Home  |   Subscribe  |   Resources  |   Reprints  |   Writers' Guidelines

January 2018

Best Practices in Coding Audits
By Susan Chapman
For The Record
Vol. 30 No. 1 P. 10

Various methodologies, accuracy standards, and productivity criteria complicate the process.

In an effort to ensure coding accuracy and regulatory compliance, the Centers for Medicare & Medicaid Services (CMS) and the Office of Inspector General (OIG) recommend that health care providers routinely enlist independent parties to audit coding practices. Many health care organizations welcome these audits as a way to obtain feedback regarding their coding programs, in large measure to improve quality.

Although the goals may generally be the same, coding audits can differ in several ways, including methodology type, accuracy definitions, and cost.

Coding audits can be retrospective, which is a review of submitted claims, or prospective, an analysis of prebilled claims. They can also be random, targeted, or a mix of both, and a code-for-code or full-record analysis.

"Prebill audits take more focus and dedication to the timing of the audit to avoid a negative impact on accounts receivable. However, no rebilling is necessary, and the learning opportunity is immediate," explains Lisa Marks, RHIT, CCS, director of client audits at nThrive. "Postbill audits allow more breathing room with the audit workflow timing, but then necessary corrections require rebilling and delay learning. Additionally, rebilling for higher reimbursement must be performed within a certain window depending on the payer, so there is a time constraint when correcting mistakes resulting in higher reimbursement."

Shannon O. DeConda, CPC, CPC-I, CPMA, CEMC, CMSCS, partner and founder of DoctorsManagement and president of the National Alliance of Medical Auditing Specialists, says, "A lot of places like retrospective audits because they don't want to pend claims. But then they don't always like that type of audit because they have to correct billing errors. CMS could see it as fraud if the corrections aren't made in a timely fashion. The prospective audit helps alleviate that problem."

Once an organization determines whether an audit will be prospective or retrospective, the next step is sample selection methodology. "Some organizations conduct a random selection of all chart types a coder has coded and then select the charts to be audited from that sample," says Gabe Stein, executive vice president of GeBBS Healthcare Solutions. "Another sample selection method is a targeted random sample where they look at specific high-risk areas and high-dollar charts, for instance."

Jacqueline Thelian, CPC, CPC-I, CHCA, president of Medco Consultants, says 10 random encounters from each provider are typically selected for review. "This is generally not a statistically valid sample, and the claims selected are usually prebill to avoid the potential of any paybacks relating to the audit and/or previous claims," she says.

In a code-for-code audit, those codes that impact reimbursement are given twice the weight of those that do not. "Another perspective along the same lines is the full-record methodology, examining each record and only looking at mistakes that impact reimbursement," says Sarah Humbert, coding and compliance manager at KIWI-TEK. "If there is a mistake with the diagnosis, for example, there is a reimbursement issue. With this method, the entire record is counted as wrong."

Marks, who says each methodology offers different benefits, recommends clients employ various auditing strategies to gain the best each has to offer. "For example, for quarterly audits, quarters one and three would be random, all payers, retrospective, while quarters two and four would be targeted, all payers, retrospective and include a targeted daily prebill audit," she says.

Facility type determines the source documents required and the rules and regulations that must be followed. "When comparing different types of organizations—for instance, a community hospital with a trauma facility—a community hospital may lend itself to a random sample while a Level I trauma facility may lend itself more to a focused audit," Stein says.

Marks adds another example of how facility type changes the equation. "A teaching university facility should have larger volumes and more frequent audits performed quarterly due to the higher level of complexity and broader range of different services it offers. Meanwhile, smaller facilities with smaller volumes have less frequent audits performed annually due to the lower level of complexity and more narrow range of service types offered," she says.

Nevertheless, Humbert says facility type should not be an important factor when choosing an appropriate audit method. "I think the objective of the audit is more the key," she says.

Standardizing Methodologies
Marks says that standardizing auditing methodologies is difficult, given the variety of health care facilities. "There are so many nuances that are specific only to individual groups. Different groups have different challenges, documentation, and coding strengths and weaknesses—clinical complexities among them," she explains. "Even within one facility, they should consider different chart selection methodologies, as how chart selection is performed can determine what you learn from the audit."

But other experts believe that some standardization is possible. "Largely, it's based on the objective of the audit," Humbert says.

"You can have a consistent audit methodology—monthly, biannually—and then change up your sample selection methodology to achieve your desired results: improving coding quality and lowering compliance risk," Stein says. "With an audit methodology, the first step is to decide whether you choose to audit by coder, by service type, or by facility—whatever your grouping is for the audit."

Defining Accuracy
"The OIG recommends that physicians maintain an accuracy rate of 95%. Shouldn't we then expect coders and auditors to be able to maintain a 90% to 95% accuracy rate as well?" DeConda asks. "When you look at documentation guidelines deep down, there are no true definitions—they are guidelines and not rules. Therefore, they have now created gray areas within documentation, such as in the history of present illness [HPI]. There are no definitions of each element such as quality. Organizations need to create a standard definition for not just quality, but all of the HPI elements to create standardization within your own team. You have to have a unilateral answer and scoring technique for not only the coders and auditors but for the providers as well. Otherwise, it's like driving down the interstate without knowing the speed limit. I know in general what the speed limit may be, but I can't obey the law if I don't know what the law is. When claims go into the carrier, they say it's right or it's wrong even if they have no published standardization of these definitions and it becomes one huge argument over interpretation of a lot of people's findings. It can get convoluted if you don't have internal policies within the organization."

Humbert believes accuracy can be viewed more subjectively. "Often you have to take into consideration a facility's guidelines and procedures plus ICD-10 and CMS coding guidelines. There are some facilities that want everything captured, including secondary codes. Therefore, accuracy depends on the facility you're auditing and which items they want to be captured," she explains.

"In general terms, accuracy is precise performance of a task that is free from error. However, coding is too complex to cleanly fit into this definition," Marks says. "Coding can be accurate based on the existing documentation; however, the documentation could be incomplete, resulting in inaccurate coding based on the patient's actual and true clinical picture.

"The codes assigned by the coder may be accurate, but the coder failed to assign numerous appropriate codes, which results in incomplete coding," Marks continues. "Some would define incomplete coding as inaccurate. In this way, coding can be very subjective and open to interpretation of the guidelines and the documentation leading to differences in coding. There are times when a scenario can be coded in two different ways and both are correct."

Accuracy vs Productivity
Given that quality is an important factor in helping health care facilities keep costs under control, accuracy is an important factor in the auditing process. Yet, productivity is also critical.

"I recommend auditing for accuracy," Humbert says. "The productivity metric can be important for coders who are paid hourly, but there is no value to inaccurate records. Speed needs to be reasonable, but we've seen it over and over, one coder is achieving 98% accuracy but codes 10 records an hour. Another coder may be able to do the same thing with 12 records. Consequently, we place all of our weight on accuracy vs productivity. I don't see the value otherwise."

Thelian prefers to take a deep dive into the data. "Firstly, it is important to be sure the documentation supports the service being reported," she says. "With regard to productivity, I always look to see whether the provider is overutilizing services without supporting medical necessity to gain a higher work relative value unit, or productivity. Medicare provides physician data so you can download the productivity of other providers of your same specialty in your geographic area to determine whether your utilization for certain services is in line with the industry."

According to Marks, productivity and accuracy are two distinct measurements. "Coding accuracy is determined through coding auditing. Coding productivity is more operational. As a coding manager, I would review for both accuracy and productivity," she says. "Your strongest coders will maintain high productivity with high accuracy. In identifying your strongest coders, you can call on them to help mentor and support others. With efforts to increase productivity without impacting quality, coding costs can then be better managed."

Accurately Gauging a Situation
When measuring the accuracy of a facility's coding program, Stein recommends looking at facility type, the types of cases being coded, coder tenure, and production. "If the production is high but the quality is low, that is very telling," he says.

Auditors must take into account that guidelines differ from facility to facility. "What Hospital A wants to capture and what Hospital B wants to capture are different," Humbert explains. "You have to go back to the objectives. Get a gauge of the coder's knowledge or money left on the table by the facility. We can be looking for other opportunities of the HIM function to improve coding."

When measuring productivity, the complexity of a facility's computer system should be taken into account, as well as the degree of clinical complexity. More complex clinical scenarios can result in lower productivity and make it more difficult for coders to achieve higher accuracy.

"The documentation itself is another issue," Marks says. "If the documentation is weak, it can lead to poor quality in coding and necessitate multiple queries, which slow the revenue cycle and add burdensome steps to the workflow. For issues like computer complexity and documentation accuracy, education, training, and the current skill set of coders can drive accuracy and productivity goals."

Coding by Specialty or Payer
Whether coders code by specialty or payer varies by organization. Payers have specific guidelines, with most following Medicare rules. However, Humbert says coders work best when they focus on specialty, inpatient, or same-day surgery. "Some coders are really great cardiac coders, for instance. And there are some coders who are able to successfully be cross-trained. We get the best quality from those who specialize in certain areas," she explains.

Chart Mix
Much like coding by specialty, some organizations divide their coders between clinical and surgical coding. They also can be focused on either inpatient or outpatient cases.

"Whether they code inpatient, outpatient, or both types of cases is based on the organization, and sometimes even the specific coder," Stein explains. "In an academic Level I higher-acuity organization, the coders are most likely split. In a smaller, rural, community-based hospital, coders may do both. Currently, particularly at academic institutions, we are seeing organizations invest in additional internal training so that coders can expand their skill set. For example, outpatient coders can become inpatient coders over time with proper training."

"Inpatient and outpatient cases—they are almost Greek to one another," DeConda says. "Most coders and auditors have only one area that we tend to stay in: physician or facility."

"Coders tend to do one or the other," Humbert says. "There are a small percentage of coders who can work all types well. The norm we see is coders that specialize in certain areas. As an example, interventional radiology coders know their area very well, and it's complex. That is a huge asset in any facility—to have someone who really understands their specialty's cases well."

Marks offers a different perspective. "I would recommend coders not be boxed into one specific area," she says. "Allow cross-training where possible. This can strengthen the coders' commitment to their field and their morale. It can avoid boredom and provide coverage for such things as vacation time. The downside is the time, effort, and cost of maintaining multiple skill sets."

The Cost of a Well-Executed Coding Audit
Stein says any costs associated with an auditing program are worth it in the long run. "Ideally, your organization is performing audits on a regular basis, across all service types, locations, and payers. We've all seen claims that have been denied, resulting in additional work that may be preventable with a proper audit program. The goal is to use the audit feedback to improve coding quality and lower potential risk and cost to the organization. In a lot of ways, any organization should be investing in an effective coding audit program to make your coders better," he says.

Humbert estimates it takes twice as long to audit a chart as it does to code it, a statistic that takes into account the time to document an explanation of any coding decision the auditor is not in agreement with. "For example, in an emergency department audit, with a productivity of 12 charts per hour for a coder, our auditors would probably be able to review six per hour."

In terms of the financial impact, using Humbert's formula, the fees for auditing would equate to time spent. "This takes into consideration the auditor is documenting for educational purposes so our clients understand why something is wrong," she says.

"The time it takes to perform a coding audit is often driven by how well the coders are performing," Marks says. "A new or inexperienced coder will likely require more time to audit as issues are identified and written up and education is offered. As the same coder improves, the time required to audit the same volume of records should decrease as fewer issues are identified, less time is needed to write up the issues, and less time is needed for coder education. Overall, when conducted correctly and with findings used appropriately, well-performed coding audits can greatly benefit an organization. Benefits that you cannot apply a price tag to include confirming accurate reimbursement and improving marketing opportunities with a stronger, more accurate database."

— Susan Chapman is a Los Angeles-based freelance writer.