How AI helps hospital executives by safeguarding patients and revenue
Healthcare organizations are facing a perfect storm – lower patient volumes, declining outpatient revenues, labor shortages, wage growth, supply disruptions, and inflation to name just a few. When revenue and margins are both taking a hit – down 37% compared to pre-pandemic levels – the need for immediate action to contain revenue must be prioritized. According to the American Hospital Association, more than half of U.S. hospitals experienced negative margins through much of 2022 – the hardest year financially since the start of the pandemic. And the financial impact has real consequences on patient experience because it restricts investments in new forms of care innovation, limits the services provided, and hinders patient access. Reports of workforce layoffs and cutting back on services are becoming more common. What are healthcare organizations to do when faced with such financial pressures?
A recent survey revealed 93% of healthcare enterprises are concerned about process inefficiencies, while 92% are worried about thinning margins. The goal of AI in the revenue cycle process is to maximize patient revenue and collection speed. But for that to occur, medical records need to be coded properly and have the right clinical documentation that supports diagnosis and treatment.
Improved Medical Coding, Documentation, and Revenues
Challenges such as discharged not final billed (DNFB), productivity, and patient information monitoring continue to add pressure to mid-revenue cycles. When documentation is incomplete or inaccurate it creates medical denials and delays provider reimbursements.
Clinical documentation improvement (CDI), the process of reviewing medical record documentation for completeness and accuracy, has become critical in improving outcomes, case-mix indices, physician incentives, and claims settlement. And the market is responding to the use of AI technology. The global revenue cycle management market is projected to grow from $115.64 billion in 2022 to $246.40 billion by 2029, at a CAGR of 11.4%.
Revenue loss due to coding errors, reimbursement rate drops, and the growing volume of unstructured data are all essential growth catalysts.
Clinical Documentation and the Patient Experience
Because outcomes take precedence over all other metrics, even minor coding errors can directly impact the quality of care provided to patients. Using coding to capture the patient’s story accurately can greatly aid hospitals in profiling and quality reporting. Unfortunately, coding errors caused by insufficient medical documentation are common and have been known to negatively impact a patient's treatment.
For example, if a 2-digit modifier for a leg injury is miscoded, an MRI study may be performed on the incorrect leg. In a more severe scenario, an obstetrician could give a pregnant woman the wrong dosage of pain medicine in the delivery room. Assigning non-billable or non-existent codes and missing Healthcare Common Procedure Codes (HCPC) for separately paid drugs are common errors in the code.
Challenges with Medical Coding
Changing regulatory guidelines have posed an ongoing challenge for healthcare organizations, resulting in a scarcity of skilled coders. In 2015, the mandatory shift from ICD-9 to ICD-10 caused a drop in medical coder productivity because it increased procedure codes from 3,824 in ICD-9 to 71,924 in ICD-10. Furthermore, in ICD-10, code descriptions are longer, and tolerance for unspecified codes is lower.
With the release of ICD-11, which introduces new diagnoses, refines diagnostic criteria, and increases code volumes to around 55,000 codes, the complexity of coding is only expected to increase. Although the implementation timeline for ICD-11 in the U.S. is undetermined, the numerous changes reinforce the need to leverage technology, which will improve medical coding quality, clinical documentation, and patient outcomes.
Avoid Over Coding and Under Coding
Healthcare organizations must ensure their existing IT infrastructure strategy considers adopting computer-assisted coding (CAC). Over-coding and under-coding are two other problems that can harm healthcare organizations. Over-coding Current Procedural Terminology (CPT) and HCPC can result in erroneously high reimbursements from insurance companies, fraudulent charges, and potential regulatory fines. Under-coding, on the other hand, causes care providers to lose money. The actual scope of diagnosis and care is not accurately recorded when under-coding is used. Insurance denials will be reduced if symptoms, diagnoses, treatments, medications, patient history, and health risks are adequately documented.
Why Medical Coders Need Artificial Intelligence
According to Derek Fitteron, the CEO of Medical Cost Advocate, 80% of medical bills in the U.S. contain errors. Furthermore, an audit by Equifax, a credit rating agency, revealed that hospital bills totaling $10,000 or more tend to contain an average error margin of $1,300.
To mitigate the financial consequences of such losses, payers and healthcare providers have turned to next-generation technologies, the most prominent of which are artificial intelligence (AI) and cloud computing. This led to the introduction and widespread adoption of CAC solutions.
The right CAC software can automatically identify and extract data from documents and insert them into the system with the help of a natural language processing (NLP) engine. It can then recommend codes for the treatment in question for coder consideration. CAC also assists in developing customized dictionaries and ontologies to support specific users in recognizing the complexities of patient records and bills. As a result, care systems and facilities can expect higher medical coding accuracy, faster billing, and higher coder productivity. According to MarketsandMarkets' 2020 Natural Language Processing in Healthcare and Life Sciences Market report, the global NLP market was valued at $1.5 billion in 2020 and is expected to reach $3.7 billion by 2025, with a CAGR of 20.5%.
NLP in Medical Coding
NLP is the preferred technology for medical coding because of its ability to analyze and decode annotations from unstructured data. Users can dictate clinical notes or other information, which is then converted to text using speech recognition algorithms. NLP also helps overcome the challenges of billing code algorithms for command line interface (CLI) recognition of text that describes symptoms used to establish a diagnosis when applied to clinical narratives. It's also being used to improve imaging workflows and support value-based reimbursement programs.
Getting Past the Obstacles
Manual data entry and reconciling data from multiple systems are two of the most significant impediments to medical coding productivity. CAC can also help with this issue by automating manual tasks and suggesting potential billable codes. This would allow coders to spend less time entering data and more time validating and producing accurate and compliant reimbursements.
However, choosing the right CAC system for a hospital or health system can be difficult. Despite the countless benefits that a CAC can bring to coding processes, improper integration can create unnecessary obstacles that diminish the positive impacts on operations. For example, a system might require the use of multiple screens in order to view the EHR and CAC solution simultaneously, making the task more complicated and confusing. As such, choosing a cloud-based solution and flexible vendor partner is critical.
As a result, it's critical to use a solution that's been specifically designed to address the issues at hand. AGS health’s computer-assisted coding solution can efficiently predict accurate billable codes, making it easier for coders who then simply need to audit or review the code. Furthermore, CAC capabilities can help reduce inpatient time-to-code by at least 40% and outpatient time-to-code by 50% or more.
NLP Used in the Right Way – The Drexel Clinic Use Case
Drexel clinical researchers used NLP to examine 5,700 patient records for HIV and hepatitis comorbidity and found 1,150 relevant hits. The adoption of NLP is gradually gaining momentum as more use cases like this become public. The following are some of the advantages of using NLP in the domain:
- Transparent Documentation: Clinical analytics can be provided by NLP to identify potential gaps in clinical documentation.
- Improved Efficiency: NLP saves time and effort by automating the coding process, allowing employees to focus on maintaining quality and accuracy.
- Better Accuracy: With improved accuracy levels in coding, hospitals have the scope to reduce compliance risk.
- Enhanced Decision-making: Physicians can make accurate clinical decisions by gaining access to unstructured data from a variety of sources, including office visits, lab results, diagnostic notes, and more.
- Reduced Costs: NLP can help healthcare organizations reduce overall costs by allowing for more accurate diagnosis and better documentation.
- Operational Visibility: Real-time operational visibility can be achieved using smart analytics, dashboards, and reports.
As a result, NLP can help healthcare organizations maintain the accuracy required in clinical documentation and medical coding, paving the way for improved clinical outcomes, enhanced patient care, and, most importantly, increased revenue. Whether it's a lab report or a diagnosis test report, incorporating NLP into the system can significantly improve mid-RCM functions.
CAC must be used correctly to ensure success both today and in the future. Prior to adopting this technology, healthcare organizations should be prepared to address several common considerations, such as:
- Change Management – Organizational leaders must consider how CAC will be operationalized among their teams and across departments where applicable.
- Accuracy – Audit trails are needed to understand CAC accuracy.
- Workflow – Coders’ workflows should be evaluated and adjusted to meet the needs of the facility.
- Data Quality – The quality of data is vital when it comes to a successful CAC implementation.
- Stability and Success – Systems need to be stackable prior to measuring; pre-determine stabilization and success criteria.
Where Are We Heading?
Healthcare organizations are embracing new care paradigms as the industry continues its transition to the value-based reimbursement model. AI provides immediate benefits to stakeholders by assisting them in overcoming current and emerging challenges.
A Solution Provider Sharing Insights
AGS Health provides AI-based and integrated Computer-Assisted Coding, CDI, Coding Compliance, Quality Measures, and Enterprise Analytics to U.S. hospitals and health systems – with a proven record for providing best-in-class AI-based mid-revenue cycle support. AGS supports the mid-revenue cycle market with cutting-edge technology, including a CAC tool that helps hospitals overcome denials resulting from inaccurate or incomplete coding and documentation.
Author
