Unique Patient Identification
Fingerprint Patient Identification Solution for USA’s PEPFAR Funded Program
Context
SEJEN’s UPID (Unique Patient Identifier) is an open-source, cloud-hosted, and fully managed Unique Patient Identification platform for integration with Electronic Health Record Systems (EHR). This case study pertains to the use of SecuGen biometric solution to integrate fingerprint-based patient identification ability with SEJEN’s solution and develop a custom fingerprint recognition application.
Challenges
The Côte d’Ivoire Ministry of Health has been putting tremendous effort to prevent Acquired Immunodeficiency Syndrome (AIDS) in the country and the spread of Human Immunodeficiency Virus (HIV). One of the main challenges in doing so is the accurate identification of AIDS patients and providing treatment at the healthcare facilities designated for the implementation of the program.
Despite setting up a national-level system for patient identification with an Identification Number, issues of unique identification of patients repeatedly emerged throughout the treatment facilities.
The Unique Identification Number-based system was taken for patient identification and maintaining the healthcare records as it was easy to set up and is still used in many parts of the world. This unique identification number had 4 parts corresponding to the partner code for implementation, the treatment site code, the year of enrollment, and the client code (patient).
While using the patient identification system based on the unique identification numbers, Program Management Officers (PMOs) started encountering mainly three types of patient identification-related issues.
- Current unique patient identification numbers are not regularly used by providers during the treatment process.
- Already registered patients often forget their unique identification number.
- To avoid stigma, some patients constantly change the site of the treatment and introduce themselves as new patients at another site. Registering an already registered patient with a new patient identification code results not only in duplicate records in the system but also increases the chances of errors in providing treatment.
These difficulties in identifying patients led to double or even multiple counts in the active queues of implementing partners. It also resulted in several undesirable consequences on the reliability of the number of staff and the calculation of treatment indicators for people living with HIV / AIDS.
Some initiatives were taken to overcome these issues, which included the use of the National Electronic Medical Records (EMR) database SIGDEP for the registration and treatment of people who tested positive for HIV. However, the expected results in terms of research and identification of patients could not be achieved.
In order to overcome the persistent issues and errors in patient identification due to the unique identification number-based process, stakeholders of the program started to find a solution. They wanted a solution that can not only overcome problems in patient identification but is also not dependent on the memory of the people who tested positive for HIV or documents provided by the center to identify them.
Solutions
SEJEN CI, after several rounds of consultation with its partners, proposes a solution to the identification problem, which is based on the comparison of models derived from patients’ fingerprints. Registering patient fingerprints as identification keys were believed to avoid many issues PMOs were facing.
This approach was expected to offer benefits such as:
- There is no need to remember a unique identification number or any other information in order to be identified at the treatment facilities and fetch the patient records.
- Using fingerprints for the identification of AIDS patients also eliminates the need of carrying any document in order to verify the patient’s identity or populate the patient information from the system.
- Patients may forget their identification number or misplace any document provided for the identification, however, they will always have their fingerprints for identification.
- With fingerprints-based patient identification, maintaining patient records becomes easier. Already enrolled patients who may try to re-register themselves as a new patient will be instantly identified as “already registered” and there would be no cases of multiple registrations of the same patient.
- Chances of populating an incorrect record with a fingerprint-based records search are next to zero, so there will be no instances in which a patient was misidentified due to an incorrect identification number.
Functional System Components
The fingerprint-based patient identification system proposed by the SEJEN CI consists of three distinct modules in the system:
- The first module is the fingerprint capture platform – it handles the patient fingerprint scans.
- The second module performs the template comparison functions to determine the unique identity of the patient.
- The third module performs the functions of integrating the identification system and National Electronic Medical Records (EMR) database SIGDEP 2.
The input platform includes the following functions:
- Integration of a fingerprint reader
- Integration with PC or Smartphone
- Conversion of fingerprints into “templates/models”
- Immediate erasure of fingerprint images after use
- Ability to operate in offline mode
- Transmission of templates to a centralized patient identification system.
The central comparison platform includes the following functions:
- Encrypted storage of fingerprint templates
- 1-N matching
- Unique identification numbers
- History of meetings with the individual
- Support of the data entry platform operating in offline mode
- Various reports on system activities.
The integration module with National Electronic Medical Records (EMR) database SIGDEP 2 allows the system to:
- Merge dual identities in SIGDEP 2
- Registration of the unique identifier in SIGDEP 2
Data Security
Since the system proposed by the SEJEN CI depends on the biometric information (fingerprints) of the patients who tested positive for HIV, it becomes crucial to ensure data security as it involves patient records as well as biometric information (fingerprints) of the patients.
- In order to ensure data security of the patients registered on the system with their fingerprints, the system does not record the fingerprints of the patients but rather relies on a data model derived from the fingerprints. As a result, fingerprints themselves are never stored on any machine or system as images or similar digital format.
- The fingerprint captured is kept in the system for a very short period of time, which allows the system to register the patient, and then the fingerprint image is destroyed as soon as the model matching template is generated. It is this template that is transmitted and used in the operations of comparison made at the central site.
- These templates can in no way be used to re-generate the original fingerprint. Templates derived from fingerprints are themselves encrypted during transmission and at rest (storage). As a result, any illegitimate acquisition of the model database could not be decrypted without the decryption keys.
The design and development of the custom fingerprint reader application were supported by the DCI along with a web service interface, and a COTS application hosted on AWS (Amazon Web Services) to compare fingerprint minutiae data and obtain an anonymized correlation identification code through interoperability with the National Electronic Medical Records (EMR) database SIGDEP 2.
Since DCI was supporting to Unique Patient Identifier (UPID) program and was supposed to provide the custom fingerprint reader application, it chose Bayometric as the third-party vendor to supply the custom fingerprint reader applications for their project.
Bayometric, while working closely with the SEJEN CI, leveraged its highly efficient and incredibly user-friendly fingerprint SDK – Touch N Go to develop the application in order to meet the unique requirements of the fingerprint-based patient identification program. Understanding the importance of data security for sensitive patient information, Bayometric followed all the latest industry-standard data security practices while developing the application.
Process
SEJEN used an Agile development process based on SCRUM to elicit the requirements, design the applications, test it. The development is done in 2-week iterations and all deployments to our cloud based environment are automated using DevSecOps processes using tools such as Jenkins.
Results
Implementation of a unique identification system based on patient fingerprints replaced previous identification number-based systems, which was contributing to several issues faced by the treatment facilities.
Unlike the older identification number-based system, on which already enrolled patients could re-register themselves as a new patient, the new systems completely eliminated this possibility as the patient will have to provide their fingerprints and if a patient is already registered the system, it will let the operator know.
The new fingerprint-based patient identification system also eliminated the possibility of duplicate records by making patient fingerprints key to maintaining records instead of the identification numbers.
Patients neither have to remember any identification/registration number nor do they need to carry any document or card for identification and avail of the treatment. Fingerprints cannot be forgotten or misplaced and also reduces data entry errors.
This project created new capabilities that securely and confidentially leverage biometric fingerprint minutiae data to uniquely identify HIV/AIDS patients accessing treatment facilities throughout Côte d’Ivoire.
Key contributors
