Using risk calculators in lung cancer screening discussions

As readers may recall, I changed my mind on lung cancer screening last year and have been working on incorporating screening discussions into visits with eligible patients in my clinical practice. The American Academy of Family Physicians recently endorsed the U.S. Preventive Services Task Force (USPSTF)'s 2021 recommendation to offer annual lung cancer screening with low-dose computed tomography (LDCT) to adults aged 50 to 80 years with at least a 20 pack-year smoking history who have smoked within the past 15 years. Although a meta-analysis of 8 randomized controlled trials found that people screened with LDCT are 19% less likely to die from lung cancer (NNS = 250), it also concluded that about 20% of tumors are overdiagnosed, in line with a previous report from the U.S. National Lung Screening Trial. Unfortunately, doctors do not often discuss harms of lung cancer screening such as overdiagnosis, overtreatment, and complications of diagnostic procedures performed for positive tests.

Deciding if the potential benefits outweigh the harms of lung cancer screening for an individual patient requires a way to personalize estimates of benefit based on patients' risk factors. In a Letter to the Editor regarding a 2019 American Family Physician article on the pros and cons of lung cancer screening, Dr. Abbie Begnaud and colleagues suggested:

If an eligible patient is reasonably healthy, clinicians could consider calculating individualized lung cancer risk using one of several well-validated risk models. We and others have developed web-based tools to help clinicians incorporate individualized risk calculations into decision-making. Individualized risk assessment can be helpful because patients at higher risk of developing lung cancer are also more likely to benefit from early detection through screening. When lung cancer risk increases, uncertainty about whether to recommend screening decreases when the person has a reasonable life expectancy.

Unlike risk prediction tools for cardiovascular disease and breast cancer, however, there is no consensus on which lung cancer risk calculator should be used. A systematic review published earlier this year in the Journal of General Internal Medicine identified 10 publicly available risk calculators and assessed their performance in 16 hypothetical patients across the continuum of lung cancer risk. The calculators used varying inputs (demographic factors, cancer history, smoking status, and personal and environmental factors) to generate lung cancer risk estimates; unsurprisingly, there were substantial differences in risk estimates for 10 of the 16 hypothetical patients. The authors concluded that the lack of standardization of lung cancer risk factors and consistency in risk estimates from web-based calculators may be an obstacle to shared decision making.

Notably, the USPSTF statement "recommends using age and smoking history to determine screening eligibility rather than more elaborate risk prediction models because there is insufficient evidence to assess whether risk prediction model–based screening would improve outcomes relative to using the risk factors of age and smoking history for broad implementation in primary care." In a Putting Prevention Into Practice case study in the July issue of AFP, Drs. Howard Tracer and James Pierre explained how to apply the Task Force recommendations in clinical practice. It will be interesting to see if the Centers for Medicare and Medicaid Services decides to follow the USPSTF and waive its current requirement for a shared decision-making visit prior to lung cancer screening.

Why you should just say no to "routine blood work"

You're at your family doctor's office to have a complete physical. Maybe you're starting a new job, or have recently joined a wellness program in your community, or it's been more than a few years since you've had a checkup and you (or your spouse or significant other) just want to make sure that everything's OK. Your doctor briefly reviews your medical history, performs a physical examination, says a few encouraging words about eating a healthier diet and exercising more, and then you're done.

You picked this doctor out of the five in the practice because your friend told you he was a sharp young fellow, but now you're not so sure. What about the blood work? You don't need any blood work, he says. Not even a urine sample? This is confusing. You've always had blood work and urine tests at your other physicals, and your insurance is footing the bill, after all. You wonder if this doctor really knows what he's doing.

This is a common situation that I face in primary care practice. For years, patients have been used to having blood samples drawn even if they felt completely well. Even today, when we know better (or ought to), up to one-third of primary care physicians still perform "routine blood work" (usually consisting of a complete blood count, a chemistry panel, liver function tests, thyroid tests, and a urine analysis) at adult physical examinations. So why is this such a bad idea? In 2007, I co-authored an editorial in the journal American Family Physician about this topic. We wrote:

"Big-ticket" tests [such as CT scans and MRIs] are easy targets for those seeking to reduce waste in health care. But what about the seemingly innocuous practice of performing routine tests such as a complete blood count (CBC) or urinalysis? ... These tests would be useful only if they provided additional diagnostic information that would not otherwise be obtained during a history and physical examination. In fact, large prospective studies performed in the early 1990s concluded that these tests rarely identify clinically significant problems when performed routinely in general outpatient populations. Although the majority of abnormal screening test results are false positives, their presence usually mandates confirmatory testing that causes additional inconvenience, and occasionally physical harm, to patients.

Don't misunderstand me. There are certain situations in which targeted screening tests can provide valuable information for the early detection of diseases. To learn more about which tests are recommended for your or your family members, I recommend that you visit the excellent website Healthfinder.gov. But the next time you go to a doctor's office and he or she proposes to check some "routine blood work," be sure to ask what these tests are for and what would happen if any of them turn out to be positive, so that you can make an informed choice about what's right for you.

**

This post first appeared on Common Sense Family Doctor on September 13, 2009, and remains as relevant now as it was then.

Pharmacogenetic testing's usefulness remains limited

For patients with frequent episodes of gout, I often prescribe the drug allopurinol to lower their uric acid levels and prevent future episodes. This drug isn't for everyone, though. In 2019, Dr. Carl Bryce wrote a diagnostic test review in American Family Physician about the allopurinol hypersensitivity assay, "a blood test to detect the presence of a human leukocyte antigen B [HLA-B] genetic variant that increases the risk of life-threatening, severe cutaneous [skin] reactions in patients taking allopurinol." According to this article and a rapid evidence review of gout, testing was recommended for Korean adults with stage 3 or higher chronic kidney disease and all adults of Han Chinese or Thai descent, who have a higher frequency of the variant, prior to initiating allopurinol. In 2020, the American College of Rheumatology (ACR) simplified and broadened this testing recommendation to "people of Southeast Asian and African American descent."

Though pharmacogenetic testing holds promise for improving clinical decision-making, a recent JAMA viewpoint contended that race-based testing recommendations are problematic. Even a racially homogenous European country such as Switzerland exhibits wide genetic diversity in the frequency of the HLA-B*58:01 allele, with one city (Basel) actually having a higher frequency than the overall U.S. African American population. Further examination of the ACR's race-based guidance reveals additional complexities and contradictions:

The ACR guideline cites Han Chinese, Korean, and Thai as examples of Southeast Asian descent, even though China and Korea are not typically considered Southeast Asian countries. The guideline then states that screening is cost-effective in Asian populations generally. However, Japan is in Asia, but the allele frequency of HLA-B*5801 in Japan is even lower than that of White individuals in the US, who are not recommended for screening. In addition, the recommendation to screen all African American patients in the US before prescribing allopurinol belies wide-ranging HLA-B*5801 variation across Africa, where reported HLA-B*5801 frequencies, based on small sample sizes, range from 1% (comparable with White individuals in the US) to 10% (comparable with Thailand).

In a noteworthy editorial, Dr. Bonzo Reddick took aim at a related issue: the widespread use of diagnostic and clinical prediction tools that, like pharmacogenetic tests, incorrectly utilize race as a proxy for genetic differences. These include the atherosclerotic cardiovascular disease (ASCVD) Pooled Cohort risk calculator, equations for glomerular filtration rate (GFR), a calculator for predicting the likelihood of a successful vaginal birth after cesarean delivery, and pulmonary function testing "correction factors" for Black and Asian patients.

Recognizing that "claims about pharmacogenetic testing ... are inconsistently supported by scientific evidence, and most tests have not been examined by the U.S. Food and Drug Administration [FDA]," Drs. Wendy Rubinstein and Michael Pacanowski shared the FDA's perspective on what clinicians need to know in the July issue of AFP. In a table of selected pharmacogenetic associations, they summarized known and potential gene-drug interactions and recommendations for clinical practice. In a diagnostic test review in the same issue, Dr. Natasha Pyzocha evaluated GeneSight Psychotropic, an expensive test panel that analyses 12 genes with possible interactions with 57 neuropsychiatric medications. Dr. Pyzocha concluded that while this test may help patients who have had multiple unsuccessful trials of therapy, "because only a small population of patients are expected to have genetic phenotypes that would necessitate medication changes, ... routine genetic testing is not recommended," and "choosing antidepressants based on health history and symptoms should still be the standard initial approach."

**

This post first appeared on the AFP Community Blog.