Discussion & Implications of the MD-PIE model
Our model is complementary to the R2C2 model and would be augmented by the application of a similar standardized feedback modality5. The R2C2 model provides a schema for feedback interaction and allows for the exploration of understanding of the clinicians. The “reaction” portion of this model is where most of our model would reside. Our model would provide perspective to the interviewer on where most physicians will fit, relative to the individual practitioner’s mindset and the quality of the data system. Our model would then provide the interviewer with an opportunity to map a method by which an individual could come to reflect on the content of the feedback with an opportunity to grow, feeding directly into the Content portion of R2C2. Our model may also identify opportunities for improvement of the data presented in feedback, allowing for iterative change and improvement on the data used for R2C2. This reliance on a robust data model is not explicitly discussed in the R2C2 framework.
When compared to other group feedback work using R2C27,8, our model places similar emphasis on high quality data to enable appropriate discussion. The CAFF model provides outstanding direction for practitioners who have some modicum of interest in professional feedback loops relative to their own practice data. Our model describes how these interactions might work across many practitioners of a variety of growth mindsets. Our model also describes how robust data may contribute to bringing individual practitioners from helplessness to engaging with self-data feedback loops and allowing for more robust discussion in the Content and Coach areas of the R2C2 model.
When comparing our model to Kegan and Lahey’s Immunity to Change (ITC) model12, we find that these reconcile very well. Their published works on assessments for competing commitments should be performed most readily within individuals found in the helplessness and disengagement territory. There is also a portion of individuals in the data resistance category that may be choosing not to engage secondary to competing commitments. Those in the data invalidity category may receive a staggered approach to involvement; first, by understanding their suggestions for the improvement of practice measurement and implementing some change where warranted and second, by subsequently considering the application of the ITC framework if change remains difficult.
Lastly, a comparison of our model to the Deliberately Developmental Organizations (DDO) framework, also by Kegan and Lahey13, was warranted. DDOs operate under the inherent belief that investing in the development of each practitioner within the organization is in the best interest of the organization’s own development and growth. This speaks to an overall mindset of improvement and adaptation to the current practice environment. The application of the DDO model to our framework would likely shrink the data validity and resistance areas and push more individuals into the systemic support region, whereby the department itself has a process in place to ensure individual improvement for all of its members. A DDO organization would seek to hire and push everybody into the systemic supports area.
This model provides a rubric through which academics, researchers and administrators may understand the interplay between an individual’s growth mindset and the data systems providing quantitative practice feedback. The model consists of a “population-level” framework which should be applied to physician groups. Within our interview cohort, important themes were raised that reflected the perspectives of specific sub-groups which could provide a more nuanced approach to data discussions and improvement. These nuances were not explicitly included in the model as they likely require additional research.
This also provides some guidance with the operational deployment of audit and feedback frameworks as it informs those conducting the implementation to expect a population who will resist this change throughout yet also reinforces the need for robust data systems to augment and enable improvement. This also allows administrators to integrate the concepts of early adopters, bystanders and nay-sayers8 with the data systems that may enable downstream improvement across categories.
Discussion on expectations of practice uniformity
In discussing A&F, our work shows that there may be individual reluctance to engage with data due to perceived risk on physician autonomy and expectations for increased uniformity. This highlights both the need for appropriate metrics and performance comparisons. In using evidence-based and patient-relevant metrics, physicians may come to see the utility of performance measures beyond simply standardizing clinical decisions. Specifically, using metrics for which there is a gold standard has previously been shown to have better feedback acceptance. Prior work by Gude et. al14 also recommends using tailored performance benchmarks, as opposed to standardizing against the group mean, to make the metric relevant and comparable. Individualized feedback thus becomes an important part of the feedback cycle to help recipients set personal targets rather than seeing metrics as an externally imposed performance comparator.
Finally, our model defines the various groups that may exist along the spectrum of data engagement. There are important considerations with regards to how these groups interact within one department. The Calgary Audit and Feedback Framework7, stemming from the Social Learning Theory, emphasizes that efficient learning can occur through reflecting, sharing practices, and discussing best evidence within the group. Based on our model, we anticipate that a department with individuals who prioritize audit and feedback can positively influence others to reflect on improving their practice in the setting of best evidence.