A Modified Communication and Optimal Resolution Program for Intersystem Medical Error Discovery: Protocol for an Implementation Study

Background Preventable medical errors represent a major public health problem. To prevent future errors, improve disclosure, and mitigate malpractice risks, organizations have adopted strategies for transparent communication and emphasized quality improvement through peer review. These principles are incorporated into the Agency for Healthcare Research and Quality (AHRQ) Communication and Optimal Resolution (CANDOR) Toolkit, which facilitates (1) transparent communication, (2) error prevention, and (3) achieving optimal resolution with patients and families; however, how medical errors should be addressed when they are discovered between systems—intersystem medical error discovery (IMED)—remains unclear. Without mechanisms for disclosure and feedback on the part of the discovering provider, uncertainty remains as to the extent to which IMED is communicated with patients or responsible providers. Furthermore, known barriers to disclosure and reporting one’s own error may not be relevant or may be replaced by other unknown barriers when considering scenarios of IMED. Objective This study aims to develop and test implementation of a modified CANDOR process for application to IMED scenarios. Methods We plan a series of studies following an implementation framework. First, we plan a participatory, consensus-building stakeholder panel process to develop the modified CANDOR process. We will then conduct a robust preimplementation analysis to identify determinants of implementation of the modified process. Using the Consolidated Framework for Implementation Research as a theoretical framework, we will assess organizational readiness by key informant interviews and individual-level behaviors by a survey. Findings from this analysis will inform the implementation toolkit that will be developed and pilot-tested at 2 cancer centers, sites where IMED is hypothesized to occur more frequently than other settings. We will measure 5 implementation outcomes (acceptability, appropriateness, reach, adoption, and feasibility) using a combination of key informant interviews and surveys over the pre- and postimplementation phases. Results This protocol was funded in August 2018 with support from the AHRQ. The University of Michigan Medical School Institutional Review Board has reviewed and approved the scope of activities described. As of April 2019, step 1 of aim 1 is underway, and aim 1 is projected to be completed by April 2020. Data collection is projected to begin in January 2020 for aim 2 and in August 2020 for aim 3. Conclusions Providing a communication and resolution strategy applicable to IMED scenarios will help address the current blind spot in the patient safety movement. This work will provide important insights into the potential utility of an implementation toolkit to improve transparent communication and optimal resolution of IMED scenarios. The natural progression of this work will be to test the toolkit more broadly, understand the feasibility and barriers of implementation on a broader scale, and pilot the implementation in new organizations. International Registered Report Identifier (IRRID) PRR1-10.2196/13396

HCRT DOSSETT, L educational plan with training that will be essential for successful completion of this research and for Dr. Dossett's career development. The training includes mentored research and graduate level courses in implementation science. This career development award will lay the groundwork for Dr. Dossett to perform ongoing, innovative health services research, and to become an independent investigator. PUBLIC HEALTH RELEVANCE: This project will assess potential strategies to facilitate transparent communication and optimal resolution of errors discovered between facilities. It will also evaluate whether these strategies are acceptable and feasible for implementation. The results will have immediate impact improving patient safety and improving cancer care quality.

CRITIQUES:
The written critiques of individual reviewers are provided in essentially unedited form below. These critiques were prepared prior to the meeting and may not have been revised afterwards. The "Resume and Summary of Discussion" above summarizes the final opinions of the committee. Overall Impact: Strengths  This uniquely qualified, well trained surgeon-scientist has a stellar mentoring team with a well elaborated training and research plan. Her background as a naval surgeon informs her career and research goals in important and productive ways. Her mentors are hands on with outstanding track records of mentoring. Her history of research on cancer outcomes, health care delivery, and provider performance offer a good foundation for her planned research. The question of how physicians communicate across organizations when a clinical mistake has been made is important and not well studied. Her research includes quantitative and qualitative methods, as well as a plan to pilot an intervention. Recognizing the requirements of this work, she has assembled mentors and consultants with a unique set of skills (e.g. outcomes research, mixed methods research, implementation science, and bioethics). Her training across several and research aims to first understand barriers to physician-to-physician communication. Her work is novel and important. Her plan overall is well structured and thoughtful. The environment at Michigan is ideal for promoting the careers of K researchers.

Weaknesses
 Her approach lacks detail on the sample for the survey and quantitative analysis and does not directly address the bias introduced by Michigan's special environment. These are really the only minor weaknesses and they are not out of line from what I'd expect from someone at this career stage, and the team shows other ample evidence of strong and helpful input throughout the application. So I down-weight these weaknesses. My only concern is that she may be too ambitious, though she has packed in an impressive set of experience and training already, so this is negligible.

Candidate: Strengths
 Candidate's background as a naval surgeon offers unique perspective especially around managing risk, the topic of her training and research.  Her clinical training, prior analytic work, and research prepared her well to pursue work that reduces risk.
1 K08 HS026030-01A1 4 HCRT DOSSETT, L  She has assembled a strong team of mentors, which speaks to the promise other seasoned researchers see in her.  She already shows ample evidence of collaboration with her Michigan mentors even though she only arrived in 2016.  Her research and training plan (as a whole) fit together very well.  Since first submission, applicant and team published "Barriers to Disclosure and Feedback of Errors between Facilities," in July 2017 Annals of Surgery. Based on qualitative analysis of 30 semi-structured interviews with cancer specialists. Weaknesses  None noted.

Career Development Plan/Career Goals and Objectives: Strengths
 Developing mixed methods and implementation science skills is a good fit for her goal of "moving evidence into practice."  Good access to appropriate coursework on implementation science (2 formal courses) and qualitative mixed methods (2 formal courses and learning lab).  Good path to R01 with R01 boot camp.

Weaknesses
 Too broad across several complex areas: primary data collection via surveys, qualitative data collection, and implementation science each could be their own focus of a K alone.

Research Plan: Strengths
 Clear and strong scientific motivation for the project -errors in cancer care (as all medical care) are common yet there is no way to handle such errors when they cross facilities. The training and research aims all work towards addressing this need.  By asking clear questions (i.e. what are barriers to communication of other physicians' errors between facilities?), the PI is likely to elicit useful and actionable information. She has greatly clarified the plan for each piece of her research aims.  To modify the CANDOR process for use across facilities -applicant proposes highly structured plan combining evidence synthesis, independent review by stakeholder panel, face-to-face meetings of stakeholder panel, and summary of final best practice recommendation (Aim 1).  5 NCI-designated cancer centers targeted for 20 semi-structured interviews and 250 cancer specialist survey respondents (50 per site) to understand how inter facility CANDOR barriers/implementation differ from within facility CANDOR.  2 pilot sites, Michigan and Moffit where PI and team have deep connections, and therefore high likelihood of success with proposed research.  Aim 3 more narrowly focused around implementation outcomes of feasibility and acceptability with more clearly detailed design. Use of toolkit (aim 2) and input from stakeholders to implement at 2 sites and measure well defined implementation outcomes including adoption (month 4-6), appropriateness, reach, acceptability, and feasibility (month 12).

Weaknesses
 The work proposed remains ambitious and any unanticipated delays outside the PI's control will make it impossible to complete the proposed work as each aim flows from the previous aim(s). This is a minor to moderate weakness balanced by the significance of the work, strength of the candidate and team.

Mentor(s), Co-Mentor(s), Consultant(s), Collaborator(s): Strengths
 Exceptionally strong primary mentor (Dr. Dimick), himself a surgeon-scientist, with ample NIH funding and with five past K mentees, Fetters), implementation science (Dr. Sales), and survey research (Jagsi).  Space in close proximity (Dr. Jagsi) facilitates formal and informal meetings.  Mentoring roles are clearly laid out.  Very strong team of co-mentors with mixed-methods experience, and one with implementation science, and a consultant who developed "The Michigan Model" of error disclosure and health system claims.

Weaknesses
 Only some of the mentor letters were updated. Notably, Dimick (primary mentor) and Mulholland, chair of surgery, were not.

Degree of Responsiveness:
Strengths  The investigator clearly fits the goals of the K08 as an early stage investigator who needs additional skills and mentoring to achieve her career goals around mixed methods and implementation science work to improve physician to physician communication regarding errors. Weaknesses  None noted.

Inclusion of AHRQ Priority Populations: Strengths
 Acceptable -the ultimate target of the work, oncology patients, span priority populations, but the immediate target, is their care providers.

Weaknesses
 None noted.

Strengths
 Generally appropriate with minor concern below. Weaknesses  The primary data collection plus focus groups and coding seems like it could require more resources than available under a K, but could be supplemented (with her startup funds of $180k).

Inclusion of Women and Minority Subjects: Strengths
 Acceptable -the ultimate target of the work, oncology patients, span both sexes and minorities, but the immediate target, is their care providers.

Weaknesses
 None noted.

Resubmissions: Strengths
 The applicant is highly responsive to reviewer critiques yielding a clearer, streamlined plan for training and research. Specifically  In response to diffuse and over-ambitious training goals, she has narrowed the scope to focus on stakeholder analysis, development of an implementation strategy, and implementation outcome assessment. She has completed 3-day training in mixed methods.  To address lack of generalizability (Michigan is a unique setting with outsized attention to surgical quality), applicant proposes additional NCI-designated regional referral cancer centers to purposively sample 5 of 69 NCI-designated centers in Aim 2, and added Moffit Cancer Center in Tampa, FL for aim 3.  To address the lack of details regarding sample frame, target survey respondents, recruitment strategy and content in Aim 2, applicant proposes targeting 50 cancer specialists from 5 sites (250) with a target RR of 65% (163), in line with successful recruitment by her mentors.  To address need for specific implementation outcomes, applicant proposes primary outcomes of acceptability and feasibility and secondary outcomes of appropriateness, reach and adoption. Applicant will use dissemination and implementation Toolkit from Aim 2 to implement I-CANDOR at 2 study sites over 12 months. Weaknesses  None noted.

CRITIQUE 2
Candidate:  Innovative question, as a framework for reporting between-institution errors isn't available.
Approach adapts a well-known CANDOR process with stakeholder input to understand barriers and facilitators to its use based on CFIR, in preparation for developing and testing an intervention.

Weaknesses
 The proposal depends highly on the ability to apply CANDOR, which helps address provider and local systems factors when an error occurs, as a model for addressing errors recognized by others, from other institutions, in hindsight. I miss an analysis of the accuracy with which some errors can be identified given limited information on decision-making.

Candidate: Strengths
 Surgical oncologist with clinical practice experience, focus on patient safety and communication between providers. Obtained her MD and surgical residency at Vanderbilt, then AHRQ T32 training and an MPH with quantitative training, and pursued a surgical oncology fellowship after 4 years of military service.  Strong candidate with expertise in statistical modeling and large database analysis as applied to studies of practice variation and provider performance, clinical outcomes re: surgery for cancer; well published to date. Since the initial application, she also participated in a 3-day participatory mixed methods workshop.  Strong letters of support from her mentor team attest to an extraordinarily promising candidate for a health services research career.

Weaknesses
 None noted

Career Development Plan/Career Goals and Objectives: Strengths
 Current plan builds on quantitative expertise and preliminary study of PCP/cancer specialist communication to build formal training in implementation science to study error communication between providers at different institutions, with goal of developing and piloting an intervention that can be formally tested through an R01 in the future.  The plan includes formal coursework in implementation science (2 courses), targeted readings and seminars. Plans for monitoring and evaluation are presented with weekly meetings with Dr. Dimick, twice monthly with Dr. Jagsi, and monthly with other mentors -all will meet biannually for feedback, concerns; annual modification of timeline and action steps.  Project is of priority to AHRQ, as builds on the CANDOR model to improve communication about errors with a goal to improve patient safety. Weaknesses  None noted.

Research Strategy: Strengths
 Problem statement and rationale suggest need for inter-facility identification and communication re: medical errors, especially in the complex care of cancer patients, and current lack of professional guidelines on this topic.  The topic is highly relevant to her previous work and career goals.  Her aims include 1) to develop a guideline for best practice re: transparent communication and resolution of errors between facilities using an expert panel process; 2) conduct a mixed methods study with stakeholders to understand barriers and facilitators to achieving the best practice from Aim 1; and 3) Develop and pilot an intervention adapted from the CANDOR toolkit.  Conceptual models to be used are clear, and methods for Aims 1 and 2a are well described and appropriate where possible. 1 K08 HS026030-01A1 8 HCRT DOSSETT, L  Aim 2 was expanded to include additional NCI designated cancer centers, in addition to UMich, and greater details for the survey administration provided -this seems clear.  Though content of the intervention is pending results from Aims 1 and 2, the application now includes the framework for evaluating the outcomes of the intervention.

Weaknesses
 I have questions re: the ability of providers to accurately assess (at least some) errors of others given limited information on context and decision-making, and of the application of CANDOR, which helps address provider and local systems factors when an error occurs, as a model for addressing errors recognized by others, from other institutions, in hindsight. HCRT DOSSETT, L Privacy and Security Protections for Patients: None noted.

Resubmission Applications (formerly "revised/amended" applications): Strengths
 The resubmission is responsive to reviewer comments. She has narrowed the scope of training to focus on implementation science, as there was concern that combining with mixed methods training was overly ambitious, and she has obtained some additional mixed methods training in the meantime. Additional NCE-designated cancer centers were added to address concern of the unique nature of the UMich site, and further developed the methodologies for the survey proposed in Aim 2, and the Framework for evaluating outcomes of Aim 3. Weaknesses  None noted.

CRITIQUE 3
Candidate: Overall Impact: Strengths  The research plan has been substantially re-considered and the candidate was responsive to reviewer comments  Great candidate who will likely be successful.

Weaknesses
 Success in the research project rests heavily on I-Candor success.  There is still a fair amount left to imagination-how "errors" are defined, what the preliminary thoughts are on how an inter-hospital response teams works.  Philosophically, it is not clear what the impact of an error-reporting system is in the absence of other communication mechanisms intended to facilitate dissemination of knowledge and best practice. This seems rooted in a perceived notion of a two-tiered system of those who are knowledgeable and those who are not, rather than one in which a referral facility extends its reach to improve knowledge of cancer care system wide. Who is the "judge" of whether there is a difference in opinion or difference in practice  There is no comment on the effect on patients---care should be taken to ensure this is managed. For example, for CANDOR, when a mistake is recognized, there is a framework for immediate disclosure. When one MD/health system blames another, they may be right or wrong. How is the impact of this managed?

Candidate: Strengths
 Has demonstrated a commitment to discovery as evidenced by time spent acquiring research skills, and through continued publication despite an 8-year break to complete clinical obligations between research years and beginning a faculty job. Weaknesses  None noted.

Career Development Plan/Career Goals and Objectives: Strengths
 Revisions now include additional training and a clearer development plan.