There is tremendous pressure on health systems to serve more patients, particularly as the need for services rises and insurance disbursements decrease. Adding new facilities to accommodate demand is often not a feasible option due to budgetary/capital constraints nor is adding more doctors with nowhere to put them. As a result, healthcare leaders consistently ask whether they are getting all that they can out of existing resources.
Operating rooms are a prime example. They serve as the economic backbone for health systems, and organizations need to maximize the use of OR capacity if they want to achieve their fiscal and patient access goals. Yet effectively managing OR blocks and scheduling in the face of volatile weekly demand patterns can feel like trying to squeeze blood from a stone.
If a significant portion of OR time has been reserved as dedicated blocks for surgeons or service lines, unless time not needed is released efficiently, ORs end up both not being used during business hours and yet working late into the night. Patients wait longer than necessary for scheduling procedures and organizations lose revenue.
Additionally, poor OR utilization makes it extremely difficult to accommodate the new surgeons that health systems want to attract. Current providers have a lock on ORs far in advance, regardless of whether they may need the time or not and/or later release these blocks. Such practices leave few opportunities for new surgeons to secure time when needed.
To remedy the situation, health systems are turning to data to look for insights into what can be done to improve OR utilization. After all, small changes in utilization translate to big differences in patient access, revenue and profitability.
The Role of EHRs
As data insights move to the forefront in organizational decision-making, there is some confusion about what EHR systems do and the role that they play. In a nutshell, EHRs tell you what’s already happened. They are vital systems, absolutely necessary for describing problems within organizations and supplying the data to back up assessments.
Breaking this down further, the purpose of implementing an EHR is threefold:
Reservation system: EHRs provide health systems a way to reserve resources, whether it be rooms, infusion chairs, or clinic time. EHRs do this as well as enable scheduling of specific equipment or providers.
Single source of truth for patient encounters: Organizations need a way to maintain records for every patient encounter — a single instance for each enterprise deployment — and EHRs provide this.
Descriptive reporting: Health systems require reporting. EHRs generate reports based on what was done, sort of like how Charles Schwab tells you how your portfolio performed.
While these functions are all essential to running a successful health system, there are several things EHRs are NOT designed to do no matter how much teams may wish they could. For example:
By Brooke Faulkner, freelance writer; @faulknercreek.
Up to a fifth of patients with serious conditions are first misdiagnosed, and that leaves tremendous consequences. With the help of healthcare technology, doctors are able to diagnosis patients more effectively and easier. For example, migrating patient data from paper to online, known as electronic health records (EHRs), has greatly aided the medical world. Technology, especially using artificial intelligence and predictive analytics, has enabled doctors to make faster, more accurate diagnoses, and thus provide better care.
The volume of big data
Duquesne University estimated there to be 150 exabytes of healthcare data collected in 2011. Four years later, they reported about 83 percent of doctors had transitioned from using paper to electronic records. By now, with the ubiquity of the cloud, these numbers have assuredly gone up.
Massive amounts of data make predictive analytics possible, as trends can be spotted and analyzed. By spotting patterns, diagnosis of a disease becomes easier even for doctors unfamiliar with a specific disease or symptom. Uploading symptoms allows a computer to compare records and identify symptoms comorbid of other problems. This allows even specialized doctors to recognize issues outside of their field. Medical mistakes lead to the death of some 440,000 people each year; while misdiagnosis is only a part of this number, correct diagnosis and treatment will reduce it.
Big data can even be collected in the form of PDFs as part of telemedicine. A doctor can send PDFs to patients as part of a poll or survey or simply to collect symptom information from the patient. From there, data entered in the PDF can be collected and analyzed, generating patient data or feedback for the doctor.
Google flu trends
Google ran what can best be called an experiment from 2008 to 2014. Using artificial intelligence, the search engine recorded flu-related searches in an attempt to predict the severity of an outbreak, as well as the affected geographical area.
It was a flawed model, and tried to use big data as a replacement, rather than a supplement, for traditional data collection and analysis. It completely missed a flu outbreak in 2013, the data off by a massive 140 percent, and Google Flu Trends ended its public version in 2014. The algorithm monitoring flu-related search terms was simply not sophisticated enough to provide accurate results. While new data is no longer available to the public, historical data remains available to the Centers for Disease Control and other research groups. It’s possible that once the algorithm and predictive analysis is capable, the program will continue.
With digital information flowing from countless sources, including electronic health records (EHRs), wearable devices and digital maps that monitor global disease outbreaks, the healthcare industry is taking a big data approach to improving patient outcomes and enhancing the daily lives of countless others.
Yet one large part of the healthcare ecosystem isn’t efficiently capitalizing on the vast amount of data that is right at their fingertips: hospital operations. Many leaders and line managers are unable to take advantage of the readily available data sets from the billions of dollars invested in IT systems that would allow them to improve operational efficiency and provide an exceptional level of care. Instead, they rely largely on spreadsheets and back-of-the-envelope math along with first-hand experience to make critical daily operational decisions, such as scheduling operating rooms and reducing emergency department boarding.
Some hospitals have recognized that driving growth starts with superior planning and optimization. These forward-thinking facilities are leveraging their data using a hospital operations management software platform to revolutionize the operational and financial performance of various parts of their organization including ED, inpatient, perioperative and clinics.
The result? Higher resource utilization, better quality of care, satisfied employees and increased revenue.
Tap existing data resources to create growth
The OR suite’s multifaceted nature makes it extremely difficult to optimize for overall efficiency. As a result, millions of dollars are wasted each year. Still, it’s a primary financial driver for most hospitals and presents one of the largest opportunities for increasing profit margins through operational improvements.
In efforts to improve services while reducing expenses, perioperative leaders have been dependent on consultants, manual spreadsheets, and trial-and-error experimentation, leading to results that are often inaccurate, time-consuming, and have significant lag time to understand impact. These traditional methods should be abandoned.
Hospitals must embrace new analytics software platforms to deliver a practical application of analytics to the business of healthcare.
Predictive analytics utilizes the data gathered from existing EHRs, bed management, case management systems and external sources (such as weather) to quickly and easily see the potential impact of scheduling, staffing, and case mix changes. It empowers hospital leaders to develop optimized block and OR schedules that are easily managed and automates the staff planning and assignment process, all the while complementing traditional time and attendance systems.
At a time when hospitals cannot afford to mismanage valuable resources, analytics prevent costly trial and error and help hospitals overcome operational shortcomings in their perioperative suite. By testing “what if” scenarios, facilities can predict and manage the impact of operational changes for little expense or exposure.
Analytics software programs demonstrate measureable impact in OR operations. Impressively, analytics enabled a nonprofit medical center in Boston to increase its annual OR volume by three percent and improve utilization by more than five percent, resulting in an annual revenue increase of more than $3 million. And in New Jersey, a regional medical center used analytics to improve its OR utilization to 71 percent and improve labor productivity by 10 percent.
It’s impossible to see the future with certainty, but one branch of technology is playing a leading role in helping institutions and industries predict, on the basis of empirical research, the future behavior of participants and the outcomes of their decisions.
This relatively new branch of tech – predictive analytics (or PA) – has made inroads at a steady clip in the marketing, manufacturing and financial services industries. It is now gaining traction in healthcare as well.
Although debates around its ethical applicability to healthcare persist – the debate around data privacy, for one – the consensus emerging across the board is that with the right skills and in the right hands, PA has the power to effectively address challenges in the healthcare ecosystem in ways that human intelligence alone cannot.
Let us examine a few recent examples.
The power of PA
The Gold Coast Health Hospital in Southport, Queensland, Australia, dramatically improved patient outcomes and hospital staff productivity by applying a predictive model that was able to project with 93 percent accuracy emergency admissions before they happened. By analyzing admission records and details of sundry circumstances that led to patient admission to the ER, hospital staff were able to know how many patients would be coming in, on any day of the year, what they would be coming in for and methodically plan procedures that were now for all purposes elective rather than urgent.
Similarly, the El Camino hospital in California was able to drive a dramatic turn-around in its high rate of patient falls by collaborating with a tech company. The company, Qventus, linked patient EHR to bed alarm and nurse call light usage to derive an algorithm that was able to alert nurses in real time about the high-risk patients under their care and the exact times when they were most likely to be vulnerable. The result was a whopping 39 percent reduction in falls, improvement in patient health outcomes and a dramatically improved reputation for the hospital.
In fact, it isn’t only hospitals that are alive to the potential of analytics. Tech companies too are cognizant of how some of the newest technologies being developed under their roofs have immediate relevance to healthcare outcomes. In a paper published earlier this year, researchers associated with Google demonstrated how deep learning algorithms were able to correctly identify metastasized cancer tissue with nearly 90 percent accuracy as compared to just 73 percent when done by a human pathologist.
Predictive analytics and advanced labor management are the most important – and underutilized – methods to assure that provider organizations have the right caregivers in the right places at the right times.
A recent survey of nurse managers by AMN Healthcare and Avantas, Predictive Analytics in Healthcare 2016: Optimizing Nurse Staffing in an Era of Workforce Shortages, (available on the AMN website) brought the need for more awareness to light in just a few stats related to staffing and scheduling:
94 percent of nurse managers say that understaffing hurts staff morale
Nearly 70 percent say they are very concerned about the impact on patient satisfaction
More than 50 percent say they are very concerned about the effect on quality of care
The survey also revealed a lack of sophisticated scheduling tools being utilized:
24 percent use paper-based scheduling and staffing tools
19 percent use simple digital spreadsheets
23 percent don’t use any scheduling tools at all
Further, the survey found that while nearly 90 percent of nurse managers said that a technology that can accurately forecast patient demand and staffing needs would be helpful, 80 percent were unaware that such a solution exists.
Strategies to Fulfill the Potential Predictive Analytics
This process to predict future patient demand and strategically plan clinician scheduling and staffing is scalable, cost effective and accurate. First, staffing data are processed with advanced algorithms, then forecasting models are created and validated, customized for each unit or service area within the organization, allowing workforce projections up to 120 days prior to the shift. The forecast is updated weekly, and by 30 days in advance of the shift, the forecast of staffing need is 97 percent accurate.
Compared to how scheduling and staffing is conducted at most healthcare organizations today, predictive analytics may seem like something out of a sci-fi movie. The truth is, this sophisticated forecasting of labor needs has been leveraged in other industries with great success. And, in healthcare, it can lay the foundation for significant advancement in utilization of staff, leading to improvements in morale, quality, and financial results. The advanced labor management strategies and tools layered on an accurate projection of staffing needs – months and weeks in advance of the shift – will turn an accurate forecast into an effective resource management strategy.
Adopting Workforce Analytics Every organization’s staffing mix should be unique to the fluctuations in its patient volume. Once an organization understands its demand, it can then determine its supply – scheduling and staffing to meet patient demand in the most productive manner possible. The organization can analyze and solve the problems that reduce its available supply of core staff, such as leaves of absence, continuing education, training and other issues. This precision understanding of workforce availability is then layered with patient volume predictions, and the result is accurate insight into the core and contingency staffing levels needed to meet patient demand.
Guest post by Dan Hickman, chief technology officer, ProModel.
With six in 10 U.S. hospitals functioning at operational capacity, patient flow optimization provides one of the most cost-effective ways to increase a hospital’s bottom line.
Around 6 a.m. every day, hospital-wide “huddles” occur to discuss and determine a collective understanding of the state of operations. Most of these huddles take less than an hour and provide hospital and departmental leaders a snapshot of census status and expected discharges.
But hospitals are complex, dynamic systems. By 7 a.m. a flood of patients could hit the ED, affecting everything from staffing to the census, and carefully crafted plans disintegrate.
Consider the current state of patient flow at most hospitals.
Most health systems today have a reactionary approach to admit, transfer and discharge (ADT), patient flow, census, and staffing. Moreover, there is no way of accurately predicting future patient flows to right-size staffing and optimize workflows.
Discharge processes are open loop, resulting in costly delays. Most hospital staff use spreadsheets stating the number of discharges planned for the next 48 hours. However, there is no way to look at patient census with diagnosis codes tied to the typical length of stay.
The current state of patient flow results in multiple problems:
For many hospitals, the length of stay and cost per case metrics exceed CMS value-based care efficiency measures affecting reimbursements and the bottom line.
The daily reality of hospital staff revolves around logistics — the timely and accurate flow of patients coupled with staff, equipment, and facilities needed to accommodate and provide care within the hospital. Yet most hospitals lack the tools to define, visualize, predict and optimize the logistical flow of real-time needs into the near-term future.
Compounding the problem, many patients cite time spent ‘waiting’ as an issue affecting their experience, and ultimately patient satisfaction and the hospital’s HCAHPS scores.
Hospitals are really good at examining what’s happened to a patient in the past. The staff knows where they’ve been, but they haven’t taken the next leap, which other industries have, at projecting out where they think patients will “flow” during their stay and how the next 24 to 48 hours could affect the status and the census. There are parallels with other highly complex industries where accuracy and logistical management are critical to safety and success. One example — air traffic control.
Founded in 2012 by former home health agency owner Dan Hogan, Medalogix has been recognized by Harvard University, HIMSS and Fierce Healthcare IT as an innovative solution that’s improving America’s Healthcare system. Medalogix currently offers two solutions, Touch and Bridge.
Medalogix offers patient outcome management (POM) solutions that use a combination of predictive analytics, workflows and business intelligence engines to improve quality and reduce costs.
We’re a healthcare tech company that provides analytics and workflows to home health providers so they can improve care and reduce costs.
Medalogix currently offers two predictive analytic workflow solutions to home health providers:
Touch, which helps ensure a patient’s successful care episode by identifying patients at risk of transferring off census (i.e. hospital readmission), falling or who would benefit from an additional care episode. After identification, Touch then automates a clinical team’s patient touchpoints to encourage the best care and cost outcome.
Bridge, identifies which patients would benefit from hospice care and then offers a workflow to help clinicians have the right end of life conversation with the right patient at the right time.
Prior to founding Medalogix, Dan Hogan owned and operated a home health agency. As the industry moved toward digital patient records, he realized there was an opportunity to analyze those streamlined data sets to detect health risks the human eye might miss. He researched to find a tool that could do this, and came up empty handed. He decided to take matters into his own hands and incorporated Medalogix in February 2009.
Currently, Medalogix integrates with leading home health and hospice EMR. We market to their users through webinars, industry conferences and referrals through existing satisfied Medalogix users.
RightPatient is a division of M2SYS Technology, an ISO 9001:2008 certified company and biometric technology solution provider. M2SYS has more than a decade of biometric technology experience, with more than 300 million enrolled users in more than 100 countries.
RightPatient is the industry’s most advanced biometric patient identification, patient engagement, personalized healthcare and healthcare intelligence platform to reduce costs and liability, improve quality of care, monitor population health and enhance the patient experience. With features for wearable and biosensor integration, health games, medication alerts, proactive health management and predictive analytics, the platform also integrates with major electronic health record (EHR) systems such as Epic, Siemens, Cerner, McKesson, Meditech, IBM and many others. RightPatient is already deployed across hospitals and health systems that collectively maintain the health data of over 10 million patients.
RightPatient is the industry’s most advanced patient identification, patient engagement, personalized healthcare, and healthcare intelligence platform to reduce costs and liability, improve the quality of care, monitor population health, and enhance the patient experience. Our healthcare ecosystem unifies clinical knowledge through data aggregation, deep learning, and predictive analytics to personalize medicine, improve outcomes, and reduce re-admissions.
Our founder and CEO Mizan Rahman immigrated to the U.S. in the late 1990s seeking to turn some of his ideas, education, and experience into tangible products that solved problems for different verticals. He has successfully shepherded two companies from startup to a multi-million dollar companies that were eventually bought out.
Mizan now oversees the strategic and operational interests of the company worldwide combining his software engineering experience and entrepreneurial leadership with comprehensive international market intelligence to solve customer problems through identity solution ingenuity. He has successfully shepherded the growth of M2SYS as a global force in identity management, pioneering the development and launch of M2SYS’ Bio-Plugin biometric middleware and Hybrid Biometric Platform – both of which were recognized by Frost and Sullivan with prestigious awards for their design innovation.
Mizan continues to be recognized for his innovation and leadership in the field of biometric identification technology, most recently as recipient of “Technology Innovator of the Year” by InfoWorld. He is a frequent speaker in many US and international conferences, symposiums and universities such as the International Biometric Conference and MIT.
To better understand the fundamentals of predictive analytics — and why it has the potential to transform healthcare — it can be helpful to use Netflix as an illustrative example.
Let’s say, for example, you’re sitting around the house one rainy October Saturday and decide to view a few movies using Netflix’s streaming service. First you watched Field of Dreams then you decided, hey, this rain isn’t letting up any time soon, and that dog doesn’t want to go out in it any more than I do. So, after a brief backyard sojourn during which you and the dog confirmed that 38 degrees and rainy is in fact unpleasant — you reconvened your Netflixing and ended up watching Bull Durham, as well. Further, let us also assume that you enjoyed both movies and watched them all the way through.
As the credits rolled on Bull Durham, the critical question for Netflix was the same it always is: What would you enjoy watching next — specifically, what should Netflix recommend? Based upon the day’s viewing you may have a soft spot for baseball movies. Though it could just as easily be the case that you’re Kevin Costner’s biggest fan and the fact that you queued up two of his baseball movies was pure coincidence.
Given the uncertainty orbiting these pieces of information, maybe the best prediction would be Dances with Wolves, starring Kevin Costner. Or maybe the right pick would be Moneyball, the story of how the Oakland A’s leveraged data-driven, evidence-based sabermetrics to remain competitive against much more highly capitalized MLB teams. But what if neither Kevin Costner nor baseball is the most important correlation—what if the best predictor of whether you’ll like a film is simply whether it’s a sports movie from the late 80s?
As with all forms of predictive analytics, the question of what to recommend multiplies in complexity as overlapping variables (often in the form of unstructured data) are added and subsequently considered within algorithmic equations that power, in this case, Netflix’s recommendation engine. Further complicating the matter, it’s likely that you’re not the only person to have watched both of these films in close proximity and there are likely to be numerous “motivations” for such viewings across the population. It becomes apparent rather quickly the inherent challenge of something that seems, on the surface, as straightforward as a recommendation engine.
In healthcare we face these same kinds of challenges, just in a different form. The questions we ask are which gaps in care create the greatest risk for the patient, or which specific combinations of gaps in care correlate with readmissions—so that clinical outreach coordinators and other staff can prioritize whom to contact right away. We ask which types of claims are most likely to be under-coded or missing charges—so that organizations can make best use of finite resources like staff time and ensure the greatest positive impact on overall financial performance.
Hospital readmissions continue to be a major contributor to soaring healthcare costs and a drain on the U.S. economy. According to the Robert Wood Johnson Foundation, 4.4 million hospital readmissions account for $30 billion every year, while 20 percent of Medicare patients are expected to return to the hospital within 30 days of discharge. The Affordable Care Act of 2010 requires the U.S. Department of Health & Human Services to establish a readmission reduction program.
This program provides incentives for hospitals to implement strategies to reduce the number of costly and unnecessary hospital readmissions. Centers for Medicare and Medicaid Services (CMS) has created quality programs that reward healthcare providers and hospitals with incentive payments for using electronic health records (EHR) to promote improved care quality and better care coordination. The reasons for hospital readmissions include adverse drug effects (ADE), lack of a proper follow-up care, inability of patients to understand the importance of their medications and diagnoses, unidentified root causes, and misdiagnosis. Technology could play a vital role here by properly documenting, tracking, diagnosing, monitoring, and enabling better communication between patient and provider.
Adverse drug events constitute the majority of hospital readmissions. A cohort study, including a survey of patients and a chart review, at four adult primary care practices in Boston (two hospital-based and two community-based), involving a total of 1202 outpatients indicated that 27.6 percent of these ADEs were preventable, of which 38 percent were serious or fatal. Human error was the leading contributor to these ADEs, followed by patient adherence. Additionally, patients who screened positive for depression were three times as likely to be readmitted compared to others.
Our analysis indicates that 28 percent of adult hospital stays involved a mental health condition. A study of Medicaid beneficiaries in New York State determined that, among patients at high risk of rehospitalization, 69 percent had a history of mental illness and 54 percent had a history of both mental illness and alcohol and substance use. We know that a properly implemented mental health screening protocol can lead to effective diagnosis, and that proper management of these issues can positively impact the reduction of hospital readmissions.
Further studies show that most cases of readmissions for certain chronic conditions have an underlying mental health issue, which appears in patients who have not been previously diagnosed for a mental health condition (i.e., anxiety, bipolar disorder or depression). For example, anxiety and/or depression increases the risk of stroke and decreases post-stroke survival, and plays a key role in diabetes treatment as 33 percent of this patient population is found to be depressed and patients with bipolar disorder have reduced life spans. Other cases where depression affects the patient’s survival and treatment cost include hypertension, stable coronary disorder, ischemia, unstable angina, post myocardial infarction and congestive heart failure.
An important point to note: congestive heart failure is the major driver of hospital readmissions in the U.S., accounting for 24.7 percent of all readmissions. Another study concluded that patients with severe anxiety had a threefold risk of cardiac-related readmission, compared to those without anxiety.