowledge & beliefs), and the organizational and process factors that are critical in planning the implementation of PROMs.

Interestingly, both frameworks provided similar results with respect to healthcare providers’ knowledge and behaviour. These included the importance of considering clinician’s attitudes and perceptions of the benefits of using ePROM scores, such as believing they will have a positive impact on patients’ outcomes. The CFIR captured more barriers in relation to knowledge and beliefs including clinicians’ beliefs that PROM scores may not be comprehensive [of patients needs] and the perception that using PROMs may decrease patients’ satisfaction with care. Intention and motivation to use PROMs were identified in the TDF and CFIR study, respectively. These individual clinician characteristics identified in our studies are similar to those reported in previous work [34‐38]. Both frameworks were also consistent in identifying clinicians’ knowledge and experience in using PROMs, and difficulties with interpreting PROM scores as barriers, similar to previous work [35, 39, 40].

Where the TDF and CFIR frameworks differed was in identifying organizational-level determinants, which the CFIR was able to capture because of the explicit questions asked linked to the CFIR domains. This was reported in previous studies that contrasted the use of both frameworks for implementing healthcare interventions [41‐43]. Some consensus among coders in each of our studies required discussion, as some statements from the qualitative data were not easy to classify in either framework. However, this was more frequent for the TDF. For example, we felt that statements related to the relationship between clinician and patient were harder to map to the TDF framework. In the CFIR we classified “PROMs will be filled by patients if clinicians motivate patients to do so” as best fitting under ‘process’ in the CFIR, and there is no equivalent domain in the TDF.

In all settings we started with evaluating the current use of measures, what improvements in measurement were required, and asking clinicians and patients about their needs in terms of patient care to link use of PROMs to these needs [11, 44]. This was a key first step to understand the baseline practices of the sites and to tailor the knowledge translation interventions for ePROM implementation and training content to the readiness of the sites to begin using ePROMs. We also had knowledge exchange sessions with clinicians regarding the measures they were currently using and whether there were more efficient ePROMs with more comprehensive coverage of domains, stronger psychometric properties, and available evidence for interpretation of ePROM scores.

The resulting KT and training intervention elements will form a multi-component intervention, which can address multiple gaps and barriers [45], increase clinicians’ knowledge and skills, and foster best practices. A previous study [11] identified the most important PROMs as perceived by clinicians and patients that will form the core set of PROMs that will be included across all sites. As clinical programs learn from their experience with this core set, the use of ePROMs will be continually improved to include additional measures, linkages between ePROMs and electronic medical records, and optimization of the timing of administering ePROMs. The ongoing evaluation will be essential for informing ongoing improvements and sustainability of implementation and use of ePROMs.

Lastly, the importance of ePROMs feedback reports was emphasized across all our studies. All clinicians, to save time and facilitate the sharing of ePROM results among clinicians and patients, highlighted the format and integration of these reports in the electronic medical record. Feeding back the scores to clinicians and patients may increase their belief in value of using PROMs in clinical practice, improve the communication between patients and clinicians, and empower individuals with chronic pain to actively participate in managing their health condition, which may lead to better treatment adherence [46, 47]. We will qualitatively evaluate the determinants of ePROM implementation over time to iteratively improve ePROM implementation elements including feedback reports, KT and training, and integration of ePROMs across the integrated chronic pain network to facilitate uptake by each care setting that is added as we scale up the use of ePROMs to support patient-centred care.