Process medical invoices digitally - private health insurance company benefits from automation

The reliable processing and checking of medical invoices and their error-free transfer to the systems of health insurance companies and insurers is an important task for companies in the healthcare sector. Using the example of a private health insurance company, we show how Konfuzio reliably processes even large volumes of data and what advantages companies gain as a result.

What is a doctor's invoice and how is it billed?

A doctor's invoice is a formal document that reflects the services provided by a doctor or medical facility. While the costs of statutory insurance benefits are determined by the uniform scale of assessment (EBM) are billed directly between doctors and health insurance companies, private health insurance costs are billed to the patient. The patient in turn settles the doctor's bill with their private health insurance company and submits it to their health insurance company in paper form or as a scan. 

In addition, the billing of medical services is subject to clear legal requirements. In addition to the uniform valuation scale (EBM) for the services of statutory health insurance funds, the scale of fees for physicians (GOÄ) is the basis of calculation for private health insurance companies and self-paying patients. Invoices submitted and services used must comply with this fee schedule in order to ensure that the invoice is paid legally and the associated costs are covered by the private health insurance company.

Initial situation

Checking and processing the invoices submitted is therefore a key task for private health insurance companies and creates an enormous internal workload. The private health insurance company receives an abundance of invoices - patients submit medical bills for their medical services and want to be reimbursed in accordance with the cost reimbursement principle of private health insurance. The task of the private health insurance company is to check the submitted doctor's invoice for authenticity and conformity with the requirements of the German Medical Fee Schedule (GOÄ). After the manual check by health insurance company employees, the reimbursement of the doctor's invoice is approved and the referral to the patient or doctor is ordered. If the check is unsuccessful, the patient or the issuing doctor is contacted and medical reports are requested in order to clarify open questions and to examine the facts of special cases more closely.

Challenge

This process ties up enormous resources - employees invest a lot of time manually checking invoices and medical reports and correctly assigning the services to the various numbers of the German Medical Fee Schedule (GOÄ). Due to this time-consuming and repetitive task, submitted invoices are often processed slowly and employees often lack the time to devote the necessary attention to more complex cases. As a result, human errors occur time and again when checking medical invoices, delays in reimbursements and multiple billings of the same invoices. However, doctors' invoices and medical reports in particular contain extremely relevant and sensitive patient information and should therefore always be processed correctly and reliably.

Objective - Digital processing of medical invoices

In order to reduce this manual work, solve the challenges identified and make the company's work processes more modern and efficient, those responsible at the private health insurance company decided to implement a digital and intelligent document processing solution. In addition to modernizing processes, the private health insurance company wanted to make the handling of sensitive health data more secure and optimize the verification of submitted invoices and medical reports. By preventing multiple billing of invoices, health insurance companies can avoid incorrect referrals and the avoidable additional work required to clarify these cases.

Requirements of private health insurance:

In order to achieve these goals, the private health insurance company is aiming to implement a solution that reliably automates the checking, processing and management of medical invoices. The process aims to extract essential information from a volume of data at high speed, categorize it into predefined classes and transfer it into a systematic and orderly structure without losing the details of the original material. 

In this way, the solution should enable relevant information to be extracted quickly and precisely, categorized and put into a structured form. Only in this way can all relevant information be transferred to the private health insurance company's systems without errors and the authenticity and conformity of the invoice with the German Medical Fee Schedule (GOÄ) be checked. Automated extraction, checking and processing should speed up processes, increase data quality and minimize the number of errors

Proof of Concept

A key component of this project is the Proof of Concept (POC)which shows how the automatic reading of doctors' invoices and reports can be implemented. The aim of this POC is to extract specific fields from the documents so that they can be used efficiently in further project planning - for example in the creation of billing statistics or the historical tracking of medical records. The fields to be extracted from a doctor's invoice include

1. Patient data
   • Name of the patient
   • Date of birth
   • Insurance number / patient number
2. Doctor and biller data
   • Name and address of the practice
   • Doctor ID or practice number (e.g. LANR or BSNR)
   • Contact details
3. Invoice details
   • Invoice number, date and amount
   • Payment term / payment deadline
   • Itemized amounts (net, gross, VAT)
4. Performance data
   • Treatment date
   • GOÄ/EBM figures and service text
   • Number of services and fees per service
5. Diagnosis and insurance
   • ICD codes
   • Name of the insurance company (statutory or private)
   • Note on co-payments / deductible
6. Bank details
   • IBAN and BIC of the practice
   • Intended use
7. Additional info
   • Notes on reminder fees or refunds
   • Document number for subsequent billing

Implementation - automation through digital invoice processing

Project preparation:

A well thought-out project structure is required for the implementation of the comprehensive Konfuzio solution for the automated reading of medical invoices from private health insurance companies. The first step is to understand the use case of the private health insurance company and the relevant documents and fields of the doctor's invoice as well as possible. Once the medical invoice has been recorded and structured, various training sessions and tests are carried out with artificial intelligence (AI) in order to train it to extract the data in the best possible way. Once the training is complete, the Extraction AI automatically read the defined fields from the doctor's invoice and provide the information in a structured manner.

Following extraction, the Konfuzio system takes over the annotation of the data. In this step, the software marks specific data and information in a document. In AI-supported data processing, these markers are used to show the AI which data is relevant. After annotation, it is able to automatically recognize and extract similar information in other documents. Konfuzio transfers the information obtained via a web-based interface (REST API) into the private health insurance company's systems. This enables seamless integration of the data into downstream processes, such as checking submitted invoices. 

Project process:

This approach not only improves the performance of the AI, but also significantly increases data quality - a key factor for precise and reliable analyses. Clean data minimizes sources of error and enables private health insurance companies to read medical invoices efficiently and fully automatically. Below is a brief overview of the process:.

1. Definition of the project structure
   At the beginning, a clear structure is defined for the project that identifies all relevant fields of the doctor's invoice and organizes them systematically. This structure enables efficient and orderly processing of the documents.
2. Data preparation
   For data preparation, the medical invoices are uploaded to the private health insurance company's system and prepared for training. Careful preparation and annotation ensures high data quality, which is crucial for AI training.
3. Model training
   The model is then trained. Several runs allow the AI to be optimally calibrated to the specific requirements of the doctor's invoice.
4. Tuning the model
   After the initial training, the model is refined. By adjusting the Hyperparameters and optimizations based on previous evaluations, the extraction accuracy will be further increased.
5. Validation
   The trained model is then validated with test data comprising around 30 % of the entire data set. This validation shows how well the AI can read the medical bills in practice.
6. Optimization
   Based on the validation results, final adjustments and optimizations are made to prepare the extraction AI for productive use and maximize the readout accuracy.

Result - Digital processing of a doctor's invoice with Konfuzio

Instead of manually reading and checking invoices, Konfuzio now enables private health insurance companies to digitize and automate these processes. Invoices are automatically read and checked by implementing Konfuzio and the extracted data is then transferred directly to the private health insurance company's systems. In this way, Konfuzio helps to detect multiple invoices and errors, validate GOÄ/EBM figures and identify missing data. The private health insurance company benefits from many advantages as a result of this modernization.

1. Time saving - The automated processing of incoming invoices in various formats (PDF, scan, image) and the digital checking of GOÄ and EBM figures saves valuable time. Invoices can be processed and paid more quickly and employees can devote more resources to clarifying special cases.

2. Better data quality - Konfuzio's automated system reduces human errors that can occur when entering data manually. This results in improved data quality, which, among other things, prevents multiple billing of the same medical invoice.

3. Traceability - Digital systems make it easier to track invoices and payments and to generate billing statistics. Processed invoices also remain in the system and can be retrieved quickly if required. Legal retention periods are also complied with automatically.

4. Improving patient satisfaction - A smooth and transparent billing process increases overall patient satisfaction and creates trust in the services provided by health insurance companies, insurers and doctors.

5. Cost efficiency - Digital processing over longer periods of time can reduce the costs for paper, printing and mailing. This is also good for the environment and reduces the health insurance company's ecological footprint.

Conclusion - process medical invoices digitally

Digital processing of medical invoices with the AI software Konfuzio brings numerous advantages for private and statutory health insurance companies. The automation of the Capture, checking and validation of invoices as well as the seamless Integration of the data into the Insurance systemserrors are reduced and processing times are shortened. Automated invoice processing also improves data quality and prevents multiple invoicing. 

These optimized processes contribute to better traceability, compliance with legal regulations and increased patient satisfaction. In the long term, digitization leads to cost savings and more sustainable administration by reducing the need for paper, printing and mailing. Overall, Konfuzio is a forward-looking solution that supports the healthcare sector with high-quality data and thus ensures greater efficiency and transparency in companies. In the case of the private health insurance company, expectations have been met and internal processes have been significantly optimized. Invoices are now checked and processed much faster and the error rate has fallen significantly.

In addition to private health insurance companies, the solution is therefore also of interest to other players. Hospitals, for example, also have a great need for fast and error-free processing of a large number of invoices. Statutory health insurance companies or insurance companies that offer private supplementary insurance also have to overcome the challenge of checking the services submitted. For both cases, Konfuzio offers a reliable and future-proof solution.