Seventy percent of health plans surveyed rank provider AI tools among their top three cost drivers. Hospitals and health systems are investing large sums of money, deploying AI-driven CDI platforms that are billed as improving revenue yield per inpatient discharge, with CDI specialists issuing queries based on the identification of diagnoses to be queried for CC/MCC Capture, clarifying a principal diagnosis. This software is also being deployed for post-discharge post-coding use where the AI software scans the record, identifies opportunities for additional revenue capture by suggesting queries for CC/MCC Capture that often lead to clinical validation denials. The risks of deploying these software platforms are multiple and include issues of upcoding through what Medicare refers to as manufactured diagnoses , defined below.
Manufactured Diagnosis (CMS-Aligned Definition)
A manufactured diagnosis is a coded condition that:
- Does not have sufficient clinical evidence in the medical record to support that the condition was present, evaluated, or treated, and
- Is introduced or reinforced through CDI queries, templates, or software logic rather than originating from the provider’s independent clinical judgment, and
- Fails clinical validation under CMS audit standards, even if it meets coding guideline syntax
Plain-English Translation (for executives and auditors)
A manufactured diagnosis occurs when:
The record contains isolated clinical indicators, but the total clinical picture does not support the condition—yet a diagnosis is still captured through CDI processes to increase reimbursement.
How CMS Evaluates It (Key Principle)
Under CMS Program Integrity Manual and UHDDS Official Guidelines for Coding and Reporting:
- Clinical validity overrides coding compliance
- A diagnosis must be clinically supported, not just query-supported
- Contradictory evidence in the record negates reportability
- The condition must meet UHDDS secondary diagnosis criteria (evaluation, treatment, or impact on care)
Red Flags CMS Auditors Look For
- Diagnosis appears only after a query
- No documented treatment, monitoring, or clinical impact
- Normal or conflicting clinical indicators
- Diagnosis driven by software logic (e.g., lab thresholds) rather than provider synthesis
- Repetitive query patterns across cases (programmatic upcoding risk)
Compliance Framing
A manufactured diagnosis is not just a documentation issue—
it is a program integrity risk that may be interpreted by CMS and Medicare Advantage plans as:
- Upcoding
- False claims exposure
- Pattern-based audit targeting
Manufactured diagnoses for reimbursement purposes create patterns of aberrant coding and billing that payers are increasingly identifying and addressing with sophisticated AI data analytics and data mining tools, used to track and trend provider submission of claims over time. AI-driven physician documentation platforms increase compliance risk with payers, including Medicare and other payers profiling hospitals using coding and billing data, targeting hospitals and health systems with aberrant patterns of documentation, coding, and billing. Profiling is used by payers, including Medicare contractors and the OIG, to identify hospitals where prospective review of records must be completed to validate coding and billing. Medicare contractors can require hospitals to undergo Targeted Probe and Educate initiatives where initial probe reviews indicate a high error rate of coding and billing, including upcoding through manufactured diagnoses. Sudden increases in CMI, consistently high CC/MCC Capture Rates, large number of cases with only one CC/MCC supporting a higher weighted DRG, and a large number of claims rebilled for a Higher weighted DRG under the Higher Weighted DRG Program are red flags for Medicare to intensify review of cases at identified hospitals’ prepayment. Prepayment payer reviews slow down reimbursement for post-discharge bills by at least sixty days, thereby slowing cash flow and contributing to higher costs to collect.
These AI driven CDI physician documentation and coding platforms, coupled with CDI programs, not only increase administrative costs to healthcare but also increases cost to payers, emploter sponsored healthcare premiums, and patient’s costs with higher premiums and out of costs expenses such as deductibles and copays and additional cost sharing.
Healthcare costs are projected to rise 9% in 2027, reaching their highest level in nearly two decades as health plans contend with mounting pressure from AI adoption, specialty drug spending, behavioral health utilization and regulatory changes.
This was among the results of a PwC medical cost trend report identifying several long-term forces reshaping healthcare spending and creating new financial challenges for payers.
One notable finding involves the growing role of AI, with 70% of health plans surveyed ranking provider AI tools among their top three cost drivers.
While AI has been widely promoted as a tool for improving efficiency and reducing administrative burdens, the report notes some applications may also contribute to higher reimbursement costs.
For example, AI-powered clinical documentation tools can generate more detailed patient notes, leading providers to submit increasingly itemized reimbursement claims. As a result, health plans are being forced to invest in monitoring and validation technologies to ensure claims accuracy and identify potential overbilling.
Administrative costs in the U.S. healthcare system account for approximately 15% to 30% of total healthcare spending, often estimated around $1 trillion annually. While broad national studies estimate the share at 15% to 30%, some institutional analyses—including figures from the American Hospital Association—estimate these expenses can consume between 25% and 35% of all health dollars spent.
The U.S. devotes significantly more to healthcare administration than comparable nations. Multiple payers, billing complexities, and insurance-related processes drive these elevated expenses.
The exact administrative percentage varies depending on the specific facility or data tracked:
- Hospitals: Studies analyzing CMS cost reports note that hospital administrative expenditures can exceed direct patient care by a 2:1 ratio, with some systems citing admin expenses approaching or exceeding 40% of their operational costs.
- Physicians: For individual practices, administrative functions can represent 25% to 30% of total practice revenues.
- National Level: Broader Peterson Center on Healthcare reports place the per-capita administrative cost at roughly $1,055—more than triple the administrative spend of similarly wealthy countries like Germany.
Costs for administrative staff and technology now consume an estimated 25% to 35% of all healthcare spending in the U.S., according to a June 2 report from the American Hospital Association.
Hospitals and health systems will never win the arms race against payers when it comes to reimbursement. Instead, hospitals must focus on improving overall revenue cycle processes with an emphasis on holding physicians accountable for defensible, proactive preventive denials avoidance documentation that is sustainable over time. Physicians must be equipped with the skill sets, knowledge base, and core competencies to be able to consistently integrate best practice standards and principles of documentation into their regular practice of medicine with charting and communication in the HER. Hospital CFOs must recognize that the CDI 1.0, reactive, retrospective, and transactional, have proven ineffective in driving operational improvement in physician documentation. What is needed is clinical transformation to CDI 2.0