Every medical code is revenue opportunity and risk

The revenue cycle, beginning at patient registration and ending with (hopefully) full payment, can take anywhere from a few days to several months, variability that can make all the difference to providers’ financial stability.

During and after encounters, clinicians and other staff detail diagnoses, treatments, supplies, medications, equipment, procedures and pre-existing comorbidities in the patient’s medical record and explain their decision-making so as to justify optimal reimbursement.

If a service, stay, procedure or other charge is not sufficiently documented, the organization could face a claim denial and potentially a write-off — or worse, a fraud investigation.

Depending on the payer(s) involved, the care setting, and the stage of billing, several code sets are used, including:

ICD-10-CMDiagnosis codesICD-10-PCSInpatient proceduresCPTGoods and services billed to private plansHCPCSGoods and services billed to selected plans, Medicare, and for all items not covered by CPTs, including ambulance transportation; durable medical equipment, prosthetics, orthotics, and supplies (DMEPOS) when used outside a physician’s officeChargemasterRates before negotiated discounts for all possible charges, including: CPT/HCPCS codes, payer-specific codes, professional services and facility/room chargesProfessional and facility codes

Professional codes: services provided by physicians and other clinicians that are facility employees

Facility codes: based on the overhead for services, and costs of drugs, consumables and equipment

Back-End Billing Relies on Fast, Accurate Coding

Medical coding does not simply review one record at a time. Related charts must all be reviewed, to add relevant modifiers, explain clinical decision-making, and document compliance with policies and regulations.

Medical codes and modifiers are used to create the superbill, which also includes units of time and the number of tangible units used (e.g. whole blood units, boxes of gauze). Authorizations for care and notes to support charges are also included.

Billers then pull information from the superbill to prepare claims, which are generally submitted electronically. During the payer’s adjudication, claims can be accepted, denied, or rejected, the latter generally due to lack of information.

Payers respond electronically as well, showing what services were paid, and explaining why a claim was denied or rejected. Unpaid claims may be corrected either through changing codes or adding documentation, further delaying payment. However, billers often calculate that “correcting the record” would cost more than it gained, effectively ending adjudication in the payer’s favor.

If/when claims are paid correctly, medical billers create statements of the patient’s financial responsibility. Providers generally cannot charge in-network patients for the difference between what the payer reimburses for covered services, and the chargemaster. However, providers can seek payment for non-covered charges, e.g. for a private room or a cosmetic procedure, and for co-pays and deductibles.

Fathom’s Medical Coding Process: Deep Learning for Deep Savings

1. The Fathom deep learning platform receives patient charts via PDF, RPA or direct EHR integration.
2. The platform codes all charts and makes an internal accuracy assessment.
3. Charts that can be automated are coded in the preferred system or interface. Lower confidence items are sent to human coders for verification, the results of which are integrated into the platform for future use.
4. All charts are double-checked by the system.

After new clients provide remote access to their EHR platform, Fathom’s system quickly learns how to deal with the specifics of the client’s installation. At launch, the system codes in parallel to human coders until results can be compared for accuracy verification, with gradual ramp-up to become the primary means of coding.

Our de-identification technology masks personal health information (PHI) in real-time, avoiding security compromises and HIPAA violations.

Toward Fully-Automated Coding: A Natural Evolution?

Challenges to the adoption of automated medical coding include resistance to change, which can be particularly strong in health care.

The rise of ACOs, joint ventures, and other mechanisms that involve multiple health systems is a strong impetus for optimizing the coding and billing functions among health record and revenue managers.

Moreover, EHR interoperability and standardizing data repositories are increasingly important to protect revenues in value-based reimbursement models, while furthering activities such as risk stratification, health outcomes research, and clinical trial recruitment.

Reducing the time and financial costs of medical coding, while improving its accuracy in order to drive prompt and optimal reimbursement, is just one step toward improving the accessibility, quality and financial sustainability of health care services. Fathom is proud to play a part in this evolution, and we welcome your questions.

Photo: Daniel Kalman