University Health System Consortium Data

The University Health System (UHC) Consortium is a data assessment tool utilized by many hospitals, and especially by quality improvement initiatives, for both benchmarking and performance improvement. UHC is a member-owned consortium representing 120 academic medical centers, including over 300 hospitals, and contains self-reported data that are evaluated with risk adjustment. UHC data often serve as a proxy for Centers for Medicare & Medicaid Services (CMS) risk-adjustment data, data that are utilized by nationally recognized external rating systems, and their data serve as a tool for hospitals to compare patient metrics and identify areas of potential improvement. UHC comparison measures include health assessment metrics, such as a 1–4 rating of risk of mortality (ROM) and severity of illness (SOI) on admission and discharge, which are reflections of patients’ risk of inhospital death and degree of illness, respectively. Case mix index (CMI) is a metric which assesses the complexity of care and accounts for patient comorbidities. Mortality index (MI), a measure calculated from the observed or actual mortality and expected or predicted mortality based on documentation and risk assessment, is a UHC health assessment metric emphasized across specialties as a representation of quality of care. Finally, diagnosis and treatment provided at admission is defined by a diagnosis related group (DRG) code that is generated at discharge. CMS determines reimbursement based on DRG codes, which are adjusted according to both ROM and SOI. In practice, these metrics, especially MI, are reflective of the overall success of quality improvement programs in improving patient care.

Collaboration and Standardization of Care

Improvement of interprovider and interhospital communication has been shown to improve patient safety through reduction of adverse events and standardization of patient care plans. Quality improvement programs’ strategies to improve communication have included initiation of multidisciplinary morning rounds and conferences to foster collaboration between hospitals in the same region and in different states. Through improved discharge planning and standardization of care protocols, these efforts have been shown to improve patient satisfaction and reduce mortality. On surgical services, widespread implementation of comprehensive time-outs and the World Health Organization (WHO) surgical safety checklist has demonstrated success in reducing occurrence of adverse events, including wrong-site surgeries and intraoperative complications including thromboembolic events and retained sponges and needles. Initiatives implementing checklists, protocols, and educational sessions to standardize medical management in the emergency department and on inpatient floors have also resulted in complication and mortality reduction.

Electronic Medical Record

With the recent shift to electronic medical records (EMR), quality improvement programs have also emphasized enhanced utilization of the EMR to stratify the patient’s risk of morbidity or mortality both in the emergency department, on inpatient units, and before surgery. In particular, many initiatives have focused on improving documentation of health assessment metrics such as SOI, ROM, CMI, and MI to improve risk stratification and better capture the severity of patients’ illnesses. As a result of improved documentation of these metrics, initiatives have led to substantial increases in hospital reimbursement as well as reductions in avoidable complications and mortality. Furthermore, quality initiatives have also demonstrated significant mortality reductions after utilizing the EMR to create order sets to standardize medical management and decrease time to treatment. For example, multiple initiatives have demonstrated significant mortality reduction after instituting treatment with “sepsis bundles” including predefined order sets to standardize care of patients with symptoms concerning septic shock. The use of standardized order sets was also a component of a comprehensive quality improvement intervention for ischemic stroke, which resulted in significant reductions in time to tPA administration and subsequent mortality across multiple hospitals. The use of order sets to standardize care in the postoperative setting has also been demonstrated to reduce mortality. Clearly, the recent shift to EMR has created opportunities for improved quality of care across specialties.

Review of Adverse Events

Finally, a major target of many of these interventions is improvement in departmental and hospital-wide review of mortalities and adverse events. Initiatives have implemented review processes including departmental morbidity and mortality (M&M) conferences, designated mortality and adverse event review committees, as well as educational programs about national databases’ retrospective analyses of avoidable complications and hospitals’ overall performance. These interventions have led to demonstrable improvements including reduced incidence of postoperative complications, increased awareness among residents and physicians about common avoidable complications, as well as significant reductions in hospital-wide mortality indices and overall mortality rates.

National Quality Databases

In addition to institution-specific interventions, the American College of Surgeons (ACS) National Surgical Quality Improvement Program (NSQIP) is the largest nationwide quality improvement program targeting mortality reduction and quality improvement across surgical services. To date, over 600 surgical programs in the United States participate in the NSQIP. Review of adverse events is a major component of the program; the NSQIP collects institutional data from all participating hospitals including variables such as preoperative risk factors and postoperative morbidity and mortality. The NSQIP then provides participating hospitals with reports to assess individual performance that enable risk-adjusted comparisons with a surgical quality standard. The NSQIP has also formed calculators to stratify patients’ risks of adverse events and mortality based on institution-wide trends. For example, the NSQIP mortality predictor (NMP) 9–11 is calculated from 35 variables including patient demographics, physiologic status, past medical history, laboratory values, and American Society of Anesthesiologist (ASA) scores. Based on these variables, the predictor calculates a probability, ranging from 0 to 1, of patients’ overall ROM. Furthermore, the NSQIP surgical risk calculator has demonstrated applicability in the preoperative setting to predict patients’ risks of adverse events specific both to their demographic and comorbidity profile as well as type of surgery. In addition to emphasizing preoperative risk prediction, the program also fosters interinstitutional collaboration at national conferences. Participation in this program has become the basis for quality improvement across many surgical departments; through NSQIP participation and subsequent development of institution-specific quality initiatives, departments across surgical subspecialties have reported significant reduction in both morbidity and mortality.

Similar databases focused on achieving quality improvement in neurosurgery through accrual of outcomes data have been developed, such as the National Neurosurgery Quality and Outcomes Database (N ² QOD) and Spine Surgery Improvement Collaborative. The development of mortality registries is of particular importance to neurosurgical quality improvement initiatives. Mortality rates, while underreported in neurosurgical literature, are frequently used as outcome measures by family members, insurance companies, and health care providers to assess quality of care. Furthermore, the high-risk nature of neurosurgical procedures increases the incidence of mortality and adverse events compared to other specialties, therefore necessitating better characterization of neurosurgical mortalities.

Based on these prior projects’ experiences, comprehensive initiatives involving combinatorial approaches to improve interprovider and interinstitutional collaboration, standardization of care, preoperative risk stratification, utilization of the EMR, and careful review of adverse events show the most promise for achieving clinically significant mortality reduction. These approaches have demonstrated potential in emergency, medical, and surgical settings, warranting their widespread application across specialties.

Mortality in Neurosurgery

Baseline mortality rates on neurosurgical services are high due to the prevalence of life-threatening pathology. Across the literature, reported neurosurgical mortality rates are similar among different institutions and departments: Sandeman reported an overall mortality rate of 2.7% of 6006 admissions, Chen et al. reported a mortality rate of 4.52% in 531 consecutive patients, Hammers et al. reported a mortality rate of 1.7% in 3650 neurosurgical procedures, and Houkin et al. reported a 3.3% mortality rate among 643 neurosurgical interventions over 2 years. In our single institutional study, we reported a similar average mortality rate of 2.41% over 2 years. These studies and ours have also reported a similar distribution of mortalities among neurosurgical subspecialties, with trauma and stroke services having higher rates of mortality compared to spine, tumor, and pediatric services. Hammers et al. reported that trauma, stroke, tumor, spinal disease, and infection were the most common etiologies in descending order underlying mortality. Sandeman similarly reported that trauma was the most common mortality etiology, followed by subarachnoid hemorrhage, tumor, and infection in a general neurosurgery practice. In this study, 94% of mortalities occurred in patients with cranial pathology, and only 6% stemmed from spinal pathologies. Similarly, while Chen et al. reported a mortality rate of 6.53% among patients with cranial pathology, no spinal surgery mortalities were observed in this cohort. In our study, traumas such as subdural hematomas and vascular pathologies including hypertensive and aneurysmal bleeds accounted for the majority of mortalities, with a smaller proportion stemming from tumors during resection or biopsy or embolic or thrombotic strokes. The smallest proportion occurred during elective spinal procedures or due to systemic causes such as sepsis or infection.

In addition to anatomic location of pathology, type of case (elective vs emergent) also influences mortality on neurosurgical services. An increased mortality rate inherent to emergent neurosurgical operations has been previously reported. In a retrospective review of 531 consecutive neurosurgical cases, Chen et al. reported a mortality rate of 12.41% for acute and emergent cranial neurosurgery compared to 2.73% for elective cranial cases. In another retrospective review of 4904 consecutive neurosurgical cases at a single institution, Zygourakis et al. reported that elective admissions were significantly associated with decreased mortality compared to emergent admissions. Moreover, Hammers et al. reported that the majority (85%) of their single institutional neurosurgical mortalities stemmed from emergent, nonelective cases. In Hammers’ study, trauma was the most common etiology underlying neurosurgical mortality. Finally, in a prospective, longitudinal analysis of 6006 neurosurgical admissions seen in one general neurosurgery practice over 15 years, Sandeman reported that 90% of mortalities were admitted from the emergency department and were documented with a preoperative surgical aim “to save life.” Similarly, in our single institutional study, 82% of our observed mortalities over 2 years occurred after emergent admissions. Owing to the unavoidability of many mortalities after emergent admissions or trauma, elective cases have become the target of many neurosurgical mortality reduction initiatives.

Neurosurgical Quality Improvement Initiatives

Neurosurgery-specific quality initiatives have included efforts to improve documentation of health assessment metrics, and improve coordination of treatment through care protocols, multidisciplinary care coordination committees, and safety checklists, as well as an enhanced review of adverse events through M&M conferences and nationwide databases such as the NSQIP. These initiatives have demonstrated success in improving quality of care and reducing mortality on neurosurgical services: implementation of the UCLA Clinical Quality Program led to a reported decreased length of stay and reduction in overall postoperative complication and readmission rates, the University of Missouri’s documentation improvement intervention led to a sustained reduction in departmental mortality rates, the Memorial Hermann/University of Texas quality improvement initiative led to decreased UHC mortality indices and infection rates, and implementation of the Seattle Spine Team Protocol led to a significant reduction in complications as well as a trend toward lower mortality rates. The success of all of these initiatives was multifactorial. The UCLA Clinical Quality Program involved the creation of multidisciplinary Care Coordination Committees for all patients preoperatively, intraoperative introduction of a comprehensive time-out including a safety checklist and care protocols for operative techniques, and postoperative communication templates, standardization of discharge procedures, and more multidisciplinary approaches to morning and afternoon rounds. The University of Missouri’s documentation intervention involved educational in-services about coding metrics, the implementation of a new neurosurgical progress note template, and designation of a “nurse-reviewer” to provide immediate feedback to physicians about medical record coding. Preoperative, multidisciplinary conferences for care coordination were also a component of the Seattle Spine Team Protocol, which also involved a patient education course as well as modification of the operative team to always include a neurosurgeon, orthopedic spine surgeon, and a dedicated spine anesthesia team. Similarly, physician roles were refined in the Memorial Hermann/University of Texas quality improvement initiative, which involved mandated subspecialization of neurosurgeons as well as increased utilization of the EMR for risk stratification and reminders, the implementation of care protocols, and the creation of multidisciplinary treatment teams to foster improved coordination of care. Thus, mortality reduction in neurosurgical departments is both necessary and attainable through implementation of comprehensive, multifactorial quality improvement initiatives.

DUMC Neurosurgical Quality Improvement: A Single Institutional Review

The Duke University Medical Center (DUMC) Neurosurgery Department implemented a multifactorial quality improvement initiative designed to reduce overall mortality on their neurosurgical service. On July 1, 2015, a multifactorial quality improvement intervention was launched involving the following components: (1) monthly and year-to-date detailed metric reports in comparison to departmental and national neurosurgery peers; (2) monthly departmental presentations displaying department- and surgeon-level data; (3) initiation of standardized, best evidence-based care-protocols for common neurosurgery diagnoses and appropriate measures to assess impact and utility; (4) education for all attending surgeons, residents, physician assistants, and nurses on accurate documentation for neurosurgery diagnoses and patient cases; (5) communication with attending surgeons on all mortalities in patients with an ROM and SOI < 4 and MI > 1.0; and (6) initiation of one-on-one meetings between surgeons to further define and explain the quality data results. Through this combinatorial approach, we sought to improve preoperative risk stratification of patients too sick to operate on, reduce avoidable mortality in elective cases, and promote standardization of patient care.

An institutional adverse-event review was performed on assessing the impact of the quality intervention on overall mortality on the neurosurgery service, which included 5434 consecutive patients on the neurosurgical service and consulted at DUMC 1 year prior to intervention ( preintervention: n = 2793 ) and one year after ( postintervention: n = 2641 ) were included in the study. After the intervention, the average SOI, ROM, and expected mortality for neurosurgical patients were higher, which reflects a sicker patient population and improved documentation. However, while expected mortality is influenced by documentation, observed mortality rates speak to the quality improvement potential of interventions and not simply to improvements in documentation. Overall, when compared to the preintervention cohort, the postintervention cohort had a decreased mean-observed monthly mortality by one patient per month, demonstrating the impact that a multifactorial quality initiative may have on reducing the overall mortality rate in neurosurgery.

Application to Other Neurosurgical Departments

Based on our own initiative as well as reports of prior institutions’ experiences, mortality reduction in neurosurgical departments is both necessary and attainable through implementation of comprehensive, multifactorial quality improvement initiatives. We suggest that future initiatives involve (1) improved coordination of care between individual providers and neurosurgical departments across the country, (2) care protocols to achieve superior standardization of patient care, (3) enhancement of documentation to more accurately reflect patients’ SOI and improve preoperative risk stratification, and (4) implementation of retrospective reviews of adverse events. In detailing these prior initiatives as well as our own, we hope to provide a guide for neurosurgical departments, as well as departments in other specialties, to employ similar initiatives aimed at improving quality of care and achieving sustained reductions in overall mortality. As a specialty and as a field, the greatest quality improvement and mortality reduction can be attained through detailed risk stratification prior to an intervention, preoperative health optimization, and enhanced inter- and intradepartmental collaboration of quality initiatives to foster innovation in quality improvement and better overall patient care.