top of page

Risks In Medicine Part #2:

Updated: Jul 4, 2020

In the previous post I discussed an unfortunate case of a woman riding along in her car when she hears a commercial for a smart-heart CT scan. This sets off a chain of events that significantly threatens her life. Although she recovers fully from these unfortunate events without any permanent harm, her experience was much maligned and significant intellectual and financial resources were used.

This discussion is going to be focused on the thought process of the subject. I will not be commenting on the care provided. The focus will be on the thoughts that led to action, which led to predicament, which led to an outcome. To explore this, I will take the liberty of providing a thought narrative that I was not privy to. So, in truth, this may have not been the actual thoughts, but I imagine the internal debate is close.

Brief summary of events: Access to cardiovascular testing offered by institution —> patient decides to have test —> Risk assessment provided —> Patient decides to pursue consultation —> Physician desires further testing —> Available tests considered based on information gathered —> Cardiac catheterization chosen —> Next follows the series of unfortunate events from which the episode is judged.

Let’s start with discussing the problem she wished to solve. The realization of her own mortality was highlighted by the death of her father. A common reaction by someone whom recently lost a parent from and unexpected illness.1 Unfortunately, none of us can predict the future or foreshadow the cause of our demise. Medicine has predictive tools we can use to provide some clarity, but they are limited. If she could see inside her body and identify a harbinger of illness, maybe she could do something about it before it strikes. Hence her pursuit. When the commercial came through the airwaves an opportunity arose that could help.

Was the test offering a solution to her problem? The heart-smart scan is a decent test, but it is not a crystal ball. Based on the amount of calcium in your coronary arteries the test can place you within a risk quartile compared to your peers that will estimate your risk for a future cardiac event. That is the value of the test. Calcium scoring (the calculation used to risk stratify the subject) can reclassify your risk, but its precision and accuracy are limited. In fact, the highest quartile risk had a 10-year event rate of 13-25% across different genders and races. Those with a calcium score > 100 had an event rate of 7.5% over 10 years. The event rate with a calcium score of zero was 1.3-5.6% 2. So, its discriminatory value is limited, therefore forecasting is limited. We do not know how our subject would have interpreted these predicted ranges of risk. Would she consider them valuable?

If a score of zero carries a high-end risk of 5.6% and the highest quartile carries a low end risk of 13% (worst discriminatory case), is there really a significant difference in the quartiles. Statistically there is. However, the battlefield is the influence the results will have over the patient. Behavioral research has shown that individuals have difficulty differentiating between narrow probabilistic ranges, especially as they venture from 0% and 100%. This is the concept of weighted probability functions. Individuals are willing to risk more for certainty as it approaches 0% or 100% as opposed to a difference between something like 10-15%. Since certainty is not provided by the prediction tool it is difficult to appreciate the value. 3 The best discriminatory case (1.3% vs. 25%) is more robust and moves closer to certainty. The truth is somewhere in-between. The next question would be how much value does the best case scenario (1.3% rate vs. the 25%) rate provide to the patient. How would have this information effected our subject’s decision making?

A binary result would solve her problem. One wonders if this was the expectation, or if this was her thought (“I have heart disease, I need to do something about this right away”). As we can see with the tests performance above there will be limitations to the knowledge gained. We would need to know the predictive range that would be worth the pursuit. This would need to be answered by the patient prior to the test, or pooled data from a survey of patients. An established action threshold and reassurance threshold would enhance the utility of this test. An action threshold would be the level of risk that would result in action and a reassurance threshold would be a level that eased her concern. The statistical discriminatory value of the test is arbitrary without a action-reassurance spread. Now the test has applicable value.

We really don’t know what the subject’s intentions were. If her intentions were to mitigate any potential cardiovascular risk. Did her heart smart scan provide any knowledge beyond the statement, “My father just died of a massive heart attack. I want my cardiovascular risk to be as low as possible.” Is any physician going to withhold a baby aspirin, statin, or blood pressure medicine from this person? In my practice alone I have had patients jump at the chance to take a medicine for 10 years to lower their risk 0.5% and others that have refused in the face of an absolute risk reduction of 10%. In this case the test set off a chain of events that almost killed the patient. One wonders, was it even necessary.

I’m sure many healthcare providers would consider this degree of detail painful. This was a one in a million case. The tool is available and should be used. Medicine is interesting in the sense that the product the patient pays for has a momentum off its own. Also, the tools and medicine we use develop momentum that carry patients in all kinds of directions. You never really know what you are paying for. The clinical research used to approve tests and interventions do not observe the entire process. It would be too expensive. They have specific objections, measurements, and benchmarks to reach. When prespecified goals are accomplished, they enter regulatory approval and then released. In the case of the heart smart scan the objectives were: 1) does it detect the outcome desired, 2) does it reasonably predict events, 3) does it reclassify patients based on a standard risk score, 4) does it discriminate different groups of patients. These are characteristics of interest. 4 In this case the performance is likely average to above average, which is a pretty good result in the sciences.

How should we prevent incidents like this from happening? First, should we allow health system to indiscriminately offer services to patients without appropriate consultation? Recommendations for USPSTF state data is insufficient for cardiovascular screening with calcium scoring, and other imaging tools. 5 Based on this recommendation, it probably should not be done. Health systems are businesses that need to generate revenue and profits to remain viable. A heart smart scan is a good way to get them in the system. It appears relatively harmless. There’s demand for the service. If abnormal, you get set up with a cardiologist and a primary care doctor. However, it exposes large populations of people to information we don’t fully understand. The area under the curve for calcium scoring when added to standard cardiovascular risk prediction tools is .81. This is an improvement over the risk prediction tool itself (.76) but the benefit is marginal. The test should really be used only in patients of intermediate risk to determine if aggressive risk control should be implemented. High and low risk individuals shouldn’t mess with this because the marginal value is so low.

Second, should have more consideration of the end user. Did the user find the test valuable? How did the user feel before and after the test resulted? Does it lead to unnecessary procedures and testing?2 6 Does it actually result in reduced event rates after results are available? These questions should be answered before these services are offered haphazardly.

Finally, as researchers, we need bench science, we need clinical trials, and we need to observe the impact of innovation on society. Observational data is not robust enough to be relied upon in the discovery phase of innovation. However, it is invaluable in understanding this impact. 7The intention of any innovation is altruistic. The hope is it will behave as it did in the clinical trial and provide benefit to society. The innovator and society benefit. Too often this is not the case. There are unexpected effects, costs, and behaviors that arise. The story of this subject would not be seen a study evaluating cardiac CT, but it happened and we should know about it.

1. Alberts NM, Hadjistavropoulos HD. Parental illness, attachment dimensions, and health beliefs: testing the cognitive-behavioural and interpersonal models of health anxiety. Anxiety Stress Coping. 2014;27(2):216-228. doi:10.1080/10615806.2013.835401

2. Budoff MJ, Young R, Burke G, et al. Ten-year association of coronary artery calcium with atherosclerotic cardiovascular disease (ASCVD) events: the multi-ethnic study of atherosclerosis (MESA). Eur Heart J. 2018;39(25):2401-2408. doi:10.1093/eurheartj/ehy217

3. Kahneman D, Tversky A. The copyright to this article is held by the Econometric Society, http://www. :30.

4. Randolph AG, Guyatt GH, Calvin JE, Doig G, Richardson WS. Understanding articles describing clinical prediction tools. Crit Care Med. 1998;26(9):1603–1612.

5. US Preventive Services Task Force, Curry SJ, Krist AH, et al. Risk Assessment for Cardiovascular Disease With Nontraditional Risk Factors: US Preventive Services Task Force Recommendation Statement. JAMA. 2018;320(3):272-280. doi:10.1001/jama.2018.8359

6. Shaw LJ, Goyal A, Mehta C, et al. 10-Year Resource Utilization and Costs for Cardiovascular Care. J Am Coll Cardiol. 2018;71(10):1078-1089. doi:10.1016/j.jacc.2017.12.064

7. Boyko EJ. Observational Research Opportunities and Limitations. J Diabetes Complications. 2013;27(6). doi:10.1016/j.jdiacomp.2013.07.007

7 views0 comments

Recent Posts

See All

You Really Should Go to Church

Last week I was in New York City. My first time. It is a massive place. Far larger than I imagined. I’ve been to “large” cities before. Chicago, San Francisco, Miami, but nothing compares to New York.

Everyone is Not Going to Like You

It’s the first day of spring break.  My wife scheduled our vacation to start on a Tuesday since she was coming off a week of call and wanted an extra day to wrap things up. I’m home with the kids. To


bottom of page