Audrey Fraizer and Becca Barrus
People connect through data. It’s generated everywhere by everyone, in the private and non-profit organization sectors, by state and federal government agencies, and on an individual basis to plan, evaluate, and monitor—for example, to track spending and establish a budget. At this stage of an international pre-occupation with quality improvement and customer satisfaction, it would be difficult if not impossible to operate without quantifiable evidence. The world can no longer spin on best guesses and gut feelings.
Data collection is also the expected building foundation in public services. Agencies rely on data collection to drive policy-making decisions, justify budget proposals, and generate cost-effective operations. For EMS, devising strategies for optimizing patient care through selective data gathering is a departure from past reliance on military-based trauma care models for the general public and a “what’s always worked” philosophy.
As prehospital medicine has evolved, so has the recognition of data as a quantitative source for making a difference, particularly in areas of EMS education, outcomes, research, and reimbursement.1
Collection of 911 data complements EMS goals, not only to assess the current state of emergency communications, but also to help measure and improve performance and reach people in ways unique to technology. For the Priority Dispatch Systems, data can validate a change that was made to any of the systems based on internal data collection done at the International Academies of Emergency Dispatch® (IAED™). Finding the same or similar results across agencies speaks to the consistency of the protocols, regardless of where they are used.
Data builds better interpersonal connections when we put it into human context: What type of data do I need to influence actions of the people I want to reach? What approach for collecting data will meet my goal?
As every good quality assurance program recognizes, paying close attention to day-to-day operating data leads to discovering new insights about the agency and personnel. Once discovered, these insights can result in actions that can improve working relationships and customer satisfaction. As pointed out in Beyond EMS data collection: Envisioning an information-driven future for Emergency Medical Services: “Decisions are more likely to be effective and efficient when they are based on data.”2
And the sky’s the limit as long as you know what you’re looking for, where to find it, and how to apply it.
Data and algorithm
An algorithm is an approach to building a model from data. It is a series of steps designed to solve a problem. In computer science applications, an algorithm is a sequence of actions designed to show us how to perform a task. In daily life, this can be as simple as a cake recipe or assembling a bookshelf. In computer science, Pedro Domingos offers this definition in his book “The Master Algorithm”3: “An algorithm is a sequence of instructions telling a computer what to do.” Algorithms, he explains, are reducible to three logical operations: AND, OR, and NOT. At the core, algorithms are built out of simple rational associations.
Data and text messaging save teens on the brink
Crisis Text Line (CTL) is a confidential text message service for helping teens and young adults experiencing a mental health-related problem. It’s available nationwide. How does CTL find people in crisis, or how do people find CTL? The developers created an algorithm. Text messages sent to the around-the-clock crisis counseling hotline (741741) are assessed based on messages received over the service’s history (these are the parameters in the algorithm and they could, conceivably, expand with time). The texts are routed automatically to a three-tier queue according to severity, with the highest priority given to messages that use specific words such as “cut,” “hurt,” or “kill,” indicating the sender is actively suicidal.
Trained crisis counselors available on the tier one queue connect with all potential suicidal texters in 13 seconds on average, according to Baylee Greenberg, CTL’s Director of Operations. Counselors further evaluate the situation through direct texting, and if there is clear indication the person is pursuing a plan to cause personal harm, Crisis Text Line will initiate an active rescue.
“Texting is very effective in reaching people in a moment of crisis,” Greenberg said. “The word ‘today’ is the most often used word, according to our database, and it doesn’t take more than three messages for the counselor to know the issue at hand.”
CTL was piloted in 2013 to address gaps in crisis communication for individuals uncomfortable with making calls to other hotlines or calling 911. Apparently, the hotline hit a nerve; so far, 45 million text messages have been exchanged in 295 area codes.
The hotline has also opened a window into the possibilities of NG911.
In 2015, five text specialist Public Safety Answering Points (PSAPs) partnered with CTL for location support on Active Rescues, and several more PSAPS have since joined the community. Counselors refer to their PSAPs’ contact lists when circumstances indicate immediate action is necessary and the PSAP, in turn, can request a ping on a cellphone. Major mobile carriers have agreed to remove the hotline’s short code from user records to ensure privacy and waive text message charges for the hotline.
In the first few months of operations, the joint venture reduced the time for Active Rescue by nearly 50 percent, from an average 60 minutes to 27.7 minutes.
“The partnership proved the value in creating relationships using an increasingly popular means of communication,” said Michael Spath, Manager, Sunnyvale (California, USA) Communication Center, who first heard about the hotline in a TED talk podcast (April 2016). “Technology gives us the ability to reach people in the moment of crisis, and that’s powerful.”
Data-based decision-making is an essential element in quality improvement, helping to assess the efficiency and effectiveness of current processes and modify the processes according to findings. Numerous methods exist for collecting data, including focus groups, surveys, reviews of internal records, and numerical tracking of events.
Survey: Data promotes mentoring relationships
Northwell Health Center is a “data-driven organization”—an understatement for every department in the New York (USA) state-based mega health complex boasting 22 hospitals, more than 550 outpatient facilities, and a single secondary PSAP keeping the Northwell Center for Emergency Medical Services (CEMS) in orbit.
Staffed by Emergency Medical Technicians (EMTs) and paramedics trained in Emergency Medical Dispatch (EMD), the CEMS communication center emergency dispatchers handle nearly 850 calls each day for intrafacility transport among the system’s huge network of hospitals and continually track nearly 35 ambulances handling close to 150 assignments each day. The center also initiates EMS care for patients arriving to any Northwell hospital from local municipalities. In times of disaster, the center doubles as Northwell’s Emergency Operations Center.
It’s no wonder that data collection figures prominently in not only improving use of the Medical Priority Dispatch System™ (MPDS®) in relation to prehospital patient care and response but also the efficiency of the EMD trained Emergency Medical Technicians and paramedics.
“We’re running data all the time,” said Anthony Guido, EMT, former Performance Improvement Coordinator, Communications, Northwell Health CEMS. “The same goes for all high-performing centers. We are always looking at ways to improve what we do.”
New hire training was an area Guido turned his attention to in 2016, but not because he had heard any particular complaints about the program. It was a matter, he said, of staying ahead of their needs. To assess current practice as the starting point for development, Guido decided on a survey format in which students rate various aspects of the program—such as ProQA® training and quality improvement processes—and comment on what worked and what didn’t work in learning about how the center operates. Did they understand what was expected of them? Did they understand the standards set by Northwell CEMS? For example, the results of the first survey showed that nearly 65 percent were uneasy with ProQA after completing their training time.
“This told me to bring back the software instructors to address these concerns,” he said. “Since the survey is anonymous, no one feels intimidated by admitting something needs more explaining.”
The survey also pointed out the importance of trending across multiple classes, said Venessa Vangroski, Northwell CEMS Trainer. For example, survey results strongly favored a structure offering greater hands-on training following each module and less book and lecture time. Learners want an instant pairing of seat time in class to seat time at the dispatch console.
“They want a live environment up front,” Vangroski said. “They want live calls once they certify to get the experience of working with the Medical Protocol and ProQA.”
This survey, while in its infancy, is gaining support among the administrative team and, in addition, connecting new hires to longer-term EDs through a mentoring program.
Data collection boils down to identifying what needs fixing and what doesn’t and quantifying the effect of a resolution designed to fix a problem. It’s that simple.
“We feel more accomplished at the end of the day,” Guido said. “It’s all about communicating and making those connections.”
Review of internal records: When EMS isn’t the answer
Lt. Jamie Baltrotsky’s 17 years as a paramedic with Montgomery County Fire & Rescue Service, Gaithersburg, Maryland (USA), drew attention to an issue troubling communication centers everywhere: EMS super-users, the relatively small group of people within an EMS service region accounting for a disproportionate number of EMS resource use, beginning at the dispatch center.
“It’s not only a matter of time and wear and tear on our resources, but also the help they need, EMS can’t provide,” she said. “We didn’t have the ability to sit down with them for a few hours to figure out exactly what they did need. We simply didn’t know what we could do for them.”
In 2013, Baltrotsky left the streets for an inside job as an executive officer. The issue stayed with her and, as she learned, it was also a concern of the agency’s EMS Medical Director, Roger Stone, M.D. He gave her the go-ahead to collect data and pilot efforts for appropriate referral within the community. She created an EMS partnership with the county’s Health and Human Services Agency (HHS).
Working with the communication center, Baltrotsky identified the top EMS super-users during the first quarter of 2015 (January through March), finding that the group had generated 128 total calls. The individual super-users identified were referred to HHS (at the time of the next call to 911) and, as the source of quantifiable comparison, Baltrotsky retrospectively reviewed the call volume before and after HHS intervention. Their call volume decreased to 64 percent (from 128 calls the first quarter to 47 calls the next quarter).47 (64 percent). A second phase conducted over a 14-month period (April 2015 to July 2016) evaluated the potential effect of referral to HHS on EMS services. Using electronic patient care reporting (ePCR) records, the study identified 265 super-users accounting for 4,393 ePCR records and of those, the top super-users accounted for 797 responses.
The study indicated vulnerable adults as a prominent group among super-users.
“They don’t know what to do, so they call EMS,” she said.
Baltrotsky is optimistic about the possibility of decreasing super-user reliance on EMS through community partnerships and referral. She is confident about the application of similar studies and consequent referral programs beyond Montgomery County.
“This is something any community can do,” she said. “The data is relatively simple to collect and the agencies set up for referral are genuinely interested in participating.”
Baltrotsky submitted a poster abstract of the research for NAVIGATOR 2017.
CAD review: Extrication research in Butler County
Chris Davis, the 911 director for Butler County Emergency Communications in El Dorado, Kansas, USA, has been thinking about his current research project for going on 20 years. For as long as Davis has been with Butler County, the fire departments they dispatch for have been asking for more complete information regarding traffic incidents that require extrication.
The information the fire departments want specifically is whether or not a semi-truck or tractor-trailer was involved and whether or not the incident was a head-on collision. The MPDS Protocol 29: Traffic/Transportation Incidents is intended primarily to identify possible medical response needs, and its focus is on mechanism of injury and symptoms such as unconsciousness and injuries.
The Protocol does identify whether any person is reported as pinned or trapped, but the fire department uses different criteria, such as the involvement of a semi-truck or the report of a head-on collision, to decide whether to bring extrication equipment to the scene. The presence of these conditions may make it more likely that they’ll need the equipment, but they don’t get that information until they get to the scene.
To better understand the issues, Davis attended a special research workshop at NAVIGATOR in April 2017 in New Orleans, Louisiana (USA). There he worked with workshop faculty, FirstWatch staff, and IAED representatives to define his methods and goals for getting the data from each agency in his region. He is collecting data about calls dispatched as 29-D-5 from CAD records in Butler County, Johnson County, and Sedgwick County. He is also looking at rescue records coupled with the calls from fire departments and whether or not extrication actually took place.
The project is still in its infancy, but Davis is optimistic about its benefits. The fire department wants Butler County to make an addition to the Protocol making it mandatory for dispatchers to specify when a semi-truck or tractor-trailer is involved. They could file a Proposal For Change (PFC), but Davis isn’t convinced that such a change needs to be made unless data proves it beneficial to the MPDS.
Download and transcribe: Finding the snags
The importance of call review to improve the dispatch process was a finding produced in a study involving the download and transcription of emergency medical calls.
The study, conducted by a team of U.K. researchers in cooperation with the Scottish Ambulance Service communication center, reviewed suspected Out-of-Hospital Cardiac Arrest (OHCA) calls in which EMDs provided CPR bystander PAIs. Calls were selected by MPDS classification from all OHCA calls taken over a three-month period in 2011. Of the 50 calls downloaded and transcribed, 47 were confirmed cases of OHCA and of those, CPR was performed in 39.4
Researchers divided PAIs into 12 stages, from the EMD requesting the caller’s address and phone number to CPR instructions (both mouth-to-mouth and compressions). [Note: The stages outlined for the study do not correspond with the MPDS process (Case Entry, Key Questions, and PAIs).] Time to progress through each stage, and move to the next, and the number of caller-dispatcher interactions, were calculated to evaluate possible holdups in performing dispatch-assisted CPR.
Audio call transcription was used as a novel approach to analyze interactions and identify factors affecting the likelihood and success of bystanders performing dispatch-guided CPR.
According to results, Stage 9 (determining if the patient is breathing by giving airway instructions) took the longest time to complete. Stage 11 (giving CPR instructions) also took a relatively longer time to complete compared to the other stages. Stage 5 (establishing the patient’s age) took the shortest time to complete.
Results showed a high degree of accuracy using transcription for data analysis, with subsequent comparison to scripted protocols (MPDS) providing factors that might delay bystander CPR. Stages in which holdups occurred could be targeted during dispatcher training and become the focus of future dispatch research.
The IAED Research Division welcomed the study and in a Letter to the Editor piece published in the same journal [Resuscitation] they wrote that [the IAED] “is always interested in reviewing and advocating for high-quality research that will help to improve any of the various aspects of our system, and in particular those process elements that involve dispatcher-caller interactions during time critical emergencies, such as cardiac arrest.”5
Interestingly enough, the Academy’s own unpublished data analysis discovered exactly the same results through a somewhat similar audit of CPR audio cases from multiple centers: “Since the completion of the study done by Clegg, et al. in May of 2011, the IAED Medical Council of Standards approved and released version 12.2 of the MPDS in June of 2012. This newer version eliminates the instruction steps for opening the airway and checking breathing in cases where the patient is not breathing, breathing ineffectively, breathing agonally, or where breathing is uncertain, as determined in the initial patient description and assessment stage, known as Case Entry. As Clegg, et al. point out, eliminating these steps saves nearly 1 minute of elapsed time and prevents the caller from experiencing the frustration of attempting to find respirations in a patient who is likely to be in a cardiac arrest state already and may be gasping agonally only—if presenting with any respiratory activity at all.”6
Wide-open playing field
Miyoshi Carstaffin caught the data collection bug as a natural extension of her work at Fulton County EMS, Douglasville, Georgia (USA). She is the agency’s Quality Assurance Officer and, among other responsibilities, she reviews calls and develops continuing dispatch education training.
MPDS Protocol 26: Sick Person (Specific Diagnosis) was a nemesis to Fulton County EMDs and similar to many centers shared the dubious reputation as the most used and most misused protocol. Carstaffin couldn’t help but ask: When an EMD coded the call as a 26, was it a 26 according to response? Did the EMD miss a stroke or a heart attack? How far were they from the real medical problem?
“I wanted to learn how to collect data,” she said.
A brochure advertising a three-day pre-conference research course at NAVIGATOR 2017 seemed like the place to start. Carstaffin registered and scheduled her trip. She took a seat close to the podium, but after a few minutes of listening to preliminary remarks, she figured she had made a mistake.
“I thought I was in the wrong class,” she said. “I had no idea what I had walked into.”
The speaker, David Page, Director, Prehospital Care Research Forum, University of California, Los Angeles (UCLA), was used to that sort of hesitation. He is a disciple of research.
“I was a curious paramedic, and research developed into a passion,” said Page, field paramedic with Allina Health EMS in Minneapolis/St. Paul, Minnesota, USA. “Research is what I do when I want to know more about what I’m doing and why.”
Page floods a room with resources whenever teaching an introductory course to research and, in the class Carstaffin was attending, the mentors were impressive: Prehospital Care Research forum representatives, the IAED Research Division, and a team from FirstWatch, a company that turns raw data into information agencies can use for situational awareness, operational performance, and clinical patient outcomes. For the next three days, the relative newcomers to EMS research partnered in groups to design a research project, analyze relevant data, review associated scientific literature, and present their initial findings to the class as a whole.
Carstaffin worked within a group studying the Protocol 26 question, led by Isabel Gardett, IAED Director of Academics. FirstWatch provided the data (provided by agency permission and cleansed of identifiers), and by the end of course—by the beginning of NAVIGATOR—Carstaffin was ready to conquer the world of EMS research and Protocol 26.
Well, that may be an exaggeration. The course, however, gave her confidence.
“I can do this,” she said. “I know how the process works.”
Page said it was the same with him when designing his first research project looking at the relevance of required paramedic training in reaching a minimum level of competence.
“Once you’re connected with the right resources and the right people, you learn how to move forward,” he said. “There’s inspiration in knowing you can make an impact on people’s lives.”
IAED research possibilities
Watch for an announcement of the next IAED Research Forum at NAVIGATOR 2018. In the meantime, start getting your ideas together for the IAED’s Annual Poster Contest.
The IAED is accepting research abstracts through Feb. 23, 2018, for the annual poster presentation at NAVIGATOR 2018. This year’s conference is scheduled from April 24-26 in Las Vegas, Nevada (USA). Topics must relate to dispatch in any discipline (police, fire, ambulance, nurse triage), published or unpublished.
The IAED Research Division will review all abstracts and, no later than March 9, announce the abstracts accepted for presentation. Posters are due by March 23. All posters will be displayed at NAVIGATOR. The author of the winning poster will be invited to the conference’s Closing Luncheon, and the poster will be published in the Annals of Emergency Dispatch & Response, the Academy’s peer-reviewed research journal. Learn more at www.aedrjournal.org about AEDR at. Go to https://www.aedrjournal.org/cfpp/ for more information about the contest.
1 Brogan J. “What’s the Deal with Algorithms?” Future Tense: The Citizen’s Guide to the Future. 2016; Feb. 2. http://www.slate.com/articles/technology/future_tense/2016/02/what_is_an_algorithm_an_explainer.html (accessed Aug. 18, 2017).
2 Becknell J, Simon L. “Beyond EMS data collection: Envisioning an information-driven future for Emergency Medical Services (Report No. DOT HS 812 361).” National Highway Traffic Safety Administration. 2016; December. https://www.ems.gov/pdf/ems-data/Provider-Resources/812361_Beyond-EMS-DataCollections.pdf (accessed July 21, 2017).
3Domingos P. The Master Algorithm: How the Quest for the Ultimate Learning Machine Will Remake Our World. First Edition. Basic Books; 2015. New York.
4 Clegg G, Lyon R, James S, Branigan H, Bard E, Egan G. “Dispatch-assisted CPR: Where are the hold-ups during calls to emergency dispatchers? A preliminary analysis of caller–dispatcher interactions during out-of-hospital cardiac arrest using a novel call transcription technique.” Resuscitation. 2013; Aug. 21. http://www.resuscitationjournal.com/article/S0300-9572(13)00451-6/pdf (accessed Aug. 15, 2017).
5 Scott G, Olola C, Gardett I, Clawson J. “Shorter dispatcher-assisted CPR time-to-compression using the latest dispatch protocol.” Resuscitation. 2014; Feb. 6. http://dx.doi.org/10.1016/j.resuscitation.2014.02.034 (accessed Aug. 16, 2017).
6 See note 4.