To screen or not to screen? That is the question
Volume 10, Issue 5 | August 8, 2024
Also in This Issue
Faster testing to choose antibiotics for sepsis
Lackluster accuracy in ED pneumonia diagnosis
Diagnosing concussion rink or court or field side
Bird flu update
New and Noteworthy
Can this test find the right antibiotic in time to cure sepsis?
Three weeks ago, we covered a urine test that can diagnose a bacterial urinary tract infection (UTI) in 15 minutes and evaluate the bacterium’s level of resistance to five common antibiotics within 45 minutes. That second step is called antibiotic susceptibility testing (AST). AST is finding the most effective antibiotic for an individual infection - the key to fighting antibiotic resistance.
Last week a study in Nature presented a technique named ultra-rapid AST (uRAST) in blood for sepsis - a life threatening fast moving systemic infection. By comparison with the UTI technique, uRAST doesn’t seem exactly “ultra-rapid,” as it takes somewhere between 11 and 15 hours. But that’s a vast improvement from the two to three days that conventional culture AST usually takes. The challenge in sepsis is that blood contains few bacteria to test, so you have to spend a lot of time growing more bacteria in order to then test how those bacteria respond to various antibiotics. This uRAST test uses an array of techniques to render blood and pure bacterial culture (BC and PC, respectively) unnecessary (See Figure 1a below).
Commentary: Many commercially available diagnostics exist to address the 2-3 day turnaround problem, but adoption has been disappointing. This Nature paper provides an interesting competitive overview in its Fig.1f. Which option will provide the most appealing combination of completeness, speed, and cost is yet to be clear.
We can do a lot better diagnosing pneumonia – a large VA study
Speaking of antibiotic susceptibility testing (AST) - did we mention that adoption of commercially available tests is disappointing? (Yeah, we just did - scroll up.) This week the Annals of Internal Medicine provided more evidence for that take. It reported diagnostic accuracy for 320,000 patients who visited emergency departments (EDs) at Veterans Affairs (VA) hospitals and were hospitalized for pneumonia as a result. It’s not a pretty picture.
Just 38% of patients who received a pneumonia diagnosis in the ED had that diagnosis confirmed by chest x-ray and were treated appropriately while they were in the hospital. Another 36% of ED pneumonia diagnoses proved to be false positives, while 22% were false negatives (see chart for details - the statistics here are tricky). The problem was compounded by high variability in accuracy across the multi-hospital VA system.
Commentary: Interestingly, tests that help distinguish bacterial from viral infections (measurements of C-reactive protein and procalcitonin) were widely used in the ED. But that’s not enough, because they are not highly specific for pneumonia. Microbiology was performed, but it’s unlikely that rapid AST was used (details were not reported).
The overall picture here is an unfortunate one (to say the least). Older populations like this one have higher rates of lung comorbidity, which can confuse and delay accurate differential diagnosis. It’s an environment in which effective use of AST diagnostic tests is critical. We are capable of better.
Diagnosing acute concussions during the game
Both professional and amateur athletes fear concussions (and so do their parents, even if the athletes aren’t little kids). The challenge is diagnosing them at the point of need - which most often isn’t a hospital or doctor’s office.
The standard of care is a five-question test called the Sport Concussion Assessment Tool (SCAT), but its accuracy is inconsistent. Research published in June in JAMA Network Open showed that 45% of athletes who are eventually diagnosed with a concussion tested normally on the part of the test that looks at their ability to recall a set of words. The most accurate part of the test asks whether the athlete has symptoms of a concussion - but athletes who just want to get back in the game have no incentive to answer those truthfully.
Enter a rash of new technologies to more accurately diagnose concussion / traumatic brain injury (TBI). One of the newest technologies uses a portable instrument to test whole blood. It takes 15 minutes to look for two markers that indicate a TBI.
Commentary: The test requires an EDTA tube of whole blood, which the average amateur-league coach can’t gather from a player on the sidelines. However, multiple other approaches exist, from impact-sensing helmets to diagnostic eye-tracking devices. We believe that acute TBI testing is a growth field.
Bird flu update: When you test, you find
Two weeks ago, Colorado’s state veterinarian published an order requiring dairy farms in that state to submit weekly samples from their bulk milk tanks for H5N1 testing. Nine previously undiagnosed herds have now tested positive for the virus thanks to that surveillance. More than half of Colorado’s dairy herds are now confirmed to have been infected.
Food for Thought
(Who) to screen or not to screen? That is the question.
Screening higher-risk individuals for cancer saves countless lives. But is the benefit of screening lower-risk people worth the cost? That’s still an open question.
Two papers this week examine the cost of screening. The first examines direct costs of screening for the five most prevalent US cancers. The second takes a detailed and balanced view of the cost vs. benefit of screening lower-risk individuals for lung cancer using LDCT (low-dose computed tomography - an imaging modality).
As the first article states, $43 billion of direct medical costs are spent on cancer screening in the US. Most of those funds go toward relatively expensive tests (55% on colonoscopy alone), primarily for the well-insured (see chart). However, that amount is small when one considers the pain, suffering, and deaths avoided. It is also overwhelmed by other costs associated with cancer (source: The Cancer Atlas): treatment costs ($161 billion), lost workdays ($30 billion), and premature death ($151 billion).
The article on lung cancer LDCT screening provides a useful case study of the complexities that arise when we broaden screening to include low-risk individuals. In the US, LDCT is widely recommended for heavy smokers who are older (50-80 years old, 20+ pack-years, fewer than 15 years since quitting) and is estimated to have reduced lung-cancer-specific mortality by 20-plus percent. Incidence in the remaining population is much lower overall, but the cancers that arise in this group are more complex and often resistant to treatment. (Smokers tend to have many oncogenic mutations which increase the effectiveness of immunotherapy.) This paper advocates a broader risk calculation that includes family history, exposure to second-hand smoke or cooking flames, other occupational exposure, lung diseases, and other comorbidities, but it remains cautious about expanding screening.
The challenge with LDCT is that it identifies lung nodules, most of which are indolent, especially in older patients. Follow-up biopsy is invasive and risky. Even after biopsy, 22% of the subsequent surgeries turn out to have been unnecessary.
Commentary: At first glance, the benefits of screening for diseases as prevalent and severe as cancer seem indisputable, but on closer inspection they’re more complex. Screening can be invasive and costly and can lead to overdiagnosis factors that offset the obvious benefits. Among these factors are 1) the discovery of tumors that are indolent, benign, or otherwise have no influence on premature death or quality of life, and 2) false positives, which rise dramatically as more and more people get screened.
At the public-health level, the math that determines which screening tests make sense and for whom is relatively straightforward, though still highly contentious. More screening always saves some additional lives, but it also inevitably increases overdiagnosis and false-positive rates.
So, what about cheaper, less invasive cancer screening using blood tests? The hope is that the decrease in testing costs will allow more at-risk individuals to get screened more often, saving lives with early, regular testing. The jury is out as to what extent this promise will be realized - we have high hopes but need to see the improvements in test accuracy along with costs (and pricing) to enable frequent testing.
Liquid biopsies may be the solution but their sensitivity and specificity numbers for Stage 1 and Stage 2 cancers need to get a lot better before they are useful. The open question is does the tumor shed meaningful DNA in Stage 1 or Stage 2 cancers? Not clear