The patient demographics are an integral part of the medical transcription service. It forms the core of the data for any medical institution. It involves a diverse set of public and private data collection systems. During patient registration the personal data of the patient is collected by the medical institution.  These patient details are usually transmitted by the provider as a fax or as electronic data to the concerned Medical Transcription Company. Transcription Companies are preparing medical documents per the data received which has to be very accurate. A mistake in patient demographics could lead to multiple issues.

Apart from the patient name, there may be other data such as DOB, SSN#  etc. which is considered personal data of the patient.  

Keeping protected patient data secure

Secure token-based transmission will not pass  sensitive data to the Medical Transcriber. This is a secure token-based method which can only be compiled and read by the provider’s computer and the concerned Medical Transcription Company’s computers. This is a feature that is provided by the Transcription Company to its clients. While dictating, the provider should tag the secure token ID. When the Transcription Company receives a dictation, the transcriptionist will do a lookup on the token for the patient info and DOS. The transcribers will only have to enter the unique ID into their transcription platform. After the completion of the transcription, when the transcriber uploads the job to the client, the patient data corresponding to the unique ID is automatically entered by the servers into the document, thus completing the report with all the patient data included in it. This method will give a major boost for the Transcription Company to gain confidence from its clients with regards to the patient’s privacy.

In a medical practice, timely documentation of patient encounter is a very critical factor. Equally important is the accuracy of the data in the patient encounter notes. Apart from the medical terminologies, it is extremely important that the patient demographic information as well as the encounter specific details is entered accurately. It has medical, financial as well as legal implications. Inaccurate information can result in denial of insurance claims or even land the provider in a legal suit. Hence it is very important that due diligence is maintained to ensure that the patient demographic as well as encounter specific information is entered correctly in patient reports.

The Challenge

One of the most important encounter specific information is the Date of Service (DOS). A wrong DOS on the patient’s medical transcription report can land the provider in problems and also result in denial of insurance claims, resulting in revenue loss.

Every medical practice has its own system of managing the medical records. This includes the handling of dictations on patient encounters. The providers use different modes to dictate on patient encounters. It could be a phone-in dictation, digital records or even the latest technology of smart phones (including iPhones). Some practices and providers have the habit of dictating the patient notes at the end of the day of the encounter and uploading on the same day. However some others dictate on the same day of the encounter but they upload the dictations for transcription in bulk later. Usually it could be twice in a week or even once in a week. At times, even within the same practice, different providers could have different ways. However, in most of the cases, there will be regularity in the way dictations are handled.

It was found that the mistakes were because of inaccurate data sent to Transcription Company from the practice. At the time of sending the dictations (especially using digital recorders), the person in the medical records department, who upload the dictations have to select the DOS for the set of dictations. At times, by mistake they happen to select the wrong DOS, resulting in wrong encounter data being typed in the report.

When this is brought to the notice of the provider by the insurance company, the providers blame the Transcription Company for the errors. This leads to a loss of trust in the quality of service provided by the Transcription Company.

AI Powered Medical Transcription – The Solution

AI can play a major role in overcoming this problem. Most of the practices and/or providers have a definite pattern of sending dictations. There were slight divergences, but on the whole a pattern could be mapped out.

Based on the result of the pattern mapping, the AI could arrive at a matrix which gave us the most probable pattern of dictation for a provider. It could be same day dictation, next day dictation, twice in a week or once in week etc. This pattern can be incorporated into transcription system and the value included in the provider/practice profile.

Whenever Transcription Company receives the dictations, the transcribers would enter the DOS based on the information provided from the practice side. AI system would compare the date of dictation and DOS and match it with the value against the provider/practice profile. If there was a divergence which was more than the accepted one, the AI system would throw an alert to the next level (editor or quality analyst) to recheck the DOS. The Editor or QA could reconfirm the DOS and in the case of a doubt flag it to the attention of the practice or provider to take due care.

This artificial intelligence based approach will result in substantial reduction in the usage of wrong encounter specific data, especially the DOS. This also resulted in major confidence boost among the providers with regard to transcription service.