By David Jaslow
We have spent the last year discussing the development and delivery of an algorithmic approach to the evaluation of a firefighter entering the rehabilitation sector to determine if and to what degree that individual may be ill or injured rather than just physically exhausted. How likely is it that the fireground medical screening exam will uncover any of the fireground medical emergencies that we have listed throughout this series? As they say in academia, “When the answer is unclear, that’s a good question.”
There is no scientific research to date concerning the clinical dynamics in the rehab sector. In other words, we don’t know the incidence (new onset) and prevalence (baseline existence) of disease processes such as uncontrolled hypertension among firefighters in the acute setting nor do we have a good handle on how well any conditions are detected by EMS personnel who are attempting to examine firefighters in the field. Much of what I and all the other columnists discuss and recommend on this website is based upon what we believe to be industry best practices. We build these constructs on the best available science in fields such as emergency medicine and public health and we apply these principles to fireground medical operations.
Given the lack of peer–reviewed medical literature on rehab sector medicine, it is critical to understand some basic epidemiological definitions so that one may better appreciate how to apply the decision-making algorithms we have proposed in this series. Epidemiology is defined as the study of disease. Epidemiologists ask questions and attempt to develop answers about why disease occurs. In a nutshell, epidemiology involves the proverbial who, what, how, why, when, where approach to problem-solving. As EMS physicians we ask the same questions but we direct them towards chest pain in the middle-aged firefighter rather than Salmonella outbreaks in Mexico, for example.
Understanding the tests
The fireground medical screening exam is a type of screening test. Screening tests are used in medicine to elicit early clues about the development of a disease process. Some common examples of screening tests are stool samples for occult blood (i.e. the rectal exam in the ED) to detect internal bleeding, chest x-rays as part of a hazmat physical and the Prostate Specific Antigen or PSA to detect prostate cancer.
A good screening test has several characteristics that make it so. The test should be cost-effective (relatively inexpensive). For example, use of the 12-lead ECG in the rehab sector is cost effective if the EMS crew already owns the biomedical device. The cost per use is then limited to electrodes and the amount of energy needed to recharge the batteries after use, which are both negligible. The test should be easy to perform in terms of the amount of physical and/or technical effort. Thus, many screening exams in medicine involve drawing a single tube of blood. Screening exams should be relatively non-invasive and easy to teach to the medical providers who perform them. The easier the exam is to teach, the more likely it is to be performed correctly, which limits errors in application. The test should have validity and reliability. A test is internally valid if the results can be demonstrated to be the same or to have a very small degree of error when done over and over again. Reliability is synonymous with reproducibility. A test is considered reliable if the results are the same when performed by different labs or people.
Finally, the screening test should have both high sensitivity and specificity to detect the disease condition. The term sensitivity is defined as detection of a disease state. ST segment elevation on a 12-lead ECG is highly sensitive for acute transmural (full thickness) myocardial infarction. When certain criteria are met, ST segment elevation is a highly sensitive test for heart attack, i.e. that test alone will detect a large percentage of the worst heart attacks. However, there are other types of heart attacks that do not cause ST segment elevation or will not show ST segment elevation early on.
Specificity is defined as detection of a non-disease state. A 12-lead ECG showing normal sinus rhythm is highly specific for the lack of an acute transmural myocardial infarction, but there are rare cases of persons having small heart attacks who have normal sinus rhythm and a completely normal appearing ECG during their initial evaluation. Thus, we say that a normal ECG is highly specific for ruling out a heart attack, but not perfect.
The best screening tests in medicine have a high degree of sensitivity and a high degree of specificity. In other words, we want to use screening tests that are very likely to find those persons who have the disease in question (true positives) and very likely to find those persons who definitely don’t have the disease in question (true negatives). We hope to avoid positive tests in persons who actually do not have the disease (false positives) as well as negative tests in persons who do have the disease (false negatives). In the latter case, this is akin to telling the firefighter that he is fine when the ECG is misread and really does show cardiac ischemia that has been missed by the paramedic
Testing on the fireground
Let’s stop here for the moment and present a quick analogy to fireground medical operations. The RAD-57 is the world’s first device to detect carbon monoxide poisoning non-invasively by use of co-oximetry rather than a blood test. The device is FDA approved. The RAD-57 is a diagnostic device but not a therapeutic device. It detects a clinical condition and then the practitioner must do something about it (apply oxygen). The RAD-57 is an excellent example of a screening tool on the fireground because it allows the EMS provider to gain insight into a clinical condition that can be very difficult to detect and/or prove by history, physical exam or use of other diagnostic testing alone.
In order to gain FDA approval, Masimo, the manufacturer, had to deliver enough data to show that the device had a high degree of sensitivity and specificity in its measurements. This device has to accurately detect the presence of carbon monoxide when it is actually in someone’s bloodstream and it has to register 1% when there is no exposure to CO. (The reason it registers 1% rather than 0 is that there is always a small concentration of this gas floating around in the air and the device only shows whole numbers. Thus, any percentage greater than zero in someone’s bloodstream, no matter how small, will show up as 1%.) More importantly, there needs to be a very small margin of error because different values suggest different treatment. In the case of the RAD-57, there is less than a 1% margin of error when a patient has significant CO poisoning and the values are up in the 35 or 40% range. Thus, when used correctly in patients who are not subject to the few constraints about any co-oximetry, such as proper application of finger probe in an environment where the fingers are not cold, the device is highly sensitive for picking up actual CO poisoning and highly specific for detecting its absence.
The Masimo example brings us to the next epidemiological teaching point. Not only is it important to understand sensitivity and specificity when applying screening tests, the practitioner must also understand the concepts of positive and negative predictive value. Positive predictive value is defined as the likelihood that a positive test actually indicates the presence of disease and is represented by the mathematical equation True Positives/True Positives + False Positives. How could a positive test not indicate the presence of disease—if it is a false positive! Similarly, negative predictive value is defined as the likelihood that a negative test actually indicates the absence of disease. It is represented by the mathematical equation True Negatives/True Negatives + False Negatives. Negative tests that are false negatives will miss the presence of disease. The risk here is the firefighter is told he is okay when this may not be the case.
The RAD-57 has been engineered to avoid false negatives which could occur if ambient light interferes with the sensors in the finger probe that detect wavelengths of light bounced through the fingernail. There are side flaps on the probe and even a black rubber “condom” that is available which slips over the probe to place it in complete darkness. However, it is still important to recognize the limitations of the device if the patient has poor peripheral blood flow, hypothermia or any other clinical condition that could affect the ability to properly measure the carboxyhemoglobin.
By understanding the technology used in the rehab sector, the EMS provider using the device will further limit the chances that there are false positives or negatives obtained. Three-lead ECG does not detect heart attacks reliably nor does the absence of PVCs indicate the absence of illness. Clear lung sounds via a stethoscope do not preclude hypoxia from a variety of conditions. If one understands the sensitivity and specificity of the history, physical and diagnostic testing for predicting illness and makes conservative decisions based upon the results, he or she will increase the positive and negative predictive value of the applied screening tests and be more likely to make better decisions. Scientific research yet to be conducted will show us how much better and how much safer these decisions will be.