Clinical coding is a meticulous task that translates patient notes into specific International Classification of Diseases, Tenth Revision (ICD-10) codes, crucial for billing and insurance claims. Coupled with this coding, clinical validation ensures that documented diagnoses correlate with the patient’s presenting symptoms, signs, investigations, and treatment. While these processes are foundational for medical billing, they are highly complex and are used by both healthcare centers and insurance companies.
Insurance companies have their own set coding and clinical validation guidelines which are adapted from national guidelines and customized to a very large extent, such that they may deviate from the nationally accepted definitions of disease processes.
This strategic adjustment introduces substantial ambiguity into the clinical validation process and raises concerns about the entities responsible for this function, including artificial intelligence (AI) systems, non-clinical personnel, and outsourced firms.
AI and automation offer the potential to streamline the labor-intensive coding process, yet accurately capturing the complexities of medical diagnoses and treatments remains a challenge. Additionally, these technologies and the non-clinical staff or third-party companies involved in validation are primarily focused on insurance directives, which aim more at controlling costs than at improving patient care or documentation accuracy.
Clinical practice guidelines are developed through a systematic process to ensure they are evidence-based, enabling clinicians to make the best decisions for their patients. Typically, a panel of experts convenes to create these guidelines for diagnosis and treatment. Each author is required to disclose any conflicts of interest, providing transparency about potential biases that could influence the guideline. In addition, the guideline-development process is explicitly outlined for each nationally recognized guideline, including the level of evidence for each recommendation.
In contrast, the process of health insurance guideline development is often less transparent. It is frequently unclear how these guidelines are formulated, who is involved in their development, and whether conflicts of interest exist. This lack of clarity can raise concerns about the objectivity and reliability of insurance guidelines, which is evident in the differences that arise between insurance guidelines and clinical standards.
The divergence between the insurance-driven criteria and the established clinical standards, such as those from the Centers for Medicare and Medicaid Services (CMS), leads to frequent denials of coverage for reasons pertaining to diagnosis-related groups (DRGs), causing significant revenue losses for healthcare practitioners. Among the millions of accounts audited, instances where insurance companies adjust DRGs to increase hospital reimbursement are virtually non-existent. In other words, a fair audit that results in a DRG change that benefits the hospital financially simply does not happen; instead, these audits invariably lead to downgrades, reducing reimbursements rather than increasing them.
This discrepancy emphasizes the need for a harmonized approach that bridges the gap between insurance company guidelines and clinical practice standards. Establishing such alignment is essential, not only for ensuring that clinical validation accurately documents patient care, but also for supporting the financial sustainability of healthcare institutions in a complex and evolving regulatory environment.
Common diagnoses frequently denied and downgraded by insurance companies include the sepsis spectrum (A41), acute kidney injury (N17), acute respiratory failure (J96), severe malnutrition (E43), and type 2 myocardial infarction (I21.A1). Sepsis and respiratory failure are among the most common diagnoses for hospital admissions under inpatient status. Most hospitals have set protocols for the diagnosis and management of such conditions. Not only are such protocols adapted from nationally and internationally published guidelines, but hospitals also include those aspects in their customized protocols due to oversight by regulatory authorities, such as CMS, to ensure timely recognition and intervention are in place.
The guidelines authored by insurance companies are only for specific diagnoses, and not the entire spectrum of health issues. Some insurance companies, such as United Healthcare (UHC), have a specific document they term a “guideline,” which is a partial adaptation of those already published by scientific organizations. These payer documents include only specific diagnostic components of the condition, with minimal aspects related to treatment. Other insurance companies quote published scientific articles as their justification for denial. Denials are largely based on diagnosis and not on the treatment offered, even if the treatment resulted in improvement of the diagnosis that is being denied.
It should be underscored that insurance companies only recognize the diagnosis component of any condition in their guidelines and do not have any comprehensive treatment directives. This renders such a document no longer a guideline. Such a document could be best termed a rulebook. If the diagnosis lags by one decimal point from a lab value criterion for a specific diagnosis, despite clinical and professional diagnosis by a clinician, and if treatment is offered that results in improvement, it may still not qualify for the relevant DRG, due to strict adherence to this rulebook. The review process will not accommodate any degree of deviation from the rulebook lab values.
A thorough evaluation of sepsis criteria by UHC and Blue Cross Blue Shield indicates that they use the Sepsis-3 guidelines for the diagnosis of sepsis, with strict adherence to the numbers outlined for determination of the minimum Sequential Organ Failure Assessment (SOFA) score for a diagnosis of sepsis. Any deviation from that SOFA score is denied. Furthermore, one insurance company considers the SOFA score only after fluid resuscitation, which is not a part of the originally published Sepsis-3 criteria. If the initial SOFA score met Sepsis-3 criteria, and a subsequent SOFA score after administration of IV fluids and antibiotics resulted in a reduction of that score below the minimum required for diagnosis, it may be denied. Given the variability of interpretation and adoption by different insurance companies, a case of sepsis validated by one company may be invalid for another.
It is important to note that Medicare Advantage (MA) patients must be treated along the guidelines recognized by Medicare. For the diagnosis of sepsis, CMS follows the Sepsis-2 guidelines and expects all participating programs to be in compliance with the sepsis-care bundle. Despite adherence to the sepsis-care bundle, commercial insurance companies covering Medicare patients still refuse to recognize sepsis as a diagnosis if the basis for diagnosis used was Sepsis-2.
The CMS-4201-F rule stipulates that Medicare Advantage plans are obliged to adhere to traditional Medicare coding policies and the nationally standardized ICD-10 and CPT/HCPCS codes, including associated coding guidelines. Yet, major insurance companies continue to deny such accounts despite all levels of appeal, including a direct peer-to-peer meeting with their medical directors. The directors express that they understand the basis for appeal but confirm they are to abide by the rulebook created by the insurance company strictly without exception, despite knowing full well that sepsis is a clinical diagnosis and what the limitations are of applying Sepsis-3 as a blanket criterion. In other words, the authority given to them to deny an account overpowers the clinical authority they have achieved as practitioners, by virtue of training and experience, to question the rulebook in a specific context.
A related issue is that the Healthcare Association of New York (HANY) in 2019 determined that it would not recognize UHC’s use of Sepsis-3 criteria when reviewing claims to validate sepsis for payment. New York State law defines sepsis by systemic inflammatory response syndrome (SIRS) criteria, otherwise known as Sepsis-2, thus making it mandatory for UHC and other companies to use Sepsis-2 criteria if they are to provide insurance to New York residents.
Similarly to the situation with sepsis, the diagnostic criteria for acute respiratory failure vary from one insurance company to another. One company argues that a patient needs to be consistently on 5 L of oxygen or use a bilevel-positive airway pressure machine and have an arterial blood gas test on file with partial oxygen pressure of less than 60 mmHg to validate acute respiratory failure. Venous oxygen saturation or indirect measures via pulse oximetry are often not accepted as a surrogate marker of acute respiratory failure. Oftentimes, the criteria for diagnosis change, and it is unclear if the change is communicated to the contracting team or the hospitals. Coders have noticed that the criteria may change for different levels of appeal in the same case.
In the case of severe malnutrition, most insurance companies recognize the Global Leadership Initiative on Malnutrition criteria and not the American Society of Parental and Enteral Nutrition (ASPEN) criteria. Both are nationally and internationally accepted criteria for the diagnosis of malnutrition, with an explicit letter from the U.S. Office of The Inspector General stating that hospitals may use either of the criteria. Despite the argument that ASPEN criteria detect acute malnutrition cases that can be easily detected and treated in a hospital setting, insurance companies refuse to accept them even if treatment is initiated, thereby downgrading the DRG or removing it from the list and recouping the amount relevant to the omitted or downgraded DRG.
As described above, large hospitals and healthcare centers are at the mercy of the oligopoly of select insurance companies to have their work recognized and reimbursed. The burden of additional work involved in appealing the DRG denials initially involves three levels of appeals. With hospitals dedicating a team for this work, and with rapid turnaround of denial-appeal cycles, insurance companies have devised new ways of determining what constitutes an appeal and what does not. One insurance company allows for multiple appeals before determining that all three levels of appeals are exhausted, thereby prolonging both time and the number of appeals required to reach the end result, if the denial is to be upheld. Additionally, it is unclear if the initial appeal letters are reviewed by board-certified physicians, non-physicians, or simply by an automated process. Certainly, the language used in denial letters when the initial denial is upheld points to a possible use of AI. Some insurance companies have a third party involved in validating codes and clinical diagnosis. When all levels of appeals are exhausted with the third party, there is no clear path forward to the next level of appeal with the parent insurance company.
Thus, healthcare institutions are left with the difficult task of defining specific diagnoses due to the risk of denials. Since the survival of healthcare centers is directly related to reimbursement by payers, should medical students and trainee residents learn medicine based on insurance company rulebooks?
This pressing issue underscores the urgent need for regulatory bodies to investigate insurers’ rulebooks and their denial-management process. The ideal outcome would be to achieve uniformity for diagnosis and management among insurance companies in line with national and international guidelines. Furthermore, there should be a close relationship between the medical team and the contracting team of healthcare centers so that the language used for diagnostic criteria and treatment is reviewed by the medical team before contracts are signed between the healthcare center and an insurance company.
While it is understood that denials are issued to less than the majority of the accounts submitted, it is certainly a major issue for the hospital industry complex, which is surviving on razor-thin margins. As insurance companies continue to post billions of dollars of profits each quarter, shouldn’t a piece of the pie be shared with the hospitals that take care of their beneficiaries, rather than withheld by denying legitimate DRGs? Awareness by medical providers and advocacy by medical organizations should be at the forefront of efforts to ensure that hospitals are paid appropriately. In addition, the concept of clinical documentation integrity should be introduced in medical schools, so that future practitioners understand the critical relationship between clinical documentation and reimbursement.
Dr. Poonacha
Ms. Chamoun
Dr. Poonacha is a staff hospitalist, clinical associate professor of medicine, and medical director of utilization management at the University of Minnesota Medical Center in Minneapolis. Ms. Chamoun graduated from the University of California, Los Angeles with a bachelor’s degree in psychobiology and biomedical research. Following an internship in North Africa where she provided psychosocial support to victims of human trafficking, she will be working in Lebanon on health advocacy for refugee communities. Aspiring to become a physician, Ms. Chamoun is also preparing to attend medical school. She is dedicated to advancing healthcare and contributing to the medical community through clinical practice and biomedical research.