Lung cancer screening judgement strongly influenced by clinicians

Key takeaways:

• The probability of an inappropriate for screening verdict differed based on the clinician.
• A high competing risk for non-lung cancer death raised the odds for clinicians deciding inappropriate screening.

During lung cancer screening selection assessments, clinicians influenced the decision of screening inappropriateness more than patient factors, according to results published in American Journal of Respiratory and Critical Care Medicine.

“Lung cancer screening has the potential to save lives by catching lung cancer at an early stage with definitive treatment, yet only 5% to 10% of people have gotten screened,” Eduardo R. Núñez, MD, MS, assistant professor at University of Massachusetts Chan Medical School - Baystate, told Healio.

“Screening is effective, and we need multi-level strategies to increase the amount of people getting screened, especially those with net benefit, while guiding those who are less likely to benefit and more likely to experience a harm toward health care decisions that align with their goals,” Núñez said.

In a retrospective analysis, Núñez and colleagues evaluated clinician assessments of 38,487 patients (mean age, 64.8 years; 96.6% men; 81.8% white) eligible for lung cancer screening (LCS) from 30 Veterans Health Administration facilities between January 2015 and February 2021, to find out what factors mark an individual inappropriate for LCS.

Researchers found a patient’s odds for this outcome using hierarchical mixed-effects logistic regression analyses that adjusted for 3-year predicted probability of non-lung cancer death.

Clinicians ruled that 1,671 (4.3%) patients from the total cohort were inappropriate for LCS, and this included 570 patients who had a 3-year predicted probability for non-lung cancer death greater than 20%. The total number of patients with this competing risk was 4,383, or 11.4% of the total cohort.

Researchers observed increased odds for a clinician assessment concluding that LCS was inappropriate among patients:

• with a 3-year competing risk for non-lung cancer death over 20% (OR = 2.66; 95% CI, 2.32-3.05);
• aged 75 to 80 years (vs. 65-69 years; OR = 1.45; 95% CI, 1.18-1.78);
• with interstitial lung disease (OR = 1.98; 95% CI, 1.51-2.59); and
• who stayed in a long-term care facility for an extended time (OR = 1.12; 95% CI, 1.03-1.21).

Patient factors that instead lowered the odds for a LCS inappropriate clinician assessment included current smoking (OR = 0.65; 95% CI, 0.58-0.73), a history of major adverse cardiovascular event (OR = 0.82; 95% CI, 0.69-0.96) and substance use disorder (OR = 0.79; 95% CI, 0.68-0.91).

When divided by U.S. Census region, researchers found a reduced likelihood for an LCS-inappropriate assessment among those residing in the South vs. the Northeast (adjusted OR = 0.31; 95% CI, 0.16-0.6). Although facility location impacted the odds for an LCS-inappropriate clinician assessment, facility volume did not impact these odds.

Researchers further noted “a large amount of variation” when assessing the probability of a patient with average cohort characteristics receiving a clinician assessment that marked them inappropriate for LCS.

“We generally say expect normal variation to be around 10%,” Núñez said. “In our study, we found that probability of being deemed inappropriate for screening (eg, due to comorbidities, less likely to benefit from screening) varied from 1% to 74%.”

Núñez explained that this means “the likelihood that a patient is deemed suitable for screening is more strongly associated with the clinician making the assessment than patient characteristics.”

An additional analysis including only patient-level variables revealed that the clinician conducting the assessment explained more variation in the likelihood for an inappropriate LCS outcome than the care facility (35.2% vs. 8.2%).

“Future studies will look to seamlessly integrate patients’ relevant data and characteristics into determining the probability that a patient will derive a net benefit from screening that is easily available to clinicians at the point of care,” Núñez told Healio.

Screening is not always appropriate for all patients, according to Núñez.

“There are patients with multiple comorbidities and limited life expectancy (eg, oxygen-dependent COPD) where lung cancer screening has little benefit as they are much less likely to experience meaningful life-year gains with LCS and more likely to experience harm (eg, overdiagnosis, distress from false-positive results and complications of downstream testing),” he said.

Núñez told Healio findings from this study demonstrate how important it is that clinicians have extensive discussions with their patients about the trade-offs of screening and work to align patient values with a decision to screen.

“For example, using the American Geriatrics Society’s framework for approaching care decisions for older adults with multiple chronic conditions and using their sample phrasing that resonates with patients (eg, ‘Your other health issues should take priority’) could lead to more informed and values-based decisions,” he said.