Context is crucial when deciding what to do with abnormal test results

Question: What do you do when presented with abnormal lab results?

Answer: Ask lots of questions.

The nursing home just sent over a urinalysis on a patient of Dr. Carlyle. I am covering his practice for a few days. The test showed that an 82-year-old woman had 3+ white blood cells in her urine. “NKDA” was written in the margin, indicating she had no allergies.

I sighed internally and called the nursing home. The charge nurse seemed a little surprised at all my questions.

“What are the symptoms? What is the patient’s kidney function? Is she on blood thinners or any other medications that might interact with an antibiotic?”

The presence of bacteria or white blood cells in the urine should not usually be treated if there are no symptoms. That’s not always been our belief, but most doctors agree with this approach today.

Looking at a test result without knowing the story behind it, we cannot decide whether or how to act.

Last week, we got a critically high potassium result on a patient with normal kidney function and no prescription medications in her profile. I did nothing about it, except order a repeat test that was normal. The obvious explanation was hemolysis; red blood cells contain more potassium than the serum that transports them and if the cells break during blood draw or handling of the vial, serum potassium will be falsely elevated.

A seizure patient of Dr. Carlyle had a high phenytoin level. I pestered the nurse to give me several past results and to track any previous dose changes. It turned out this patient had stable levels for a year and a half and suddenly had a low level last month. Dr. Carlyle raised the dose. In retrospect, the patient probably had missed a few doses, and would have been fine staying on the same dose. I dropped the prescribed dose back down and expect the patient to do fine.

A hypothyroid patient was hospitalized with abdominal distention and constipation. She is non-verbal, and fearful of medical procedures. The hospitalist checked her thyroid function, as undertreated hypothyroidism can contribute to constipation. The test suggested she needed a higher dose, so she was discharged on a substantially increased dose of levothyroxine. As soon as I saw her again, I reversed the medication change; her TSH had been normal one week before her admission, and a severe illness or traumatic experience can affect thyroid values. I figured the hospitalist did not notice her old TSH result in the hospital computer.

Context is crucial when deciding what to do with abnormal test results. But doctors are often pressed for time, and finding the story behind the results takes time. Even when all the data is in our electronic medical records, it takes time to see the patterns: The test results are usually in one place, the prescriptions in another, the office notes in a third, and the phone messages in a fourth. My own EMR can produce flowsheets with lab results, but each test is identified by the date it was ordered instead of the date it was performed, so correlating lab values with prescription dates becomes confusing, for example when following thyroid cases.

In times past, when solo practice physicians cared for their patients in the office, hospital and nursing home, they kept the threads of context and continuity together more easily. Today, with more providers sharing the care, and with other office staff also interacting with patients and their families, there is more room for errors, gaps and confusion. The tools we have right now are not always as effective as we would like, and they certainly can be cumbersome and slow to use. Reading each other’s notes can take a while, as the EMR format is primarily built for coding and not for ease of following the clinical story.

A few words doctor to doctor, doctor to nurse or doctor to patient can sometimes do what half an hour on the computer might not. Treatment without context is essentially just random reflex actions: Killing the innocent bacteria, lowering the falsely elevated potassium, treating the lab value and not the patient — none of it does anybody any good, and probably will cause harm to some unfortunate patients.

Our temptation to view test results as obvious facts in a predictable process instead of possibly misleading clues in a complex mystery reminds me of these words from a Sherlock Holmes novel:

There is nothing more deceptive than an obvious fact.
- Sir Arthur Conan Doyle

“A Country Doctor” is a family physician who blogs at A Country Doctor Writes:.

View 4 Comments >