Improving Communication
Four planning grants (Carilion Medical Center, University of Washington, University of Utah, and Sanford Research) addressed improved communication by assessing attitudes toward error and harm disclosure and by implementing communication interventions in clinical environments.
Carilion Medical Center. This project examined patient, family, and clinician attitudes about the disclosure of individual, team, and system failures associated with four adverse obstetric outcomes. Analysis of focus group and survey data revealed that patients and their family members tended to agree on which system failures were important to disclose, regardless of the type of adverse obstetric outcome. Clinician responses were more varied and were affected by type of adverse event, but they did agree with patients in some instances. The highest correlation in patient/family member and clinician responses was found in regard to failures that occur during delivery and result in intrapartum (the period from the onset of labor to the end of the third stage of labor) fetal death due to group B strep.
In another component of the grant, a customized TeamSTEPPS® training was tested with focus groups and implemented with participants recruited from mother–baby education classes at the Carilion Medical Center. Class participants reported that the training was useful, and they had very positive reactions to the curriculum. The training did not change participant attitudes about communication and teamwork skills, but it did significantly increase participant knowledge about medical communication and teamwork and their effect on quality and safety in patient care.
University of Washington. This planning grant intended to improve communication between patients/family members and providers through the testing and use of shared decisionmaking (SDM) processes and tools, in effort to more actively engage the patient in his or her clinical decisions. Numerous barriers to adoption and implementation were identified and addressed as the result of implementing the SDM model in orthopedic spine surgery clinics. Results based on the observation and assessment of physician–patient encounters indicate that physicians “were most effective in discussing the nature of the clinical condition and less effective in engaging patients as partners in decisionmaking.”1 The elements of SDM most often lacking in clinical encounters were seeking input from others (65%), establishing the patient role in decisionmaking (53%), using teach-back to assess patient understanding (42%), eliciting patient preference for treatment choices (24%), communicating uncertainty (24%), and discussing treatment alternatives (18%).
The project culminated in the development of an SDM train-the-trainer toolkit that integrates the processes developed and lessons learned. Available through the Association of American Medical Colleges MedEdPORTAL (http://www.mededportal.org/publication/9413), the toolkit can be implemented more broadly in other settings. In addition, decision aids have been developed for use with anesthesia patients but have not yet been implemented.
University of Utah. This project aimed to facilitate the disclosure of unanticipated medical outcomes through use of a “systemwide, evidence-based, ethical, and legally sound standardized recommended process.”2 A review of medical records, focus groups with families and risk managers, and other activities resulted in the development of a 10-step disclosure protocol that researchers implemented within a large, regional health care system. The research team found that implementation of a disclosure protocol across a large regional health system is not only possible, but it can also produce successful results. Disclosure protocols support improved communication among patients and providers and suggest a linkage with reduced liability claims. As of the final report, the grantee was establishing the Center for Medical Communication and Conflict Resolution at the University of Utah School of Medicine to promote the inclusiveness of medical communication.
Sanford Research. This project prepared for and implemented Vanderbilt’s Patient Advocacy Reporting System (PARS) in the Sanford Health System, a large, multistate, not-for-profit health care provider. PARS is a tool for identifying and reducing unnecessary variation in a targeted safety/quality indicator—in this case, using patient complaints to identify and intervene with physicians who stand out from peers because of high complaint levels. As part of PARS implementation, a “Project Bundle”a readiness assessment was conducted to assess the presence of 10 elements deemed critical to success and to address any gaps. A bundle is a structured way of improving the processes of care and patient outcomes: a small, straightforward set of evidence-based practices that, when performed collectively and reliably, have been proven to improve patient outcomes. In this case, the “Project Bundle” refers to a set of 10 intervention elements associated with PARS. Using a variety of methods, the health system was able to ensure it met requirements in these 10 areas. The researchers concluded that through key stakeholder engagement and leadership commitment, a large multistate health care system was able to successfully implement PARS and doing so was dependent on key stakeholder engagement and leadership commitment.
Observations
As a group, these grantees learned that the beliefs, preferences, and behaviors of physicians play a key role in facilitating or impeding the adoption of new practices and processes. Taking the time to identify areas of shared agreement and concern regarding communication between patients and providers can help hone communication improvement efforts.
Improving Patient Safety
The five projects in this category (Washington State University, Johns Hopkins University, North Carolina Department of Health and Human Services, The Ohio State University, and Jackson Memorial Hospital) sought to improve patient safety by measuring safety problems, characterizing adverse events, and conducting clinical safety interventions.
Washington State University. Pharmacists examined data on medication discrepancies collected for patients ages 50 and older transitioning from hospital to home care to determine the potential for discrepancies to result in an adverse drug event (ADE), the severity of the ADE (i.e., serious, significant, minor), the potential health consequences (e.g., death, permanent or temporary disability, abnormal lab results), and any anticipated additional health care services needed. Of the medication discrepancies examined, pharmacists found that 41 percent may contribute to an ADE, of which 69 percent could be considered serious or significant. Discrepancies were found to occur across all classes of medications. None of the discrepancies involved permanent disability or death, but almost half of the discrepancies involved symptoms or temporary disability, of which 16 percent were judged to require an office visit, emergency department visit, or hospitalization.
This grantee also explored hospital-to-home medication discrepancies in 10 focus groups of diverse stakeholders, including patients and family members, care providers, pharmacists, and lawyers. Barriers included patient factors (lack of understanding about medications and how to manage them, retaining and using old prescriptions, and problems with access to medications) and health system factors (poor communication and coordination of care, lengthy and complex discharge processes, and staffing and time constraints). This planning grant demonstrated that risk might be minimized with solutions that integrate medication risk management efforts into transitional care models.
Johns Hopkins University. This project assessed closed malpractice claim files over a 10-year period to explore the frequency of and factors related to suboptimal care during hospital discharge. Cases involved discharge from one of four hospital sites affiliated with the academic medical center. Of the 230 claims examined, 13 (5.7%) were determined to involve a potential transition of care event, defined by a multidisciplinary team as occurring after discharge and clearly involving the suboptimal transfer of equipment, information, or components of the management plan. For each of these 13 cases, the researchers collected and reviewed the hospital’s clinical and administrative documents to identify potential failures occurring during the transition.
Themes and subthemes emerging from the analysis suggest failures may occur in multiple domains of the hospital work system (tasks, organization, patient/caregiver, provider, technology and tools, and environment) and care transition processes. This project also found that of the 13 claims judged to have potentially involved a suboptimal care transition, only 3 (23%) had associated event reports noted in hospital reporting systems, suggesting that most of the cases would not have been identified without the claim being filed by a patient.
Following the claims analysis and a literature review, two predischarge surveys were developed to identify patients who are at risk for safety problems after hospital discharge: one for patients and their families and caregivers, and one primarily for inpatient providers. In a feasibility test of the two surveys, hospital case managers reported that the provider surveys were easy to use and not disruptive during multidisciplinary rounds. Patients reported that the patient survey’s length and response burden were acceptable. The survey and processes were revised based on the findings from the pilot. These instruments respond to a need in clinical safety and can be further tested and revised in preparation for broader use.
North Carolina Department of Health and Human Services. This project designed an electronic system allowing clinical and administrative staff to anonymously report near-miss events (errors that do not result in patient harm) and implement remediation efforts. This approach was pilot tested in seven diverse primary care practices across North Carolina. Over the 12-month period, the sites reported a total of 770 near-miss events (with a range from 43 to 177 each) and 34 practice improvement projects to remediate near misses (with a range from 1 to 15 each). An assessment of the 632 near-miss events reported over a 12-month period and included in the analysis revealed the frequency of various types of near-miss events.
Notably, almost half (299 of 632) of reported near misses involved office processes. “Common human error” and “not taking time to do the task correctly” were the two most commonly identified causes of medical errors by reporters of the near-miss events. The researchers were surprised to find that electronic medical records, designed in part to prevent errors, were a contributing factor in 14 percent of errors in the sample.
The primary care practices continued to report near-miss events after the 12-month pilot when they were no longer receiving any compensation for participation, and some of the practices reported plans to continue using the system.
The Ohio State University. This project successfully developed a statewide pregnancy-associated mortality review (PAMR) system in Ohio in an effort to understand the factors associated with maternal death cases and identify opportunities to improve patient safety and health care quality. A retrospective analysis of maternal death cases covering a 2-year period prior to the receipt of grant funding provided data on the frequency and causes of pregnancy-associated and pregnancy-related deaths, as well as the factors involved in these deaths. In 2008, 39 percent of cases were categorized as pregnancy related or possibly pregnancy related, resulting in a pregnancy-related mortality ratio of 9.4 per 100,000 live births. In 2009, 58 percent of cases were categorized as pregnancy related or possibly pregnancy related, resulting in a pregnancy-related mortality ratio of 20.8 per 100,000 live births.
In the assessment of the factors involved in pregnancy-associated and pregnancy-related deaths, the top three individual factors identified were chronic medical conditions, substance abuse, and delay and/or failure to seek care; the top three system factors were lack of case coordination and management, poor patient–provider communication, and lack of continuity of care; and the top three clinical factors were delay or lack of diagnosis or treatment, risk screening, and consultation with another provider. After the end of grant funding, project leaders continued to refine the system and began collecting data for a third year. In 2012, Ohio’s maternity licensure authority began requiring mandatory maternal death reporting.
Jackson Memorial Hospital. This planning grant was used to develop and pilot test the Initiative to Reduce Inpatient Suicide (IRIS), an enhanced suicide screening, assessment, and psychosocial intervention for hospitalized medical/surgical patients who have been identified as vulnerable for attempting or completing suicide. Although researchers were not able to compare the hospital’s medical/surgical patients in the IRIS condition with those receiving the hospital’s usual care as they intended, they did find that the hospital’s medical/surgical patients have a high rate of suicide risk.
In addition, “patients who participated in the study received a psychiatric evaluation, on average, 3 days after admission and have an average length of stay of 8 to 9 days,”3 indicating that screening and intervention need to be conducted earlier in the hospital stay. Data collected from participating nurses suggest the need for more training for nurses in detecting and managing patients at risk for suicide and more research on obtaining reliable and valid data from nurses. Including nurses and other clinical staff in the formal root cause analysis (a common error analysis tool in health care) of the underlying problems after an attempted suicide event at the hospital led to the development of a Suicide Risk Advisory Committee (an inter-disciplinary safety committee) at Jackson Memorial Hospital.
Observations
In general across these projects, efforts appear to be effectively identifying the causes of and contributors to medical errors, and there appear to be some promising interventions and strategies available to prevent or minimize them.
Exploring Resolution Methods
The third category, Exploring Resolution Methods, includes four projects (Beth Israel Deaconess Medical Center, Multicare Health System, Wishard Health Services, and the Office for Oregon Health Policy and Research) that focused on medical liability interventions—variations of a disclosure, apology & offer (DA&O) model, as well as a safe harbor model.
Beth Israel Deaconess Medical Center. Findings from the stakeholder interviews conducted through this planning grant laid the foundation for developing a guide on how to start a DA&O program in Massachusetts, which is now being used by other States interested in DA&O. The findings also led to the establishment of the Massachusetts Alliance for Communication and Resolution following Medical Injury (MACRMI), which comprises a variety of stakeholder groups. Since forming, MACRMI renamed the approach Communication, Apology, and Resolution (CARe); developed clear policies, procedures, algorithms, and guides for facilities implementing CARe; helped develop projects to implement CARe in six hospitals in the State; and created a resource Web site (http://www.macrmi.info).
Importantly, this planning effort resulted in a historic and unprecedented partnership among physicians and attorneys from the Massachusetts Medical Society, Bar Association, and Academy of Trial Lawyers. These three groups have held traditional and, in some cases, opposing viewpoints on tort reform policies. The grant’s work culminated in the passage of a law in Massachusetts that allows all health care delivery organizations to develop DA&O programs to settle medical malpractice claims.
Multicare Health System. This project conceptualized an ACEx2 system, which is used to identify Avoidable Classes of Events that would also be Automatically Compensable Events through disclosure, apology, and early offer of compensation following a standardized compensation schedule. Experts in health care quality assurance and peer review developed criteria for such events that warrant an offer of compensation (e.g., they are preventable and reliably identifiable, and the harm is measurable), and they identified 18 safety events meeting the criteria as well as others worthy of further investigation.
An implementation plan was recommended that outlines the process for identifying and disclosing events to the patient or patient’s family, apologizing, and having early discussions with patients or families about compensation. A standardized approach to compensation for Avoidable Classes of Events was also developed. The grantee offered recommendations for implementing a voluntary ACEx2 program in lieu of the current tort system in the Seattle metropolitan area.
Wishard Health Services (now subsumed under Eskenazi Health). This study aimed to evaluate the Wishard Health Services Reformulated Medical Claims Model (RMCM), which was instituted in 2008 based on the successful University of Michigan Claims Management Model. In the first part of the study, researchers surveyed patients and medical staff involved in 41 claims closed since the RMCM was instituted. Overall, staff members rated the RMCM favorably, with 100 percent of respondents reporting being “very satisfied” or “satisfied” with the risk management/peer review process, the RMCM process, and the claims committee process, and 75 percent stating that the RMCM was “very efficient” or “efficient.” Although one-third of staff respondents indicated the RMCM improved on the claims system used prior to 2008, two-thirds of respondents were unsure due to limited knowledge about the previous process.
In the second part of the study, researchers compared the same 41 claims closed under the new system to 125 cases processed through the old system and found that the RMCM approach fared better or the same in the length of time for processing cases, the size of settlement awards made, and the amount of legal fees encumbered by Wishard. Also, a significantly larger percentage of cases under the RMCM reached a settlement than under the old claims model (48.8% vs. 12.7%).
Office for Oregon Health Policy and Research. This grantee concluded that diverse stakeholders in Oregon view the safe harbor approach as valuable for improving patient safety based on a preliminary analysis of medical claims closed over an 11-year period; a retrospective analysis of selected closed claim files to determine if (1) better adherence to medical guidelines could have prevented some patient harms that led to medical liability claims and (2) a safe harbor rule could have altered the course or outcome of medical liability claims; and discussion with stakeholders. There is less clear support, however, for the legal safe harbor approach as a means for reducing liability costs; in Oregon, it would likely yield savings of only 5 percent of total annual medical liability costs. The grantee found that a significant challenge to the safe harbor approach would be the development, adoption, and use of evidence-based medical guidelines addressing clinical situations that result in significant numbers of patient injuries and medical liability claims. Without such guidelines, a safe harbor rule would be ineffective.
Observations
These grantees learned that engaging diverse stakeholders to assess the strengths and weaknesses of the current system and creating vehicles for ongoing stakeholder involvement (e.g., subcommittees that meet regularly) help to identify key issues, obtain buy-in, and effectively design solutions. The findings from these planning grants represent promising strategies and models for reducing liability costs and, at the same time, improving patient safety.
