AI & assessment of medical necessity

The assessment of medical necessity – what is the purpose of this? The question of medical necessity is a key component in the assessment of claims and the calculation of benefits in private and statutory health insurance. 

This involves professional assessment by experts in the billing department of a health insurance company as to whether the form and extent of medical treatment has been calculated appropriately and in proportion to the patient's diagnosed symptoms. The purpose of this medical necessity assessment is to ensure that the treatment costs incurred by the community of insured persons are always kept within reasonable limits and to identify incorrect treatment and over-treatment in good time.


What is the current status of assessing medical necessity?

This important task is performed by experts in the claims departments of both statutory and private health insurance companies. In most cases, the determination of whether or not there is a medical necessity for treatment is assessed manually by experts who have many years of experience and extensive medical knowledge. The difficulty here is that these experts must decide solely on the basis comparing the specified diagnosis with the list of doctor’s fee codes, which represent the description of the services provided. Due to different levels of knowledge and the fact that these are always individual decisions, different assessments and different decisions based on the same facts may occur when assessing medical necessity. Furthermore, the billing process as a whole requires a high proportion of professionally trained personnel and is very time consuming.

However, this problem is not as serious in the case of statutory health insurance companies, because these companies have closed claims catalogs which strictly regulate the treatments and the associated billing processes between doctors and statutory health insurance companies. Nevertheless, statutory health insurance companies can also benefit from a greater use of AI in their billing systems.

In private health insurance, the situation is completely different. Because private health insurers use open claim catalogs with individual rates, there is no agreement on fixed treatment guidelines, billing procedures or fee rates. However this means that doctors and hospitals have greater freedom and flexibility in terms of both treatment and billing. For example, private health insurers have to re-examine the conditions for reimbursement for each claim submitted by the customer on the basis of the existing circumstances and must also assess the medical necessity of the treatment.


AI can support the review process

Health insurers can integrate AI support into their billing systems in order to increase both the standardization of individual decisions made by experts when manually assessing the various types of claims and to increase the speed and quality of the processing of claims applications.

In addition to the already established fee-based assessments of the scanned billing data, modern, powerful and fully-automated systems are able to compare the submitted and scanned customer receipts with a health insurance company's existing fee system and can calculate claims on a case-by-case basis. In addition to a significant increase in productivity, the fixed algorithms allow for a higher degree of uniformity in decision-making.

Medical necessity assessments can also be integrated into the automatic billing process. AI-supported OCR software allows the readout quality to be extended to invoice tables, unstructured data and free texts. On the one hand, within an invoice, the claims descriptions of the invoiced fee codes can be compared with the specified diagnosis according to a fixed set of rules and a decision can be made regarding reimbursement. On the other hand, surgery reports can even be checked against the corresponding surgery invoices to determine the plausibility of the billed services.

Taking one step further, treatments which have been provided can also be compared with the treatment guidelines of the German Association of Statutory Health Insurance Physicians in order to assess whether the treatment was performed in proportion to the given diagnosis.

The abovementioned AI support relieves the burden on claims experts in terms of both quantity and quality, and also standardizes decision-making processes.