Article Text
PDF
Statistics from Altmetric.com
Introduction
Health systems worldwide are focused on improving patient and population healthcare outcomes. Data are increasingly leveraged to support these aims, including clinical performance feedback (CPF) initiatives that provide clinicians with data on how often their patients are receiving evidence-based care compared with a specific improvement goal, which may reflect a standardised target or how they perform in comparison to their peers.1 2 Unfortunately, engagement with and impact of providing data is variable. We need to tailor data to the nature of its clinician recipients in the same way we tailor the nature of care to the individual patient. This will require continued investment in producing and analysing data and a better understanding of the clinicians whose behaviours we are trying to optimise. CPF initiatives that truly help clinicians achieve their goals in this regard are more likely to be well received. Imagine a future in which CPF is tailored for clinicians in the same way that Amazon, Google and Netflix personalise interfaces, informed by a range of data points and usability heuristics.
CPF initiatives focus their efforts on telling clinicians what the outcome should be instead of better understanding them as recipeients and the resources or skills they need to improve those outcomes. It should not be surprising then that upwards of one-third of clinicians who sign up to receive voluntary CPF do not engage with it.3 In order to effectively tailor CPF we need to couple ongoing investments in producing and analysing routinely collected data with insights about the people whose attention we are trying to capture and whose clinical behaviours we are trying to optimise. Such insights would form the foundation for tailoring our strategies that seek to improve quality of care.
The continued implementation of CPF initiatives with a one-size-fits-all mentality is a design flaw that is at odds with the theoretical underpinnings of CPF, outlined in Clinical Performance Feedback Intervention Theory (CP-FIT). CP-FIT—the first healthcare-specific feedback theory developed through a qualitative meta-synthesis of 65 feedback studies—emphasises interaction, perception, verification and acceptance of the data are foundational elements in the feedback cycle.4 We believe, and CP-FIT supports,4 that tailoring CPF to improve clinical performance requires those involved in the science and implementation of CPF initiatives to focus on three things: (1) clinician skills and characteristics; (2) the impact of organisational and professional culture; and (3) design feedback with existing workflows in mind. These foci map to the three variables that influence the feedback cycle as outlined in CP-FIT clinician variables, contextual variables and feedback variables, respectively.4 Table 1 provides an overview of the key challenges we have identified within each area of focus and points to potential evidence-based strategies to address them which will be discussed in greater detail below.
Challenges limiting the impact of clinical performance feedback and potential strategies to overcome them
Clinician skills and characteristics: understand the mindset of those who you need to engage
Our work in Ontario, Canada has repeatedly shown that most primary care physicians (PCPs) lack two key capabilities: (1) the knowledge around how practice-level CPF could help them and (2) how to identify an action in response to these insights.5 6 Clinicians’ inability to respond to CPF is a product of more than insufficient skills and supports,6 it is also a product of how clinicians think. In primary care, there is a disconnect between the prevailing mental model (where PCPs think about data on a patient-by-patient basis5 6) and how CPF is designed (a practice-level, retrospective view of performance); therefore, the relevance of an average outcome to a single patient is not readily apparent. Put another way, PCPs do not readily recognise the value in using retrospective data to determine what needs to be done prospectively for the current patient in front of them. Across all clinical practice areas, including specialty care, clinicians are often sceptical of the data and justify observed variation as a product of external factors beyond their control.7
Systematically categorising the knowledge and skills of clinicians is the first step towards achieving a targeted understanding of what needs to be addressed and how to ensure CPF is positioned to drive practice change. Some clinicians may be facile with considering organisational, proactive responses to aggregated, population-level data. Rather than expecting all other clinicians to shift their mindset to the needs of those designing CPF initiatives, it is incumbent on those designing the initiative to demonstrate how engagement will help clinicians achieve their own patient care goals. Framing CPF as a helpful mechanism to identify subconscious ingrained habits and heuristics might support clinicians in seeing how population-level data could inform patient-by-patient actions. Understanding how to tailor such framing could help increase engagement with—and the impact of—the initiatives. Once values align and clinicians are engaged, supports are needed to help close the intention-to-action gap. Action planning can be facilitated by identifying common barriers and proactively providing goal-oriented actions to overcome them—a strategy that has proven to significantly enhance the impact of CPF on clinical behaviour and patient outcomes.8 9 Leaders and those implementing CPF purposefully design and facilitate opportunities for clinicians to engage with their colleagues to share practices and identify those who may have discovered more effective strategies or approaches to care.10
While much attention has been given to how clinicians interact with CPF and understanding individual reactions to feedback,11 12 whether and how these reactions impact engagement has yet to be explored. Consider that the perception of feedback delivery (eg, whether the feedback is delivered in a non-judgemental manner) has a positive association with reactions to feedback among older clinician recipients, while feedback quality (eg, whether the feedback is relevant, specific, consistent and detailed) has a positive association with reactions among younger clinician recipients.11 Furthermore, studies have shown that in some settings, how care is provided and patient outcomes differ between male and female physicians.13–15 While the driving force behind these differences remains unclear, it suggests that gender may be a confounder for an unknown latent variable—perhaps attitudes, beliefs or clinical habits. If so, what is that variable, what is its mechanism of influence, and can that latent variable be influenced in turn? The goal is understanding what characteristics to tailor for, how and for which clinician subgroups (eg, tailor CPF differently for male vs female clinicians or by stage of career). Those implementing CPF can take action to tailor existing reports using codesigned cointerventions which can be balanced with embedded testing of different elements of design and delivery using simple A/B testing or more complex factorial designs to advance our understanding of what is most effective and for which clinician subgroups.
The impact of organisational and professional culture: culture will eat strategy for breakfast
A well-designed intervention is necessary but not sufficient to improve the quality of care; contextual factors such as organisational culture and climate,16 organisational mental models (cognitive representations of concepts)17 and organisational and system leadership18 all play a role. CPF interventions must ‘fit’ properly within their context to optimise impact. Additionally, given that audit and feedback interventions may lead to unintended outcomes, including reduced confidence and cognitive interference,19 20 it may be helpful to prospectively use a theory-driven approach to identify possible unintended outcomes and mitigate the likelihood these occur.20 Although common contextual factors limiting the uptake of evidence are well documented (eg, available resources, values, individual beliefs, past experiences),21–23 how to adapt for contextual variations remains unclear.
Culture strongly influences beliefs about feedback—specifically, the values, beliefs and priorities. Both credibility (trusting and believing in the source of the feedback) and constructiveness (the extent to which the data is valued and therefore perceived as useful and actionable) are influenced by medicine’s feedback culture,24 which creates a disconnect with the way system-level CPF initiatives are currently designed and delivered. The current emphasis on direct observation of clinical skills as the primary source of credible feedback and the distrust of system administrators and decision makers as a source of CPF will inherently undermine CPF initiatives. This raises the question of whether the interaction between clinician characteristics and culture is a missing link that helps to explain the heterogeneous effects of CPF.
Another contextual factor is the degree of psychological safety and its impact on feedback-seeking behaviours. Defined as the extent to which the clinician views the social climate as conducive to interpersonal risk,25 the presence of psychological safety is variable in healthcare settings,26 yet high psychological safety positively influences knowledge sharing,27 continuous quality improvement28 29 and learning from performance. Adopting a commitment-based approach to CPF whereby leaders model the desired behaviour(s) and encourage participation in quality improvement activities may positively impact psychological safety.30 Specifically, when leaders model openness and accessibility in their interactions, it facilitates learning from failure and improved performance at the team level.28 In contrast, low psychological safety may undermine feedback-seeking behaviours and the desire to engage in dialogue or activities that are perceived to monitor performance or challenge assumptions related to CPF. For example, highly-skilled musicians proactively seek out feedback on the specific aspects of a performance where they are not performing up to standards so that they can deliberately practise these aspects in pursuit of improvement—a practice that stems from the belief that feedback is vital to ongoing development.24 In contrast, healthcare does not always have a positive culture of improvement that focuses on how success is achieved rather than understanding errors.31 Feedback design requires an approach that considers both the goals of the clinician and broader culture in which learning occurs, specifically the degree to which feedback is valued and how.32 The medical profession has historically valued physician autonomy, independence and self-directed learning,33 and unlike other professional training environments, further work is needed to shift away from the perception that CPF is punitive to create the conditions where medical professionals are supported in interpreting and responding to feedback.24 34
Design with existing workflows in mind: understand how clinicians work and what they value
Tailoring helps account for individual differences but requires an understanding of who we are tailoring for, what matters to them and how the goals of CPF align to their professional goals. Once equipped with this information, we can then tailor content and messaging to help clinicians understand how CPF both fits with and supports them in achieving their goals. It is important that this be designed for and embedded within existing workflows and care pathways35—guidance that is rarely put into practice. Additionally, design should account for clinician preferences which can be achieved using a participatory design approach to help illuminate values and beliefs and inform cointerventions.8 9 36 For example, understanding that the initial reaction to poor performance is shame can inform how feedback is framed and communicated to clinicians.36 Investing time upfront to understand potential barriers to higher performance is also a worthwhile endeavour that can directly inform the nature of cointerventions implemented alongside CPF.8
CPF best practice also encourages the use of cointerventions such as social interaction (providing opportunities to discuss feedback with others)35; however, CPF is commonly deployed as a stand-alone strategy perceived to create accountability rather than support improvement. This may be an artefact of the earlier quality assurance approach, where practices were focused on using data for judgement, instead of a continuous quality improvement approach where practices are focused on using data for improvement.37 When cointerventions are deployed, they often attend to individual differences at the expense of contextual factors, or vice versa. Understanding how clinician characteristics and cultural context interact is a key element of effectively tailoring cointerventions. A cointervention responsive to both the clinician characteristics and cultural context might include coaching, a form of social interaction that focuses on helping the clinician grow their practice skills as a primary objective with the intended downstream effect being improved quality of care. Cointerventions could also be operationalised upstream to promote engagement with CPF, including leveraging opinion leaders, using reminders/prompts or providing clinicians with checklists.
Individuals and teams working to design and implement CPF interventions to improve patient outcomes should focus more purposefully on tailoring. CPF can be positioned as the type of strategy clinicians want and need to engage with by understanding who they are and what they believe and tailoring strategies to those insights. Similarly, understanding what clinicians are (and are not) capable of and the culture they are working within informs the type of cointerventions that will lead to action and improved patient outcomes. While we have outlined many factors here, understanding whether these factors are important, in what settings and what to do about them should be a central consideration. It is no longer about the problem of taking someone to water or making them drink, but rather the opportunity of understanding their thirst and having the right type of cup on hand.
Ethics statements
Patient consent for publication
References
Footnotes
Twitter @lauradesveaux, @ZRS_Rosenberg
Contributors LD, ZRSR-Y and NI all participated in the conceptualisation and writing of the manuscript. All authors read and approved the final manuscript.
Funding This study was funded by the Canadian Institutes of Health Research and the Canada Research Chairs.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.