Please Correlate Clinically

Clinicians hate this phrase in a radiology report. I’ve heard fellow clinicians say “this means nothing” or “radiology doesn’t want to take any responsibility” or “that can mean anything”.

I used to be like that. But since I’ve had the opportunity to delve deeper into clinical decision making, I’ve changed my mind.

What this dreaded phrase means is that the radiology findings can mean different things under different circumstances. Radiology is not giving us the answer, radiology is giving us another piece of the puzzle. We as clinicians still have to put the pieces of that puzzle together. We have to see how this new data fits into the big picture, and try to make the oftentimes scattered data into a cohesive picture of useful information.

Let’s use an example that uses the exact same concept, but without the words “Please correlate clinically” attached. You order a CBC and the white blood cell count result returns at16,000. We know 16,000 is above normal and that’s about it from this single data point. We as clinicians have to put this data into a clinical context to give us some useful information.

Can this mean the patient is acutely sick?

If a patient chronically has this white blood cell count, then no.

Can it mean the patient is getting worse?

If the patient had a white blood cell count of 13,000 yesterday, then it indicates that the patient is getting worse.

Can it mean the patient is getting better?

If the white blood cell count yesterday was 22,000, the it indicates the patient is getting better.

Does it mean the patient has a serious infection?

If the patient has no symptoms of infection and has been on corticosteroids for an asthma attack for 3 days, then it points away from infection.

The same data can have different meanings depending on context.

Most clinicians won’t get frustrated and say “why didn’t the lab just tell me what the patient has?”. We’ll take the white blood cell count of 16,000, put it in it place in the clinical context, take a step back, and use that information to help figure out what’s going on with our patient. Basically, we correlate clinically and it’s no big deal.

A radiology reading gives us data just the same as the white blood cell count. We should treat it just the same. We should put all pieces of data about our patient together into the puzzle that is our patient. Our job as clinicians is to put those pieces together in a sensible, cohesive way to achieve our diagnosis.

Let’s see a couple radiology examples:

We have a chest x-ray with an ill defined, left lower lobe opacity. Radiology tells us, “Please correlate clinically”.

In one scenario, the patient has 3 days of fever, worsening productive cough, pleuritic chest pain, and a white blood cell count of 14. Sounds like the patient’s got pneumonia.

In another scenario, the patient has 3 months of weight loss, generalized weakness, and a scant, productive cough that is occasionally tinged with a small amount of blood. Here it Sounds like the patient’s got cancer.

Let’s see yet another example:

The patient had a CT of the abdomen. The small intestine shows an area of bowel wall thickening that could be consistent with inflammatory bowel or poor distension due to incomplete filling with oral contrast. Radiology tell us, “Correlate clinically”.

In one scenario, the patient has a week of worsening diarrhea, nausea, low grade fever, diffuse abdominal tenderness on repeat exams, and a white blood cell count of 15. Sounds like inflammatory bowel disease.

In another scenario, the patient has vomiting and diarrhea that started yesterday resolved just prior to arrival to the ER, has decreasing tenderness on repeat examinations despite only receiving fluids, feels hungry, and has a normal WBC. Sounds more like incomplete filling.

The take home message here is that “correlate clinically” just means you should put the pieces of the puzzle into place, and make the decision on the patient’s diagnosis as a clinician. Most of the time, you should be making the diagnosis, not the radiologist. The radiologist is there to help you make the decision, with one piece of the picture… not the whole picture. Sometimes radiology’s piece of the picture will be bigger, sometimes it will be smaller, but it will rarely be the whole puzzle.

So the next time the radiologist writes, “Please correlate clinically”, take this as a gentle reminder that you’re the one putting the puzzle together.

If you’re interested in learning more about clinical decision making (either a student, resident, faculty, or lifelong intern), please keep an eye out for my upcoming book tentatively titled A Guide To Making Clinical Decisions.

You may also be interested in my book on medical charting for clarity and medicolegal defensibility: The Handbook of Medical Charting

Pay Attention to Your Format AKA Use Paragraphs

As chart writers, we often overlook a medical chart’s formatting. What I mean by formatting is simply how your chart looks, how the words are laid out on the screen in front of you. How a chart looks affects how easily and quickly someone else can read it.

We do not have control over what some parts of our chart will look like. For example, many medical record systems lock us into a particular format for the review of systems and physical exam. The medical record software often controls aspects of the HPI and medical decision making sections as well.

However, most electronic medical record systems allow clinicians some control over the free text sections of the HPI and medical decision making. These parts of the chart contain much of the critical, clinical information. I’ll show you how small adjustments to our prose in these sections can make a big difference in our readers’ ability to quickly understand what we are trying to tell them.

I’m going to demonstrate two related concepts:

1. Separate different ideas.

2. Group similar ideas together.

They are essentially opposite sides of the same coin.

Functionally you end up writing paragraphs. Paragraph writing a skill that we probably all learned early in our education. This skill still applies when we write medical charts.

When you clump all of your ideas together, it’s more difficult to read. Simply making space at the appropriate place makes a huge difference in ease of reading. Reading a complicated history is much easier when you ‘ve divided the ideas into logical paragraphs. The paragraphs don’t (and usually shouldn’t) be lengthy either. You’ll see this in the examples below.

In the following examples the spacing choices could potentially be a little bit different. However, these examples only illustrate the principle. Sometimes we have multiple ways to format that make sense.

Our first example:

The patient is a 57 year old female with a history of hypercholesterolemia and arthritis. She complains mainly of dizziness, nausea, vomiting, and palpitations. Dizziness started about one week ago. The other associated symptoms worsened when her dizziness gets worse. Moving her head triggers her dizziness. She had similar symptoms in the past when she received a diagnosis of vertigo. She has no diarrhea, headache, blurry vision, or loss of consciousness.

Let’s see how breaking this up a little bit can make it easier to read:

The patient is a 57 year old female with a history of hypercholesterolemia and arthritis.

She complains mainly of dizziness, nausea, vomiting, and palpitations. Dizziness started about one week ago. She had similar symptoms in the past when she received a diagnosis of vertigo.

When her dizziness is severe, the other symptoms appear. Moving her head triggers her dizziness.

She has no diarrhea, headache, blurry vision, or loss of consciousness.

Here I separated the background information from the primary complaint. I also separated exacerbating factors, and ROS type symptoms. Making paragraphs allows the readers to focus on different parts of the patient’s story without interference from the other information.

Another example:

CBC, electrolytes, and cardiac enzymes negative. EKG shows NSR, normal intervals, no ST segment deviation, and no T wave inversions. The clinical picture does not fit MI. I discussed the results with the patient. Will discharge to followup with PMD.

When it’s broken up, we get:

CBC, electrolytes, and cardiac enzymes negative.

EKG shows NSR, normal intervals, no ST segment deviation, and no T wave inversions.

The clinical picture does not fit MI. I discussed the results with the patient. Will discharge to followup with PMD.

I separated the results by type, and then led into my conclusion and plan. Hitting the “Enter” button a couple of times made my chart much easier for my reader to read.

Now one more example that’s a little bit more complicated:

A 60 year old female with a history of asthma, HTN, hypercholesterolemia, arthritis, and cholecystectomy comes in with SOB. She has wheezing, dry cough, and chest tightness with cough. Exposure to smoke from a kitchen fire triggered her symptoms. Taking several doses of her albuterol inhaler improved her symptoms mildly. Symptoms feel similar to her asthma. She denied fever, palpitations, chest pain, abdominal pain, nausea, vomiting, and dizziness. However, she noted diffuse joint pain that is chronic and unchanged.

With spacing and a little re-arrangement to keep similar ideas together:

A 60 year old female with a history of asthma, HTN, hypercholesterolemia, arthritis, and cholecystectomy comes in with SOB.

Symptoms feel similar to her asthma. She has wheezing, dry cough, and chest tightness with cough.

Exposure to smoke from a kitchen fire triggered her symptoms. Taking several doses of her albuterol inhaler improved her symptoms mildly.

She denied fever, palpitations, chest pain, abdominal pain, nausea, vomiting, and dizziness. However, she noted diffuse joint pain that is chronic and unchanged.

In this example, I’ve given each of the following their own small paragraph: the patient’s medical history, the sentences focusing on the asthma symptoms, the triggering/relieving factors, and a mini review of symptoms.

When we pay attention to our formatting, we can make a big difference in how easy our chart is to understand. Something as easy as hitting “Enter” 3 or 4 times improves the readability.

As you write your chart, consider how your chart will looks when you’re finished. Consider what is easy to read and what is difficult to read. Small changes can make a big difference.

If you’re interested in learning more about medical charting for clarity and medicolegal defensibility, check out my book, The Handbook of Medical Charting.