Why AI/ML is not a panacea to Payment Integrity?
Health insurances need to ensure all the claims are processed accurately to keep their Medical Loss Ratio (MLR) low. Currently, the health claims MLR is at 60% or higher for retail and 80% to >100% for groups health insurance products, most of the claims are processed manually for Inpatient and daycare services in India. As the market size increases and addition of comprehensive OPD coverage (Pharmacy, Diagnostics, Consultation etc) can increase the MLR as the current approach is prone to improper payments, fraud, waste & abuse with large volumes of claims.
Are the existing solutions scalable and efficient without increasing the operational cost as the market size increases and add more products like OPD, disease-specific programs?
Let’s understand how healthcare claims are handled in developed countries (See below image) and how India can adopt the best practices.
What is Medical Standardization (Medical Coding / Digitization)?
Medical standardization comes under a major gamut of healthcare data digitization. In healthcare data digitization, converting hard copy records into digital records alone will not be helpful. Digitization should also standardize the medical or clinical data for various business, social, research, care improvement, and various other purposes. Medical standardization for insurance and majorly EMR/EHR also is achieved with Medical Coding.
“Medical coding is the transformation of healthcare diagnosis, procedures, medical services, and equipment into universal medical alphanumeric codes. The diagnoses and procedure codes are taken from medical record documentation, such as transcription of physician’s notes, laboratory and radiologic results, etc. Medical coding professionals help ensure the codes are applied correctly during the medical billing process, which includes abstracting the information from documentation, assigning the appropriate codes, and creating a claim to be paid by insurance carriers.” – American Academy of Professional Coders (AAPC)
Medical codes translate that documentation into standardized codes:
Patient’s diagnosis
Medical necessity for treatments, services, or supplies the patient received
Treatments, services, and supplies provided to the patient
Any drugs used or prescribed
Any unusual circumstances or medical condition that affected those treatments and services
History
Medical coding derives from public bills of mortality posted in London in the 18th century. It was through correlating these that doctors determined the cause of a cholera epidemic.
Quality of care can only be measured and benchmarked through standardization. In a country like India, healthcare should speak one language to improve outcomes and enforce Standard Operating Procedures.
Medical coding India
In India, the medical codes that are captured are
ICD-10 PCS for Inpatient stays
ICD-10 for Diagnosis.
Many hospitals do not code them and even if coded, they are coded without following the ICD-10 coding guidelines which is very important for capturing essential information for making payment decisions. Many insurance companies fill these codes themselves for IRDAI reporting purposes. ICD-10 codes alone are not sufficient for any automation in downstream processes. Many more codes are required for usability and effectiveness of the data which can be used for multiple purposes (to be discussed in a later section) like the US or other developed countries.
What are the major medical codes in active use today?
Listed below are the medical codes in USA and their usages:
ICD-10-CM (International Classification of Diseases, 10th Edition, Clinically Modified)
Diagnosis codes, encounters, external causes etc.
ICD-10-PCS (International Classification of Diseases, 10th Edition, Procedural Coding System)
Procedure codes for inpatient reporting
CPT (Current Procedure Terminology)
Procedure/services codes for outpatient and professional billing
HCPCS Level II (Health Care Procedural Coding System, Level II)
Procedure/services/supply codes for outpatient and professional billing
Modifiers
A modifier is a code that provides the means by which the reporting physician can indicate that a service or procedure that has been performed has been altered by some specific circumstance but has not changed in its definition or code
CDT (Code on Dental Procedures and Nomenclature)
Dental procedure/service codes
NDC (National Drug Codes)
Pharmacy codes
MS-DRG (Medical Severity Diagnosis Related Groups) and APC (Ambulatory Payment Categories)
Inpatient and outpatient visits grouper
Many more
Why Medical Standardization?
A patient’s diagnosis, test results, and treatment must be documented, not only for insurance claims but to guarantee high-quality care in future visits. A patient’s personal health information follows them through subsequent complaints and treatments, and they must be easily understood. This is especially important considering the hundreds of millions of visits, procedures, and hospitalizations annually happening across the globe.
The challenge, however, is that there are thousands of conditions, diseases, injuries, and causes of death. There are also thousands of services performed by providers and an equal number of injectable drugs & supplies to be tracked. Medical coding classifies these for easier reporting and tracking. And in healthcare, there are multiple descriptions, acronyms, names, and eponyms for each disease, procedure, and tool. Medical coding standardizes the language and presentation of all these elements so they can be more easily understood, tracked, and modified.
This common language allows hospitals, providers, and payers to communicate easily and consistently.
The results may be submitted to insurance companies for reimbursement (Payment), but the data derived from the codes also are used to determine utilization, manage risk, identify resource use, build actuarial tables, and support public health and actions.
Applications of standardized data
The standardized data can be used across the company by various teams for data driven decision making, senior management reporting, strategic initiatives, IRDAI reporting.
Auto-adjudication for claims team: with the quality of data, which is generated, we can work towards Efficient & Scalable, auto-adjudication. Benefit design mapped to code & code to rule
Payment Integrity for Risk/ILM/SIU teams: Rule based (policy, clinical, utilization & behavioural) & AI/ML models are now possible, as the data is structured and machine readable (Discussed in the next section)
Network benchmarking for network team: Standardized network pricing, benchmarking the cost, quality etc
New product innovation: Structured data for Underwriting
Clinical decision support systems: like Population health and Disease management program
Preventive healthcare – Screening and wellness exam based on patient’s health profile
Better risk stratification for care management
Predictive Analytics, AI/ML Models to detect diseases at early stage to bend the cost curve and better patient health outcomes
Research, and Quality measurements (quality of care and outcomes)
Interoperability: Seamless communication between various healthcare system and codes sets
Many more
What is Payment Integrity (PI)?
Health care payment integrity: The process of ensuring the health care claim is paid accurately by the health insurance/Payer, for eligible/covered members, as per the Policy/contractual/clinical benchmarks, not in error or duplicate, and void of fraudulent, wasteful & abusive practices.
Health Insurers are seeking to uncover the signs of overcharging, false reporting, errors, wasteful, non-compliant practices or detect novel schemes from claims and billing processes. Claims standardization will immensely help automated payment integrity solutions. Risk/ILM (Investigation Loss and Minimization) /SIU (Special Investigative Unit) teams can now focus on the cost containment efforts and help mitigate risk proactively. This can be achieved by following the below, as many things are well-defined rules which will curtail the majority of the improper payments.
Rules driven (Proven and working in many developed countries).
Data analytics: Statistical or AI/ML (AI/ML needs standardized and labelled data). AI/ML alone cannot be seen as a panacea for this humongous and multi-layered problem.
What are the regulations in India for Medical Standardization and Fraud management systems?
In the USA, the Government enforces hospitals to standardize the medical data before submitting to insurance companies for reimbursements. Similarly in India, there are also guidelines to follow but not strictly enforced. On the brighter side many government initiatives towards health insurance schemes, claims standardization, claims automation, and fraud management have been rolled out or in development. Listed below are few guidelines:
IRDAI/NHA wants ICD-10 Diagnosis and ICD-10 PCS to be used for Primary and additional procedure/Dx for each claim
NHS – Policy Engine, Claims engine, fraud management systems.
NDHM/NHS – SNOMED CT/LOINC/RXNORM/ICD-10 Dx
FHIR R4 – Claim format
NHA for PMJAY, Health Benefit Package (HBP) 2.0 is aligned with International Classification of Health Interventions (ICHI) and International Classification of Diseases (ICD) coding of the WHO
What is the current state in India?
Even though, IRDAI proposed medical coding, very few hospitals are adhering to that. Also, the coding guidelines are insufficient to understand the patient’s illness in its entirety. Lack of medical codes and incomplete codes, insurance claims are manually verified for authenticity and necessity.
There are thousands of conditions, diseases, injuries, and causes of death. There are also thousands of services performed by providers and an equal number of injectable drugs & supplies to be tracked. It is humanly impossible to verify the claims accurately and efficiently.
Also, through this manual adjudication, there is no combined learning which can be taught to a machine for automation. There is data but there is no data.
There has to be a large and serious push towards medical standardization in a cost-efficient manner. In a country like India with growing insurance penetration (Recent distribution side Insurtech start-ups are being well funded to increase the market size), it is paramount to set up these guidelines and adopt automation to catch up with the fast-growing market size.
OPD accounts for ~64% of the healthcare spend in not just India but almost similar across the world. The current approach will never allow OPD to be part of the health insurance. OPD needs are different as the unit prices, volumes and touchpoints are wide & varied. Compare the number of hospitalizations with the number of consultations/prescriptions/diagnostic tests. They are not comparable at all for any country. Manual adjudication is not possible. Not only the operations but also the Fraud, Waste & Abuse (FWA) are unmanageable. FWA can only be addressed if there is any real data. Otherwise, nobody understands the utilization and what for.
Conclusion
The COVID-19 is testing the healthcare infrastructure both at healthcare providers and health insurance. The beds must be freed sooner to the needy. Holding a bed is not an option, especially quoting claims approvals are pending.
Having said that there is an unprecedented fraud in COVID cases and Insurance companies are forced to do physical investigations, which could be addressed in an automated fashion to a larger extent by payment integrity solutions. Again, due to the lack of standardization and so the legacy decision systems at insurers’ side, the bills are always higher from an insured person as opposed to the non-insured person.
With recent activity in the distribution space by InsurTech start-ups will increase the market size and create new opportunities for innovative product requirements (like cashless and comprehensive OPD etc).
Standardization and automation in claims processing will bring greater transparency, scalability, control, and accountability to the most important stakeholders of our healthcare system. Unless the merits and demerits are understood intimately, a better thing cannot be developed. Otherwise, costly mistakes will be repeated, and advancements are delayed.
About CoverSelf:
CoverSelf is a global SaaS payment integrity provider, specifically built to focus on health care claims and payment inaccuracies. We are incorporated to address these challenges and bring decades of learning from the various healthcare systems. We are pioneering and customizing various payment integrity solutions with the next generational architectural shift by empowering the domain experts with data-driven decisions that will reduce the burden of labor-intensive tasks and help implement better scalable practices and eventually increase savings and reduce operational costs. We help to improper payments, identify anomalies, and potential FWA based on claims patterns and benchmarking (Peer/pricing, etc) through standardized data for risk/PI teams to perform a targeted investigation on ground.
Sources:
https://www.aapc.com/medical-coding/medical-coding.aspx
https://www.aapc.com/blog/26995-medical-coding-is-vital-to-healthcare-data-analysis/
https://www.aafp.org/fpm/2001/1100/p28.html#:~:text=The%20HIPAA%20transactions%20and%20code%20set%20standards%20are%20rules%20to,to%20computer%20without%20human%20involvement.
https://niti.gov.in/writereaddata/files/document_publication/NHS-Strategy-and-Approach-Document-for-consultation.pdf
https://www.nrces.in/ndhm/fhir/r4/index.html
Author: Raghavendra Pawar & Rajasekhar Maddireddy, Co-Founders, CoverSelf
Disclaimer: The opinions expressed within this article are the personal opinions of the author. The facts and opinions appearing in the article do not reflect the views of IIA and IIA does not assume any responsibility or liability for the same.