Medical auditing processes used at hospitals across the country are manual and error-prone. In an experiment with two large hospitals, we used AI to identify documentation errors that were missed in audits. There were more coding and documentation errors than anyone expected.
The amount of healthcare data being collected every day is increasing, perhaps faster than in any other industry. Today, this clinical data is collected in electronic health records (EHR), coded, and then submitted for reimbursement to payers and various regulatory authorities to assist with health system planning, performance monitoring, quality improvement, and research funding allocation. However, a serious problem exists: dirty data. Inaccurate or missing data has serious consequences on patient care, reimbursements, and the reliability of data for secondary use cases for health systems.
Data quality issues are particularly important in healthcare, where inaccurate data introduces risk for both patients and health systems. Further, medical coding is directly responsible for the revenue and compensation that health providers receive. That makes coding fundamental to the long-term capabilities of hospitals to deliver top-quality patient care. Medicare found that 2.7% of the records have coding issues that impact on the accuracy of insurance bills. Many of these data quality errors (76%) occur in high-stakes data entry points like procedure coding, claims filing, and medical records. Errors can result in claim denials, which lengthens the time period for reimbursement. When hospitals cannot account for regular cash inflow, managing operations becomes increasingly difficult.
However, the financial implications of medical coding errors are not the only longstanding challenges. In addition to providing evidence of service for reimbursement, hospital data is also used to make improvements in healthcare, health system performance, and population health. To do so, they require access to high-quality coded data that complies with national and organizational standards in order to ensure the validity of the information they share and the efforts they take. Submitting miscoded data inhibits the strengthening of the larger healthcare system and risks the continuation of preventable medical practice issues.
A study from the Healthcare Information and Management Systems Society (HIMSS) also found that poor quality data has an impact on medical errors that can cause patient death. An AHIMA paper further supports this difficult reality stating that “it can lead to medical errors, which can kill or cause long-term damage to the health of patients.” This paper also cites an Institute of Medicine report that estimates 44,000 and 98,000 lives are lost every year due to medical errors in hospitals alone. While not all errors are attributed to inaccurate data, AHIMA states that “a number of studies have shown a link between poor quality data (in databases) and medical errors and subsequent poor quality care.” This risk is further heightened in increasingly technology-reliant health systems, where physicians may have to act quickly with the available information. Challenges like this contribute to the hesitancy to adopt new technology solutions that would improve the speed and quality of care, administration, and decision making.
High-quality, accurate data is more important than ever for health systems and patients. That is, of course, why measures are in place to verify the accuracy of coding and data quality. However, the implementation of these checkpoints at major hospitals has left gaps where mistakes can be missed.
Auditing is a fundamental stage of day-to-day operations in medical practice to comb through the large quantity of clinical documentation and claims information. That’s why most health providers employ teams of specialists to manually identify and rectify data quality issues. Auditors are responsible for ensuring the accuracy of the information, like medical codes, to ensure compliance with ICD-10 codesets, mitigate risk, and streamline revenue cycle management.
When an auditor performs a medical coding audit, they primarily seek to identify the following:
This manual process is time-consuming and error-prone. With so many priorities, most medical coding audits rely on samples to test accuracy rather than complete documents. Additionally, due to the time constraints of a small and manual team, most audits are performed retrospectively (after billing), which misses the opportunity to identify documentation gaps and miscoding before claims are rejected. Between the sampled approach and retrospective auditing style, it is not challenging to see where medical coding errors may be missed.
At Semantic Health, we’ve been working with top hospitals to help with their coding and auditing practices and have identified a pattern of missed medical coding and documentation errors. We recently ran a brief investigation with two large hospitals to test the quality of their coded data after audits had been done. We deployed our AI-powered auditing tools to identify data quality errors and better understand the effectiveness of current processes. The results emphasized that hospitals have tremendous room for improvement in their auditing practices:
Across these challenges, there were a number of common threads. Here are the most frequent issues that our algorithms discovered that auditors missed.
Most AI suggestions were clinical in nature, allowing the health providers to identify high-impact and relevant documentation and coding gaps - many of which were used to enhance clinical documentation improvement (CDI) initiatives.
AI-powered auditing algorithms, if trained and deployed correctly, can streamline and augment existing medical auditing processes in hospitals. This will help health providers to improve their medical auditing processes and catch common errors that are often missed in manual audits. These algorithms are especially helpful because they can be trained or tuned on existing and retrospective hospital data, to considerably reduce deployment time and improve efficiency.
An AI platform audits coded charts to evaluate if assigned codes are supported by clinical documentation. It then identifies missed or underspecified codes and points the auditor to the exact location in the documentation that requires further review. These insights help health providers to develop data-driven, high-quality clinical documentation programs.
By implementing an AI medical auditing platform to address manual auditing failures, coding and documentation issues can be resolved proactively, saving time and money. If you are interested in learning more about how AI in medical auditing can improve efficiency and accuracy, and enhance CDI initiatives, feel free to contact our team of experts at firstname.lastname@example.org.
Semantic Health helps hospitals and health systems unlock the true value of their unstructured clinical data. Our intelligent medical coding and auditing platform uses artificial intelligence and deep learning to streamline medical coding & auditing concurrent with patient admission, improve documentation quality, optimize reimbursements, and enable real-time access to coded data for secondary analysis.
Want to get actionable healthcare data, medical coding, and auditing content straight to your inbox once a month? Sign up below to be the first to know about new posts, webinars, content, and events that will help you improve data quality and boost HIM team efficiency.