When it comes to the healthcare space, the term “big data” may seem needlessly redundant. After all, each and every patient generates a staggering amount of information each second they’re in the clinical setting.
Even a quick trip to the ED for a broken toe or a few stitches will involve a stream of data from a pulse oximeter, a blood pressure cuff, and a thermometer, plus a few paragraphs of clinical notes, an imaging study, or a consult with a specialist - not to mention a list of used supplies that need to be tracked and replenished, a room that needs to be cleaned, a billing process that needs to be started, prescriptions that need to be filled, home care instructions to be delivered, and follow-ups that have to be scheduled.
Providers may be starting to emphasize the notion that the patient should be at the center of care, but they may not fully realize that he or she is also at the focal point of an enormous constellation of data, procedures, conversations, and processes that must all align in order to deliver a positive outcome, a satisfactory experience, and cash in the bank for the healthcare organization.
What are some of the key steps along a patient’s journey through this big data landscape, what can go wrong during each one, and how can providers ensure that they have the tools and strategies in place to address these potential pitfalls?
Patient matching during arrival and intake
As soon as a patient steps into the waiting room, the data avalanche begins. Whether the patient is new to the facility or has been a regular for the past twenty years, the administrative staff will ask many of the same questions to ensure that insurance information and demographic details are up to date.
But front desk staff must be careful from the very first stroke of the keyboard. Patient matching – or patient mismatching – is a critical data integrity and patient safety issue that is drawing attention from many of the top professional and advocacy organizations in the industry.
AHIMA, The Sequoia Project, CHIME, and WEDI have all spoken out in recent months about the critical importance of ensuring that patient records are not unnecessarily duplicated, confused, or incorrectly merged.
“Severe patient-care issues can occur and resources are wasted when systems are inundated with duplicate records,” AHIMA said earlier this year in its online journal.
“Patient safety is a major concern for many organizations, yet it is necessary to increase awareness of the safety, legal, financial, and compliance concerns created by duplicate and overlaid medical records.”
AHIMA recently found that more than half of health information managers routinely face patient matching issues that must be manually addressed, yet only 48 percent have quality assurance programs in place at their facilities to tackle data integrity deficiencies in a standardized way.
The College of Health Information Management Executives (CHIME) has taken on patient matching as one of its primary advocacy issues, and is working on developing an industry-wide patient matching strategy to prevent unintentional harm from missing or incorrect patient records, especially as health information exchange increases across the care continuum.
“As we surveyed members and asked them what was really holding them back from making huge strides with interoperability, safety, quality, and cost cutting, almost universally, the answer was how hard it is to identify the right patient and aggregate their information,” said CHIME President and CEO Russell P. Branzell, FCHIME, CHCIO to HealthITAnalytics.com in February.
“You don’t want to have a one percent chance of not walking out of the hospital because someone looked at the wrong record,” he added. “That’s just not acceptable.”
CHIME, along with several other organizations, has advocated for the development of a national patient identifier that would function like a Social Security number or driver’s license number. This standardized code would help to prevent typos and overly-similar names from causing confusion at the front desk, and would make it easier for health information exchange organizations to transfer records between participants.
But the concept of the national patient identifier has raised alarms among privacy advocates, and the development of such a system has been blocked by Congress since 1999.
Despite a growing industry consensus that a national ID is crucial to the continued development of the health information exchange ecosystem, that situation has prevented regulators like CMS and the ONC from pursuing the idea on a federal level.
“You don’t want to have a one percent chance of not walking out of the hospital because someone looked at the wrong record.”
Groups including WEDI, NATE, MGMA, and the Sullivan Institute for Healthcare are working to develop a “virtual clipboard” that would work around the lack of a national ID system to smooth the patient intake process, but healthcare organizations are still facing a significant big data problem in the meantime.
AHIMA suggests that providers develop data integrity programs focused on the unique challenges of patient matching and record duplication.
“Initial steps that can be taken to lessen the burden of duplicate records include partnering with colleagues in patient access to establish standard policies and procedures, such as patient searching protocols, standard name entry conventions, and questions that registrars can ask the patient in order to determine if the patient has ever been to the facility or practice before,” the organization explained.
Providers should establish a regular schedule for reviewing and updating key data fields, including changes to names, addresses, phone numbers, and insurance carriers.
Staff members should receive training and periodic review of data collection procedures, such as what to do about middle names, hyphenated names, or non-standard address data, and should make a concerted effort to develop health IT systems that flag potential duplicates without automatically merging files that seem similar on the surface before manual review.
Clinical documentation during the treatment process
Once a patient has been properly identified and sent to a room or a bed, the delivery of care can begin. A nurse or physician assistant may be the first provider to greet the patient, and is likely to reconfirm his or her identity before taking basic vital signs, recording the chief complaint, and reviewing critical safety items such as current medications or recent procedures.
Thus begins the clinical documentation process for that encounter – and thus begins a significant opportunity for big data to get the better of a healthcare organization.
The perils of clinical documentation are very familiar to any provider that has navigated the transition to an electronic health record. While “clinical documentation improvement” was a phrase that hit its peak during the implementation of ICD-10, the principles of data integrity, accuracy, and completeness have not diminished in importance after the largely uneventful go-live date for the new code set.
Busy physicians flitting from one patient to the next may be tempted to use shorthand, workarounds, or acronyms to sketch the gist of a patient encounter, and might not sit down to complete their full documentation until hours after that patient has left the building.
Combined with well-known shortfalls in EHR usability, including the frustration of using standardized text fields and scrolling through drop-downs – or being given nothing but a blank text field with no guidance – physicians often produce clinical notes that lack key elements required for accurate coding, a detailed diagnosis, or an optimal level of patient safety.
Furthermore, this incomplete record often presents problems for data scientists who are trying to develop algorithms that can predict future events, aid population health management, or understand gaps in clinical or operational efficiency.
Documentation that is incomplete or lacks specificity may be subject to review by health information managers before they can code for billing purposes, which slows down the revenue cycle, requires a huge investment in time and manpower, and frustrates physicians.
Some organizations, like Southeastern Ohio Regional Medical Center, address this roadblock by conducting concurrent reviews of physician documentation. Using health IT tools that eliminate the need to deliver a sheaf of paper queries to providers after a claim has been denied for insufficient documentation, Southeastern Ohio uses real-time feedback to boost the integrity and quality of physician notes.
“By doing concurrent documentation reviews, we’re getting into those charts and really seeing how we can make improvements to work towards our goals,” said Denise Stephens, Director of Documentation Integrity and Utilization Management. “Because we’re in there concurrently reviewing the chart, we can put a query in and ask if there’s any additional information the clinician would like to include before we submit the claim.”
Regular documentation audits, along with consistent reinforcement of best practices for clinicians, can help to speed up the payment process – and create a data pool that is ready to be used for analytics projects.
Providers may wish to engage in the following big data integrity activities to make certain they are ready to dive into clinical analytics:
- Create a data integrity task force that includes members of the HIM department, as well as clinical staff, IT experts, and executive leaders, to define best practices for documentation, create a wish list for optimizing the EHR interface, and generate commitment to high documentation standards across the organization.
- Understand the requirements and role of the data warehouse for big data analytics, including the need for complete and standardized data, the opportunities – and limitations – of traditional warehousing techniques, and the skills necessary to perform advanced big data analytics.
- Develop an analytics roadmap that outlines the organization’s strategic goals for big data use. The roadmap should include input from across the organization and address specific use cases and include a step-by-step plan for achieving measurable results.
- Invest in standardized, open technologies that leverage accepted standards for generating and exchanging health data across disparate systems. This will help to avoid data siloes and ensure that information can be used and reused for future projects.
Patient education, communication, and care coordination
Patient care always involves a dialogue. Providers must communicate information and instructions to patients and their caregivers, and this information should be included in the electronic record for future reference.
Healthcare organizations are also increasingly responsible for extending that dialogue to other members of the provider community. The advent of value-based care and financial risk sharing has created an imperative for communication across the care continuum, which in turn places an onus on providers to invest in health information exchange technologies that can shuttle data back and forth.
This multi-faceted conversation requires providers to reassess their ability to inform patients and their care teams about their activities, place a premium on patient education, and develop care coordination techniques that satisfy new partnership arrangements.
Patient portals have become the basic technology of choice when it comes to giving patients access to their own information. Adoption of these online tools has been spurred on by the EHR Incentive Programs, which started off with ambitious benchmarks for their integration into the care process.
While Stage 2 of meaningful use quickly backed off its original 5 percent threshold for patient engagement, the momentum of the marketplace has continued to promote the value of engagement, open data access, and bidirectional online conversation.
Healthcare providers are using a variety of technologies to meet their growing internal and external communication needs. In addition to patient-facing portals and open notes systems, they are adopting secure messaging tools to exchange information with business partners, joining health information exchanges that open pathways for sharing and receiving data, and developing patient education tools that use big data to personalize care.
At San Francisco General Hospital, which serves a high proportion of socioeconomically challenged patients, as well as many patients who do not speak English as a first language, a data-driven approach to patient education and home care instructions helps to mitigate the language barrier.
“When a patient leaves the hospital and they get pages of instructions and summaries in English, it sets the stage for significant issues with medication adherence, because they don’t always understand what happened or what they’re supposed to do,” said David Smith, PharmD, while explaining the significant impact of personalized educational technology on the hospital’s patient population.
As part of a study on medication adherence, SFGH identified a group of patients at high risk of being readmitted to the hospital due to the complexity of their conditions. These patients were given a tailored medication calendar, written simply in their native language, which helped them to understand their home care needs.
“When it comes to high risk patients, it’s important that you bring as many healthcare providers to the patient as possible, instead of just assuming that the system is going to take care of them,” said Smith.
Among the small group of patients, the intervention produced a 70 percent decrease in the likelihood of a 30-day readmission, and increased participants’ confidence and understanding of their treatment regimens.
Communication with other providers can also benefit from a big data makeover. Health information exchanges, like Maine’s HealthInfoNet, don’t just provide a simple way to share discharge summaries and clinical notes with local partners, but also deliver aggregated population health analytics, alerts, and trending data unavailable to individual organizations.
The HIE, which is one of the few in the country to secure large-scale participation from local providers, creates a single, unified health record for patients receiving care at the majority of Maine’s providers.
“We also provide alerts,” explained Shaun Alfreds, Chief Operating Officer. “We can alert providers when a patient has an inpatient admission or ED visit or discharge; we alert providers when patients have final lab or radiology results available, and when their HbA1C is above 9.”
“That’s one of the exciting things about being in partnership with these organizations as they try to achieve these targets. They’re bringing us into the conversation about how to create their strategies and asking us how we can help them meet their goals, whether those are related to CMS requirements, or the ACA, or their relationships with private insurers.”
Alfreds suggests that providers who are interested in bolstering their care coordination capabilities through a local health information exchange should start with a self-assessment of their goals and needs, particularly when it comes to interoperability.
“What we’re seeing with providers is a lack of understanding, starting out just with their contract with their EHR vendor,” he said. “They need to make sure the contract spells out interoperability in a way they can clearly articulate. If they don’t understand what that contract says, they need to get an expert in. That’s where we see a lot of challenges.”
Billing insurers and collecting patient financial responsibility
The big data burdens don’t stop once a patient is safe and sound at home. For the billing department, the fun is only beginning.
Revenue cycle management often includes two major activities: receiving reimbursement from insurance plans and collecting the remaining financial responsibility from patients themselves.
For payers, it is becoming increasingly difficult to meet value-based care benchmarks, avoid quality penalties, and scoop up bonuses or shared savings as part of an ACO.
On the patient side, the growing popularity of high deductible and high copay plans have complicated this endeavor. A recent study found that 74 percent of providers are seeing a rise in the amount of money required of patients directly, requiring healthcare organizations to invest in new strategies to recoup the costs of care.
“Missing or late patient payments directly impact the bottom line of a provider’s business, potentially threatening the future of the organization,” said the authors of the InstaMed study. “To maintain cash flow, providers need to look at other industries that meet a high level of payment assurance through a more consumer-friendly payment experience.”
“With the focus shifted to the consumer, both payers and providers will be challenged to overhaul their payment processes or face lost revenue and poor customer retention.”
Electronic payment options, conducted through the patient portal or elsewhere on a provider’s website, are becoming increasingly popular as a way to provide a convenient and familiar service to patients.
“Both payers and providers will be challenged to overhaul their payment processes or face lost revenue and poor customer retention.”
Providers are also exploring the use of financial analytics and data-driven revenue cycle management (RCM) tools to enable them to collect patient responsibilities at the time of service.
“Now with high deductibles, because many more people are forced into a situation where they owe far more money than they used to for the same services, their share has grown so significantly that a lot of providers have decided we need to try to collect this money at or near the time of service,” said Mark Owen, Director of the Division of Emergency Medicine for Payor Logic.
“But we need to know what that amount is in order to do that. We have further complicated the whole collection process by saying not only do you now need more information faster than you used to, you now have to qualify to get the resources that provide that information. Once you get data, it's about the ability to consume that data. Data doesn't just turn into money.”
Once again, the availability of complete and accurate clinical documentation becomes a critical must-have for providers. When organizations can code their case files, submit bills to payers, and receive reimbursement in one clean pass, the revenue cycle can operate without any damaging hiccups – and providers can avoid the bad debt that threatens their operations.
Automated revenue cycle tools are becoming very popular among providers who are looking for efficient and effective strategies for speeding up collections.
In 2014, Black Book Research found that more than half of Chief Financial Officers believed their jobs were on the line if they didn’t find the right RCM tools to rescue their organizations from sluggish performance.
“Generally speaking, in order to strengthen the revenue cycle management, embracing technology within the revenue cycle is key,” Chad Sandefur, Director and Healthcare Analyst at AArete, told RevCycleIntelligence.com.
“Having the platforms to seamlessly facilitate provider-payer interactions are really integral. In many cases, it’s mostly about bad debt avoidance. With that in mind, there are a few specific points. Some of these specific five might not be the most glamourous, but certainly on the element of embracing technology, they are critical.”
Sandefur suggests that providers create an open dialogue with their patients about the financial impacts of their care, and even try to estimate the consumer’s ultimate responsibility to help them make care decisions and prepare for the bills they will receive.
They should also focus on training staff members to embrace RCM technologies, including data analytics tools that can forecast anticipated revenue, identify weaknesses in the revenue cycle, and even reduce claims denials by flagging documentation or coding errors.
“We use analytics for predictability - whether that denial or that payment is based upon a particular code, particular specialties, particular locations, or particular accounts,” he said. “What we see with analytics now is not just a rear view mirror look but ultimately, when you think about machines, we have a specific team of data scientists.”
Regular review of these processes and increased reliance on big data tools will make it easier for providers to close out the last step in the patient encounter while ensuring that both parties involved are prepared for the aftermath of an appointment or treatment.
Good data is the key to making each piece of the healthcare encounter fit together seamlessly, from the initial greeting to the last rubber stamp on a paid claim. A strong emphasis on data integrity, open communication, robust education, and a seamless revenue cycle will forge meaningful relationships across the care continuum while protecting patient safety and providing a welcoming environment for the delivery of high quality care.
This article was originally published on July 25, 2016.