Many claim that people in the United States have access to the best healthcare in the world. And while that is easily refuted with data from multiple sources, there is no question that the United States has one of the most complex healthcare delivery systems in the world. The combination of public and private sources of financing, with the varying healthcare structures across state borders, has created a challenging system to navigate. As a result, the health record, initially used as a tool to communicate among providers, now serves a multitude of purposes. Among those purposes is the ability of payers to find information that allows them to validate the care provided and ensure accurate payments.
The denial of payment by payers is monitored closely by organizations. In many cases, that denial is appropriate. If a claim contains incorrect demographic information and cannot be matched to a covered person, that claim should be denied. If a claim is submitted for a “medically unlikely” service, the payer is within their rights to seek justification prior to paying the claim.
If the claim is submitted properly and the rules are all followed, it should be paid. While HIPAA is best known for its medical record privacy provisions, the law also sets national standards for claims submissions. This seems to be increasing in frequency and, in fact, in late June 2021, the Centers for Medicare and Medicaid Services (CMS) issued a press release asking providers to report violators so they can be investigated. Denials due to these improper modifications of the standard transaction set rules can increase the frustration of both patients and providers.
One area of ambiguity in the US health system is the issue of the medical necessity denial. Physicians may perceive medical necessity as having a very clear definition: a service that is necessary for the evaluation and treatment of an injury or illness. But in the denial world, medical necessity has many meanings. Medical necessity denials are issued if the ICD-10-CM code does not match the “approved” codes, as in a national coverage determination (NCD) or local coverage determination (LCD). An admission may be denied for medical necessity because, according to the payer, the patient was placed in the wrong status. An emergency department visit may be denied for medical necessity because the visit was not deemed an emergency—after the fact, of course.
Many healthcare professionals will say that the best way to manage denials is to avoid them. Physicians don’t wait around for heart attacks and then treat them as they occur. Patients are urged to know their cholesterol, eat properly, exercise regularly, and treat with medications when indicated to prevent the heart attack. But despite physicians’ best efforts, patients still have heart attacks, and when they do, they are treated aggressively to prevent damage and hopefully reduce the risk of heart attacks in the future. Denials fit into the same paradigm. But knowing the number of denials is not nearly enough. Every single denial is an opportunity to prevent a future denial.
In the utilization review world, hospitals often look at length of stay as a measure of proper utilization of hospital services. Each year, Medicare publishes a list of every diagnosis-related group (DRG) along with the corresponding geometric mean length of stay (GMLOS). This number is based on the claims data from services provided in the year prior to the year of publication, meaning that the list published in late 2020 for fiscal year 2021 is based on data from 2019 admissions. GMLOS is based on a varying number of admissions. For organ transplantation, it may be in the hundreds, and for common illnesses like heart failure, it could be tens to hundreds of thousands. Hospitals then use the GMLOS per DRG to compare their performance to the “expected” length of stay (LOS). If their LOS exceeds the GMLOS, an edict comes down to “fix our length of stay.”
This logic has opportunity for improvement. First, it is more than likely that the hospital’s LOS data is based on the average LOS per DRG and does not take into account the effect of outlier admissions. And even if outliers are removed, can data from tens of admissions be accurately compared to tens of thousands of admissions? Medicare bases its LOS data on the date of inpatient admission and the ending date in the “statement covers” field on the claim, but most hospitals use some permutation of dates from their registration system that rarely match CMS’ dates. For example, the CFO and director of utilization review never have the same average LOS for the same period of time.
Rather than use LOS, a more accurate portrayal of opportunities may be to improve efficiency and utilization by tracking avoidable services and days. Every day in hospitals around the country, patients occupy beds who do not require hospital care or receive services that could safely be done as outpatient. For example, patients often stay one extra day due to the onset of evening. Perhaps the patient’s spouse does not drive in the dark. Perhaps the physician has not yet rounded to place the discharge order. Perhaps the durable medical equipment (DME) company has not yet delivered the home oxygen. Note that while each of these delays leads to an avoidable day in the hospital, each one has a different attribution. The lack of a ride home may be due to the hospital staff not planning ahead for the patient’s discharge needs. The physician who does not round until evening is clearly the root cause of that delay. And the delay due to the DME supplier who only delivers before 4 p.m. should be attributed to the DME company or even to the payer who contracts with that provider.
The varieties of avoidable delays are myriad—the hospital that does not do MRIs or stress tests on Sunday or have peripherally inserted central catheter (PICC) nurses available. The payer who will not approve a nursing home stay until 72 hours after notification of patient stability. The patient whose family is slow to choose a nursing home. The doctor who orders extraneous services that do not need to be done during the hospital stay. It is only when one accumulates such discrete, highly attributed data that action can be taken to address the root causes of the delays.
Denials are similar. First, what denial rate is appropriate? Some would argue that the fewer claims denials you have, the better you are performing. But with payers more aggressively denying claims and hiring outsourced audit companies to perform audits months and years after the claims are paid, a low denial rate could be a sign of being overly conservative in an organization’s processes. If an inpatient admission is never denied, how many observation stays should have actually been inpatient admissions? If there are 10 more inpatient admission and two get denied, the eight approved admissions more than compensates for the two denied ones.
What may be better is to look at are denial patterns. If Dr. Smith’s admissions are denied more than any other doctor, Dr. Smith could receive targeted education, not the entire medical staff. If there are a flurry of denials of lab tests, where the lab HCPCS code must be matched to an “approved” ICD-10-CM code, address the system that checks that prior to performing the service. If there are suddenly denials for total joint surgeries, someone should determine if there is appropriate documentation in the health record to show the patient received and failed conservative measures.
In many of these cases, the organization may be able to start taking action even before the denials arrive. For example, if a single payer starts requesting a series of charts for similar diagnoses, such as joint replacements or cardiac procedures, get the utilization review staff to review the cases and determine if the record is complete prior to submitting the records. And at the same time, determine if missing documentation from the physician’s office should have been obtained preoperatively instead of scrambling to get it after the surgery.
Guidelines and standards for medical necessity also change with time. Many hospitals are facing recoupments of hundreds of thousands of dollars for placement of cardiac defibrillators because, in 2018, CMS added a requirement to the national coverage determination that a shared decision-making visit must occur prior to the placement of the device. When audits found records that had no evidence of such a visit, it was not possible to simply call the office and get the notes because no such visit took place. The patient now had a defibrillator, and the hospital had to pay back the total payment.
In summary, tear down that wall between health information and utilization review when talking about denials. Use chart requests and denial data to confirm patients are receiving services that are not only medically necessary but also are covered by the insurer.
Ronald Hirsch, MD, FACP, CHCQM-PHYADV, CHRI, FABQAURP is vice president of the Regulations and Education Group at R1 Physician Advisory Services. Dr. Hirsch’s career in medicine includes many clinical leadership roles at healthcare organizations ranging from acute-care hospitals and home health agencies to long-term care facilities and group medical practices. In addition to serving as a medical director of case management and medical necessity reviewer throughout his career, Dr. Hirsch has delivered numerous peer lectures on case management best practices and is a published author on the topic.