Learning from lawsuits: Using malpractice claims data to develop care transitions planning tools

Abstract

Objectives:

Our understanding of care transitions from hospital to home is incomplete. Malpractice claims are an important and underutilized data source to understand such transitions. We used malpractice claims data to (1) evaluate safety risks during care transitions, and (2) help develop care transitions planning tools and pilot test their ability to evaluate care transitions from the hospital to home.

Methods:

Closed malpractice claims were analyzed for 230 adult patients discharged from four hospital sites. Stakeholders participated in two structured focus groups to review concerns. This led to the development of two care transitions planning tools—one for patients/caregivers and one for front-line care providers. Both were tested for feasibility on 53 patient discharges.

Results:

Qualitative analysis yielded 33 risk factors corresponding to hospital work system elements, care transitions processes, and care outcomes. Providers reported that the tool was easy to use and did not adversely affect workflow. Patients reported that the tool was acceptable in terms of length and response burden. Patients were often still waiting for information at the time they applied the tool.

Conclusions:

Malpractice claims provided insights that enriched our understanding of suboptimal care transitions and guided the development of care transitions planning tools. Pilot testing suggested the tools would be feasible for use with minor adjustment. The malpractice data can complement other approaches to characterize systems failures threatening patient safety.

INTRODUCTION

Despite more than two decades of efforts to improve patient safety, care transitions in health care remain a point of increased risk in patient care. Care transitions refer to the movement of patients from one provider or healthcare setting to another,¹ such as from hospital to skilled nursing facility. The hospital-to-home transition represents a key hazard point for lapses in coordination and communication, which serve as important drivers of adverse outcomes and malpractice claims. Effective care transitions require thorough and precise coordination and communication of information among healthcare providers, patients and caregivers. All of the individuals involved may fail to execute tasks successfully during care transitions^2,3 which can lead to adverse health events, unplanned hospital readmission, poor patient and family satisfaction, and malpractice claims. ^3–11

Malpractice claims as a data source for improving care transitions

Despite substantial research in the area of care transitions, readmissions rates continue to be high, adverse events associated with care transitions remain an issue, and patients are often dissatisfied with the quality of their care.^4,12 Many of the adverse outcomes associated with suboptimal care transitions go unrecognized by the hospital system, in part because several occur after hospital discharge. It falls upon other actors within the system– post-acute care sites, home health agencies, primary care providers, patients, and caregivers—to recognize and “pick up the pieces.” A number of models have been developed to improve transitions from hospital to home.^13–18 However, the best practices by which to improve care transitions are not clear, in part because intervention efforts are often disease-^19–24 or setting-specific^15,25,26 and vary widely. Additionally, previous studies of suboptimal care transitions have used epidemiologic methods to identify clinical risk factors. Less is known about system-level risk factors, such as care processes within a hospital, which contribute to suboptimal care transitions.^27,28 This suggests there is a gap in our approach for how to improve transitions.

Medical malpractice claims are a potentially rich source of information about suboptimal care transitions. Malpractice claims related to suboptimal care transitions represent egregious mishaps and/or marked patient dissatisfaction, and claims can provide important information regarding underlying system-level risk factors. Many claims involve sentinel events which the health system can learn from about how to improve healthcare quality and reduce future risk of liability—an opportunity to “make medical errors into medical treasures.”²⁹ Malpractice claims analyses have made important contributions to patient safety in other areas of health care,^30–34 and claims analyses offer strengths for evaluating the quality of care transitions: (1) claims integrate documentation from both formal legal inquiries and confidential internal investigations, and claims can provide detailed descriptions of care, including information that is not always available in medical records;^30,31,34 and (2) claims are likely to have information regarding an entire episode of care that spans across care settings and time.

Combining malpractice claims with clinical documentation and reports of adverse events can inform the development of a conceptual model to guide the development of new care transitions planning tools. New tools are needed because current tools to measure and improve care transition quality^35–37 are incomplete and do not typically provide real-time feedback. Thus, the objectives of this study were to (1) evaluate safety risks during care transitions from an acute care setting to the home that have led to malpractice claims, and (2) use the claims to help develop care transitions planning tools and to pilot test their ability to evaluate care transitions from the hospital to home.

METHODS

Overview

We undertook a multi-step process to achieve the stated objectives. First, we conducted a content analysis of malpractice claims, and used the results to inform development of a conceptual model of safety risks during care transitions associated with malpractice claims. Then we conducted structured focus groups in which participants rated elements of the conceptual model, and used the results to create two care transitions planning tools – one to be used by patients/caregivers and one for front-line care providers. Finally, we conducted initial feasibility testing of the tools. This study was approved by a Johns Hopkins School of Medicine Institutional Review Board.

Content analysis of malpractice claims related to care transitions

We developed a process to extract and analyze malpractice claims related to care transitions. Summaries of closed malpractice claim files for the past 10 years involving patients 18 years of age and older (n=665) were obtained from the institution’s malpractice insurance carrier (MCIC Vermont). Claims are closed for a variety of reasons, including closure due to settlement, mediation, or if the statute of limitations has passed. A multi-disciplinary research team of patient safety experts, health services researchers, primary care and hospital-based providers, and administrative leaders developed a case definition that was used by nine expert reviewers to review these claims for relevance to care transitions. A transition of care event was defined as follows: (1) an event occurring after discharge from an acute care facility in our health system; and (2) an event clearly involving the suboptimal transfer of equipment (dentures, hearing aids, glasses, prosthetics, assistive devices, and others), knowable information (diagnoses, test results, complications), or components of the management plan at the time of discharge (aftercare, monitoring, follow-up). Two hundred and thirty claims met these criteria. The claims that achieved 75% or more inter-reviewer agreement and deemed likely to involve a sub-optimal care transition were selected for full review of the litigation record and associated documents (medical records, patient complaint documents, risk management files, and incident event documents from patient safety reporting systems, n=17).

Closed malpractice claims were analyzed for patients, aged 18 and older, hospitalized and discharged from four hospital sites affiliated with an academic medical center (n=230). Two of the four hospitals were teaching hospitals. One hospital had approximately 1000 licensed beds, and the other three had between 200 and 400 licensed beds each. All claims were opened between January 1, 2000 and December 31, 2009 and included an adverse event that occurred after discharge from the hospital. To protect confidentiality and privilege, claims were redacted of sensitive or identifying information prior to analysis (e.g., financial information, details of legal strategy).

A qualitative approach was used to extract a broad set of risk factors related to a suboptimal care transition from the malpractice data. We defined risk factor to be any attribute or characteristic at the patient or system level that was present in our review of cases and that could potentially increase the likelihood of a suboptimal care transition (i.e., care that was not safe and/or did not meet the patient’s or informal caregiver’s expectations). First, three investigators (AIA, EMK, RTB) reviewed a sub-sample of cases to develop a preliminary coding template based on concepts emerging from the review.³⁸ The coding template was reviewed by the entire multi-disciplinary research team and refined for use in subsequent case reviews. Resulting coding changes were applied retroactively to all previously coded cases. All of the cases were reviewed independently by at least 2 investigators and differences were reconciled by consensus. The qualitative research software, ATLAS.ti was used to facilitate data analysis (ATLAS.ti Scientific Software Development, Berlin, Germany).

Development of care transitions planning tools through focus group ratings of elements of conceptual model

Risk factors within the conceptual model were rated and prioritized for inclusion in the tools through a series of two structured focus groups with key stakeholders. The care transitions planning Tool Development Group (TDG) consisted of inpatient physicians, ambulatory physicians, patient safety officers, and senior health system executives in safety and quality. The steps completed by TDG members were as follows: 1) conducted a series of facilitated discussions to prioritize the conceptual domains in terms of importance to improving patient safety during care transitions and reducing risk for litigation associated with hospital discharge processes; 2) proposed a list of candidate measures for monitoring the quality of care transitions planning; 3) narrowed the list of proposed measures and assembled them into two brief survey instruments that served as care transitions planning tools: one intended for use with front-line inpatient care providers and one intended for patients and their families/caregivers at hospital discharge (see online supplement).

Feasibility testing of care transitions planning tool

Feasibility testing of the tools was conducted on 53 patient discharges from an inpatient general medical nursing unit over a two-week period. To test the feasibility of applying the tools in operational settings, we recruited staff from a 17-bed inpatient general medical patient care unit at one of the hospital sites, which was already engaged in process improvement efforts to improve care transitions at the time of discharge. This unit was chosen because of the presence of institutional champions and strong staff awareness of the importance of improving care transition quality. This allowed for frank feedback that facilitated feasibility testing.

Each morning, except for weekends and holidays, physicians, nurses, social workers and other hospital staff on this unit participated in a 90-minute “multi-disciplinary rounds,” (MDR) led by the unit’s nurse case manager and attended by the unit’s social worker, utilization nurse, pharmacist, dietician, physical and occupational therapists, and each patient’s nurse and hospitalist physician. The purposes of this meeting were to 1) review each patient’s care plan daily, 2) promote early discharge planning, and 3) identify and address issues that might arise after hospital discharge. Typically unit staff spent less than 5 minutes discussing each patient.

To identify potential barriers to administering the tools, project investigators attended multi-disciplinary rounds (MDRs) to observe hospital workers using the front-line provider tool, and separately interviewed discharged patients and their caregivers to obtain responses to the patient tool independent of those obtained by unit nurses. The investigators observed MDR workflow and patient-nurse interactions for usability concerns related to the tools, and recorded independent responses to the provider and patient tools. Differences between responses recorded by hospital unit staff and those recorded by project investigators were compiled and discussed with the unit case manager afterwards.

The front-line provider tool was administered for each patient who was planned for discharge during the hospital unit’s MDR (N=19). The provider tool was integrated into the MDR workflow by enlisting the unit case manager to verbally administer the tool to all MDR attendees as a group whenever a patient was identified as planned-for-discharge that day. The unit case manager recorded the responses of the group as a whole. Unit nurses administered and completed the patient tool at the time of discharge when they provided patient/family education (N=53). Data were collected on paper and later submitted to the project investigators.

RESULTS

Content analysis of malpractice claims related to care transitions

The content analysis of claims yielded 33 risk factors to suboptimal care transitions corresponding to domains of the hospital work system, care transitions processes, and outcomes of care. Table 1 represents a list of these factors and their definitions. For each litigation claim, we reviewed the associated medical records, patient complaint documents, risk management files, and incident event reporting documents from patient safety reporting systems. The factors in Table 1 from the claims analysis represented a distinct set of factors from malpractice claims related to suboptimal care transitions, revealing little overlap across the different reporting methods (litigation claims, patient relations reports, and patient safety reporting systems).

Table 1.

Risk factors to suboptimal care transitions and definitions resulting from qualitative analysis

Risk factors related to suboptimal transitions	Definition of risk factor
Additional interventions by others	Person received additional interventions/advocacy from others (e.g., family, friends, caregiver, personal connections) to get needs met.
Adverse event--Diagnostic error	An error in diagnosis occurred.
Adverse event--Medication error	An error related to medications occurred (e.g., error in administration or choice of medication).
Adverse event—Procedure	An error related to a procedure occurred.
Adverse event or complication	Person experienced an adverse event or complication during hospitalization or after discharge. Please use this code if the other more specific “adverse event” codes are not applicable. Please also specify what type of adverse event or complication occurred.
Breach of confidentiality	Breach of confidentiality of personal health information.
Care fragmentation	Care fragmented across providers or settings. Examples may include person receiving conflicting information from multiple sources, providers unaware of what other providers are doing, or lack of unified source of information.
Complaint to staff, patient relations, risk management	Person made a complaint to staff, patient relations office, or risk management prior to claim being filed.
Complicated medical history	Person has a complicated medical history that is contributing to their risk of adverse events.
Complication--Nosocomial infection	Person experienced or is concerned about the acquisition of a nosocomial infection.
Concern--Health-related goals not met	Person felt that their health-related goals were not met (e.g., did not experience expected recovery, did not receive explanation for symptoms).
Concern--Lack of apology	Person concerned that they have not received an adequate apology for harm done.
Concern--Not understanding follow-up or self-care plan	Person did not understand follow-up or self-care plan (e.g., type of monitoring needed, medication management, additional tests/procedures, follow-up appointments).
Concern--Premature discharge	Person concerned that discharge occurred too soon.
Concern—Suboptimal care	Person concerned about the delivery of suboptimal or substandard care.
Concern--Undue suffering, lost productivity, costs	Person concerned about adverse event or complication resulted in undue suffering, lost productivity, high costs.
Confirmation In Medical Record	Notation in medical record that confirms a statement made elsewhere in the claims file.
Delay--Care or follow-up	A delay in care or follow-up occurred.
Delay--Notification of important information	A delay in notification of important information occurred (e.g., test results, diagnoses).
Frequent encounters with healthcare system after discharge	Person has experienced frequent encounters with the healthcare system after discharge. This may include visits to the emergency department, other outpatient/post-acute sites of care (clinics, rehabilitation, skilled nursing facilities), or to the hospital.
Harassment by billing office	Continued requests to the patient to pay outstanding bills.
Lack of documentation	There was a lack of documentation describing needed information.
On high-risk medications	Person is on medications that require frequent monitoring or changes in dose. Examples may include blood thinners (warfarin), diuretics (furosemide), psychiatric medications.
Previous experience with adverse events	Person has personally experienced an adverse event or has witnessed another experience an adverse event (e.g., family, friends, colleagues) prior to hospitalization.
Prolonged length of stay or recovery	Person has experienced a hospitalization or recovery that has taken longer than anticipated.
Special status	Patient works in health care, law, or is connected to JHMI
Suboptimal communication--Across care settings	There was suboptimal communication across sites of care (e.g., emergency dept., hospital, skilled nursing facility, rehabilitation facility, home health care agency, ambulatory clinic).
Suboptimal communication--Among providers	There was suboptimal communication among healthcare providers (e.g., nurses, consultants, primary care physicians, discharge planners, home healthcare nurses).
Suboptimal communication--Discharge or follow-up plan	There was suboptimal communication about the discharge or follow-up plan (e.g., medications, durable medical equipment, monitoring, appointments, services needed).
Suboptimal communication--Findings/test results	There was suboptimal communication about findings or test results.
Suboptimal communication--Lack of adequate informed consent	There was suboptimal communication about anticipated complications of a treatment or procedure.
Unresolved medical issues/symptoms	Person has medical issues or symptoms that are unresolved (i.e., no definitive explanation for cause or still ongoing).
Unsuccessful attempts to access care	Person tried to access care (e.g., request to speak to healthcare professional, request for appointment), but was unsuccessful.

Development of care transitions planning tool through focus group ratings of elements

Using a series of 3 structured rating exercises, the TDG reviewed all 33 risk factors converted the risk factors into 2 care transitions planning tools: The front-line provider tool is an 8-item instrument designed for administration during hospital multidisciplinary discharge rounds prior to hospital discharge. It is intended to identify concerns of front-line, inpatient care providers. The patient tool is a 4-item instrument intended for administration to patients and their families/caregivers immediately prior to hospital discharge. The purpose of these tools is to be used as both (1) an intervention to raise awareness of care transition planning issues in real time (as the patient is being discharged), and (2) a measure for monitoring the quality of care transition planning. Copies of both instruments are included in the Appendix.

Feasibility testing of care transitions planning tools

Feasibility testing yielded several insights. Hospital case managers who administered the provider tool reported that the instruments were easy to use and did not affect the flow of multidisciplinary rounds. However, case managers noted that it was important to actively seek responses from all participants, as some later reported that members of the multi-disciplinary team did not raise certain issues out of concern it would delay rounding. Providers also expressed concern about identifying and documenting problems on the anticipated discharge day, as many issues they identified (e.g., chronic substance abuse or absence of social support at home) were problems that they were unable to address directly. If the care team had concerns about patients’ readiness for discharge (e.g., patient needing additional tests, consultations, or education prior to discharge), the team at times did not know during morning rounds if the issue had been addressed. The use of the tool prompted team members to communicate about these issues prior to discharge and act upon unresolved issues.

Patients who received the patient tool reported that the instrument was acceptable in terms of length and response burden. The primary concern identified during feasibility testing of the tools is that patients and their caregivers often were still waiting for information (e.g., about post-discharge appointments or about test results) at the time they encountered the tool. When they did have the information, patients specifically reported that the care plan seemed appropriate.

DISCUSSION

We used malpractice claims deemed likely to involve suboptimal transitional care to guide the development of care transitions planning tools. The pair of tools captured the experiences of healthcare providers, patients, and families, to identify which patients are at risk for safety problems after they leave the hospital. The purpose of these tools was to be used as both (1) an intervention to raise awareness of care transition planning issues in real time (as the patient is being discharged), and (2) as a measure for monitoring the quality of care transition planning. Implementation of these instruments in an initial pilot study yielded lessons about feasibility and further modifications.

Suboptimal care transitions represent a broad concept in healthcare delivery. Many patients and providers may not identify suboptimal care during transitions as readily as they might identify other suboptimal outcomes that arise in care delivery, such as post-procedure wound infections, for example. Moreover, some risk factors identified in our study may seem more obviously related to care transitions than others, yet the common theme is that the risk factors were identified in care transitions-related claims and can contribute to patient harm. In other words, while the presence of an individual risk factor in isolation may not necessarily lead to a malpractice claim, several risk factors taken together could lead to harm and a subsequent claim. It is also important to note that the under-recognition of suboptimal care transitions also may stem in part from a lack of a standardized classification system for malpractice claims. Malpractice insurers and healthcare organizations can further modify the criteria used in this study to identify claims related to suboptimal care transitions for quality improvement purposes.

The results of this study were important as they led to a novel approach to combine claims data and clinical data for use in patient safety analyses, and to the development of new tools to aid in hospital discharge. There were several important lessons learned from this study. First, in analyzing the claims, we found little overlap among litigation claims, patient relations reports, and patient safety reporting systems. This suggests that the examination of claims helped to identify a unique set of risk factors for suboptimal care transitions. Second, these risk factors corresponded to not only patient-specific clinical factors, but also to factors related to the hospital work system, care transitions processes, and outcomes of care. This suggests, as others have,^27,28 that broader targets potentially can be helpful in care transitions planning. Broader targets identified in this work include identifying the following situations: patients whose expectations have not been met; patients who required additional interventions by others (family, friends, network) to meet their needs; patients experiencing an adverse event; patients with unsuccessful attempts to access health care in the organization; or providers experiencing difficulty coordinating or communicating about care plans. These broader targets highlight the importance of providing feedback to, and soliciting feedback from, patients, informal caregivers, and “sending” and “receiving” healthcare providers before, during, and after the care transition. Third, the collection of real-time information using our care transitions planning tools identified potential suboptimal care during transitions at a point well before a claim might be filed. Feasibility testing of these tools yielded several insights that will shape further development of the care transitions planning tools.

Our study had some limitations. Claim files represent highly selected cases from which it can be difficult to generalize. They may also be subject to hindsight bias, as well as other biases.³⁹ However, claims file analysis also has some strengths relative to other methods, most prominently its ability to detect latent errors, as well as including information from several different perspectives.⁴⁰ Given the high rates of rehospitalizations and post-discharge adverse events that have been reported previously, our examination of malpractice claims and associated documents likely underestimated the prevalence of care transitions issues in our organization. One reason for this may be that certain kinds of failures, such as suboptimal care transitions, are reported through mechanisms other than the legal system. Also, many suboptimal care transitions are likely not reported at all since robust mechanisms for reporting are often not in place after discharge, a time when many care transition deficiencies are detected. This suggests that using hospital documentation systems to comprehensively examine suboptimal care transitions will require combining and integrating multiple sources of event reporting. The processes developed for this project may also be used to examine other kinds of non-structured administrative documents, such as patient complaints. Finally, some malpractice claims may not be related to the quality of care delivered. Nonetheless, we believe that in the context of the purpose of this work, it is reasonable to assume that a claim represents adverse outcomes for those involved. Our goal with this paper was to begin to understand why claims related to care transitions are filed. We do believe study findings demonstrate important insights learned from claims that can improve quality and safety. This paper is not suggesting that health systems necessarily undertake a review of their claims, some of which may not be related to quality of care, but rather to review the risk factors identified in our study, and to consider using the care transitions planning tools that emerged from our analysis.

Potential usefulness of the care transitions planning tools

There are several areas of potential usefulness for the care transitions planning tools developed in this study. First, the paucity of practical measures for monitoring suboptimal care transitions processes poses a barrier to effectively address safety risks during care transitions. The care transitions planning tools developed by this project add to the few tools available to clinical teams, hospital executives and other healthcare leaders.^27,41

Second, the tools could be used to provide real-time feedback regarding care transitions processes. Providing real-time feedback to healthcare providers is a critical step in improving patient safety and ensuring sustainability of improvement efforts.^42–45 In healthcare settings, the critical nature of real-time feedback has been demonstrated in areas of service improvement, reduction of bloodstream infections, and operating room safety, among others.^43,46–49 Particularly relevant to care transitions, real-time feedback was the most important driver of error reduction during care transitions within a hospital in a recent study.⁵⁰ Providing real-time feedback would allow organizations to develop and implement action plans and build “recovery scenarios”²⁷ to address challenges identified by the tools.

Third, the care transitions planning tools provide information from multiple perspectives. A consensus document by the U.S. National Transitions of Care Coalition (NTOCC),⁵¹ outlines three perspectives from which information needs to be obtained in order to fully address optimal care coordination and transitions: (1) patient/family; (2) healthcare professional; and (3) healthcare system.³⁵ The tools complement existing measures examining care transitions and incorporate the three perspectives outlined by NTOCC. Next steps for this effort will be to revise the tools and administration procedures based on lessons learned from feasibility testing, retesting the revised tools, and further dissemination of these instruments as part of efforts to improve care transitions planning processes.

Finally, lower-than-expected case identification in malpractice claims, and limited overlap between claims documents and those from patient relations, patient safety, and risk management sources suggest the need to not rely on one source of information and instead combine multiple sources of event reporting when studying suboptimal care transitions. The process developed for this study may be applied to any of these non-structured records.

We conclude that use of patient reports, including malpractice claims, can be an important driver to promote quality and safety. Indeed, the Centers for Medicare and Medicaid Services has been promoting the use of consumer survey, such as the Hospital Consumer Assessment of Healthcare Providers and Systems (HCAHPS),⁵² to obtain feedback about quality of care of hospital and healthcare systems. Our study findings support prior research demonstrating that patient-reported adverse events can enhance the assessment of organizational safety problems, in particular by raising awareness and providing feedback about unsafe systems and providers.^53–57 Zhu, et al. described that while traditional methods of error detection rely primarily on provider input or administrative data, patients were able to reliably identify unique adverse events that may not have been identified by other methods.⁵⁸ They found physician reviewers agreed that patient-reported events represented a true clinical adverse events in > 70% of cases. Others may use our study methodology to greatly augment this figure by combining multiple sources of event and patient reports when studying suboptimal care transitions.

CONCLUSION

Care transitions represent a point of heightened risk in the care of patients and an opportunity to prevent harm. Strategies to identify increased risks of suboptimal care during transitions may complement efforts to reduce readmissions, lower healthcare costs, and decrease patient adverse events. The tools developed in this study can be used to complement other approaches to characterize and monitor systems failures during care transitions at hospital discharge. This study confirmed observations that noted a need for practical measures for monitoring care transitions processes within healthcare organizations. The care transitions planning tools developed during this project promise to add to the limited number of tools currently available to healthcare leaders.

Evaluating and synthesizing concerns drawn from malpractice claims provided insights that enrich our understanding of suboptimal care transitions and the development of care transitions planning tools. Pilot testing of the tools suggests they would be feasible for use with minor adjustment. The malpractice data approach can be used to complement other approaches to characterize systems failures threatening patient safety.