The Myth of Predictive Analytics--Really??

by Karen Wolfe

Mark Walls recently posted an article entitled, “The Myth of Predictive Analytics”[1] where he says he has yet to see cost savings from predictive analytics. Such a statement will surely generate a firestorm of comments. This is one of them—and it specifically addresses predictive analytics-informed medical cost management in Workers’ Comp.

The power of predictive analytics to mitigate medical costs is entirely dependent on the operational design of the delivery system. Predictive analytics is the information vehicle that creates knowledge for claims adjusters and others in the medical management process. System design determines how the information gained from predictive analytics is operationalized. How the information is implemented makes all the difference.

Walls goes on to say, “However, the potential for cost savings doesn’t come from the flag, but what you do in response to it. You need to take action and do something differently than you would have done without the flag.”[2] To a point, that’s true. However, the timing and method of delivery and format of the “flag” is critical. The conditions of information delivery that drive cost savings are timeliness, accuracy and efficiency, ease of use, and structured protocols. All are functions of delivery system design.

Timeliness for early intervention
Information must be delivered in the form of alerts as concurrently as possible. The claims rep should receive the information very close to the time of the risk occurrence. To achieve that, the data must be monitored continually with alerts sent immediately. Factors unknown early in the claim can occur at any time throughout the claim. Timely notification activates early intervention, before further damage is done and before the situation becomes more complex. That saves time, money, and leads to better outcomes.

Accuracy and Efficiency
An alert is useful only if it contains all the information the claims rep needs to make an informed decision, to adjust reserves, and to initiate measures that will prevent further medical loss. Predictive analytics is used to calculate and portray projected costs based on history, differentiated costs, and expected time lines. The alert also displays a medical summary of the claim. All the information is portrayed for the claims rep and requires no data look-up and no data entry.

Accuracy and efficiency are cost savers because they are time savers. Even less-experienced claim reps can take accurate steps when all necessary information is provided.

Easy
The information generated by predictive analytics notifies and informs the claims adjuster at the appropriate time without additional effort on the part of the adjuster. The system automatically portrays all pertinent information without need for searching or data entry. At that point, the adjuster can take appropriate and informed action.

Structured protocols
Medical management in Workers’ Comp is traditionally designed and delivered by individuals in the organization in one-off situations. That means processes are inconsistent. Good system design that draws from predictive analytics infuses structure and measurability into the process. Those situations in claims that should be referred to a nurse case manager are tagged in the system by senior management in advance, so they are referred automatically. The claims adjuster is relieved of the problem of when to refer.

The system is designed to make referrals automatically, thereby making them timely and consistent. Pre-determining what kinds of conditions will be referred and to whom, is how the organization sets up standardized medical management protocols. Such consistent, intelligent process management generates measurable results.

Measure results
Cost savings are objectively and accurately measured in a predictive analytics-supported system. On case closure, actual medical costs for the claim are compared to predicted costs based on documented history. Because of predictive analytics and continuous data monitoring, interventions are executed early, making them more effective. Appropriate referrals are made automatically according to protocol rather than intuition. The medical management team collaborates to improve on projected costs.

Documented process
Medical management alert activity on the claim has been documented by the system throughout the course of the claim. Therefore, costs can be appropriately allocated to the claim, the client, or policy-holder including activity detail, thereby creating transparency and trust among constituents. The organization enjoys increased profitability and strategic competitive advantage.

Walls also states, “In the end, good old-fashioned claims handling skills are still the best way to achieve superior outcomes on claims.”[3] However, when the claims handler is supported by a well-designed, predictive analytics-informed intelligent assistant, claims handling the old way is simply obsolete.

Karen Wolfe is the founder and President of MedMetrics®, LLC, a Workers’ Compensation, predictive analytics-informed medical loss management and technical services company. MedMetrics offers intelligent medical management systems that link analytics to operations, thereby making insights actionable and the results measurable. karenwolfe@medmetrics.org

---

[1] Walls, M. The Myth of Predictive Analytics. Leaders Speak. WorkCompWire. 9-19-2017. http://www.workcompwire.com/2017/09/mark-walls-the-myth-of-predictive-analytics/

[2] Ibid.

[3] Ibid.

T

How to Avoid Insuretech Disruption

By Karen Wolfe

In insurance, those who hold the data, hold potential power. Those who analyze the data and apply the knowledge control their destiny.

“The abundance of data – and the technology used to capture it - is driving profound disruption in the relationship structure of the insurance industry. As the traditional gatherers and guards of massive amounts of data, insurers face threats from new, tech-savvy competitors who can adapt to changes more quickly.

There are very powerful trends coming together to cause serious industry disruption. That can be a big threat, but if insurers start responding now and embracing the change, it could also be a big opportunity.”[i]

What is Insuretech?
Google defines Insurtech as referring to the use of technology innovations designed to squeeze out savings and efficiency from the current insurance industry model. Insurtech is a portmanteau of “insurance” and “technology” that was inspired by the term Fintech (Financial Technology).

The problem
Longstanding reluctance to change is preventing many organizations in the Workers’ Comp industry from embracing new technology, especially technology that streamlines processes and worker performance. However, this approach is no longer sustainable. Simply stated, those who cling only to the old-time culture and ways will be disrupted. Think Amazon or Uber.

“Big Data and analytics are forcing insurers to adjust their processes when it comes to collecting and using data. With the expansion of the Internet of Things, sensor technology, machine learning and artificial intelligence, there is more information available than ever before.”[ii]

Data, the asset
Organizations that continue to ignore the facts will wonder why they are no longer competitive. Others who are open to new approaches using new technology will experience positive results. It is a matter of attitude and willingness to try newer methods.

Nevertheless, insuretech need not be invasive or costly. To make a positive impact on processes and outcomes, an organization must first take the position of data-centeredness, believing data is its valued asset.

Step ahead of disruption
Having accepted the position that data is the organization’s valuable asset, accepting and incorporating new technology requires focusing on only three basic initiatives: data quality, data analysis, and intelligent application of the intelligence gained through analytics.

Data quality
If data is an asset, then its quality must be valued and protected. Using poor or erroneous data never ends well. Information gained from poor quality data will not improve an organization’s processes or outcomes, will lead to poor decisions, and detrimental actions. Therefore, resources must be applied to guaranteeing quality data input. Moreover, considerable resources may be needed to improve historic data.

Over the last twenty-five years organizations have been focused on collecting data, but little attention has been paid to insuring the data is accurate and complete. Now that must change. A data-centered organization will also guarantee its data is pristine.

Analytics
The second initiative needed to avoid insuretech disruption is to analyze the organization’s data. Collect and analyze all data over the previous five years. Methods such as predictive analytics can be applied to gain greater understanding of the organization, how well it operates, and what are the cost drivers both operationally and at the transaction level. This is simply a matter if analyzing historic data and monitoring concurrent data to reveal trends, threats, and possibilities. Know thyself.

Intelligent knowledge application
Having quality data and analyzing it leads to the next critical step of designing intelligent solutions to problems identified during the analysis phase. Apply the knowledge gained to specific areas of need by creating “apps” that solve problems and improve processes in the organization.

Alert the right person when conditions or events pose a risk to the organization or work product identified in the analysis phase. Deliver key intelligence to specific individuals or groups at the exact time they need it for decision support. Facilitate timely communication within the organization. Knowledge assistance provided at the right time to the right persons saves time, creates accuracy, efficiency, and increases profitability.

Stepping into the world of insurtech and avoiding disruption is largely a matter of perspective and attitude. It requires a view that data is an asset and when properly managed, lets the organization define its destiny.

---

[i] Will You Survive the Great InsureTech Disruption? James Dodge, Senior Consultant, Advanced Analytics & Data Solutions, Milliman. Risk and Insurance. 6-27-17

[ii] Ibid.

 

MedMetrics, a WC insuretech company, leverages predictive analytics to provide WC claims reps knowledge assistance by projecting probable ultimate medical losses, automatically integrating claims and medical management to contain predictive medical threats, and offering additional intelligent tools such as provider performance analysis and diagnostic severity scoring. MedMetrics clients are easily and affordably more accurate, efficient, and profitable.

 

How to Monetize Intelligent WC Medical Management

by Karen Wolfe     

”If data is the currency of the new digital economy, then organizations that know how to monetize data will generate the highest returns. They will make better decisions that lower costs, grow customer loyalty, and increase revenues.”[1]

 

Over the past twenty-five years, the Workers’ Comp industry has collected vast amounts of data. Moreover, organizations within the industry have easy access to their most valuable asset: their data. Their challenge now is to monetize the data and profit from it.

 

Experts say medical costs now amount to 60% of claim costs in Workers’ Comp. If true, organizations should be charging ahead to find ways to optimize medical loss management and monetize their data for profit.

 

Data integration

The first step toward monetizing medical data is to integrate data from disparate data silos. All bill review, claims system, pharmacy (PBM) and other relevant data should be integrated at the claim level to gain a full picture of individual claims. Once integrated, predictive analytics methodologies are applied to covert the data to usable information.

 

Past is prologue

What happened in the past is a good indicator of what will occur in the future when similar conditions appear. Organizational culture, protocol, and individual preferences are consistent influencers. Consequently, data gathered from other organizations may offer inaccurate results.

 

Analyze historic data using predictive analytics to discover conditions that are cost drivers or cost accelerators. Uncover trends. What conditions or combinations initiate or perpetuate high cost situations? Where are the gaps in timing in operational flow? What actions encourage positive or negative claim resolution? Finally, the information must be made actionable.

 

Inform stakeholders

Portray the predictive analytics-informed information for claim stakeholders in timely, informative notifications when risk situations occur. An example of this is a diagnosis of a comorbidity such as diabetes appearing in the data long after the date of loss. Predictive analytics has determined the comorbidity of diabetes adds complexity and cost to claims, therefore an alert is generated and key Information is conveyed to appropriate persons.

 

Probable ultimate medical costs

Based on predictive analytics, the probable ultimate medical costs for the claim are portrayed for the claims rep along with other key information regarding the claim in question. The claims rep adjusts medical reserves accordingly and moves on. Time is saved and accuracy is optimized.

 

At the same time, the predictive analytics-informed system automatically notifies the nurse case manager based on the organization’s referral protocol. The claims rep is informed of the referral but is not required to take action.

 

Similar claim information is presented to the nurse case manager for quick review, thereby integrating and coordinating claims and nurse case management initiatives.

 

Monetize medical management

Data is made intelligent and can be monetized through predictive analytics combined with a timely information delivery system. Searching for decision-support information takes time and is inefficient. Manually entering data is time-consuming, annoying, and often inaccurate. On the other hand, intelligent information delivered appropriately is monetized as claim stakeholders make informed decisions quickly, effortlessly, and accurately without need for data gathering and data entry.

 

Projected probable ultimate claim cost with comprehensive supportive information displayed for claims reps does not require data search or data entry. Even less-experienced adjusters are accurate and efficient. Accuracy and efficiency is optimized, productivity is increased, and profitability follows. Moreover, early intervention through timely alerts allows for action before further damage is incurred.

 

Medical loss management is also monetized by the ability to objectively measure claim cost savings. Having projected the ultimate medical costs for a claim, quantifiable cost savings are available at claim closure due to coordinated medical management initiatives. Monetization is realized through client satisfaction, customer loyalty, and client retention. Moreover, the story is proof of value serving the organization’s strategic competitive advantage.

 

Intelligent medical management

Organizations that monetize their data have greater returns, including return on investment. The intelligent medical management system is monetized internally and externally, thereby paying for itself. Such statements are familiar as sales platitudes, but with intelligent medical management, positive results are objectively measured. Savings are greater than the cost.

 

---

[1] Eckerson, W. How to Monetize Data: Strategies for Creating Data-Driven Applications. Eckerson-How-to-monetize-data.pdf Zoomdata. March, 2016.

Karen Wolfe is the founder and President of MedMetrics®, LLC, a Workers’ Compensation, predictive analytics-informed medical loss management and technical services company. MedMetrics offers intelligent medical management systems that link analytics to operations, thereby making insights actionable and the results measureable. karenwolfe@medmetrics.org 

Intelligent WC Medical Management, a Process for Efficiency and Measured Results

by Karen Wolfe

Technology in Workers’ Comp is hardly new, but new ways to infuse technology and predictive analytics into the claims and medical management processes can significantly improve accuracy, efficiency, outcomes, and, importantly, profitability. Well-designed technology that streamlines operational flow, provides key knowledge to the right stakeholders at the right time, promotes efficiency, and generates measureable savings is formidable. The system is intelligent and includes these key components:
1.     Predictive analytics
2.     Data monitoring
3.     Knowledge for decision support

Predictive analytics
Predictive analytics is the foundation for creating an intelligent medical management process. Analysis of historic data to understand the risks and cost drivers is the basis for an intelligent medical management system. For the risks identified, the organization sets its standards and priorities for which stakeholders are automatically alerted to those specific conditions in claims as they occur.

The stakeholders are usually claims reps and nurse case managers but others inside or outside the organization can be alerted, such as upper management or clients, depending upon the situation and the organization’s goals. Upper management establishes specific action procedures for specified conditions or situations, thereby creating consistent procedures that can be measured against outcomes.

Data monitoring
Incoming data must be updated and monitored continuously. Random or interval monitoring leaves gaps in important claim knowledge that is overlooked until the next monitoring session. The damage may have escalated by then. With continuous data monitoring, everything is reviewed continually so nothing is missed. When the data in a claim matches the conditions outlined by the predictive modeling, an alert is sent to the stakeholder so action or intervention is initiated.

Some say the stakeholders will not comply with such a structured program because they resist being directed. To solve that problem, accountability procedures in the form of audit trails in the system act as overseer. At any point, management can view what alerts have been sent, to whom they were sent, for what claim, and for what reason, thereby observing participation and supporting accountability.

Knowledge for decision support
The alerts sent offer collected knowledge about the claim needing attention so the stakeholder is not forced to search for information before deciding upon an action. The reason the alert was triggered, detailed claim history including medical costs paid to date is displayed for alert recipients. Importantly, the projected costs for a claim with similar characteristics are portrayed, making reserving adjustments easy and accurate.

The projected ultimate medical costs for the identified claim is portrayed for the claims rep based on the analytics, thereby providing decision support for adjusting reserves. Data entry into the system is never needed, therefore, accuracy and efficiency is optimized.

At the same time, a nurse case manager is automatically notified of the situation if indicated by the organization’s rules in the system and is informed with the same claim detail. Now the case manager and claims rep are collaborating to mitigate the costs for this claim. They know the projected ultimate medical cost for the claim and the projected duration of the claim so they have a common and concrete target to challenge. Moreover, improvements on the projections offer objective and defensible cost savings analysis.

Predictive analytics combined with properly designed technology to create an intelligent medical management process establishes a distinct advantage. Knowledge made available at the appropriate time for the right people leads to efficiency and accuracy. Early, intelligent intervention drives better results.  Stakeholders coordinate efforts to mitigate the claim, working toward a shared goal. Finally, knowledge provided for decision-support positions for measureable, objective, reportable savings at claim closure.

Karen Wolfe is the founder and President of MedMetrics®, LLC, a Workers’ Compensation, predictive analytics-informed medical management and technical services company. MedMetrics offers online apps that link analytics to operations, thereby making insights actionable and the results measureable. karenwolfe@medmetrics.org

 

Will Watson Replace Workers' Comp Professionals?

by Karen Wolfe

A recent post in LinkedIn noted the Japanese firm, Fukokui Mutual Life Insurance has replaced more than thirty office workers with artificial intelligence.[1] The Artificial Intelligence (AI) replacement in this case is the famed IBM Watson. Watson, or one of its doubles, is in fact impacting nearly all industries in multiple ways. Eliminating workers is the paramount goal. But could Watson replace workers in Workers’ Comp?

AI
AI has been around for decades but now with advanced technology, it has fully caught on and its applications are widely varied. AI is what drives driverless vehicles and operates machinery sans human involvement. Short of that and more practically, AI machine technology is used to enhance worker productivity, accuracy, and efficiency. Importantly, AI should never reach Workers’ Comp if more pragmatic, technology-based strategies are implemented now.

Wake-up call
Replacing Workers’ Comp professionals with Watson or its double is not feasible at this point or hopefully, ever. The possibility of replacement by the likes of Watson should not panic anyone in the Workers’ Comp industry, at least not now. Yet, it is a wake-up call to the industry.

Watson in WC
Imagine injured workers navigating the Workers’ Comp system without claims adjusters and medical case managers. Picture Watson managing claims. It could make payments without difficulty, and even review the bills effectively. Watson could also determine which claims are the most challenging and refer them to medical case management.

Stop there!

Watson as case manager
Envisioning Watson as medical case manager is a real stretch. Human interaction is central to medical case management effectiveness. Likewise, Watson delivering claim management services without dialogue with the claimant would be spotty and unpleasant at best. Accuracy and efficiency under Watson management could be nearly perfect, but claim adjusting relies heavily on human interaction. Injured workers managed by Watson would feel victimized in a heartless system. The only recourse would be to litigate. Watson might have trouble with that.

While replacing professionals with technology like Watson is going too far, it should prompt Workers’ Comp payers to actively engage current technology to improve processes and outcomes—just to keep up. Clearly, the momentum in every industry is more technology in order to gain efficiency and Workers’ Comp cannot afford to lag behind. To stay in the game, technology designed to assist workers with task-relevant knowledge and decision support that makes them more accurate, more efficient, and, yes, smarter is crucial.

Dodging Watson
Watson will replace health insurance industry administrative workers fairly easily. Essentially, bills are paid if they match the benefit plan and the treating doctor is in the PPO. However, the Workers’ Comp industry is very different from general health and much more complex. The question is how can the Workers’ comp industry optimize efficiency and productivity without discarding its professionals and alienating injured workers? The answer is to apply currently available predictive analytics technology to make WC professionals smarter, more accurate, and highly efficient. Of course, that also spells profitability for the organization.

Knowledge assistance

Apply predictive analytics to understand historic data and the cost drivers inherent in it. Monitor the data continuously to identify risk conditions as they occur. Create apps that inform claims reps of conditions and events in claims that need attention in real time so action is early and proactive.

 

Assist claims reps by providing information for decision support such as the probable ultimate medical reserve amount for a claim. Time and effort is saved, while accuracy and efficiency is gained. Rather than laboring with decisions such as adjusting reserves, a timely and accurate projection is presented, thereby optimizing efficiency.

 

Similarly, relevant information should be available for medical case managers so they can avoid searching for claim information and status. Timely alerts and shared information promote collaboration and integration of efforts between claims and case management decision-makers in the organization. Watson is thwarted.

Karen Wolfe is the founder and President of MedMetrics®, LLC, a Workers’ Compensation, predictive analytics-informed medical management and technical services company. MedMetrics offers online apps that link analytics to operations, thereby making insights actionable and the results measureable. karenwolfe@medmetrics.org

 

 

 

 

---

[1] Japanese company replaces office workers with artificial intelligence. LinkedIn. 1-25-17.

https://www.theguardian.com/technology/2017/jan/05/japanese-company-rep-laces-office-workers-artificial-intelligence-ai-fukoku-mutual-life-insurance?CMP=share_btn_link

Analytics-powered WC Medical Management

by Karen Wolfe

“I firmly believe based on extensive qualitative and quantitative research and the many interviews and discussions I’ve conducted with insurance industry leaders—that analytics represent the industry’s best path to success and survival in a rapidly transforming world.”[1] 

This is not a controversial statement. Most people now acknowledge the importance of analytics. The question is what is the best approach to analytics in order to achieve optimum results? It’s not enough to prepare analytic graphs and pin them to the wall or publish them in reports for the C-Suite. The information might be eloquent, but nothing will happen or change until it is pushed into operations where action can be taken. Analytics should solve problems. 

Solve problems 
In Workers’ Compensation, a major problem is our inability to contain continually increasing medical costs. Cost containment initiatives implemented to date have helped, but the job is far from complete. Happily, a new methodology is now available—analytics-powered medical management. 

Analytics defined 
Analytics is basically a fancy term for data analysis and there are many forms. A good place to begin is with descriptive analytics, a preliminary stage of data processing that creates a summary of historical data and integrates data from different sources to yield useful information. It also prepares the data for further analysis. 

Developing knowledge 
The data can be re-organized to make it easy to identify patterns and relationships, not otherwise obvious. The data can be queried and reported for more insight. It can also be re-packaged for better understanding or to initiate an action. To an individual in an organization, analytics can make the information derived from the data easy to locate, access, understand, and act upon. 

Real time intelligence 
When the data is monitored continuously, business units gain the advantage of near real time intelligence. Concurrent knowledge of conditions and events in a claim offers the opportunity for early intervention. Early intervention means the damage can be curtailed, thereby reducing claim costs and complexity. Outcomes are improved. 

Actionable information
Information is most powerful when it is current. What occurred two months ago may not be accurate or even relevant now. Conditions about events may have changed substantially so time is wasted manually updating the information before taking action.

One of the most important benefits of analytics is the ability to push timely information to the people who need it and can act on it most effectively. In the case of Workers’ Compensation, those persons are most likely claims adjusters and medical case managers who are in the trenches with claims. When these business units are alerted to pre-defined conditions in claims, their responses are more authentic and better outcomes result. 

Structured notification 
Specific information in claims pushed to the appropriate person in near real time is powerful. That person might be a claims adjustor, medical case manager, supervisor, or medical director who will act on the information based on procedures developed by the organization. The organization determines what situations in claims should be addressed, those potentially the most costly or disruptive based on historic analysis. Moreover, the organization determines what actions should be taken, and by whom. 

Structured response 
Structured notification and responses in the form of standardized procedures lead to consistency. As with any organizational procedure, allowances are made for professional authority, nevertheless, cost savings gained through consistent processes can be reliably measured. 

Spon
Measureable results 
Measures of medical management savings in Workers’ Compensation have been elusive. Without structured problem identification and response, apples to apples comparisons and analyses are impossible. A major benefit of analytics-powered medical management is the ability to dependably measure the benefits derived through data monitoring, analysis, notification, and response.

Karen Wolfe is the founder and President of MedMetrics®, LLC, a Workers’ Compensation, analytics-powered, medical management company. MedMetrics analyzes and scores medical provider performance and offers online apps that link analytics to operations, thereby making them actionable. karenwolfe@medmetrics.org

---

[1] Applebaum, S. Analytics and Survival in the Data Age. LinkedIn Pulse. news@linkedin.com. February 18, 2016.