Letters

Coding for diagnostic tests

To the Editor:

In "Seven Tips to Improve Your ICD-9 Coding for Diagnostic Tests" [Getting Paid, February 2002, page 16], you advise reporting unrelated and co-existing conditions/diagnoses as secondary diagnoses, for example, in the case of a patient with chronic hypertension and diabetes who presents with a cough and whose subsequent X-ray indicates pneumonia. In the example given, would you report pneumonia, hypertension and diabetes? Obviously not at the initial visit if the patient's presenting symptom is a cough. Does your example pertain to follow-up visits or diagnostic testing? Would you code co-existing conditions if you are not treating them?

Janet Stanbro
Mentor, Ohio

Author's response:

Consistent with the first point in the article ("Code the confirmed diagnosis whenever possible"), you should code pneumonia at whatever point it is confirmed. Thus, if the patient has a chest X-ray while in the office for the initial visit, and your interpretation of the X-ray confirms that the patient has pneumonia, then you could report pneumonia at that encounter as a confirmed diagnosis.

If you simply order the X-ray at the initial visit, then, as you noted, you could not code pneumonia at that encounter since it has not been confirmed. Rather, you would have to code "cough" as the sign or symptom that prompted ordering the X-ray, which was the second point in the article ("If there's no confirmed diagnosis ... code the signs and symptoms that prompted you to order the test"). Use of the diagnosis of pneumonia would have to await the results of the X-ray and would be applicable to follow-up visits or subsequent diagnostic testing, as you suggested.

In either case, you could report hypertension and diabetes as unrelated, co-existing conditions even if they are not being treated during the encounter. For example, you might want to report them to highlight the complexity of the patient's situation, help justify the plan of treatment ordered for the pneumonia, etc.

Kent J. Moore
AAFP
Leawood, Kan.

Practice Diary: the real deal

To the Editor:

As a young physician, I find the practice described every month in Practice Diary very appealing. Dr. Brown shows that being a solo doc, operating with a low overhead and spending time with your patients can happen.

What I appreciate most about Dr. Brown's insights is what I perceive they give to students of family medicine. We are running businesses, but nowhere does it say that a medical practice has to be run like IBM instead of a mom-and-pop grocery store. Students can appreciate that and can get a feel for the interactions with patients that make our days rich.

Tom Sommers, MD
St. Louis, Mo.

Anticoagulation device accuracy

To the Editor:

The accuracy of the devices discussed in "Improving Anticoagulation Management at the Point of Care" [February 2002, page 35] is of prime importance. How can I assure myself of the accuracy of these tools?

Joseph J. Baum, MD
Salem, Va.

Authors' response:

Questions of diagnostic and laboratory accuracy often assume a widely accepted, reliable, reference test or "gold standard." Unfortunately, many standards contain more tin than gold. Such is the case with International Normalized Ratio (INR) anticoagulation assessment, as the thromboplastin reagents used to derive prothrombin times (PTs) vary across laboratories. This variance should be ameliorated by use of the International Sensitivity Index (ISI) to calculate the INR. However, even when using reagents with similar ISIs, resultant inter-laboratory INR values may still vary (Hirsh J, Dalen JE, Anderson DR, et al. Chest. 2001;119:8S-21S). Such variance may result in clinically relevant INR differences (Finck KM, Doetkott C, Miller DR. Am J of Health Syst Pharm. 2001;58:684-688).

Though several studies, including the study by Finck et al, have demonstrated that whole blood point-of-care anticoagulation monitors correlate highly with laboratory based serum INR assessments (CoaguChek to laboratory correlation coefficients of 0.92-0.95 and ProTime to laboratory coefficients of 0.87-0.90), it is unclear how much inter-laboratory correlation affects outcomes that matter to patients ­ the safe reduction of thrombotic events.

More important than choosing which device to employ is the method of employing that device. Patients who receive regularly scheduled provider-present anticoagulation assessments maintain their INRs in target ranges more often than those who use other methods (Jones RM, Pedersen CA. Hosp Pharm. 1998; 33:66-69). The clinical advantage of point-of-care devices is that they replace reliance on a single laboratory value with complete patient assessments.

Anne E. Caffee, PharmD
Peter G. Teichman, MD, MPA
Harpers Ferry, W.Va.