Automated Patient Registration using OCR & NLP
Care is provided by guidance and understanding; but to start care one has to be open for it. Registering with the clinic or a hospital is the first path to cure and care.
Registrations are done almost universally at every hospital. The in-patient and out-patient registrations aren’t just a statistical component in hospital management but also testament to the hospital’s service. They also serve as an identification card combined with retrievable record of the patient’s medical history.
They are generally paper based and even if they are digital; they are merely just for storage. Bigger hospitals database them properly; but most clinics use digital cataloguing as a only a method of filing. These registration forms can sometimes contain obscure information and handwriting, creating simple but massive obstacles.
New Delhi, India
Client
Healthcare Technology Solution Provider
CLIENT
Our client is a healthcare-oriented IT services and consultancy service from India. Embebo provides hospitals, clinics, and healthcare units with state-of-art Customer Relationship and Records Management.
In-patient and out-patient data management and analysis is a crucial task for Custom Relationship Management for which we were tasked to create an in-house OCR and NLP tool to extract information and automate the process of registration.
CHALLENGE
There are certain challenges that need to be resolved first, they are listed below:
-
Compilation of Data - As mentioned earlier most of the data present in registration forms are filled by the patients by hand. These forms might be rife with corrections, changes, and sometimes incomprehensible writing. This is a major problem for compiling data from these forms.
-
Conflicting and changing registration data - When an already existing patient (or in some cases a new patient) enter their data for registration or re-registration has to be changed as per their health progression. This could create a conflict within the existing and new data.
-
Loose archives - Most of these registration archives are not standardized and compartmentalized.
-
Wide data range - A variety of data is present in registration forms; from the patient’s height and weight to a concise description of their medical history.
SOLUTION
The solution that we devised for the varying challenges present by creating a system that contains an integrated, specially developed in-house Optical Character Recognition. Used for detecting printed or written characters on any document, a OCR is a smart tool that could identify and reproduce the text from any document digitally with proper formatting.
-
Coupled with Natural Language Processing; our system of automating the induction, creation, and securing of registrations could be made easy since NLP systems can identify written texts on any style with ease.
-
Any and all discrepancies can be quickly rectified with no issues.
-
NLP is also very use for extracting critical fields from the registration form if needed.
RESULT
-
NLP and OCR units working in tandem can boost and definitely improve the registration rates, which can induct many new patients and help them get their treatment quick.
-
There’s also the aspect of proper consolidation and security for these important forms that contain the core details of a patient.
-
Corrections can also be done at a moment’s notice as well.