Wednesday Recap: 2016 Healthcare Analytics Summit™ Kicks Off

has16-logoIn front of more than 1,000 attendees, the third annual Healthcare Analytics Summit launched today, on September 7, 2016, in downtown Salt Lake City, Utah. It began with a reminder of healthcare’s shared purpose: to scale outcomes improvement. And to do so using data and analytics. Dan Burton, Health Catalyst CEO, kicked off HAS using data to show that majority of the audience is in the midst of analytics adoption, demonstrating that we are all in the middle of our outcomes improvement journey.

Burton explained that the summit would focus on essential elements for meaningful, sustainable change. He noted topics including:

  • Scaling outcomes in meaningful ways
  • Realizing ROI analytics investment
  • Training and developing analysts
  • Governance to improve outcomes and scale

Burton also made a point of honoring the collective level of expertise among HAS attendees. According to polling before the summit, the largest group of participants has between 10 and 20 years of experience in healthcare. Burton also touched on other pre-event polling questions—such as how rapidly we’re moving toward value-based care and the effect of this shift on quality of care—and reported that the answers were mostly optimistic. He added that 2016 attendees were largely healthcare providers (those with direct impact on healthcare delivery), and that participants projected that social determinants of health would emerge as the most important new data set.

Keynote: Criminal Justice Analytics and Insights for Healthcare

Anne Milgram, former New Jersey Attorney General, Senior Fellow at NYU School of Law

If you hadn’t previously seen similarities between analytics in healthcare and in the criminal justice system, Anne Milgram’s keynote left no doubt that the two share many characteristics and challenges. Health Catalyst Executive Vice President Dale Sanders introduced the former New Jersey Attorney General and Senior Fellow at NYU School of Law with a comparison between criminal recidivism and unplanned hospital readmissions. He proposed an overlap between social determinants of crime and health, alluding to shared strategies (namely, analytics) for improving outcomes in both sectors. In addition, both health and crime are traditionally driven by addressing the problem, not preventing it: just as criminal justice thrives on crime and not innocence, healthcare thrives on sickness and not health.

Milgram expanded on these topics but outlining what’s been lacking in the criminal justice system. In summary, she pointed to underutilization of technology and data and a system that was failing to make us safer. The system, Milgram said, has needed to progress from anecdote-driven decision making to a data driven approach. “What are the trends? Who are the outliers?” she asked—questions that could are only accurately answered with big data.

To further strengthen parallels between fighting crime and healthcare, Milgram noted that they are both massive systems that impact countless lives and are incredibly expensive. The opportunity is clear for each entity to learn from the other—with the support of analytics. Areas included:

  • The criminal justice system focuses on backend activities—arrest, booking, trial, incarceration—rather than prevention; healthcare focuses on treatment rather than prevention.
  • Both are massive systems, with the criminal justice system presiding over countless lives and healthcare presiding over almost everyone in the U.S.
  • Both are very expensive systems with criminal justice system costs running in the billions and healthcare costs running in the trillions (or as Milgram puts it, “I have a B and you have a T.”)
  • Both need to focus on avoidable readmissions (whether it’s jails or hospitals).
  • Both need to be data driven and, in doing so, must convince their service providers (clinicians or law enforcement) that presenting data isn’t a game of “gotcha.”
  • The social determinants of health have a large overlap with the social determinants of crime.

Keynote: Reducing Waste at Intermountain Healthcare: The Vision, Mission, and Tools to Change Everything

Jay Bishoff, MD, FACS, Director of the Intermountain Urological Institute

What sets one healthcare organization apart from all the others? Facilities? Providers? ERs? Helicopters? In reality, the vision sets health systems apart. And while the vision is often clear to administrators, it’s not clear to the frontline.

When it comes to reducing healthcare waste, a majority of the audience believes 40 percent of healthcare is wasteful. At Intermountain Healthcare, the focus is on measured experience – using data to tell the story. Delivering the measured experience is the gateway to excellence. Unfortunately, healthcare has a long way to go. Health systems can start by focusing on what causes the most frustration, waste, or poor outcomes, while remembering three important things about processes and people:

  • Every process is perfectly designed to deliver the results you’re getting.
  • 85 percent of the time the frustration is caused by your process.
  • Only 15 percent of the time it’s caused by people.

Systems need to look at their biggest opportunities for cost savings using a four-pronged approach:

  1. Align incentives
  2. Appropriate utilization
  3. Shared Accountability
  4. Guaranteed Insurance Price

It’s not enough to show providers the data and expect them to change; data needs to be the basis for dialogue and the focus needs to be on the improvement plan. Don’t beat providers up with data; data is the secret to success, but can be demoralizing. Stop thinking in terms of my experience; start thinking in terms of measured experience.

More specifically, Bishoff looked at reducing waste in the healthcare system and posited that by focusing on quality (tests, instruments, and more that are producing the best outcomes), systems could reduce cost while also improving care. In addition to data (what’s working, what’s not), transformation requires system-wide understanding. This means that everyone from cleaning and cafeteria staff to top-level executives must be able to communicate and understand the same vision

The road to improvement, Bishoff said, relies on doing things the same way instead of the right way. We have to learn through this process, he explained, in order to determine what the “right” is. This is why measured (data-driven) experience is more useful to determine quality than experience alone.

Keynote: Population Health: Lessons from One of the Nation’s Most Innovative Rural Community Models

Documentary: The Story of New Ulm—A Population Health Transformation

Dale Sanders, Executive Vice President, Health Catalyst and Toby Freier, MBA, FACHE President, New Ulm Medical Center

According to Dale Sanders, we are overcomplicating population health, which boils down to personalized care and social determinants. New Ulm Medical Center is leading the country in changing individual behavior by leveraging the power of community to significantly improve population health. New Ulm’s ten-year project, The Heart of New Ulm, had the vision to reduce heart attacks by focusing on simple interventions: being active, eating healthy, and engaging every facet of the community to make it easy for members of the community to be healthy. New Ulm has stabilized obesity, improved nutrition, reduced hypertension by seven percent, and lowered admissions by 20 percent.

It is a challenging time in healthcare because systems are tied into a reimbursement system that pays for people to be in the hospital. But our mission is to prevent illness; a mission that requires data and the support and resources of entire health systems. Population health is a collaboration between systems and the community; it takes a village.

Moving from a disease basis to a prevention basis turns the industry model upside down. Communities have to make it easy for people to do the right thing. Population health is a quality of life issue. The industry can learn from small, 14,000-person, salt-of-the-earth communities such as New Ulm to learn what works. Population health is working in New Ulm, and it can work in large communities too. New Ulm’s community-based approach is a national model for population health that centers on the fact that all change happens at the community level.

This year’s Healthcare Analytics Summit featured the debut screening of The Story of New Ulm, documentary film. A community that prides itself on the three “Bs” (beer, bratwurst, and butter) might seem like the least likely place to succeed with a healthy lifestyle initiative. But improving health, largely through behavior, is exactly what happened in the Minnesota town of New Ulm.

Breakout Session 5: UPMC’s Systemwide Change to Service Lines—Supported by Activity-Based Costing: The Blueprint to Healthcare Improvement Efforts

Robert Edwards, MD, Milton Lawrence McCall Professor, Chair Department of Obstetrics, Gynecology, and Reproductive Services, Magee-Womens Hospital; and Paula Lounder, Director, Payer-Provider Programs, UPMC

UPMC believes using a service line approach is the best approach for improvement efforts. Using service lines, defined as an aggregation of services provided to patients with similar medical conditions (along with the revenue and expense components), UPMC focuses on three solutions for initiatives: 1) an evidence-based clinical pathway; 2) monitoring progress through measurements; and 3) engaging physicians. This case study session looked closely at the Women’s Health Service Line and showed how UPMC used the methodology to produce a 20 percent reduce in length of stay for hysterectomies and a 200 percent increase in same-day hysterectomies, the latter of which led to an estimated cost savings of $250,000.

Breakout Session 6: Predictive Analytic Models – a Must in the Journey to Reducing Readmissions

Karen Tomes, RN, MA, PHN Vice President Care Management & Coordination, Allina Health

As part of their initiative to improve quality care measures, Allina Health used data analytics to determine that if they could eliminate a potentially preventable readmission (PPR) within the seven-day window after the patient was discharged, they would prevent a PPR from happening in the months to come.

After further analysis, they discovered that 60-80 percent of their patients had a complex need during care transition, leading to an increased likelihood for a PPR. To address this care transition need, Allina Health developed complex discharge planning teams, composed of nurses, social workers and pharmacists. The multidisciplinary teams evaluate the care management resources appropriate for complex discharges and hold “transition conferences” with the patient and the patient’s caregiver(s).

These conferences are heavily focused on the patient, with approximately 70 percent of the time allocated for patients and caregivers to share their barriers to care with clinicians, allowing clinicians to then explore alternate care options such as the hospital’s transitional care unit (TCU) or home health services that might better fit the patient’s needs and reduce the risk of PPR.

The care transition conference has been a vital tool in reducing PPRs for Allina Health, and at the same time, it is a significant change in culture that involves some heavy lifting by the care team. As we’ve seen with Allina Health’s success, it’s a cultural journey worth taking!

Breakout Session 7: Actionable Analytics: From Predictive Modeling to Workflows

Chad Konchak, MBA, Director, Clinical Analytics, NorthShore University Health System

NorthShore University HealthSystem, a four-hospital, 950-bed organization, was intent on integrating predictive models and analytics tools into clinical workflows to support data-driven decision making. NothShore’s analytics team, led by Director Chad Konchack, successfully added analytics to workflows by taking several components into account:

  1. Data sources (EMR, patient satisfaction, claims, etc.).
  2. Standardization and normalization (e.g., security flags).
  3. Data enrichment (patient registries, predictive analytics, and data grouping).
  4. Workflow (point of care CDS, physician portal, care coordinator portal, patient portal, administrative outreach, and administrative portal).

Analytics governance is a crucial element of successfully integrating analytics into clinical workflows. The NorthShore analytics team created a 200-page predictive analytics manual as part of its systemwide effort to democratize analytics.

NorthShore’s models and tools support a wide range of initiatives. For example, the analytics team created a heart failure model to generate a list of high-risk patients for the advance care planning team. The team also developed a logistic regression model to generate a list of patients at a high risk of readmission, with the goal of scheduling follow-up appointments within two days for these patients. NorthShore’s flagship tool, a provider dashboard called MyPanel, has dramatically improved quality scores across several care gaps (e.g., mammography and colon cancer screening).

Health systems should start small with a specific use case in which the needs for analytics is clear. They should get commitment from operations to test and iterate new tools, and have a clear plan for helping new users use these new tools.

Breakout Session 8: Predictive Analytics: The Power to Predict Who Will Click, Buy, Lie, or Die

Eric Siegel, PhD, Best Selling Author and Founder of Predictive Analytics World

Eric Siegel began his breakout session with this question: How do we establish credibility for a field that has the audacity to call itself predictive analytics? He then proceeded to provide multiple examples of how they were used to turn the tide in the 2008 presidential election. The Obama campaign broke ground by applying predictive analytics to predict whether voters were persuadable. Then they knocked only on those doors of those voters, a method known as persuasion modeling.

Predictive analytics are the information age’s latest evolutionary step and they have lots of applications in healthcare. Predictive analytics improve patient care, reduce cost, and bring greater efficiencies via a wide range of applications in clinical practice, marketing, insurance, and administration. They work by learning from data (modeling).

Breakout Session 9: Building an Enterprise Analytics Organization

Joseph M. Dudas, Division Chair, Enterprise Analytics, Mayo Clinic                                     

When Joseph M. Dudas, Division Chair, Enterprise Analytics at the Mayo Clinic, looked at driving change within his organization, he asked the following questions:

  • Is my organization ready to take advantage of these technologies and drive change, or do I need to change the culture within the organization?
  • Are we keeping up with massive amounts of information we have? Are we in a position to take advantage of it?

The difference between top and underperforming organizations, Dudas said, is that top performers use analytics to determine future strategy, research and development, and sales and marketing. He explained that to be successful with a strategic plan, you need to know where you are and where you want to go.

Other organizational strategies Dudas outlined included:

  • Resetting expectations of analytics.
  • Repositioning data.
  • Focusing on data quality
  • Designing for self-service: easy to use tools for all staff
  • Determine potential economic impact, or ROI. The goal is to increase revenue and decrease cost.
  • Shift resources.
  • Tap into discretionary funds.

Breakout Session 10: FHIR’D Up About Clinical Data Intelligence: Cleveland Clinic’s Real Time Decision Support System

Andrew Brookhouse, Lead Systems Analyst, Cleveland Clinic; and Suzanne Fink, MSN, RN, CCRN, Clinical Analyst, Clinical Solutions Center, Cleveland Clinic

A lot of worthy focus is placed on EMR interoperability and new web data standards for healthcare. But without data fidelity, all these efforts are moot, said presenters Andrew Brookhouse and Suzanne Fink of Cleveland Clinic. The solution is elusive for most health systems, in part due to the data challenges of EMR’s that are notoriously onerous to customize—as well as a lack of understanding of the clinical nuances of stored data.

Cleveland Clinic has made some breakthrough discoveries enabling the development of a real-time decision support system that uses new data standards like FHIR to deliver some significant results. Cleveland Clinic started by constructing a team of developers and clinicians with overlapping clinical and technical knowledge that use agile development methods rather than traditional waterfall processes to drive innovative new approaches to old problems. The new Clinical Solutions Center has demonstrated success by persistently and patiently navigating a bureaucratic structure of policy surrounding its legacy enterprise EMR, and internally developing new EMR capabilities.

Keys to the success of the Clinical Solutions Center include their commitment to hiring the right team and basing them close to where clinical work is done. The Clinical Solutions Center is on the main campus of Cleveland Hospital so team members are within walking distance of the nurses and physicians they serve. A focus on “user empathy” has also been a critical success factor. “User empathy is the foundation of all we do – ideas are overrated,” said Brookhouse. The team takes as much time as possible to understand end-user problems before generating ideas and only then turns to development. “Create and create often,” is another mantra of the group, which shares solutions with end-users at several points throughout the development process to gather feedback and encourage end-users to feel part of the process. As a result, the end-users tend to become evangelists when the solution is rolled out.

Breakout Session 11: Patient Don’t Measure Quality Care—They Experience It

Glenn Drayer, Director of Enterprise Business Intelligence, Stanford Health Care; and Mysti Smith-Bentley, RN, MBA, Administrative Director, Service Excellence, Stanford Health Care

When Stanford Health Care saw that 59 percent of patients want to be able to use search results to inform decisions about healthcare providers, it saw an opportunity to increase transparency and feedback for its physicians. Using comments from Press Ganey patient satisfaction surveys, Stanford has created online profiles and rating for physicians. Additionally, physicians can use the comments to improve patient care. This initiative has shown remarked results. In 2011, Stanford was in the 63rd percentile for likelihood to recommend (according to the Press Ganey surveys). In 2015, that jumped to the 89th percentile. A key takeaway from the program is to be empathetic to physician concerns about comments by outlining (then using) an appeals process. Physicians and patients alike feel heard—and it shows.

Breakout Session 12: From the Boardroom to the Bedside—Using Analytics to Drive a Culture of Continuous Improvement

Chris DeRienzo, MD, MPP, FAAP, Chief Quality Officer, Mission Health System; Jon Brown, Chief Information Officer, Mission Health System; and Michael Creech, BSEE, LSSBB, LSSMBB,Vice President, Process Engineering and Applied Analytics, Mission Health System

Mission Health believes it is dead in the water without a culture of continuous improvement grounded in analytics that permeates everything we do and all that we are. The need for improvement in today’s healthcare industry is immense: 70 percent of hospital strategic initiatives fail.

Mission Health, one of the largest ACOs in the country, is motivated by its AIM to get every person to their desired outcome, first without harm, also without waste, and always with an exceptional experience. On its quest to achieve this AIM and continuous quality improvement, it launched process improvement activities in 2012 (e.g., clinical programs with standard algorithms of care) but had difficulties accessing data due to many unintegrated systems (13 EMRs, 12 billing systems, and 350 IT applications).

Mission overcame these difficulties by remaining focused on and dedicated to process transformation after early wins (which always lead to more difficult and complex follow-up projects) and clinical program development—both require a great degree of empirical, data-driven evidence to compel difficult process changes.

Mission’s continuous improvement success is due to having the right data infrastructure in conjunction with process improvement efforts:

  • EDW
  • 30 Care Process Models
  • Advanced Analytics Teams

Ultimately, Mission learned that tools without purpose won’t be adopted, physician leadership is essential, clear ownership of processes are critical, and tailored analytics tools are more likely to be used. Creating a culture of data-driven continuous improvement requires having the right tools, the ability to use these tools, and the ongoing leadership support.

Breakout Session 13: An 85% Prediction Model? Advances in Sepsis Prediction at Johns Hopkins

Suchi Saria, PhD, Assistant Professor, Johns Hopkins University and Nishi Rawat, MD, Assistant Professor, Johns Hopkins University

Suchi Saria, PhD, an Assistant Professor at Johns Hopkins University Bloomberg School of Public Health, and Nishi Rawat, MD, an Assistant Professor at Johns Hopkins’ Armstrong Institute for Patient Safety and Quality, described the development of a sepsis early warning tool that predicted septic shock with 85% accuracy on a test data set. They began their 6-year study by engaging systems engineers to analyze the barriers to sepsis detection. They determined that a successful sepsis program should have three components: an alert that is accurate vs. existing alerts, which have high false positives; real-time tracking including decision support telling providers what to do and when; and more lead time to comply.

The team then analyzed 8 years of data on over 10,000 patients using natural language processing to find all of the early signs and symptoms differentiating patients with sepsis who went into septic shock from those who didn’t. They reviewed hundreds of markers, all routinely collected so any algorithm they eventually developed could work in the background in real time without the need for further data entry. The study revealed that current sepsis models fail to consider enough context to provide reliable predictive alerts. For instance, a patient who has not yet developed certain trigger conditions for a sepsis alert may be taking medications that would prevent those trigger conditions.

Based on this research, the team developed a tool called TREWS (targeted real-time early-warning System) that monitors sepsis risk in real time using routinely-collected data; contextualizes the signals to reduce false alerts; helps providers understand why an alert is going off to help with decision support; and provides actionable information leading to collaborative decision making. TREWS was validated on a data set of 14,000 patients, finding that it detected 2/3 of the sepsis cases before organ dysfunction developed. Compared to routine screening protocols, TREWS identified 60% more patients prior to sepsis, most of them 24 hours prior to development of septic shock, and with 85% accuracy. The algorithm and the study results have been published. In its last stage of testing in collaboration with clinical teams, these tools may soon become available to others.

Breakout Session 14: The Geisinger Hedged Unified Data Architecture

John Kravitz, MHA, CHCIO, Senior Vice President and CIO, Geisinger Health System; and Alistair Erskine, MD, Chief Strategic Information Officer, Geisinger Health System

Consumerism is driving the need to improve in healthcare because patients now have choices. But healthcare is unprepared for how to deal with this. Geisinger confronted this challenge by displaying physician performance statistics online for all the world to see. They also created a refund application to guarantee satisfaction. This allowed the organization to learn what was wrong and to ask patients how much of a refund they wanted. Novel ideas, but they demanded completely different sets of data that existing transaction systems couldn’t yet handle.

Geisinger has hundreds of facilities and affiliates, covers two thirds of Pennsylvania, and generates $6.7 billion in annual revenue. Geisinger had to look at their existing data warehouse and other systems to see how they could meet these new demands.

Challenges with their existing EDW were similar to what many healthcare systems experience: too many pockets of data; too many undocumented data sources; too much time spent entering data manually; clinical data quality problems related to patient safety exist; hierarchies at many levels; no data dictionary.

Geisinger needed a new enterprise data strategy, so they chose a unified data architecture to move from data silos to a unified data platform; move from data sprawl to data integration; decrease costs from $500K to $15K per 10 TB; and move from no scalability to scalability.

Kravitz and Erskine gave a live demonstration of their architecture and impressed the audience with the amazing speed of their Hadoop real time data warehouse.

Breakout Session 15: Improved Outcomes and a Proven ROI Model for Quality Improvement: Transforming Diabetes Care

Charles G. Macias, MD, MPH, Chief Clinical Systems Integration Officer, Director of the Evidence-Based Outcomes Center / Center for Clinical Effectiveness, Texas Children’s Hospital, Associate Professor of Pediatrics-Emergency Medicine, Baylor College of Medicine

During his work in the emergency department, Charles G. Macias, MD, MPH, of Texas Children’s Hospital, identified a critical need within his organization to improve quality and consistency of care for diabetic ketoacidosis (DKA) within his organization. He observed such variability in as length of stays (LOS) between 2 to 5.5 days, depending on the provide. TCH clearly needed to an enterprise-wide campaign to drive diabetes improvement and a dedicated diabetes care team.

Firstly, all teams involved in diabetes care needed to integrate (many teams didn’t even know each other). They also need to address inefficiencies in current care delivery that affected LOS.

With his improvement initiatives, Macias saw improvement in the following areas:

  • Standardization of care
  • Hospital throughput
  • LOS
  • Patient satisfaction
  • Financial

Breakout Session 16: Security Frameworks in Data Warehousing and Their Interplay with Healthcare Analytics

Patrick Nelli, SVP of Analytics, Health Catalyst

Although some healthcare analysts might consider data utilization and security/privacy a zero-sum game, that simply isn’t true with the right strategies. Below we share our lessons learned for balancing data utilization and security/privacy.

  1. Data is useless if you don’t put it in the hands of analysts, operators and clinicians. Health systems need to strike a balance between security/privacy and data exposure.
  2. Logging is not enough. One needs to make the data actionable through monitoring, with search and business intelligence (BI) on top of logs. This can lead to multiple benefits including:
    • Improved security/privacy
    • Increased performance efficiencies
    • Enhanced product development
  3. Data de-identification is typically not a good balance of utilization and security. Instead, health systems should focus on storing the personal health information (PHI) in a secure environment.
  4. Cloud environments, if set up correctly, assist with the balance of data utilization and security. Cloud providers offer cost-effective, diverse environments with advanced security that enable quick and easy implementation, even for the more complex distributed environments. However, health systems should be aware of what their responsibilities are during an audit as opposed to those addressed by the cloud provider.

In short, by following the aforementioned best practices, health systems can find the center of its seesaw of data utilization and security/privacy, resulting in optimal data analytics and actionable insights that support the quest for better patient outcomes.

Keynote: The Price Is Right 2.0: Designing Population Health to Thrive in a Value-Based World

Tom Burton, Co-Founder and Executive Vice President, Health Catalyst

Analytics needn’t be dry; it can be fun and entertaining, as demonstrated by Health Catalyst’s version of The Price Is Right 2.0: Designing Population Health to Thrive in a Value-Based World, which showed how analytics can improve population health in an innovative, game-based way.

The Price Is Right was powerful learning session. Contestants plays for great prizes, using healthcare improvement knowledge, and a willingness to learn, to conquer a patient–healthcare plan matching game, the Plinko board, and the chance to rank improvement opportunities (to win at this round, each opportunity had to encompass all components of the three systems approach). The gist of each challenge was going from no data, to some date, and finally to as much data as possible. With more data, players were able to make accurate choices and win big. Each game clearly demonstrated that without data, we can’t make accurate decisions and improve outcomes.

Principles that the games covered included:

  • Informed contracting
  • Engaging all stakeholders
  • The challenge of predicting costs.
  • Ways of engaging patients.
  • Specific interventions needed to improve payment structures.
  • How to use analytics to thrive under system of value-based payments.
  • High-risk patients make up a vast majority of costs, so healthcare organizations need to identify them; predictive and retrospective analytics are required to do this.
  • Prescriptive analytics are required to identify and prioritize opportunities to improve systems of payment.
  • Need to look for ways to reduce variation in care delivery; look at high opportunity areas that have high volume or high costs as well as high variability.
  • Three systems are needed to succeed in managing population health: analytics, best practice and adoption.

The game began with contestants bidding on various populations (diabetes, back and neck pain, etc.). In an industry where systems are entering into at-risk contracts without much information, systems can use analytics to evaluate how per member, per month costs are trending. Retrospective analytics is as important as predictive analytics, which enable systems to identify rising risk patients and predict future utilization.

Contestants also tried to assign patients to the right care plans. At first, contestants had no information, demonstrating how difficult it is to make good decisions without data. As contestants learned more about each patient, they more accurately assigned patients to care plans, demonstrating the importance of collecting as much data (claims, financial, clinical, etc.) as possible about patients.

Health systems are challenged by the fact that healthcare costing methodologies are inaccurate. Most systems use cost-to-charge ratio or RBU methods, which are less precise. Activity-based costing, on the other hand, is patient-centric and assigns true costs to patients based on actual utilization.

The game concluded with a game of Plinko, demonstrating barriers to patient engagement and the importance of social determinants on how well patients live. Systems need to remove barriers, such as access to care and financial constraints, to engage patients.


Loading next article...