February 19, 2007

Swinging for a More Perfect Approach to Coding
By Carlton M. Cottrell
For The Record
Vol. 19 No. 3 P. 16

Electronic medical records power the engine that can allow healthcare organizations to optimize case mix and improve documentation.

When hospitals want to enhance their revenue cycle, the focus often turns to cash flow, even though the impact of optimizing case mix can be more significant. Why? Although the industry unanimously agrees that improving cash flow is a worthwhile goal, conversations regarding increasing case mix can introduce discomfort. Will the hospital risk committing fraud? How can increasing case mix be open and objectively measured like cash flow? Will focusing on case mix merely put added pressure on the already beleaguered coders?

The reality is that everyone in the coding process works hard and to the best of their abilities; it’s the process that needs work. Electronic records facilitate this change by improving coding quality, the most important determinant of case mix.

The Imperfect Science of Coding
Stripped of all pretenses, coding is a knowledgeable judgment. The coder is asked to come up with a maximum of nine diagnoses and six procedures that describe the totality of the patient’s clinical experience.

Coding is the translation of an inpatient record that is typically roughly 100 pages. The 350,000 or so unique clinical concepts that might be present in the patient record are translated into the 16,000 possible ICD-9-CM codes that describe diagnoses and procedures. Their translation must be consistent with the volumes of rules that govern the code use, order, and combinations.

Coders must glean and translate these codes from documents created by physicians who steadfastly refuse to speak ICD-9-CM and may act as if creating accurate and complete documentation is not the most important use of their time.

Coding, in turn, determines payment. Most payments to hospitals and physicians are based on the coding of the patient’s medical record. The extent to which the payment is correct is entirely determined by the quality and supportability of coding. The financial implications of correct coding are immense—every 1% change in case mix translates directly to a 1% increase in prospective payment system (PPS) revenue.

For a 100-bed hospital with $70 million in annual revenue where 50% of the payments are from PPS payers, a 1% increase in case mix translates to a $450,000 windfall—45 times the effect of decreasing accounts receivable by one day. This underscores the importance of case mix, which is based entirely on coding, to financial viability.

Almost all facilities, when asked, will testify that their coding quality is superb. However, just how well is coding performed in the average hospital? Little has been published that dispassionately analyzes coding quality, but those that exist generally conclude that coding quality as well as case mix have room for improvement.

Three past studies by the Department of Veterans Affairs and the Institute of Medicine (1977-1980) found only two-thirds agreement on the coding of medical records, and even expert coders disagreed on patients’ principal diagnoses 19% of the time.

More recently, three studies conducted in 2000 with Agency for Healthcare Research and Quality grants found only 84% accuracy in coding 28 complications. One study, at William Beaumont Hospital and Wayne State University, both in Michigan, found that two coders agreed on the evaluation and management (E&M) coding of a set of emergency department charts less than 40% of the time—despite the fact that there are only five E&M codes from which to choose.

The largest ongoing study of coding quality has been conducted by the Centers for Medicare & Medicaid Services to assess PPS accuracy. It routinely finds that between 2% and 9% of the charts they sample do not support the assigned diagnosis-related group.

Why do we still see significant variation in coding after 20 years of intense scrutiny under PPS reimbursement? The answer: because the process of coding has changed little in the last 25 years. Coders still work basically alone; they code as fast as they can from whatever documentation is available to them; they are under intense pressure to code as soon as possible after discharge; and they receive little ongoing feedback on their accuracy.

If coding accuracy and case mix are going to improve, significant parts of this process must change.

The Coding Ideal
Optimal case mix comes from the accurate specification of the principal diagnosis—the diagnosis responsible for the patient’s hospital admission. It comes from identifying every existing complication and comorbid condition that could affect the patient’s treatment. It also comes from finding and correctly coding every procedure.

Achieving an optimal case mix is about getting below the surface in the patient’s record to find the exact duration of a coma, to find how long the patient was on a ventilator, to see whether the record indicates pneumonia or a respiratory infection, to determine whether the chest pain at admission was really a myocardial infarction, or to see whether the dietitian’s notes and diet indicate that the patient had malnutrition.

This hunt is the coder’s life.

Coding quality depends on three separate factors: documentation, coder skill, and coding time frame. If any of these factors is deficient, coding quality will suffer, and nearly all can be positively affected by electronic records.

Electronic Records and Documentation
The necessity for clear, complete, and unambiguous documentation cannot be overemphasized. Traditionally, the chart dies shortly after discharge when the physicians have dictated all their reports and moved on to new cases. The road to document clarification and reassessment, where ambiguity exists, requires effort on the part of physicians and coders alike if paper is involved. However, coders cannot code what is not clearly specified. Ambiguous documentation leads to ambiguous coding and is one cause for reduced case mix or denials.

Electronic records allow coders and physicians to work together and communicate to clarify ambiguity in the patient’s set of diagnoses and procedures without significant effort on either’s part. This process also provides ongoing feedback to physicians on their documentation practices. A key to both optimizing case mix and minimizing payment denials is coding that conforms to documentation.

Electronic Records and Coding Skill
Coding is a skill built on training, experience, and feedback. Without training, coders can miss significant content. Without experience, they can’t learn to find the subtleties in real patient records. Without ongoing feedback, they can’t tell whether their work is accurate and thorough.

Electronic records cannot immediately make coders more skillful. They can, however, allow skill level variations to be managed by “triaging” more straightforward charts to less skilled coders and complex charts to more highly skilled coders. Remember, these coders do not have to be at the hospital or even in the state. With electronic records, hospital staff can include some of the most highly skilled coders in the country. Also, because several coders can see a chart (or specific document) at the same time, coders can easily collaborate on questions before sending the codes to billing.

Ongoing feedback is essential to coding quality assessment and improvement, yet most hospitals rarely audit. Even those that audit tend to do so infrequently—semiannually or annually—because little is budgeted for auditing, and travel often consumes a significant portion of the audit budget. The feedback provided by these audits, often months after the records were coded, does more to annoy coders than help them learn.

Electronic records can change this. Because several coders can access one record at the same time, concurrent auditing is possible. External auditors need not travel to the facility, so audit dollars stretch much further. Feedback to coders can be immediate, while charts are still fresh in their minds, so real and significant improvement can take place. Electronic records can put charts in the hands of appropriately skilled coders as well as help improve the skills of a hospital’s own coders. With coding skill comes improved case mix and reduced denials, especially if coders are given sufficient time to code.

Electronic Records and Coding Time Frame
The old saying “Do you want it fast or do you want it right?” also applies to coding. Even the most skilled coders with perfect documentation will not do an adequate job if they are not allowed sufficient time to thoroughly evaluate charts, including documented complications and comorbidities in other volumes. Pressure is intense to get bills out and keep discharged, not final billed days down.

When all coding must be done by a hospital’s own captive staff, there is little recourse but to spend less time on each chart when backlogged—a tactic that jeopardizes case mix. Also, coders often do abstracting at the same time as coding, a task that doesn’t require coders’ skills but is performed by them simply because they have the chart.
Electronic records can help reduce the time pressure because coders can have immediate and unlimited access to charts.

Electronic Records: The Coder’s Most Important Tool
Even the best worker cannot do well without the correct tools. The pressures of insufficient documentation, knowledge, and time can thwart even the most diligent coder’s best efforts.

Electronic records become the ultimate tool to enhance coding quality because they alleviate these pressures without placing additional burden on coders. With coding quality comes an “accurate” case mix—in other words, a hospital will receive all the reimbursement it deserves for the actual services and treatments it provides, no more or less than this.

— Carlton M. Cottrell, a 20-year veteran of HIM, is senior director of marketing at ChartOne, Inc.