Where did all the money go? Exploring medical coding and billing errors
Improper healthcare spending can add up quickly when clinical information is lost or distorted across documentation, coding, and reimbursement.
EHRs sit at the center of this process. Physicians capture what is clinically happening, while hospitals and insurers need that information translated into standardized documentation and codes. When these layers do not align, the system compensates with reviews, queries, corrections, and audits, creating more administrative work and financial uncertainty.
EHRs Make Clinical Documentation Part of Reimbursement
EHRs serve two different purposes. For physicians, the medical record captures the patient’s condition, clinical evidence, diagnostic reasoning, treatment, and follow-up. For hospitals, insurers, and regulators, that same record must support coding, billing, quality reporting, compliance, and reimbursement.
This creates tension. Physicians think clinically, but reimbursement systems require clinical decisions to be documented with enough specificity for administrative systems to interpret them.
Health IT can support diagnosis and clinical care, but poor workflow integration and documentation requirements can increase cognitive burden and interfere with clinical reasoning.
Medical Coding Translates Clinical Decisions into Billing Data
Medical coding bridges clinical documentation and reimbursement.
Diagnoses, procedures, treatments, and other aspects of care are converted into standardized codes used for claims, reimbursement, quality measurement, and risk adjustment.
The difficulty is that clinical medicine and coding do not describe patients in exactly the same way. A physician may describe a condition accurately in clinical terms, while coding requires additional specificity about anatomy, severity, laterality, complications, or procedures.
With extensive and continuously updated code sets, physicians cannot reasonably master every coding distinction. Their primary responsibility is determining what is clinically happening.
Accurate coding therefore depends on collaboration. Physicians establish diagnoses and document clinical decisions, while coding professionals translate that documentation into standardized codes. Problems arise when the information needed for that translation is incomplete, inconsistent, or ambiguous.
Coding Queries Correct Gaps After the Clinical Decision
Consider a patient hospitalized after a traffic accident.
At discharge, the physician records “left lower leg crushing injury with necrosis.” Clinically, the description may appear reasonable. But when the diagnosis enters the coding process, it may not contain enough specificity to map confidently to the appropriate code.
The coding team therefore sends a query asking the physician to clarify the documentation. The physician must reopen the case, reconstruct the clinical context, review the evidence again, and update the record while managing other patients.
Queries are an important safeguard, but they expose a limitation of traditional coding improvement: the information gap is detected downstream, after it has entered the medical record.
Making queries more efficient improves correction, but it does not prevent the gap from occurring.
Billing and Payment Errors Create System-Wide Financial Risk
The consequences grow once documentation and coding gaps reach reimbursement.
Some medical billing advocates have estimated that errors may appear in up to 80% of medical bills, although published estimates vary substantially depending on what is classified as an error. Problems can include incorrect codes, duplicate charges, missing information, and other discrepancies that affect payment.
In 2020, the Centers for Medicare & Medicaid Services estimated $0.93 billion in improper payments for Medicare Part D and $86.49 billion for Medicaid.
Improper payments are broader than coding errors and are not synonymous with fraud. They can result from insufficient documentation, eligibility issues, inaccurate information, coding problems, overpayments, underpayments, and other failures to meet program requirements.
Their scale shows how consequential information integrity becomes once clinical information enters reimbursement systems. Patients may receive inaccurate bills, providers may face reimbursement or compliance risk, and payers may act on incomplete information.
What begins as an information gap at the point of care can become a financial and administrative problem downstream.
Fixing Coding Downstream Does Not Fix the Clinical Foundation
Hospitals use coding review, clinical documentation improvement, physician queries, and audits to identify documentation and coding gaps. These functions are necessary, but most intervene after the physician has already made the diagnostic decision.
If a diagnosis is missing, outdated, insufficiently supported, or inconsistent with the available evidence, the consequences extend beyond reimbursement. The same information can influence treatment decisions, care coordination, quality measurement, risk adjustment, coding, and future interpretation of the patient record.
A coder can request clarification, a CDI specialist can identify missing specificity, and an auditor can find inconsistencies. But each intervention attempts to reconstruct clinical context after the fact.
The more fundamental opportunity is therefore to strengthen the connection between clinical evidence and the physician’s diagnostic decision at the point where the diagnosis is established or reassessed.
DxPrime Strengthens Diagnostic Integrity at the Point of Decision
This is the broader problem DxPrime is designed to address.
DxPrime analyzes longitudinal patient information together with evidence from the current encounter to help physicians determine whether a condition is current, clinically supported, and relevant.
It reconstructs clinical context across diagnoses, medications, laboratory results, procedures, treatments, and other available information, then brings the relevant evidence together for physician review.
The physician remains responsible for the diagnostic decision. DxPrime supports that decision by making the evidence easier to evaluate and identifying potential integrity gaps while there is still an opportunity to act. Physicians can confirm, update, resolve, rule out, or follow up on a condition before that diagnosis propagates further through the healthcare system.
A physician-validated diagnosis creates a stronger foundation not only for documentation and coding, but also for clinical care, care continuity, quality measurement, reimbursement, auditability, and downstream use of health data.
The difference is therefore broader than moving coding intervention upstream.
Traditional downstream processes ask:
How do we correct the record after a discrepancy appears?
DxPrime asks:
How do we ensure the patient’s disease status is clinically supported, current, and traceable before that information becomes consequential across care, documentation, coding, and reimbursement?
Medical AI should not simply automate administrative correction. Its role should be to strengthen clinical reasoning where decisions are made, preserve the connection between evidence and physician judgment, and create a more reliable clinical foundation for every process that follows.
| Reference |
1. Medical billing errors growing, says Medical Billing Advocates of America
3.2020 Estimated Improper Payment Rates for Centers for Medicare & Medicaid Services (CMS) Programs