Authors:<\/strong><\/p>\n Sebastian Wittor, Project Manager <\/strong>at BAYOOMED<\/p>\n Julia Schliesch, Marketing <\/strong>Generalist at BAYOOMED<\/p>\n Patient transportation is an important part of our healthcare system. It enables insured individuals to access necessary medical services even when they are unable to travel to the treatment location on their own for health reasons\u2014whether for dialysis, outpatient surgery, radiation therapy, or chemotherapy, or via emergency medical transport in the event of an emergency. Every year, millions of trips are organized and billed for insured individuals: in 2023, there were over 55 million transportation services (1)<\/a>, which amounts to roughly 150,000 trips per day. Statutory health insurance spent approximately 9.6 billion euros (2)<\/a> on this in 2024\u2014nearly twice as much as just five years earlier.<\/p>\n However, as the number of trips increases, so does the challenge for health insurance companies: How can they identify incorrect billing, unnecessary costs, or suspicious patterns without having to manually review every single transaction?<\/p>\n This is precisely where the use of artificial intelligence opens up new possibilities. Modern data analytics help identify anomalies early on, manage audit processes more effectively, and reduce costs in the long term.<\/p>\n<\/div><\/div><\/div><\/div><\/div> The assessment of travel expenses is significantly more complex than it appears at first glance. Unlike many other benefits, travel expenses cannot be evaluated in isolation; rather, they must always be considered in the context of the underlying medical care and the patient\u2019s circumstances.<\/p>\n In addition, various factors make it difficult to assess individual cases:<\/p>\n<\/div> Different cost-bearing entities, exceptions, and approval requirements make it difficult to correctly classify and assess cases.<\/p>\n<\/div><\/li> A wide range of stakeholders: insured individuals, doctors, hospitals, and transportation service providers<\/p>\n<\/div><\/li> High number of trips and billing transactions<\/p>\n<\/div><\/li> Delays in data availability make it difficult to review and track information in a timely manner.<\/p>\n<\/div><\/li><\/ul> In practice, this means that even minor anomalies are often difficult to detect. At the same time, seemingly minor deviations can quickly add up to significant costs over thousands of trips.<\/p>\n<\/div><\/div><\/div><\/div><\/div> An anomaly does not automatically mean that there was intentional misconduct. Unnecessary expenses often arise simply from incorrect allocations or suboptimal procurement processes.<\/p>\n Typical examples include:<\/p>\n Incorrect Cost Centers<\/strong><\/p>\n Transportation costs are billed to the health insurance provider, even though another payer would actually be responsible\u2014for example, the pension insurance provider for certain rehabilitation measures or a hospital for transfers.<\/p>\n Excessive or Unnecessary Services<\/strong><\/p>\n The mode of transportation billed does not always correspond to the actual need. For example, an ambulance ride may be billed even though a taxi ride would have been sufficient.<\/p>\n Opportunities for Billing Optimization<\/strong><\/p>\n In addition, there are cases in which services are billed incorrectly, though this does not necessarily indicate any intent. These include, among others:<\/p>\n<\/div> Discrepancies in recorded trips that cannot be clearly explained in retrospect<\/p>\n<\/div><\/li> Inconsistencies Caused by Trips Recorded Multiple Times<\/p>\n<\/div><\/li> Discrepancies in mileage figures compared to the expected route<\/p>\n<\/div><\/li> Discrepancies between the billed mode of transportation and the mode actually used<\/p>\n<\/div><\/li> Individual cases involving signatures that cannot be clearly attributed<\/p>\n<\/div><\/li><\/ul> Precisely because individual discrepancies often involve only small amounts, they frequently go unnoticed when viewed in isolation. Taken together, however, they represent a significant opportunity to optimize billing processes.<\/p>\n<\/div><\/div><\/div><\/div><\/div> Manually reviewing travel expense reports is time-consuming and resource-intensive. With tens of thousands of transactions per day, it is hardly economically feasible to examine every single case in detail.<\/p>\n In addition, much of the relevant information is scattered across various data sources. Prescriptions, insurance data, diagnoses, medical devices, medications, care levels, and travel data must first be consolidated and evaluated.<\/p>\n It is precisely at this point that traditional testing methods often reach their limits.<\/p>\n<\/div><\/div><\/div><\/div><\/div> Artificial intelligence enables health insurance companies to achieve significant cost savings. It does not replace claims adjusters, but rather helps them filter out, from large volumes of data, precisely those cases that warrant further review.<\/p>\n Document and Free-Text Analysis<\/strong><\/p>\n Important information is contained in the regulation in free-text fields. AI systems can automatically extract this content, interpret its context, and make it available for further analysis.<\/p>\n Big Data Analytics<\/strong><\/p>\n By comparing large amounts of data, statistical outliers can be identified. This reveals notable trends in costs, unusual trip volumes, or discrepancies between different ride-hailing providers.<\/p>\n Pattern Recognition<\/strong><\/p>\n By linking policyholder data, diagnoses, care levels, travel information, and other characteristics, it is possible to identify typical patterns that indicate incorrect or unusual billing.<\/p>\n Smart Prioritization<\/strong><\/p>\n Scoring models evaluate anomalies based on priority level, probability, and economic relevance. This allows case workers to focus specifically on cases that require intervention and take action before these trips take place.<\/p>\n<\/div><\/div><\/div><\/div><\/div> The greatest value often lies not in tracking amounts that have already been paid, but in avoiding future costs from the outset. Simply knowing that a more thorough review is underway is often helpful in itself.<\/p>\n Intelligent analytics can be used, for example, to:<\/p>\n<\/div> Detecting Abnormalities Early<\/p>\n<\/div><\/li> Providing Targeted and Appropriate Support to Service Providers<\/p>\n<\/div><\/li> Optimize Supply Processes<\/p>\n<\/div><\/li> Identify more cost-effective supply chains<\/p>\n<\/div><\/li> Reduce unnecessary transportation costs in the long term<\/p>\n<\/div><\/li><\/ul> Artificial intelligence thus not only helps identify discrepancies, but also contributes to more efficient and tailored management of care.<\/p>\n<\/div><\/div><\/div><\/div><\/div> Experience from ongoing projects demonstrates the potential that artificial intelligence offers health insurance companies in the review of travel expenses. By combining big data analytics, AI-powered pattern recognition, and intelligent review processes, annual savings in the high seven-figure range have already been achieved.<\/p>\n Instead of manually reviewing every single trip, AI helps prioritize anomalies in a targeted manner and allocate available resources where the greatest economic benefit is expected.<\/p>\n<\/div><\/div><\/div><\/div><\/div>Why transportation costs are particularly challenging for health insurance companies<\/h2><\/h2><\/div>
Where unnecessary costs arise<\/h2><\/h2><\/div>
Why traditional testing methods are reaching their limits<\/h2><\/h2><\/div>
How artificial intelligence helps health insurers save on transportation costs<\/h2><\/h2><\/div>
More than just anomaly detection: Prevention and cost control<\/h2><\/h2><\/div>
Potential savings on travel expenses in the millions<\/h2><\/h2><\/div>
Conclusion<\/h2><\/h2><\/div>