Advances in Patient Safety and Medical Liability

Abstract

Introduction. Near-miss events, where no actual harm comes to the patient, represent a lower‑risk opportunity to improve patient safety and address patient expectations regarding disclosure of medical errors, both factors associated with medical malpractice claims. The Institute of Medicine and others have called for the creation of voluntary reporting systems to detect near-miss events to allow for analysis of patterns of errors; widespread adoption of near‑miss reporting systems in primary care could improve safety in that setting where more than 70 percent of medical encounters occur. Barriers to reporting near-miss errors include the additional workload burden imposed by a reporting system, concern over punitive action arising from a report, lack of confidence that positive change will result from such reporting, and psychological barriers to admitting an error. Furthermore, unless practices find this information useful to correct errors, near-miss reporting will be unlikely to become a routine procedure in ambulatory practice.

Objective. A primary goal for this project was to better understand barriers and facilitators of implementing a near-miss reporting and remediation system in primary care.

Methods. We designed and implemented an anonymous, near-miss reporting and improvement tracking system in seven diverse primary care medical practices to demonstrate that such a system could be successfully adopted in a variety of settings.

Results. In this paper we describe the reporting program, how it was implemented in the practices, the measured impact on practice and safety culture, and how practice leaders used the near-miss reports to improve care processes.

Conclusion. Near-miss reporting within primary care practices can help guide performance improvement activities and lead to meaningful changes in policy and practice. Successful implementation requires leadership commitment, incentives for staff, periodic reminders, and a system that allows reporting to be easy and anonymous.

Introduction

Near-miss events are errors that occur in the process of providing medical care that are detected and corrected before a patient is harmed. The Institute of Medicine (IOM) and others have identified near-miss reporting and analysis as vital to understanding and correcting weaknesses in the health care delivery system and to preventing actual adverse events (AEs) that harm patients.^1,2 Although the majority of health care encounters take place in ambulatory settings, most attempts to record and address near-miss events to date have been carried out in hospitals. Few examples of such efforts can be found in the United States to date, even though reporting systems have been successfully implemented in other countries.^3-5 At least one voluntary, anonymous reporting system has been established in U.S. primary care settings;⁶ others have demonstrated that near-miss reporting systems can increase awareness of patient safety.⁷

There are many potential barriers to reporting near-miss events in other care settings, such as hospitals, including the additional workload burden imposed by a reporting system, varying perceptions of staff responsibility for reporting, concern over punitive action arising from error reporting, lack of confidence that positive change will result from such reporting, and psychological barriers to admitting involvement in an error in patient care.^8-12 Factors cited to facilitate reporting include improved confidentiality, evidence that the reports would be used to improve patient care and not for punitive action, reporting forms that can be completed in 2 minutes or less, coupling reporting with an explicit process of quality improvement, and reminders to stimulate recall of near-miss events.^13-14

The primary goal of this Agency for Healthcare Research and Quality (AHRQ) planning grant (Medical Liability Reform and Patient Safety Planning Grant RFA-HS-10-022 Regional Ambulatory Near-Miss Reporting and Tracking to Improve Patient Safety) was to better understand the barriers and facilitators of implementing a near-miss reporting system in primary care practices. The North Carolina Office of Rural Health and Community Care (ORHCC), which was the grantee for this AHRQ pilot project, proposed to test a voluntary near-miss reporting and improvement tracking system across a variety of primary care practices and settings, including safety net providers, family medicine residency training programs, rural clinics, and small private practices. The goal was to narrow gaps in knowledge regarding near‑miss events—including patterns of near misses in different types of primary care practices—and factors that contribute to successful implementation of near-miss reporting; another goal was to observe how practice leaders use near-miss reports to address identified safety issues.

The specific aims were to measure differences in reporting rates between practices after a standardized orientation and common small incentive program, to measure changes in the practice and safety culture after near-miss reporting had been implemented, and to assess how practice leaders engaged in the near-miss reporting and subsequent practice improvement activities and their perceived value of this type of activity. By selecting a range of practices in terms of ownership, size, and urban and rural settings, the pilot was designed to be generalizable to a wide range of outpatient practices. A better understanding of these factors should inform efforts to adopt near-miss reporting and quality improvement more widely in the ambulatory care setting.

Methods

We recruited seven diverse practices to participate in the 1-year pilot project. The intervention itself had four components: (1) develop and conduct a standardized orientation for each practice with regards to reporting near-miss events; (2) collect and analyze near-miss reports from each of the practices for 6 to 9 months; (3) facilitate a practice improvement collaborative with the participating groups that included regular reminders and feedback to practice leaders about their reporting performance; and (4) provide a platform for the practices to establish and track efforts to remediate near-miss events.

The seven practices selected all belonged to the Community Care of Western North Carolina (CCWNC—formerly Access II Care) network and included two residency practices (which included three small rural satellite clinics), two safety net providers (one rural), and three private practices (two rural and one urban pediatric group). Together, these practices and sites employed more than 70 medical providers and 200 clinical support staff, provided over 2,000 office visits per month, and represented the full scope of primary care, including pediatric, geriatric, adult, and obstetrical care. Details of these practices are reported elsewhere.¹⁵

Prior to implementation, we attempted to collect anonymous surveys from all staff involved in clinical operations using the AHRQ Medical Office Survey on Patient Safety Culture¹⁶ to establish baselines. Staff were asked to complete the same surveys 6 to 9 months later, at the end of the official reporting period.

Each practice was enrolled in a rolling 3-month process that brought on one to three practices each month. The standardized orientation for each practice included a detailed overview of the project that was presented to leadership personnel at each of the participating practices, the goal of which was to develop an individualized implementation plan for their practice. A second presentation included all clinical and administrative staff and was intended to instruct staff how to recognize a near-miss event (as opposed to an adverse event or an error that had no potential for patient harm) and how to use the system to report these events as close to when they occurred as was possible. These included video instruction vignettes that were posted on the Internet and available for review from within the reporting system.¹⁷

After the initial orientation and each practice began reporting near-miss events, we sent electronic monthly diaries to practice leaders to gather ongoing snapshot views of how implementation was proceeding. We also provided monthly feedback through the electronic database and during the monthly conference calls to each practice about how many near-miss reports had been filed.

The near-miss reporting system was anonymous but allowed staff to include their name if they chose to do so. A previous trial of the system had determined that the 14-question form could be filled out accurately in most cases with minimal instruction in 2 minutes or less. Access to the online reporting system was through a desk-top icon that was installed on all work computers in each of the practices. The reporting system was structured so that the initial report was forwarded first to designated leaders in each practice who would review the report to make sure that the report was a near-miss (i.e., no harm came to the patient) and not an adverse event (AE, where the patient had suffered some harm). Only when determined to be a near-miss would the report be released into the general database. Each practice leader had access to all their own reports, but only those that originated in their practice; the evaluation team had access to the entire database, but the coders were blinded to the practice where the error had occurred.

Once practices had been reporting for at least 2 months, practice leaders were introduced to the near-miss remediation tracking program. The orientation to process improvement was based on the Plan-Do-Study-Act (PDSA) or Deming cycle; the practice was provided with a tracking tool integrated with the reporting system to allow practice leaders to assign each near-miss event (or a bundle of similar events) to an individual or group. The assignment included a “charter” outlining the problem to be addressed, expected outcomes, and a timeframe for reporting back to leadership. The program contained embedded references and links to help staff implement the PDSA model. The platform allowed the individual and/or group to keep a record of their steps and progress and provided automated email reminders to both those assigned the task and the manager(s) tracking the progress. Practices could choose to make this database open and viewable by other practice leaders in the collaborative.

We also conducted several conference calls with the practice leaders during this phase of the project to allow practices to share ideas about how to encourage staff to make near-miss reports, and how each was implementing the remediation process. Finally, at the end of the reporting period, we conducted structured focus groups with practice leaders at each of the practices.

Each of the practices was reimbursed the following amounts if they met the following monthly benchmarks:

1. Initiation of the project: Practices received $5,000 if they identified a core implementation team, participated in two planning meetings, made time for the all-staff orientation, and returned all of the baseline survey information.
2. Monthly near-miss reporting and participation goals: Practices then received $1,500 a month for 6 months if their practice reported at least 10 near-miss events a month and identified at least one near-miss event to remediate and track.

The total official reporting period for near-miss reports and initiating remediation projects was 9 months; depending on when the practice was enrolled, each practice had 6-9 months participation.

The near-miss reports themselves were reviewed by a team of six coders, all of whom went through a standardized training course. Using a previously developed taxonomy of ambulatory care errors developed in 2002 from a database of 344 near-miss reports from family practice,¹⁸ we classified each near-miss event and validated the coding with a second coder for 10 percent of the reports. The primary error was defined as “the breakdown in process, or knowledge/skill deficit that led to the problem.” Each reported error was coded with just one primary error but with up to four additional associated or cascade errors that could be included in the coding. In addition to classifying the type of errors, contributing factors, and possible preventive measures, the coders were also asked to provide a second rating of the potential seriousness of the event, the potential cost to the patient, and the estimated cost to the practice to remedy the problem. The potential cost to patients was defined as direct costs or lost wages; these were placed into one of three ranked categories: “None/minimal,” “some,” or “a lot,” with “a lot” defined as “a major financial loss for the patient.” Similarly, the estimated costs to the practice to remedy the problem were ranked, with “a lot” in this case defined as “a major investment in time or money.” After the events were coded, the coding and evaluation teams then developed a more succinct taxonomy of the errors contained in this contemporary database, eliminating codes of errors that did not appear and consolidating others that the coders agreed were equivalent events into a single code. This process was done by consensus of the primary coders, the primary investigator, the evaluation team, and the national consultant.

The pilot was reviewed by an institutional review board (IRB) and approved as a waived performance improvement project.

Results

Near-Miss Event Reporting by Practices

The practices were all largely successful in reporting near-miss events. The original target was for the combined enrolled practices to produce at least 300 near-miss reports over a 6-month period or seven reports per practice per month. The practices reported a total of 632 near-miss events during the official reporting period; most of the practices reached the minimum per month threshold. The mean number of reports per 1,000 patient visits was 9.1, with a range of 5.1 to 21.2 (Table 1).

Somewhat surprisingly, the practices continued to report near-misses even after the official reporting period and small financial incentives had ended; collectively, they reported an additional 138 near-miss events; by the end of the 12-month project period, 770 events had been logged into the database. Practices were allowed to continue to log-in and use the reporting and tracking software; reports continued to trickle in but at a declining rate for another year. Because the overall target of 300 near-miss events were reported by the group, all the practices were awarded their participation incentive as outlined in the methods section.

Table 1. Near-Miss Events Reports by Practice, per month and per 1,000 patient visits

Practices	All	A	B	C	D	E	F	G
Month enrolled	52	6	9	7	9	7	6	8
Visits during enrolled months	69,392	6,832	17,085	3,767	16,688	13,563	2,292	9,167
Near-miss reports	632	43	177	80	139	69	43	81
Reports per month	12.2	7.2	19.7	11.4	15.4	9.9	7.2	10.1
Reports per 1000 visits	9.1	6.3	10.4	21.2	8.3	5.1	18.8	8.8

The 632 reports that were logged in during the study period were coded and analyzed for type of event and seriousness. This analysis has been reported elsewhere.¹⁵

Impact of Near-Miss Reporting on Practice and Safety Culture

The participating practices baseline levels of office safety culture and those at 9 months into the project, compared with previously published cohort, are shown in Table 2. We did observe a five percentage point improvement in five of the twelve measures in the survey which corresponds to a “better than comparison” according to the scoring guide of this measurement tool. The percentage of staff completing the follow-up surveys was significantly below the baseline survey participation rate.

Table 2. Practice Safety Culture

Staff Opinions About Office Safety Culture	Percent With Positive Responseb		Comparison Dataa
	Baseline (n=141)	Follow-up (n=49)
Teamwork	73 percent	78 percent	82 percent
Patient care tracking and followup	68 percent	67 percent	77 percent
Organizational learning	66 percent	66 percent	74 percent
Overall perceptions of patient safety and quality	61 percent	56 percent	74 percent
Staff training	59 percent	57 percent	72 percent
Leadership support for patient safety	60 percent	68 percent	67 percent
Communication about error	58 percent	65 percent	67 percent
Communication openness	55 percent	59 percent	65 percent
Patient safety and quality issues	49 percent	42 percent	50 percent
Office processes and standardization	44 percent	42 percent	59 percent
Information exchange with other settings	49 percent	53 percent	54 percent
Work pressure and pace	38 percent	54 percent	46 percent

^a. Source: 2010 Preliminary Comparative Results: Medical Office Survey on Patient Safety Culture. (Prepared by Westat, Rockville, MD, under Contract No. HHSA 290200710037). Rockville, MD: Agency for Healthcare Research and Quality; November 2010. AHRQ Publication No. 11-0015-EF.
^b. Percentages in bold font differ by at least 5 percentage points from baseline. A change of 5 percentage points corresponds to ‘better’ or ‘lower’ than comparison data, per scoring guide.

Practice Leader Engagement

We observed very good practice leader buy-in to the project, based on focus group data and monthly online diaries. We did not identify any significant barriers to the implementation of this near-miss reporting system in this group of practices. Two of the seven practices with the highest near-miss reports per 1,000 patient visits appear to be those that provided direct incentives to staff for reporting, but one of the lowest also provided direct reporting incentives. The consensus from the group leader focus group was that the financial incentives did not greatly impact either the decision to participate in the pilot or to encourage staff to submit reports.

All seven practices initiated practice improvement projects based on the near-miss reports, ranging from 1 to 15, for a total of 34 near-miss-related formal projects reported in the tracking database. Of these projects, seven were followed through to completion, seven were placed on hold, four were inactivated, and the remainder were in progress at the close of the pilot.

The practice leader focus group revealed considerable variation in how practices approached remediation. Several of the practices had robust continuous quality improvement (CQI) processes in place prior to their participation in this project; near-miss reports were incorporated in some of these existing committees or efforts. Other practices had almost no formal practice improvement process in place at the beginning of the project; these practices appeared to use the near-miss reporting to begin address quality and safety in a more formal way. A common theme in the focus group was that practice leaders used near-miss reports to address and fix faulty processes immediately, based on perceived potential seriousness of the reported error. We observed that practice leaders were willing to share information with regards to remediating near-miss errors with others in the collaborative. We heard frequently from practice leaders that the mechanics of getting practices together at a particular time for a real-time conference call proved to be difficult and cumbersome.

Discussion

As previously noted, others have reported a number of potential barriers and facilitators to implementing near-miss event reporting generally. Our pilot built on a number of those facilitators, including anonymous reporting and using an online near-miss reporting form that had been previously tested and shown that it could be reliably completed by most users in 2 minutes or less with minimal training. The reporting process was coupled with explicit quality improvement activities in the practice. It was introduced practice-wide with a standardized staff training module that focused on the importance of near-miss reporting to create a safer care environment for patients and staff that should be undertaken in a non-blame management culture. Our results indicate that these facilitators appeared to overcome the other barriers of the small additional workload burden imposed by the reporting system, varying perceptions of staff responsibility for reporting, concern over punitive action arising from error reporting, and the psychological barriers to admitting involvement in an error in patient care. In fact, many of the near-miss event reports contained statements from those submitting the report that they felt personal relief bringing to light an error they had made in hopes that changes could be made in the care process to make such an error less likely to occur in the future.

The standardized training we developed for staff was focused and modular; although we presented the training in all-staff meetings (usually over lunch) aided with visual presentation slides, the training could easily be adapted to short online video training that could facilitate broader implementation.

We chose to make the reporting anonymous by default (although staff could indicate their name if they chose to do so) because we were unsure at the outset of the pilot how such reporting would be viewed and handled by practice leaders. This turned out to be a right choice because the participating practices appeared to have a marginally lower baseline culture of safety score than benchmark practices. In practices that have a well-established no-blame culture for error reporting, anonymous reporting is probably unnecessary and could restrict following up on a report with more specific inquiry. But, beginning with anonymous reporting until a culture is well established could facilitate reporting and the development of trust in the process of reporting.

We now also know that practices use the information from near-miss reports to improve care processes, even if they do so by using more informal methods. While some practices incorporated near-miss reports into existing formal quality improvement activities or created new teams to address certain flawed processes, managers reported that they also used information from near-miss reports to remedy problems identified by the reports in real time. Understanding how practices on a wider scale use near-miss reports to improve practice processes is a vital next step in the development of best practice strategies to implement safety improvements in ambulatory practice.

Reporting of near-miss events without establishing clear and transparent process improvement is not only a missed opportunity to utilize important information to improve care, but it sends a clear message that the information isn’t really desired. We observed that the practices that not only actively encouraged reporting in regular staff meetings but also included reports on what had been done in the interval to address the most serious potential errors had the highest rates of reports per 1,000 patient visits. Several practice leaders in practices that already had robust performance improvement processes reported that the energy generated around near-miss reporting appeared to invigorate their efforts to improve other aspects of care.

Practice improvement collaboratives have been demonstrated in other settings to stimulate practice transformation.¹⁹ Although practice leaders reported that the monthly conference calls were cumbersome and somewhat burdensome, we believe that the participating practices did respond somewhat to group norms established around success of the project. As the pace and scope of practice improvement work accelerates, better methods to exchange ideas quickly and efficiently should be explored.

Although this was a short pilot, and the gains in practice and safety culture were not large, implementing near-miss reporting and process improvement appeared to lead to positive changes in culture. A longer study with larger numbers of staff would be important to document if this observation holds, is durable, and to what extent it would reduce AEs, lower cost, and improve the patient experience.

Although we included a small financial incentive for practices to participate in the pilot, the incentives did not appear to be instrumental in determining if a practice participated or have a demonstrable impact on reporting. Although our sample practices were diverse in type, size, and location, the practices were more representative of early-adopter practices. Wider adoption of near-miss reporting and process improvement around safety may require more tangible financial incentives, such as those adopted by the Centers for Medicare & Medicaid Services (CMS) to incentivize outcome reporting through the Physician Quality Reporting System (PQRS).

Finally, a probable (and positive outcome) of widespread reporting of near-miss events would be to begin a conversation with patients about if and how to share this information with them. Ultimately, disclosure of errors in the care process, particularly those events where no harm came to the patient, represents an opportunity not only for care providers to become more comfortable with the process of disclosure when the stakes may be lower, but also an opportunity for patients to become partners in helping make the care system safer for them and others. Both developments would likely result in reduced AEs through improved safety but perhaps also fewer liability claims.

Limitations

The practices that participated in this pilot project were selected and not randomized. We purposely chose practices to represent a diversity of size, ownership, specialty, and range of clinical services, but these practices are not necessarily representative of “average” practices in the region. Also, because this was a pilot project, we wanted to keep the number of practices confined to our region where we could deal in person if issues arose during implementation. Finally, we allowed practices to use the modest financial incentives in any way they saw fit. While we tracked what each practice did to motivate staff, including some that used the financial incentive to reward reporting, we did not design the intervention to test which incentive plan was the most effective.

The pilot project was designed as a natural experiment to observe how practice leaders would use near-miss reports to address safety and quality issues revealed by this information. We did provide a tool and structure for tracking remediation efforts, limited training and resources for designing quality improvement activities around the near-miss reports, and provided opportunities for participants to share their experiences and ideas with other practices engaged in similar efforts. We did not provide detailed training or attempt to impose a rigid structure on the practices. Until there is further study, there does not appear to be sufficient evidence to support best practices for near-miss remediation in ambulatory practice.

Our pre- and post-surveys of practice and safety culture showed a drop in participation in the post-surveys, which may have skewed our results. Also, because there was only about 9 months between surveys, it is unclear if that is a sufficient time to measure culture change reliably. Finally, nearly all the practices experienced some turnover in staff over the course of the year. Observed changes in the culture measures could simply reflect staff turnover rather than a change in attitude among the same staff. A subsequent study should track individual staff assessments of culture over time.

Finally, we made great efforts to exclude AE reports from this database because of feedback we received in developing the pilot that practices were very reluctant or were excluded by legal counsel from voluntarily sharing information about AEs. As a result, we have no way of correlating the total of 770 near-miss reports we collected during the study with the number of AEs that occurred in the practices over the same period of time. This correlation deserves additional study.

Conclusions

This pilot demonstrated that a near-miss reporting and remediation system can be successfully implemented in primary care practices representing a range of size, ownership, and clinical mission, and that practice leaders find the information useful and appear willing to act on this information to improve care processes. Practices demonstrated consistent reporting numbers during the observation period with relatively small inducements to do so ($1,500 per month), through simple automated monthly email reminders to clinical staff, encouragement from practice leaders, and coupling reporting with a transparent process improvement process.

We believe these findings suggest that a larger, longer demonstration of near-miss event reporting should be undertaken to demonstrate the marginal, tangible benefit in lowering overall cost, reducing AEs, reducing liability claims, and improving the patient care experience. A larger, standardized database of near-miss reports could be valuable in identifying systemic problems that lead to errors and formulating solutions to reduce their incidence. A larger database of near-miss events could also be used in conjunction with root cause analysis from actual AEs to estimate systemic risk and to focus remedial efforts on the most pressing safety issues.

Finally, a large demonstration project could be designed to measure the marginal impact of practice transformation collaboratives in sharing data and information and whether this hastens innovation in practice improvement and to test if this model could increase the rate of adoption of best practices.

Acknowledgments

This project was funded by Grant No. HS19558 from the Agency for Healthcare Research and Quality. Key administrative support was provided by the North Carolina Office of Rural Health and Demonstrations, Community Care of Western North Carolina. Consultation on research methods and analytic support was provided by Generativity LLC. The Near-miss Reporting and Tracking system was designed and maintained by WindSwept Solutions (WSS).

Author Affiliations

Steven Crane, MD, Professor of Medicine, UNC-Chapel Hill; Assistant Director, Division of Family Medicine, Mountain Area Health Education Center (MAHEC); Medical Director of Primary Care, Mission Health System, Asheville, NC. Phillip D. Sloane MD, MPH, University of North Carolina, Chapel Hill, NC. Nancy C. Elder, MD, MSPH, University of Cincinnati, Cincinnati, OH. Lauren W. Cohen, MA, Duke University, Durham, NC. Natascha Laughtenschlager, MD, MAHEC, Hendersonville, NC. Sheryl Zimmerman, PhD, University of North Carolina, Chapel Hill, NC.

Address correspondence to Steven Crane, MD, Mountain Area Health Education Center (MAHEC), 709 North Justice Street, Hendersonville, NC 28792 or email.

References

1. Kohn KT, Corrigan JM, Donaldson MS, eds. To Err is Human: Building a Safer Health System. Washington, DC: Committee on Quality of Health Care in America. Institute of Medicine, National Academy Press; 2000; p 87.

2. Riley W, Liang BA, Rutherford W, et al. The Patient Safety and Quality Improvement Act of 2005: developing an error reporting system to improve patient safety. J Patient Saf 2008; 4(1):13-7.

3. Being open: communicating patient safety incidents with patients, their families and carers. National Learning and Reporting Service (Web site). National Patient Safety Agency, National Health Service England, Redditch, UK; 2009. Accessed March 14, 2017.

4. Hoffman B, Beyer M, Rohe J, et al. "Every error counts": a web-based incident reporting and learning system for general practice. Qual Saf Health Care 2008;17(4):307-12.

5. Bhasale, A. The wrong diagnosis: identifying causes of potentially adverse events in general practice using incident monitoring. Fam Pract 1998; 15:308-18.

6. Fernald D, Pace W, Harris D, et al. Event reporting to a primary care patient safety reporting system: a report from the ASIPS Collaborative. Ann Fam Med 2004; 2:327-32.

7. New York Chapter American College of Physicians. Near-miss project e-newsletter. Accessed March 14, 2017.

8. Elder NC, Graham D, Brandt E, et al. Barriers and motivators for making error reports from family medicine offices: a report from the American Academy of Family Physicians National Research Network (AAFP NRN). J Am Board Fam Med 2007; 20(2):115-23.

9. Rowin E, Lucier D, Pauker S, et al. Does error and adverse event reporting by physicians and nurses differ? Joint Comm J Qual Pat Saf 2008; 34(9):537-45.

10. Fisher MA, Mazor KM, Baril J, et al. Learning from mistakes. Factors that influence how students and residents learn from medical errors. J Gen Intern Med 2006; 21(5):419-23.

11. Chan KD, Gallagher TH, Reznick R, et al. How surgeons disclose medical errors to patients: a study using standardized patients. Surgery 2005;138:851-8.

12. Hingorani M, Wong T, Vafidis G. Patients’ and doctors’ attitudes to amount of information given after unintended injury during treatment: cross sectional, questionnaire survey. BMJ 1999; 318:640-1.

13. Garbutt J, Waterman AD, Kapp JM et al. Lost opportunities: how physicians communicate about medical errors. Health Aff 2008; 27(1):246-55.

14. Plews-Ogan ML, Nadkarni MM, Forren S, et al. Patient safety in the ambulatory setting. J Gen Intern Med 2004;19:719-25.

15. Crane S, Sloane PD, Elder N, et al. Reporting and using near-miss events to improve patient safety in diverse primary care practices: a collaborative approach to learning from our mistakes. J Am Board Fam Med 2015; 28(4):452-60.

16. Sorra J, Gray L, Famolaro T, et al. AHRQ Medical Office Survey on Patient Safety Culture: User’s Guide (Prepared by Westat, under Contract No. HHSA290201300003C). AHRQ Publication No. 15(16)-0051-EF (Replaces 08(09)-0059). Rockville, MD: Agency for Healthcare Research and Quality; June 2016. Accessed April 19, 2017.

17. How to open a near miss reporting form (video). Accessed March 14, 2017. How to complete a near miss reporting form (video). Accessed March 14, 2017. Role of the near miss project site manager (video). Accessed March 17, 2017. Role of the NM site director (video). Accessed March 14, 2017.

18. Dovey SM, Meyers DS, Phillips Jr. RL, et al. A preliminary taxonomy of medical errors in family practice. Qual Saf Health Care 2002; 11:233-8.

19. Newton W, Baxley E, Reid A, et al. Improving chronic illness care in teaching practices: learnings from the I³ collaborative. Fam Med 2011; 43(7):496-502.