Secondary Logo

IT-driven Clinical Improvement Processes

The SkunkwORks

Yuan, Jennifer J., MD, MBA*; Paganelli, William C., MD, PhD; Jacques, Paul St., MD

International Anesthesiology Clinics: January 2019 - Volume 57 - Issue 1 - p 45–62
doi: 10.1097/AIA.0000000000000213
Review Articles
Free

*Department of Anesthesiology, Stanford University Medical Center, Stanford, California

Department of Anesthesiology, University of Vermont Medical Center, Burlington, Vermont

Department of Anesthesiology, Vanderbilt University Medical Center, Nashville, Tennessee

The authors declare that they have nothing to disclose.

Address Correspondence to: Jennifer J. Yuan, MD, MBA, Department of Anesthesiology, Stanford University Medical Center, 300 Pasteur Dr, Room H3580 MC5640, Stanford, CA 94305. E-mail: jennifer.j.yuan@gmail.com

In 1977, a group of doctors and computer scientists, known as the Hardhats, collaborated to build an innovative system at the Veterans Administration (VA) network. Using new devices called personal computers, they envisioned a health care information technology (IT) program that would be safer and less costly to maintain than the disorganized, often illegible paper records system that served as the repository for patient information. They wanted to create an electronic program that would allow physicians to consolidate clinic notes, patient medications, and imaging results into a central system. The ideas were decades ahead of their time. Mirroring the growth and development of the internet, these forward-thinking individuals eventually developed VistA, one of America’s first electronic medical records (EMR). In retrospect, it was a ground-breaking system.1

The Hardhats’ model for radical innovation was not new. The concept of a small, agile team of highly trained individuals who operated independently from a traditional bureaucratic structure was initially coined by Lockheed Martin’s Advanced Development Programs. They called their program “Skunk Works”—affectionately named because of the aroma in the warehouse from a nearby plastics factory. Skunk Works famously developed a number of innovative aviation technologies during the height of the Cold War, including the U2 spy plane and the F-117 stealth fighter.2 Soon, Skunk Works developed a reputation for solving difficult problems quickly and economically.

Senior leadership at Lockheed Martin played an essential role in the success of the Skunk Works. They developed processes and structures, established the organization’s direction, and built a culture that was receptive to change. Clarence “Kelly” L. Johnson took a handpicked group of engineers to form his Skunk Works organization and implemented an unconventional approach known as “Kelly’s rules.” The rules encouraged creativity with less oversight, smaller teams, and greater independence of individual team leaders over his or her own project. His vision and unusual approach to management created space for problem solving and empowered teams to take risks, unafraid of failure.2 Indeed, Harold Brown, United States’ Secretary of Defense during this time period, remarked that the strength of Skunk Works was “the autonomy they enjoyed from management and their close teamwork and partnership with their customers.”3 This close relationship allowed the design team to remain in touch with the products’ end users and ultimately generate a product that closely fulfilled their clients’ needs. The idea that engineers and designers needed to work side by side predated the vision of another innovator, Steve Jobs, who would start his small computer company in the next decade.4

They maintained momentum through an unwavering commitment of resources and a willingness to minimize the impact of administrative bureaucracy.5 Alan Brown, who worked in Skunk Works, stated that they had a straightforward formula to estimate the duration of projects: “the time it takes to go from initial design to operational use … is directly proportional to the size of the Air Force oversight committee that’s guiding the airplane design.” In other words, the greater the bureaucracy, the slower the progress.6 Military leadership clearly recognized that a larger number of stakeholders resulted in suboptimal processes, the death knell of innovation.7 By adopting a bottoms-up approach, the engineers had full autonomy during the design and development process. Today, the Skunk Works model of innovation has become the archetype for a company seeking to innovate rapidly outside organizational constraints in the manufacturing, technology, and even entertainment industries.

Health care in the United States, it can be argued, is in dire need of a disruptive innovation. The behemoth industry has been plagued by soaring costs and inefficiencies. In 2012, the Institute of Medicine estimated that the health care industry consumes roughly $750 billion a year—in part through redundant testing, paperwork, fraud, and other waste.8 In 2016, health care expenditures reached 3.3 trillion dollars, 17.9% of the gross domestic product (GDP), and are projected to reach a shocking 19.9% of GDP by 2025.9 Although the United States spends the most in comparison with all other developed nations, the health system still has extensive issues with the quality of care and the optimization of clinical services.10 It is estimated that >250,000 Americans die each year from medical errors, making it the third leading cause of death.11

Despite the widespread awareness and urgency for change, health care transformation has largely been stagnant.5,7 The American health care system is a patchwork of numerous stakeholders, an antiquated payment system with misaligned incentives, and a complex policy and regulatory environment, all of which serve as barriers to large-scale change.7 Against a sea of bureaucratic, policy, and legal constraints, health care leaders should consider the Skunk Works as a platform for change.12,13 In this review article, we will explore examples of health care IT innovation by focusing on the interoperability challenges of current EMR systems, the analytics applied to big data, and the technology that supports clinical decision-making.

Back to Top | Article Outline

EMRs: Scaling the Tower of Babel

After the development of VistA, the widespread adoption of EMRs did not occur until several decades later for civilian hospitals. Even then, it did not occur uniformly. In a 2009 survey by the American Hospital Association of acute care hospitals, only 1.5% of the responders reported a comprehensive EMR and 7.6% had a basic functioning system. Although the largest barrier to the adoption of an EMR was inadequate capital for purchase (74%), a significant number of respondents felt that physician resistance (36%) and the lack of local technological expertise (30%) also precluded implementation.14 Indeed, many physicians, the end users, felt that the commercial EMRs were difficult to navigate, time-consuming, and ultimately, disruptive to their clinical workflows. The lack of technical support, limited customization, and long project implementation times only exacerbated the perception that the EMRs were unwieldy and unnecessary.7,15

Despite the plethora of EMR platforms, many institutions were unable to find suitable EMRs.7 At Vanderbilt University Medical Center (VUMC), the Chairman of Anesthesiology directed a maverick group, a Skunk Works, to build a complete perioperative information system in 1995. They started with a customized perioperative EMR, named “VPIMS” for Vanderbilt Perioperative Information Management System. True to the Skunk Works model, the group operated independently within the institution and the medical center provided a separate funding stream, even though there were other substantial institutional IT development initiatives already underway. Over a period of 12 years, VPIMS grew in scope and sophistication.

In the beginning, the group faced a gap between the capabilities of the software and the hardware that was available at the time. In response, they installed the first wired network of microcomputers in the medical center. The installation of the network created its own issues. For instance, additional items that would be considered trivial today, such as mounting hardware, simply did not exist at the time. Without the administrative oversight typical of software development projects, the group was able to continually refine the product and clinical services into a superior documentation system.

As the scope of the project grew, software engineers used a modular format to tackle different aspects of perioperative care. Initial projects started with the preoperative evaluation process, followed by intraoperative anesthesia care, and finally nursing documentation within the perioperative services. The Vanderbilt group created new technologies that had not been envisioned at the beginning by using an iterative process common in design projects. For instance, they built and prototyped the first distributed schedule display system in the country, eORboard. Resembling the information display boards at modern airports, eORboard allowed anyone within the institution to determine the status of any operating rooms (OR) case on any particular day (Fig. 1).

Figure 1

Figure 1

By recognizing that the EMR could also support clinical care, the group implemented several decision support algorithms, many of which were the first of their kind. These pathways supported and streamlined clinical care, billing, and documentation processes. Using real-time active and passive alerts, these clinical decision support algorithms empowered health care providers to meet and stay abreast of changing regulations and quality initiatives. For example, VUMC became the first institution to provide reminders for antibiotic administration before incision in compliance with the national Surgical Care Improvement Project (SCIP) guidelines. Eventually, the decision support software expanded to include SCIP ischemia prevention and intraoperative glucose monitoring.16–18

The American Society of Anesthesiologists awarded first place to the VUMC group for its exhibit on the Vigilance application at the 2004 Annual Meeting.19 Pushing the envelope of the eORboard technology, the application combined live video feeds from the OR, live waveforms from patient monitors, and other patient-specific data. The data and video feeds were recombined into a single screen, selectable by OR. Using either fixed computer workstations or mobile devices, supervising clinicians could monitor the status of any patient in any OR from anywhere within the institution (Fig. 2). For this work, the VPIMS development group was featured on CBS Evening News in 2005.20 In a joint venture, Vanderbilt University and private investors brought the VPIMS product to the commercial market as Acuitec PIMS. Medhost later acquired Acuitec and continues to market a commercial version of the product as MedHost PIMS.

Figure 2

Figure 2

Back to Top | Article Outline

Nonzero: The Efficiency of Information Transfer

After the Health Information Technology for Economic and Clinical Health (HITECH) Act of 2009, many hospitals quickly adopted EMRs. Limited by the time constraints created by government lawmakers, most hospital systems chose commercial EMRs already on the market, recognizing that none of the systems communicated with their competitors. In a short time, hospital-based information silos dotted the United States.21,22 Rather than a unified medical information library, patients have “frequent flier cards” that are associated with separate hospital networks.23 Unfortunately, the Hardhats’ initial vision of integrated patient health information across a common platform never came to fruition.

For decades, medical professionals have recognized that the lack of interoperability creates information inefficiency that increases duplicate testing, generates unnecessary costs for the patient, hinders communications between health care providers, and increases the probability of error propagation.24 For example, when patient information can only be accessed within a single health system, physicians have to work from scratch if a patient moves between health care systems. However, finding financially sustainable solutions has been challenging. Federal guidelines for EMRs are vague, with variable interpretations. There is little incentive for commercial EMR vendors to increase functionality after a hospital system has already implemented their program.21 Some forward-thinking independent groups have addressed these problems with varying levels of success.

In 2008, Google rolled out a personal centralized health information database called Google Health with the goal of consolidating a patient’s personal information on a single platform. Driven by personal experiences navigating a disjointed health care system, Adam Bosworth, Google’s Vice President of Product Management during this period, spearheaded the project. In an official blog post published on the company website in 2006, Bosworth wrote,

… people need the medical information that is out there and available to be organized and made accessible to all. Patients also need to be able to better coordinate and manage their own health information.25

Google Health was offered free to the public. Patients would “opt into” their services, and update the database with their health history. Google Health could import medical and drug prescriptions. This information was maintained in a central database, and patients could grant physicians access to their Google Health records. In its terms of service, Google Health stated that it was not a “covered entity” under the Health Insurance Portability and Accountability Act of 1996, and as such, the HIPAA privacy laws did not apply. The reaction from the public was varied—some felt that Google likely protected personal information better than hospitals; others felt that the program was not secure.

The purpose of Google Health was 2-fold: first, lower the barriers to information transfer and second, increase patient engagement. In the years after the rollout, the majority of early adopters were tech-savvy individuals or fitness enthusiasts, mirroring the growth of similar technologies from Silicon Valley. Because of the personal effort and attention required to maintain the records and keep them up-to-date, the platform did not gain widespread use in the general public. In addition, providers were reluctant to use an online database as a source of patient information. In a 2011 survey conducted by IDC Health Insights, 7% of consumers had tried personal health records (PHR), but less than half continued to use them. More than other health IT, PHR required the patients’ effort and motivation. Although patients have reported strong interest in having their own PHR, the actual numbers of people adopting it have been low. Dunbrack, an analyst at IDC Health Insights, described PHR as “a technology in search of a market.”26 Finally, in January 2012, the program was discontinued.27 In retrospect, Google Health might have recognized these user and feasibility issues had health care providers been involved in the initial design process. It is noteworthy that Google Health did pave the way for open-source platforms and established a legal precedent for the concept of PHR.27

Other hospitals since then have tackled the problem of information silos. In 1996, Mark Smith and Craig Feied recognized the problems associated with accessing patient health information across different IT platforms when they attempted to tackle patient delays in the emergency department. Their hospital, Washington Health Center, was part of the MedStar Health system, a network of 8 hospitals and other health service sites in the Washington, DC area. Physicians at the hospital utilized numerous EMR databases. They realized that the health care providers were spending too much time gathering information and reconciling the different information systems, even though most of the necessary patient information existed in some electronic form.

Smith and Feied eventually designed a “middleman” computer software called Azyxxi that could interface with these databases.28 With this data interface, physicians could access individual patient data concurrently across disparate EMRs and a wide variety of data types (eg, past clinic notes, laboratory studies, and radiology images) to create a composite picture of a patient’s health history. Azyxxi organized the information into a custom format to highlight crucial information.29,30 At the Society for Academic Emergency Medicine conference in 2004, Dr Feied stated that,

… we have learned that the transformative power of information in the medical arena has little or nothing to do with making doctors and nurses type new data into the system and everything to do with getting existing data out.

The introduction, implementation, and installation of Azyxxi reduced patient waiting time, increased patient throughput, and doubled their annual census. It also markedly reduced manual medical record requests, presumably decreasing the risk of medical complications from delayed treatment of disease or incomplete medical history.

Moreover, by integrating information from legacy EMRs, Smith and Feied added an additional layer of surveillance monitoring across a health care system. This software, now called a unified health-intelligence platform, was the first of its kind. In October 2011, the system detected a spike in patients with flu-like symptoms in the Washington, DC, area and alerted the physicians to prepare for additional screenings.30 Eventually, the District of Columbia Department of Health adopted the Azyxxi platform for the surveillance and management of mass-casualty incidents. In 2008, Azyxxi was acquired by Microsoft and renamed Microsoft Amalga.31 The New York-Presbyterian Hospital, Johns Hopkins Health System, and the Wisconsin Health Information Exchange have adopted the commercially available platform.32 In 2011, Microsoft spun off Amalga and a number of other health-related services into a joint venture with GE Healthcare called Caradigm.33

Back to Top | Article Outline

A Tool for Benchmarking and Compliance

In 2005, the RAND Corporation projected that the rapid adoption of health IT would lead to increased efficiency, greater productivity, and potential annual savings of $81 billion for the United States.34 Twelve years later, the annual aggregate health care expenditure in the United States has grown by $1.3 trillion. Yet, the empirical data on its impact on outcomes have been mixed.35,36 So, how did the experts fail in their predictions?

The projections made by the RAND Corporation had many problems. First, they assumed that EMRs would be interconnected and interoperable on a shorter time horizon. Second, they envisioned that EMRs would be more widely adopted and used effectively. Finally, they believed that the health care industry leaders would recognize that technology would add increased data complexity (and costs) to their EMR systems. To provide value, health care systems must continually translate that data into actionable information, which can ultimately improve outcomes.37 Not surprisingly, none of these assumptions were correct.36

When EMRs were implemented, the majority of health care institutions used the technology to improve billing practices and facilitate clinical workflows, such as patient orders or medication entries. In essence, they adapted health IT to maintain current processes. The shift to using technology as a means of supporting best practice and standardizing processes was slower to follow.23 As a result, while the databases for EMRs grew in size and complexity, hospital leaders had not leveraged the technology to improve, let alone evaluate, clinical processes. Without a methodology to measure efficiency and efficacy, there remained no meaningful method to improve health IT.38,39

Kaiser Permanente expanded the capabilities of the EMR beyond the billing efficiencies. Kaiser is a Managed Healthcare Organization (MHO), which means that the payer and the provider are vertically integrated. In other words, they provide both insurance coverage and the health care delivery for their patients and they are incentivized to deliver cost-effective care. In 1999, when Kaiser Permanente rolled out an EMR that they co-developed with IBM, the organization quickly realized that they had logistical problems. Physicians were spending an additional 30 to 75 minutes to complete their tasks because of the cumbersome interface, resulting in significantly increased operating costs. George Halvorson, who came on as Kaiser Permanente’s CEO in 2002, pushed to abandon the existing EMR platform, writing off some $452 million in losses. He knew that having a fully functional, integrated IT network that complemented the end users was critical for the platform’s success. Although the sunk costs were daunting, the benefits of an interconnected EMR would have far-reaching benefits. In the next few years, Kaiser Permanente, under Halvorson’s stewardship, would invest another $6 billion in developing a new EMR platform.40

In 2004, Kaiser began installing a comprehensive IT platform that included an EMR system across care settings (inpatient and outpatient with real-time connectivity to other ancillary systems such as radiology and laboratory) and secure patient-to-provider and provider-to-provider messaging systems.41 Kaiser integrated care pathways across multidisciplinary teams by providing disease-specific care pathway protocols, documentation templates, and other clinical decision-making support tools. Health care providers could utilize KP HealthConnect’s clinical performance monitoring and tools to benchmark a patient’s chronic disease management. This rapid feedback improved clinical performance by encouraging adherence to best-practice guidelines and treatment protocols.42

Kaiser’s IT system aggregated patients’ data into registries, and compared the statistics against national and international benchmarks. As their EMR registries grew in complexity and volume, they began tracking outcomes against comorbidities. Using specific high-risk markers gleaned from the database, nurse managers were better able to assist patients with chronic disease conditions.43 For instance, after the Colorado region implemented the patient registry for cardiac disease and coupled it with telephone reminders from ancillary staff members, cholesterol screening increased from 55% to 97% of patients, and cholesterol control increased from 26% to 73% of patients.42 Although care pathways have improved and patient engagement has increased, initial studies have shown a concomitant increase in cost.41,43,44 It is speculated that the management of chronic health conditions, such as diabetes, hypertension, and hyperlipidemia, is more resource intensive, but it can also be argued that disease prevention has future potential cost savings.

The Geisinger Health System has coopted the capabilities of a well-functioning IT system in a different manner. Against a backdrop of increasing health care costs and changing landscape of policy and politics, Dr Glenn Steele, surgeon, oncologist, and CEO of Geisinger during this time, saw the inherent problems with the current fee-for-service payment model. In February 2006, Geisinger started “ProvenCare,” a 90-day “warranty” for elective coronary artery bypass graft (CABG) surgery. Under this new warranty, there would be a fixed fee for the surgery that would cover any additional costs of hospitalizations up to 90 days after operation. Under the old system, hospitals and providers were incentivized to do more. For a patient undergoing CABG, the hospital would be paid more if a patient underwent surgery and then required a re-operation for complications, than if he or she underwent one straightforward surgery. In short, Geisinger’s experiment was essentially a bet—that they could provide perioperative care for CABG patients for less than the amount that they billed the insurance company (calculated as the initial cost of the procedure plus 50% of the historic follow-up cost for a 3-mo period). With the hospital on the hook for considerable financial risk, hospital administrators and physician leaders understood that they needed to redesign the perioperative processes and the surgical procedure itself.45

What resulted was a restructuring of care delivery into evidence-based “bundles.” Working with Geisinger’s cardiothoracic surgeons, Geisinger implemented a list of 40 “best-practice” steps for every CABG patient, from their first clinic visit to their discharge from the hospital. Geisinger also designed their EMR systems to provide real-time feedback across services, making compliance easier to track and verify, and set up a series of reminders to ensure compliance.45,46 Although the process was strictly standardized, doctors could deviate from the clinical guidelines by providing the necessary clinical justification. With the overarching goal to reduce variation in the delivery of care for all patients, information continuity was crucial and all health care providers involved in an individual’s patient care had access to the same information.

When ProvenCare was first implemented, only 59% of the checklist was followed. When the program reached its 3-month mark, compliance climbed to 100%. Between February 2006 and 2007, Geisinger found that the ProvenCare group had 16% shorter hospital stays and a 5.2% mean decrease in hospital charges when compared against historical controls.47 As of 2011, Geisinger had a 67% reduction in in-hospital mortality, a 76% reduction in deep sternal wound infections, and a 10% reduction in overall complications.48

Although the concept of a warranty was new, Geisinger’s true innovation was in utilizing the EMR to standardize a clinical process. By restructuring the reimbursement system to pay-for-performance and making accountability a priority, Geisinger improved the clinical outcomes of their cardiac patients and concomitantly reduced the cost of care. After the success of ProvenCare, Geisinger has subsequently expanded their warranty to additional procedures such as bariatric surgery, perinatal care, and hip and joint arthroplasty. The Geisinger experience shows that effective health IT development coupled with the right incentives has the potential to markedly curb costs and improve the quality of health care delivery by changing culture.49

Back to Top | Article Outline

Big Data and Data Analytics: Making Sense of the Alphabet Soup

With advancements in health IT, patient monitoring systems, and network technology, there has been rapid accumulation of electronic patient data. This was in part accelerated by the rapid adoption of EMRs by hospitals after the US Congress passed HITECH in 2009.50 A report from IDC predicted an overall increase in heath data at 48% annually and estimates that the volume of health care data will increase to 2314 exabytes by 2020.51 Further, the application of analytics to data to create meaningful information requires resource management, predictive risk assessment, and clinical decision support.52

Some hospitals have undertaken the challenge by sifting through years of information from multiple sources, including previously unlinked data.52,53 Beth Israel Deaconess Medical Center (BIDMC), a Harvard-affiliated hospital, created a database that aggregated deidentified patients’ data from their intensive care unit (ICU) from 2001 onwards until 2012.54,55 BIDMC’s database was unique in that it encompassed a diverse population of ICU patients and a broad range of associated diagnostic and laboratory information (eg, electronic clinical documentation, bedside monitor trends, and waveforms).

In 2015, BIDMC began pushing live feeds of ICU patients into a custom program, called “Risky States,” that stratified a patient’s risk level at any given time. Designed in collaboration with integrated-systems scientists from the Massachusetts Institute of Technology and human-factors experts at Aptima Inc., the program aggregates data, provides analytical models and decision support systems, and sends health care provider alerts in real time. Using machine learning algorithms, the system identifies patients who are at high risk of developing dangerous complications in the ICU setting.56

Previous public ICU risk assessment calculators were the APACHE, SAPS, and MPM II scores, and these risk-stratification scoring systems were developed using patient data from the late 1980s and the early 1990s. Here, the risk assessment was based on different treatment strategies and only assessed a patient at a specific point in time. By contrast, the data for Risky States’ database are continuously updated and the algorithms to assess patient states evolve with changing practices. Over a 2-year period, Risky States developed into a surveillance system that translated data correlations and trends into actionable information that care providers could use.56 Interestingly, the group has identified clinical scenarios associated with greater patient risk that were not traditionally considered risk factors, such as a greater number of sicker patients in the ICU or a higher percentage of nurses with <1-year ICU experience. Ultimately, BIDMC’s goal is to leverage information extracted from big data to assist in clinical decision-making for physicians and nurses in patient care.

In addition to clinical decision support in ICU settings, tools and technologies targeting the OR have entered the market. Because of the immense volume, broad variety, and rapid presentation of information, providers may find it difficult to elucidate concerning trends in real-time data.57,58 We have all noted the increase in the number and complexity of monitors in the ICU and the OR; most modern anesthesia monitors are now capable of displaying over 20 periodic or continuous points of data describing a patient’s physiological status. As the human brain can only assimilate and act on a set amount of information at a time, this overabundance of information, technology, and alarms overwhelms a health care provider’s attention.58 To combat information fatigue, several companies have introduced visual analytics that integrate patient data with real-time interactive visual interfaces.

AlertWatch, created by a team led by Dr Kevin Tremper, is a real-time visual patient surveillance monitor. The system aggregates patient health information from different EMR sources (eg, medical history, recent laboratory values, baseline vital signs) and integrates them into a visual display (Fig. 3). For individual cases, the visual display presents the patient demographics and organ systems, which are color coded to alert the provider to patient comorbidities or concerning trends.59 Effective data analytics providing real-time feedback can be a powerful tool to guide clinical decision-making. When the University of Michigan developed and trialed AlertWatch in their ORs, they found that the real-time visual alerts increased provider adherence to hyperglycemia treatments and hypotension management.60

Figure 3

Figure 3

In a model similar to a “telemedicine” surveillance system, Barnes Jewish Hospital at Washington University in St. Louis established an anesthesia “control tower” utilizing the AlertWatch intraoperative alert system in 2016. In addition to the single OR view, AlertWatch also has a “control room” display, which places the ORs of the hospital onto a single screen (Fig. 4). The purpose of the control tower is to remotely monitor OR cases, providing support to anesthesia providers, and reinforcing best-practice guidelines. Individual ORs are represented with a box containing symbols that denote the status of the case (eg, anesthesia start, surgical incision, dressing placement). To facilitate rapid assimilation of clinical information, the color of the box serves an indicator. For example, green shows that all monitored systems are within the normal range, yellow is the marginal range, and red is the abnormal range. By integrating patient information into a visually identifiable system, AlertWatch allows a clinical director to quickly access general OR flow (eg, if a case was early or delayed) and the capability of focusing on an individual case with concerning trends.61 Ultimately, it is hoped that the application of machine learning techniques using big data will facilitate the development of clinical decision support systems that predict potential risk factors associated with poor outcomes in the surgical patient population.62

Figure 4

Figure 4

Back to Top | Article Outline

The Future

Recently, a number of factors have aligned to accelerate health innovation. First, the payment system is moving away from a fee-for-service model to a performance-based model, changing the incentive structure for care delivery. Although technologies such as telehealth and remote patient monitoring are largely not reimbursed at this time, hospitals may choose to invest in these programs because they can improve overall outcomes. Second, the issues of interoperability, although still problematic, will fade as health care systems continue to integrate. These systems, which serve large populations of patients under a single network, are motivated to make patient care more seamless across care settings. Third, there are rapidly developing technologies that support statistical modeling and machine learning algorithms. Clinical decision support technologies, by studying patient trends and associations, will change the way physicians approach patient care in the future and help tackle medicine’s challenges by managing big data. Medicine may remain a fragmented industry with different management structures, cultures, and local regulations. However, IT innovation needs be driven by a Skunk Works to meet the needs and challenges at the patient, provider, and institutional level to add value.

Back to Top | Article Outline

References

1. Allen A. A 40-year “conspiracy” at the VA, Politico. 2017. Available at: www.politico.com/agenda/story/2017/03/vista-computer-history-va-conspiracy-000367. Accessed November 17, 2017.
2. Lockheed Martin. “Skunk Works Origin Story.” Available at: www.lockheedmartin.ca/us/aeronautics/skunkworks.html. Accessed December 2, 2017.
3. Rich B, Janos L. Skunk Works: A Personal Memoir of My Years of Lockheed. New York, NY: Back Bay; 1994.
4. Leone D. Two former Skunk Works members seem to know why the F-35 program is a mess. 2013. Available at: https://theaviationist.com/2013/04/08/skunk-works-jsf-mess/. Accessed December 20, 2017.
5. Lukas CV, Holmes SK, Cohen AB, et al. Transformational change in health care systems: an organizational model. Health Care Manage Rev. 2007;32:309–320.
6. Herzlinger RE. Why innovation in health care is so hard. Harv Bus Rev. 2006;84:58–66; 156.
7. Institute of Medicine. Best Care at Lower Cost: The Path to Continuously Learning Health Care in America. Washington, DC: The National Academies Press; 2013.
8. Keehan SP, Poisal JA, Cuckler GA, et al. National Health Expenditure Projections, 2015-25: economy, prices, and aging expected to shape spending and enrollment. Health Aff. 2016;35:1522–1531.
9. Chassin MR, Galvin RW. The National Roundtable on Health Care Quality. The urgent need to improve health care quality. Institute of Medicine National Roundtable on Health Care Quality. JAMA. 1998;280:1000–1005.
10. Makary MA, Daniel M. Medical error—the third leading cause of death in the US. BMJ. 2016;353:i2139.
11. Hwang J, Christensen CM. Disruptive innovation in health care delivery: a framework for business-model innovation. Health Aff. 2008;27:1329–1335.
12. Porter ME, Lee TH. The strategy that will fix health care. Harv Bus Rev. 2013;91:50–70.
13. Jha AK, DesRoches CM, Campbell EG, et al. Use of electronic health records in US hospitals. N Engl J Med. 2009;360:1628–1638.
14. Randeree E. Exploring physician adoption of EMRs: a multi-case analysis. J Med Syst. 2007;31:489–496.
15. Jacques PS, Sanders N, Patel N, et al. Improving timely surgical antibiotic prophylaxis redosing administration using computerized record prompts. Surg Infect. 2005;6:215–221.
16. Sandberg WS. Anesthesia information management systems: almost there. Anesth Analg. 2008;107:1100–1102.
17. Higgins MS, Beattie C, Jaques PS, et al. The Vanderbilt Perioperative Information Management System Forms Foundation for outcomes management. Proc AMIA Annu Fall Symp. 1997;948.
18. Rosenberg AD. Scientific and educational exhibit award winners. ASA Monit. 2004;68:21.
19. Holguin J. Goggles give docs all-seeing power. CBS News. 2004. Available at: www.cbsnews.com/news/goggles-give-docs-all-seeing-power/. Accessed February 1, 2018.
20. McDonnell C, Werner K, Wendel L. Electronic Health Record Usability: Vendor Practices and Perspectives. Rockville, MD: Agency for Healthcare Research and Quality; 2010. Available at: https://healthit.ahrq.gov/sites. Accessed January 2, 2018.
21. Singh H, Sittig DF. Measuring and improving patient safety through health information technology: The Health IT Safety Framework. BMJ Qual Saf. 2015;25:226–232.
22. Sahni NR, Robert SH, Chigurupati A, et al. The IT transformation health care needs. Harv Bus Rev. 2017;95:128–136.
23. Stewart BA, Fernandes S, Rodriguez-Huertas E, et al. A preliminary look at duplicate testing associated with lack of electronic health record interoperability for transferred patients. J Am Med Inform Assoc. 2010;17:341–344.
24. Bosworth A. Health care information matters. 2006. Available at: https://googleblog.blogspot.com/2006/11/health-care-information-matters.html. Accessed January 24, 2018.
25. Lohr S. Google offers personal health records on the web. The New York Times. 2008. Available at: www.nytimes.com/2008/05/20/technology/20google.html. Accessed January 2, 2018.
26. Pearlstein S. Innovation comes from within. The Washington Post. 2005. Available at: www.washingtonpost.com/wp-dyn/articles/A5780-2005Mar3.html. Accessed January 24, 2018.
27. Dunbrack LA. Vendor assessment: when will PHR platforms gain consumer acceptance? IDC Health Insights. 2011.
28. Plaisant C, Lam S, Shneiderman B, et al. Searching electronic health records for temporal patterns in patient histories: a case study with microsoft amalga. AMIA Annu Symp Proc. 2008;601–605.
29. Feied CF, Smith MS, Handler JA. Keynote address: medical informatics and emergency medicine. Proc 2004 AEM Consens Conf. 2004;11:1118–1126.
30. Monegain B.Microsoft gets to work on Azyxxi project. Health Care IT News. 2008. Available at: www.healthcareitnews.com/news/microsoft-gets-work-azyxxi-project. Accessed January 24, 2018.
31. Microsoft. Microsoft announces the Amalga Family of Health Enterprise Systems. 2008. Available at: https://news.microsoft.com/2008/02/13/microsoft-announces-the-amalga-family-of-health-enterprise-systems/. Accessed January 25, 2018.
32. Herper M. How Microsoft moved out of one health business to protect another. 2011. Available at: www.forbes.com/sites/matthewherper/2011/12/19/how-microsoft-moved-out-of-one-health-business-to-protect-another/. Accessed January 31, 2018.
33. Girosi F, Meili R, Scoville R. Extrapolating Evidence of Health Information Technology Savings and Costs. Santa Monica, CA: RAND Corporation; 2005.
34. US Centers for Medicare and Medicaid Services. National Health Expenditure Accounts. Baltimore, MD: US Centers for Medicare and Medicaid Services; 2016.
35. Kellermann AL, Jones SS. What it will take to achieve the as-yet-unfulfilled promises of health information technology. Health Aff (Millwood). 2013;32:63–68.
36. Norman DA. Living with Complexity. Cambridge, MA: MIT Press; 2016.
37. Lighter DE, Fair DC. Principles and Methods of Quality Management in Health Care. Frederick, MD: Aspen Publication; 2000.
38. Sheikh A, Sood HS, Bates DW. Leveraging health information technology to achieve the “triple aim” of healthcare reform. J Am Med Inform Assoc. 2015;22:849–856.
39. Goldsmith J. An interview with George Halvorson: The Kaiser Permanente Renaissance, and health reform’s unfinished business. Health Aff. 2014. Available at: www.healthaffairs.org/do/10.1377/hblog20140930.041389/full/. Accessed January 24, 2018.
40. Chen C, Garrido T, Chock D, et al. The Kaiser Permanente electronic health record: transforming and streamlining modalities of care. Health Aff (Millwood). 2009;28:323–333.
41. McCarthy D, Mueller K, Wrenn J. Kaiser Permanente: Bridging the Quality Divide with Integrated Practice, Group Accountability, and Health Information Technology. New York, NY: The Commonwealth Fund; 2009:1–17.
42. Pearl R. Kaiser Permanente Northern California: current experiences with internet, mobile, and video technologies. Health Aff (Millwood). 2014;33:251–257.
43. Fireman B, Bartlett J, Selby J. Can disease management reduce health care costs by improving quality? Health Aff (Millwood). 2004;23:63–75.
44. Paulus RA, Davis K, Steele GD. Continuous innovation in health care: implications of the Geisinger experience. Health Aff (Millwood). 2008;27:1235–1245.
45. McCarthy D, Mueller K, Wrenn J. Geisinger health system: achieving the potential of system integration through innovation, leadership, measurement, and incentives. Commonwealth Fund. 2009;6:1–15.
46. Casale AS, Paulus RA, Selna MJ, et al. ProvenCare: a provider-driven pay-for-performance program for acute episodic cardiac surgical care. Ann Surg. 2007;246:613–621.
47. Geisinger Health System. 2011 System report. 2011. Available at: https://www.geisinger.org/-/media/OneGeisinger/pdfs/ghs/about-geisinger/news-and-media/annual-reports/ar-final-no-al-111711.pdf?la=en. Accessed January 15, 2018.
48. Shih T, Chen LM, Nallamothu BK. Will bundled payments change health care? Examining the evidence thus far in cardiovascular care. Circulation. 2015;131:2151–2215.
49. Charles D, Gabriel M, Searcy T. Adoption of electronic health record Systems among US non-federal acute care hospitals: 2008-2014. ONC Data Brief, No 9. Washington, DC: Office of the National Coordinator for Health Information Technology; 2013.
50. Corbin K. How CIOs can prepare for healthcare “Data Tsunami.” 2014. Available at: www.cio.com/article/2860072/healthcare/how-cios-can-prepare-for-healthcare-data-tsunami.html. Accessed January 17, 2018.
51. Hersh WH. Medical informatics: improving health care through information. JAMA. 2002;288:1955–1958.
52. Giambrone GP, Hemmings HC, Sturm M, et al. Information technology innovation: the power and perils of big data. Br J Anaesth. 2015;115:339–342.
53. Saeed M, Villarroel M, Reisner AT, et al. Multiparameter intelligent monitoring in intensive care II (MIMIC-II): a public-access intensive care unit database. Crit Care Med. 2011;39:952–960.
54. Johnson AEW, Ghassemi MM, Nemati S, et al. Machine learning and decision support in critical care. Proc IEEE Inst Electr Electron Eng. 2016;104:444–466.
55. Bacheldor B. Beth Israel launches big data effort to improve ICU care. Health Data Management. 2015. Available at: www.healthdatamanagement.com/news/beth-israel-launches-big-data-effort-to-improve-icu-care. Accessed January 18, 2018.
56. Simpac AF, Ahumada LM, Rehman MA. Big data and visual analytics in anaesthesia and health care. Br J Anaesth. 2015;115:350–356.
57. Lowe A, Jones RW, Harrison MJ. The graphical presentation of decision support information in an intelligent anesthesia monitor. Artif Intell Med. 2001;22:173–191.
58. Kruger GH, Tremper KK. Advanced integrated real time displays. Anesthesiol Clin. 2011;29:487–504.
59. Sathishkumar S, Lai M, Picton P, et al. Behavioral modification of intraoperative hyperglycemia management with a novel real-time audiovisual monitor. Anesthesiology. 2015;123:29–37.
60. Miller B. WashU researchers to test anesthesiology control tower. 2017. Available at: https://engineering.wustl.edu/news/Pages/WashU-researchers-to-test-anesthesiology-control-tower.aspx. Accessed January 17, 2018.
61. Avidan M. Anesthesiology control tower. 2018. Available at: https://clinicaltrials.gov/ct2/show/NCT02830126. Accessed January 14, 2018.
62. Rogers EM. Diffusion of Innovations. New York, NY: Free Press; 2005.
Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved.