The Case

A 63-year-old man with history of stroke, systolic heart failure, and ventricular tachycardia with a pacemaker in place presented from a skilled nursing facility (SNF) with shortness of breath. He was treated with two days of intravenous diuretics for mild heart failure exacerbation and an echocardiogram was performed. The results of the echocardiogram were pending on discharge, with anticipation that the patient’s primary care provider would follow up the results.

The patient was readmitted from the SNF two weeks later and was found to have endocarditis and infected pacemaker wires. The admitting physician reviewed the echocardiogram done on the prior hospitalization and noted there was a vegetation on the tricuspid valve, which was an unexpected finding. On review, the echocardiogram results had populated into the electronic health record after the patient was discharged and the new vegetation was not flagged as a critical finding. No providers were contacted about this finding by the cardiologist who read the echocardiogram. The results were sent to the inbox of the ordering resident, who was not on the primary service taking care of the patient. The resident did not check his inbox. The primary service taking care of the patient at the time also did not follow up the patient’s echocardiogram given that he had already been discharged to a skilled nursing facility when the finding was noted. After the patient was readmitted with endocarditis, he had a complicated hospital course resulting in death.

The Commentary

By Noelle Boctor, MD, and Mithu Molla, MD, MBA

This case describes a 63-year-old man with multiple medical comorbidities and an indwelling pacemaker who underwent echocardiography while in a mild heart failure exacerbation during a hospital admission. The echocardiogram revealed a new vegetation on the tricuspid valve, concerning for infective endocarditis. However, this critical finding was not reported to any providers involved in the patient’s ongoing care. As such, the patient’s condition progressed while at a SNF until he developed fever and overt infection of his pacemaker wires. If the vegetation had been recognized earlier and treated aggressively with intravenous (IV) antibiotics and possible pacemaker lead removal, the patient’s rapid clinical decline and death may have been prevented.

Approach to Improving Safety & Patient Safety Target

Swiss Cheese Model

Missed diagnoses are one of the most common types of medical errors in the outpatient setting, and sometimes result when the interpretations of imaging or other diagnostic tests are delayed or misdirected. Several studies have found that 9-25% of hospitalized patients experience adverse events and about half of these are potentially preventable.^1-5 Many of these events result from the combined effect of multiple weaknesses or vulnerabilities in flawed systems. The Swiss cheese model has been used to describe how errors made by individuals, resource constraints, and other circumstances can result in harm to patients because we operate in flawed systems, which create the holes in the “Swiss cheese”. When enough of these holes or weaknesses align, it can lead to devastating consequences, such as in the patient’s case described above.

Because harms can result from complex and evolving conditions, including holes that open or close with transient local circumstances (e.g., distractions, excessive cognitive load), “deconstruction” to find the “causes” of an event is not always accurate or meaningful. Nevertheless, root cause analysis of this case revealed several figurative Swiss cheese holes that aligned:

1. the patient had already been discharged to a skilled nursing facility when the echocardiogram results populated, creating obstacles for effective communication;
2. none of the primary providers on the team followed up the echocardiogram results;
3. the resident who had ordered the echocardiogram was not on the primary team caring for the patient and did not check his inbox;
4. the cardiologist’s dictation did not draw attention to the concerning finding of a new, large vegetation; and
5. the results were not flagged within the electronic health record (EHR) and were therefore easily overlooked.

Adopting this systems-based model for evaluating adverse events, we can see how the holes or gaps in care aligned and led to the delayed diagnosis of infective endocarditis and, ultimately, an unfortunate death. We can then attempt to strengthen overlapping layers of care in the hospital environment to eliminate some of the holes and thereby minimize the number of errors that affect patients.

Adverse events during transitions of care

Perhaps the biggest hole in our figurative Swiss cheese model in this case involved the transition from hospital to skilled nursing facility. Transitions of care increase the risk of adverse events due to the potential for miscommunication as responsibility is given to new parties. The hospital discharge is a complex process representing a time of significant vulnerability for patients.⁶ Incomplete tests that require further diagnostic workup once results are available are one of several areas that contribute to unsafe transitions of care.⁷ Adverse events span a multitude of patient care domains and almost 40% of patients are discharged with test results that are still pending.⁸ In teaching hospitals where several team members can order tests, it is imperative to build systems that will easily remind and allow whoever is responsible for the patient’s ongoing care to follow up results in a timely manner.

Recommendations that can plug the holes in the Swiss cheese

Use checklists and structured handoffs

Standardized, structured handoffs have been shown to improve communication and patient safety.⁹ With the work-hour restrictions of the Accreditation Council for Graduate Medical Education (ACGME) implemented in 2011, post-graduate residents are limited in the number of hours they can care for patients. This has increased the number of handoffs, which has been associated with more medical errors.¹⁰ As a result, a structured method for handoffs is essential. I-PASS is a handoff program that decreases medical errors and preventable harm; the I-PASS mnemonic is defined as illness severity, patient information, action list, situational awareness and contingency plans, and synthesis by receiver.¹¹ While I-PASS is designed for handoffs within a healthcare organization, this structure can be extended to handoffs between organizations. If the EHR is shared between the discharging and receiving facilities, the I-PASS tool can be easily referred to in a pre-designated location within the patient’s record. When the receiving facility has a different EHR, the same I-PASS tool in written format (attached to the patient’s discharge summary) can standardize handoffs, ensuring shared responsibility for the transition of care. To ensure thorough communication, the discharging team should receive confirmation of receipt of the discharge summary from the accepting team/facility.

In addition to using a structured handoff system such as I-PASS, the discharging team in this case could have followed up on the echocardiogram results if it had been explicitly included as a task on their checklist. The bystander effect is a known human tendency to be less likely to offer help in critical situations when others are nearby.¹² Although the research on the bystander effect is based primarily on laboratory studies, the effect has also been demonstrated in real-world situations in which responsibility is diffused among multiple persons.¹² Making the expectation for follow-up of the echocardiogram explicit or assigning the task to an individual on the team can help overcome this tendency. In patients with complex medical problems requiring multiple laboratory tests, checklists can act as a useful cognitive aid in ensuring that tests are followed up on and diagnostic errors are avoided. One systematic review concluded that checklists targeting specific tasks in the diagnostic process (e.g., interpreting chest radiographs or electrocardiograms) appeared to be more consistently associated with reductions in diagnostic errors (in five of seven published studies) than checklists targeting cognitive factors (i.e., only four of ten studies demonstrated reductions in diagnostic errors).¹³ Perhaps creating a checklist of follow-up or action items that could be used at both the individual provider and team levels could have avoided the echocardiogram results being lost to follow-up in this case.

Utilize Patient Navigators

Some hospitals have attempted to bridge potential gaps during transitions of care by using providers dedicated to transitions of care, or patient navigators. Patient navigators are an example of a skill-mix innovation in which new tasks or ways of working are implemented. Navigators come from various backgrounds: they can be health professionals, such as nurses or social workers, or trained lay persons. Key roles of patient navigators vary depending on their individual skills and experience. Typical tasks might include identifying the needs and barriers to care of individual patients, educating patients and communities, and linking patients with different providers. Professional patient navigators may carry out more advanced clinical tasks¹⁴ thereby helping to ensure timely access to healthcare services as well as completion of diagnostic and follow-up care.¹⁵ Most of the outcomes measured in randomized controlled trials of the effects of utilizing patient navigators reflect the process of receiving care. Few studies have assessed patient-oriented outcomes and many studies involved only a short duration of follow-up and lacked enough power to detect effects on clinical outcomes.¹⁶ Additionally, there remains significant overlap in the functions and roles of patient navigators versus those of other healthcare providers, such as case managers. Establishing standardized definitions detailing scope of practice and allowing for flexibility across different settings may improve service delivery.¹⁷ Following up on pending tests or laboratory studies and informing providers of results may fall within the scope of patient navigator practice, a role that could have helped draw attention to the vegetation finding in this case earlier after patient discharge and transfer back to the SNF.

Adopt closed-loop communication for critical or important results

The lack of communication of critical imaging results comprised a substantial hole in the Swiss cheese model of this patient’s care. Closed-loop communication (CLC) of results, especially those that are critical or important, should be prioritized to prevent potential harm to patients. This communication model originated with military radio transmissions and is based on verbal feedback to ensure proper team understanding of a meaningful message.¹⁸ CLC is a three-step process in which 1) the transmitter communicates a message to the intended receiver, utilizing their name, when possible; 2) the receiver accepts the message with acknowledgement of receipt via verbal communication, seeking clarification if required; and 3) the original transmitter verifies that the message has been received and correctly interpreted, thereby closing the loop.¹⁸ CLC is currently utilized by health systems to effectively report critical lab results and radiologic reports but barriers remain such as inability to locate or identify the treating physician or lack of standardization of critical results that require CLC.

Hospitals should implement policies that categorize different diagnoses as critical, important but non-urgent, and routine, and include corresponding time frames for the diagnosis to be communicated to a provider taking care of the patient. Radiologists, pathologists, and other physicians interpreting laboratory studies and images should have access to the list of findings in the critical and important categories in order to reduce risk of clinical judgment error. For echocardiograms, critical findings may include large pericardial effusion with tamponade physiology, a vegetation measuring over 1cm (such as in this patient’s case), or severe valvular regurgitation. To decrease the workload of front-line care providers, process improvements that utilize standardized software programs can be implemented,¹⁹ for example using a computer or artificial intelligence program to identify terms in the radiologist’s interpretation of a study that meets the critical or important criteria. These identified findings should then be reviewed by a physician, reducing additional effort required from the interpreting provider (the radiologist in this example) and decreasing chances of overlooking important findings. To ensure timely action, critical findings should be communicated verbally via CLC to a physician caring for the patient. If the patient has already been discharged, the interpreting physician could contact the patient’s primary care provider or the patient. Confirmation of communication should be noted in the patient’s medical record. Successful implementations of this practice have been published; for example, the four-year impact of an alert notification system on CLC of critical test results documented an improvement from 90.7% to 97.9%.²⁰

Call attention to important results with documentation and EHR enhancements

In addition to CLC for critical results, documentation practices and EHR enhancements may also be utilized to help prevent these types of diagnostic errors and call attention to critical findings. Bolding or highlighting important findings in the report or including findings in a section of the report titled “Important findings,” will help draw attention to key findings and decrease chances they will be missed. For example, bolded statements such as “Large, mobile 1cm vegetation seen on tricuspid valve” draw the reader’s attention. Use of the EHR to alert the provider that there is a concerning finding on the echocardiogram, such as highlighting or placing a red exclamation point associated with the echocardiogram icon in the chart, would make this finding more noticeable without even opening the echocardiogram result. Such notation is often used in conjunction with abnormal microbiology results and should also be considered for other diagnostic tests that may need quick action.

Various models leveraging the EHR to communicate important results have been developed. In one model, the radiologist interpreting a study can dictate an important finding in a dedicated section of their report (using an automated macro for “result with recommendation for non-urgent follow-up"), which is then transmitted electronically to trained staff who page the patient’s primary provider at the time the report is posted. If the patient had already been discharged, they would contact the patient’s primary care provider and document the follow-up plans.²¹ A similar model, using an “unexpected findings navigator” program and staffed by radiologic technologists with appropriate professional experience, was used to find that follow-up imaging was scheduled for 60% of the unexpected findings, of which potential neoplasms were discovered in 11%.²² Although no baseline data were available for comparison, the authors reported that in 1054 cases (38%), the provider treating the patient chose not to follow up on the neoplastic lesion because of patient choice, because the findings did not correlate with symptoms, or because the finding did not need radiologic follow-up (e.g., small lung nodules in a patient at low risk of lung cancer).

Another method that may improve effective communication incorporates elements of CLC whereby an alert shows up on the EHR (such as an exclamation point or a red flag) and the receiving provider can click an “acknowledge” button once they have viewed the result. In cases where the result has not been acknowledged for a predetermined period of time, the reading/interpreting physician would be notified and can reach out to the provider via the EHR inbox message, live EHR chat, or pager. Applying such a system could have prevented the echocardiogram findings in this case from getting lost to follow up, although it might not have bridged the divide between the hospital and the SNF.

Empower the patient with critical information

Strategies for optimizing safety during transitions of care should also ensure that there is adequate communication with the patient and/or their caregivers, which should include informing them of pending tests and follow-ups needed with outpatient providers. Involving a patient’s family members, if permitted by the patient, helps them to understand the patient’s diagnoses and who to contact if there are any issues after discharge. In this case, informing the patient that his echocardiogram would require review by his primary care provider might have empowered him to take the initiative for his subsequent health care. Pending test results can be included in a section of the patient’s discharge paperwork where they can be reviewed by the patient’s family members and primary provider. When patients or their family members inquire about hospital test results it provides an extra layer of reinforcement to ensure that primary care providers close the loop regarding pending inpatient studies. Had the patient in this case, or his family members, been informed about the pending echocardiogram results, perhaps the new vegetation would have been caught in enough time for effective treatment to have been implemented.

Conclusion

We practice medicine in a system that has inherent flaws. Moreover, as humans, although we cannot be perfect in the care we provide to our patients, we can identify problems, innovate solutions, and respond appropriately. Knowing our limitations and using the human and technical resources available to create additional layers to diminish the number of holes comprising patient safety risks in the “Swiss cheese” model of our medical system will create a safer environment for our patients. Understanding vulnerabilities and learning from successful practices will help healthcare organizations develop reliable processes for effective communication about a patient’s current condition and follow-up requirements during patient care transfers, which will contribute to preventing harms from diagnostic delays.

Take-Home Points

• Utilizing a systems-based model, such as the Swiss cheese model, to evaluate vulnerabilities in how test results are documented and communicated can provide clues about changes that could be implemented in a healthcare system to minimize risks associated with transitions of care.
• Utilizing check lists and structured handoffs that specifically address pending results, adding patient navigators to the hospital team, and utilizing EHR enhancements to draw attention to concerning or potentially life-threatening results are among effective processes that have been identified and can be implemented to plug or decrease the number of figurative Swiss cheese holes, i.e. patient safety risks, in a health system.
• Critical findings should be communicated to the primary care team or test-ordering provider either verbally or in written form, and receipt of the communication should be acknowledged and documented to ensure that the communication loop has been closed.
• Providers should inform the patient and their family members of results that need to be followed up on after discharge, which empowers them to be involved in their own medical care and can also provide an added layer of reinforcement for ensuring follow up of critical results post-discharge.

Noelle Boctor, MD
Associate Physician Diplomate
Department of Internal Medicine, Division of Hospital Medicine
UC Davis Health
mboctor@ucdavis.edu

Mithu Molla, MD, MBA
Clinical Professor
Chief, Division of Hospital Medicine
Department of Internal Medicine, Division of Hospital Medicine
UC Davis Health
mmolla@ucdavis.edu

References

1. Levinson DR. Adverse Events in Hospitals: National Incidence Among Medicare Beneficiaries. Washington, DC: US Department of Health and Human Services, Office of the Inspector General; November 2010. Report No. OEI-06-09-00090. [Free full text (PDF)]
2. de Vries EN, Ramrattan MA, Smorenburg SM, et al. The incidence and nature of in-hospital adverse events: a systematic review. Qual Saf Health Care. 2008;17(3):216-223. [Free full text]
3. Panagioti M, Khan K, Keers RN, et al. Prevalence, severity, and nature of preventable patient harm across medical care settings: systematic review and meta-analysis. BMJ. 2019;366:l4185. [Free full text]
4. Bates DW, Levine DM, Salmasian H, et al. The safety of inpatient health care. N Engl J Med. 2023;388(2):142-153. [Available at]
5. Levinson DR. Adverse Events in Hospitals: Case Study of Incidence Among Medicare Beneficiaries in Two Selected Counties. Washington, DC: US Department of Health and Human Services, Office of the Inspector General; December 2008. Report No. OEI-06-08-00220. [Free full text (PDF)]
6. Transition of Care. Agency for Healthcare Research and Quality: Rockville, MD, June 2018. [Free full text]
7. Rennke S, Nguyen OK, Shoeb MH, et al. Hospital-initiated transitional care interventions as a patient safety strategy: a systematic review. Ann Intern Med. 2013;158(5 Pt 2):433-440. [Free full text]
8. Roy CL, Poon EG, Karson AS, et al. Patient safety concerns arising from test results that return after hospital discharge. Ann Intern Med. 2005;143(2):121-128. [Available at]
9. Starmer AJ, Spector ND, Srivastava R, et al. Changes in medical errors after implementation of a handoff program. N Engl J Med. 2014;371(19):1803-1812. [Free full text]
10. Desai SV, Feldman L, Brown L, et al. Effect of the 2011 vs 2003 duty hour regulation-compliant models on sleep duration, trainee education, and continuity of patient care among internal medicine house staff: a randomized trial. JAMA Intern Med. 2013;173(8):649-655. [Free full text]
11. Blazin LJ, Sitthi-Amorn J, Hoffman JM, et al. Improving patient handoffs and transitions through adaptation and implementation of I-PASS across multiple handoff settings. Pediatr Qual Saf. 2020;5(4):e323. [Free full text]
12. Fischer P, Krueger JI, Greitemeyer T, et al. The bystander-effect: a meta-analytic review on bystander intervention in dangerous and non-dangerous emergencies. Psychol Bull. 2011;137(4):517-537. [Free full text]
13. Al-Khafaji J, Townsend RF, Townsend W, et al. Checklists to reduce diagnostic error: a systematic review of the literature using a human factors framework. BMJ Open. 2022;12(4):e058219. [Free full text]
14. Budde H, Williams GA, Scarpetti G, et al. What are patient navigators and how can they improve integration of care? [Internet]. Copenhagen (Denmark): European Observatory on Health Systems and Policies; 2022. [Available at]
15. Budde H, Williams GA, Winkelmann J, et al. The role of patient navigators in ambulatory care: overview of systematic reviews. BMC Health Serv Res. 2021;21(1):1166. [Free full text]
16. McBrien KA, Ivers N, Barnieh L, et al. Patient navigators for people with chronic disease: a systematic review. PLoS One. 2018;13(2):e0191980. [Free full text]
17. Kelly KJ, Doucet S, Luke A. Exploring the roles, functions, and background of patient navigators and case managers: a scoping review. Int J Nurs Stud. 2019;98:27-47. [Available at]
18. Salik I, Ashurst JV. Closed Loop Communication Training in Medical Simulation. 2023 Jan 23. In: StatPearls [Internet]. Treasure Island (FL): StatPearls Publishing; 2023 Jan–. [Free full text]
19. Banerjee I, Davis MA, Vey BL, et al. Natural language processing model for identifying critical findings- a multi-institutional study. J Digit Imaging. 2023;36(1):105-113. [Available at]
20. Lacson R, Prevedello LM, Andriole KP, et al. Four-year impact of an alert notification system on closed-loop communication of critical test results. AJR Am J Roentgenol. 2014;203(5):933-938. [Free full text]
21. Dibble EH, Swenson DW, Cobb C, et al. The RADCAT-3 system for closing the loop on important non-urgent radiology findings: a multidisciplinary system-wide approach. Emerg Radiol. 2017;24(2):119-125. [Available at]
22. Schwartz FR, Roth CJ, Boardwine B, et al. Electronic health record closed-loop communication program for unexpected nonemergent findings. Radiology. 2021;301(1):123-130. [Free full text]