Submit your article of up to 500 words in length, subject to editing for clarity and brevity (please note: I run only original articles that have not appeared on any Web site or in any publication and I can’t use anything that looks like a commercial pitch). I’ll use a phony name for you unless you tell me otherwise. Thanks for sharing!
Note: the views and opinions expressed are those of the authors personally and are not necessarily representative of their current or former employers.
Paying Attention to How NLP Can Impact Healthcare
By Chris Tackaberry, MB, ChB
Unstructured clinical narrative is increasingly being seen as the primary source of sharable, reusable, and continually accessible knowledge, essential in helping providers make informed decisions, reduce costs, and ultimately improve patient care. While form-driven EHRs readily leverage and share captured structured data, the richest patient information remains locked inside EHR databases as unstructured notes.
Natural Language Processing (NLP) technology is becoming increasingly recognized in healthcare as a powerful tool to unlock this vital clinical data and turn it into analyzable, actionable information. While many have heard of NLP, there is significant confusion about what it actually means for healthcare.
In short, NLP means recovering computable data from free text. Even though most of the world’s knowledge is documented in some form of written narrative, we increasingly rely heavily on computers to analyze the world around us, and computers work better with well-defined, structured data rather than unstructured text.
Google has clearly proven that simple text search allows us access to vast amounts of information, but it still requires humans to determine meaning in the results. NLP is the science and art of teaching computers to understand the meaning in written text in order to extract data from narrative for reporting, analysis, etc.
NLP, typically embedded within other solutions, can help deliver significant benefit to providers and their patients by:
- Improved reporting and monitoring. Many administrative tasks in healthcare depend on structured data, including the submission of billing codes that describe diagnoses and procedures to insurance companies. The identification of billable concepts in clinical narrative is probably the most common application of clinical NLP because it is the most direct path to delivering financial benefits.
- Improving utilization of clinician time, resulting in more efficient care delivery. Doctors and nurses are accustomed to carefully documenting the condition and care of each patient in clinical notes. Without computable data, however, hospital operations, physician reimbursement, and patient care are all compromised. By pulling data directly from notes with NLP, even in real time at the point of care, we can save clinician time and frustration while identifying more data and detail to support clinical decision making, efficient care delivery, better public monitoring, and more.
- Improved physician understanding of patients. NLP provides the level of clinical detail necessary to provide quicker access and review of patient histories. Revealing key information in existing notes that would be invaluable for more timely, better-informed clinical decisions.
- Better research and monitoring. Existing studies have looked for correlations between patient genes or proteins and characteristics identified in the patient’s medical record. Conducting similar studies with the greater volumes of so-called phenotypic data, which can be pulled from patient records using NLP, will reveal far more about what makes our species tick – or sick.
- More efficient clinical workflow. There is an intrinsic inefficiency in EHRs because so much of the information must be documented repeatedly. As a result, there has been significant physician pushback against EHRs, despite their acknowledged advantages.
- Embedded NLP tools can facilitate EHR redesign for more efficient and intuitive documentation of patient information in a manner already natural to the traditional physician workflow.
Done well, there are countless ways NLP can be leveraged in healthcare to deliver benefit by improving efficiency, driving outcome-based performance, promoting access, facilitating research, and supporting population-based healthcare delivery models.
The application of NLP technology to healthcare will transform what we know about disease, wellness, and healthcare performance, enabling major improvement in efficiency and outcome. At the heart of this data-driven transformation is clinical narrative, a powerful and valuable asset. We need to recognize that.
Chris Tackaberry, MB, ChB is CEO of Clinithink of London, England.
Defining a Complete Patient Engagement Solution
By Jordan Dolin
A few years ago it was somewhat rare for a technology vendor to pitch the benefits of patient engagement. Today it seems that everyone is claiming to be a “leader in patient engagement technology.” This has led to a good deal of confusion in the marketplace.
Patient engagement can deliver significant financial and clinical results, but to actually achieve these benefits, organizations need to select a "complete" solution. A complete solution is one that addresses the needs of all constituents. It engages patients on their terms and also contains the content, technology, and regulatory considerations sought by providers to support care in every setting across the continuum.
Simply stated, a solution that satisfies these eight critical elements has the ability to improve clinical and financial outcomes.
- Understands how to synthesize and deliver actionable information to patients. An effective solution must impart information to a patient in a manner that will actually change behaviors and improve outcomes. Addressing a spectrum of learning styles, literacy levels, and cultural relevance requires a tremendous amount of expertise across multiple communication methodologies.
- Facilitates engagement along settings across the continuum of care. A complete solution must support the needs of the patient and the provider in care settings across the continuum as well as the transitions between them. This includes addressing clinical, operational, and regulatory needs of providers in addition to supporting new models of care such as ACOs and PCMH.
- Engages patients at their convenience. Historically, healthcare technology solutions have always targeted the convenience for the provider, not the patient. Patients must have the ability to receive information when they want, where they want, and on the devices they already own.
- Seamlessly integrates into IT systems and workflow. Organizations are no longer willing to accept disruptions to their infrastructure or existing processes. To be successful, solutions must be complementary and additive, not disruptive or distracting.
- Results measured down to the individual patient. The single unifying goal that now pervades healthcare is accountability. A solution must contain tools that allow providers to measure their impact from multiple perspectives. The ability to confirm that a patient received and reviewed information prescribed by their clinician is a fundamental measure needed to quantify impact.
- Measures and delivers an economic return. Healthcare organizations are accountable for outcomes and their partners should be as well. Clients should expect hard dollar ROI studies and vendors should impartially fund and conduct them.
- Backed by an organization with the requisite knowledge and experience. Investing in an engagement solution to support key business objectives is a critical decision. The vendor selected should have the appropriate experience and staff to support the success of their clients and their clients’ employees and patients.
- Effectively supports the near-term and long-term objectives of the organization. The partner selected must understand the challenges of health systems and have a track record of delivering solutions that effectively address them. In addition, it should be clear that investments are being made in new solutions and innovations that will continue to address the needs of an ever-changing market.
Jordan Dolin is co-founder and vice chairman of Emmi Solutions of Chicago, IL. This article contains an abbreviated list due to space limitations; the complete list is available by download.
Physician Compensation: The Accountable Care Challenge
By John C. Roy
As healthcare systems and physician groups across the country grapple with definitions and implications of “accountable health care” and “value-driven contracting,” physician compensation based on a fee-for-service model is irrational. Pioneering institutions have already incorporated quality and outcomes into their compensation plans. Similarly, payment for health care services is shifting into fee-for-value models.
As these models evolve, compensation plans must reward physicians for meaningful quality improvement and patient outcomes. Key questions emerge. How can clinical and other data help providers enhance value in the most strategic ways? What measurement strategies, and which data, can be used to reward provider teams that contribute the highest value?
In a fee-for-value world, physicians and hospitals will have to focus on quality, outcomes, and cost (or efficiency) requiring a true culture of quality improvement. Physician engagement is critical in shaping that culture. Physicians will have to assess and agree upon outcome measures and practice standards and change practice based upon valid, practice-specific data.
Today, many health systems struggle with the absence of such data. Essential data supporting such a transformation is often stored in disparate clinical and financial databases, including multiple electronic medical record systems and homegrown software solutions.
One universally challenging example is accurately attributing patients to individual physicians. Accurate attribution is central to reporting outcomes, but all too often proves extremely difficult. If physicians don’t trust that the data accurately reflect their practice, they cannot invest adequate time and energy in improving quality of care.
On the other hand, when physicians trust data that truly does reflect their practice, the data spur meaningful conversations around quality and outcomes. They see improvements in real time. The ability to correctly assimilate, align, and attribute patient data to individual physicians is a fundamental issue today and a cornerstone of reimbursement and compensation tomorrow.
As payment for health care shifts from “caring for sick” to “maintaining health,” providers will need extremely effective, efficient care management strategies for chronic disease patients. They will rely on patient data that is strategically aggregated to identify interventions around priority patient populations. They will direct sophisticated, well-coordinated management plans to help insure appropriate patient management, appropriate testing, control complications, and improve direct attention to that patient. They will have the ability to report improvements in quality, demonstrating the value of their work over time. All of these efforts deliver significant value that needs to be monitored and rewarded when achieved.
In a fee-for-value world, the provider groups who use population-level data to create and implement successful strategies for effectively managing their chronic disease patients will command higher compensation, regardless of their RVUs. Successful systems and groups will design physician compensation models around elements that matter most in a new, risk-based health care environment. To do this, patient data needs to be more physician-centric, with improving population health as the primary goal.
John Roy is vice president of Forward Health Group of Madison, WI.
Six Facts You Should Know About Stage 2 Meaningful Use and Data Interoperability
By Ali Rana, MBA, MCITP, CISSP
In the world of care delivery, having access to the right information at the right time can be a matter of life or death. Anyone who has been a patient or cared for one understands that the transfer of medical information – whether current or historical – among providers is not readily happening today.
The Stage 2 Meaningful Use requirements, which begin as early as fiscal year 2014, call on eligible providers and hospitals to increase the interoperability of clinical data and adopt standardized data formats to ensure disparate EHR systems are capable of information sharing.
The following are six high-level areas of the Stage 2 rules to consider during your preparations. These areas underscore how clinical data interoperability will change and impact IT infrastructure:
- Interoperability of clinical data is no longer optional. Hospitals are required to connect with disparate EHR systems and send clinical information electronically for at least 10 percent of its discharges.
- Vendor software certified for 2014 clinical data interoperability criteria will produce and consume a consolidated CDA (C-CDA) document (one specification). The C-CDA document must contain medications, allergies, and problem list elements as well as many other clinical data elements. The majority of the clinical data elements in the C-CDA have single, well-defined coding system requirements. For example, the SNOMED CT July 2012 release for a problem list. Thus, all vendors will speak the same language.
- Transmission specifications to other systems for Stage 2 include only “e-mail” (SMTP) and cross-domain sharing format (XDS). These do not require costly and complex HL7 interfaces and instead just configuration to make connections for data flow.
- Vendor software certified for 2014 clinical data reconciliation criteria will be able to import and reconcile home medications, allergies, and problem list elements as discrete, codified data. The ability to reconcile discrete, codified data in conjunction with the C-CDA and transmission standards nearly eliminates vendor and technical obstacles to clinical data sharing. The coding standards also eliminate some of the complexities. Vendors will likely have to map the data into their systems to support drug-to-drug and drug-to-allergy checking.
- Hospitals must have ongoing submission of reportable labs, syndromic surveillance, and immunization information unless there is no entity present that can accept and exchange this data. This bi-directional information sharing is largely at the local level, meaning the abilities on hand to perform this function in a production state will vary. The requirement of these three submission measures is a significant change from Stage1, which only required one data sharing test and failure of that was an acceptable option.
- Patients must have electronic access to their records within 36 hours of discharge. Eligible entities must provide a patient portal that enables the patient to view, download, and transmit information. This Stage 2 criteria now mandate providers to encourage patients to make behavioral changes accessing their own data. The information that feeds these patient portals must be available within 36 hours of discharge. Therefore, key workflow modifications ensuring appropriate timing are a top priority.
Ali Rana, MBA, MCITP, CISSP is manager of implementation and integration services and client services for T-System, Inc. of Dallas, TX.