In May 2018, Boston Medical Center (Boston, MA) was named a finalist for ECRI Institute's 12th Health Devices Achievement Award for its project to implement an early warning system (EWS) to more quickly identify and direct clinical resources to patients whose condition is deteriorating. Key elements of this initiative were integrating the EWS within the electronic medical record (EMR), customizing the alert-triggering clinical variables and thresholds to the unique needs of BMC's healthcare environment, and establishing multiple response mechanisms, including mobilizing critical care nurses to triage and initiate therapeutic interventions. Data collected to date suggests that these steps have led to reductions in the length of stay, the number of ICU transfers, the time to resuscitative efforts, and unexpected mortality.

The Health Devices Achievement Award recognizes innovative and effective initiatives undertaken by member healthcare institutions to improve patient safety, reduce costs, or otherwise facilitate better strategic management of health technology. For details about the winning submission and other finalists, see The Health Devices Achievement Award: Recognizing Exceptional Health Technology Management.

ECRI Institute congratulates the project team members: James Moses, MD, Adil Yunis, MD, Patrick Wisdom, and Abhinav Vemula, MD.

The Challenge

To reduce clinical deterioration and preventable death among hospitalized patients by implementing an EWS that leverages the capabilities of an EMR to:

1. Identify clinically decompensating patients through the capture and analysis of rapidly changing patient physiologic and laboratory variables

2. Alert frontline clinicians when a patient's condition is beginning to deteriorate

3. Trigger interventions from critical-care-trained "resource nurses" who can implement measures to prevent further clinical deterioration and reduce preventable mortality

The Landscape

1. Clinically decompensating patients—patients whose condition is beginning to deteriorate—are at increased risk of disease progression or death. Failures to quickly identify deterioration are associated with:

a) High rates of morbidity and mortality

b) Prolonged hospital stays

c) Increased healthcare costs

2. Timely recognition of vital sign derangements, lab result abnormalities, and other clinical changes that may precede the worsening of a patient's condition can allow for early medical intervention, thereby reducing the risk of adverse patient outcomes.

3. EWSs that use an EMR to capture rapidly changing patient variables for multiple physiologic parameters are seen as a useful tool for identifying patients at risk.

4. However, BMC's research revealed that EWS capability on its own is insufficient to significantly reduce the incidence of disease progression or preventable death. For an EWS to be effective, the healthcare facility must:

a) Select clinical variables and thresholds that will accurately reflect clinical deterioration in the facility's specific patient population

b) Reliably and swiftly direct appropriate clinical resources toward decompensating patients

The Process

1. BMC's analytics and IT teams, healthcare providers, and quality department leadership collaborated to develop an EWS that could be integrated with BMC's Epic EMR.

2. The team researched various options for EWS thresholds of activation—single-parameter systems, multiple-parameter systems, and aggregate-weighted systems—and determined that an aggregate-weighted system would best fit the organization's needs.

a) Aggregate-weighted systems calculate a score for patients by assessing multiple physiologic parameters grouped into different categories, with each category of abnormality assigned different point values that contribute to a patient's score.

b) BMC's review of the literature suggested that such systems can more accurately predict ICU transfer, cardiac arrest, and mortality.

c) BMC selected the following to include in the EWS: white blood cell count, temperature, pulse, systolic blood pressure, respiratory rate, oxygen saturation, and level of consciousness. These parameters were selected because of their use in other validated EWSs and because of the consistency of their timely documentation within the EMR.

3. The result was a dynamic EWS scoring system that helped identify patients at risk for clinical deterioration by integrating rapidly changing patient variables captured within the EMR.

4. To convert this score into action to prevent further clinical decompensation and reduce mortality, the team established a process for directing appropriate clinical resources toward the patient. The process included three components:

a) The EMR activates a Best Practice Advisory (BPA) to alert nursing assistants, nurses, and physicians when their patient's EWS score indicates that the patient is at risk of clinical decompensation.

b) Staff are able to add the EWS score to their patient dashboard within the EMR. This capability allows clinicians to view their patients' current score at any time. Color coding of the score when it achieves a threshold signifying risk for clinical deterioration (e.g., red for a high score) allows for rapid visual identification of the patient's status.

c) Additionally, BMC's EWS notifies specially trained resource nurses—clinicians who receive critical care training and who proactively review and assess patients with high EWS scores as identified in a constantly updated EMR-based list.

5. The project was piloted on two non-ICU medical-surgical floors. An associated improvement process allowed for key stakeholders to provide feedback to enhance the EWS for BMC's specific patient population and healthcare system. Through this effort, the team was able to, for example:

a) Better define the roles of different healthcare providers on the EWS response team.

b) Identify and address barriers to mobilizing the resources needed to respond to clinically deteriorating patients who trigger the EWS.

The Results

A retrospective comparison of pre- and postintervention samples of clinically decompensating patients showed the following results:

1. The time to resuscitative efforts was reduced in the intervention group by 28 minutes (n = 79, p <0.0001).

2. The time to ICU transfer was reduced by 110 minutes (n = 79, p <0.0034).

3. A reduction in mortality was also observed, with the intervention group showing an 8.61% rate of mortality compared to the control group's rate of 14.5%.

4. A reduction in healthcare resource utilization was reflected by a reduction in the length of hospital stay in the intervention group (11.6 days versus the control group's 14.1 days; n = 79, p <0.0015).

5. Although the intervention group experienced an ICU-consult rate of 19%, which was higher than the rate of 5% prior to the intervention, the early and targeted mobilization of resources led to fewer patients being transferred to the ICU, with 25% of patients transferred preintervention and 18% of patients transferred postintervention.

Key Takeaways

1. The integration of an EWS within the EMR allows for frequent and efficient EWS scoring and rapid clinician response.

a) Updated scores are calculated automatically as new vital signs readings and laboratory results are recorded in the EMR. Thus, the EWS reflects the patient's real-time clinical status.

b) The proactive alerting system allows for more rapid clinical intervention. Providers are alerted as soon as changes in objective clinical data occur.

2. BMC reports that the generalized aggregate-weighted EWS that it initially adopted became a more effective tool when the organization began to customize it for BMC's unique healthcare environment.

a) By leveraging IT and analytics resources, BMC assessed its internal clinical data and determined clinical variables and thresholds that, when combined, would accurately reflect clinical deterioration in its patient population.

b) BMC suggests that healthcare facilities seeking to implement an EWS should customize the scoring system using data from their own clinical setting to suit the needs of their specific patient population.

3. BMC likewise has found value in setting up its EWS to trigger a standardized response protocol whereby critical-care-trained resource nurses respond to clinically deteriorating patients located throughout the hospital. BMC states that this approach allows for a multidisciplinary team of providers to apply their unique skills and to coordinate efforts in caring for clinically deteriorating patients.