ARTICLE: Physician Attitudes About Using Life Expectancy to Inform Cancer Screening Cessation in Older Adults—Results From a National Survey
JOURNAL: JAMA Intern Med. 2022 Nov 1;182(11):1229-1231. doi: 10.1001/jamainternmed.2022.4316.
Methods: Using the American Medical Association Physician Masterfile in this survey study, we mailed surveys (with 2 follow-up mailings) to 1800 primary care physicians in internal medicine, family medicine, general practice, and geriatric medicine about breast, colorectal, or prostate cancer screenings and 600 gynecologists about breast cancer screening in older adults (age ≥65 years) between April and November 2021. A total of 1893 physicians received surveys to our knowledge. The Johns Hopkins School of Medicine Institutional Review Board approved the study, and participants provided informed consent. We followed the AAPOR reporting guideline.
The primary outcome was whether a less than 10-year estimated life expectancy was considered a reasonable criterion for stopping screening. We dichotomized the outcome and examined its association with covariates with the χ2 test. We included all covariates with P ≤ .05 in the unadjusted analysis in a multivariable logistic regression model. Additional details are available in the eFigure and eMethods in the Supplement.
Results: Among 991 respondents (response rate 52.4%), 791 were eligible, and 75.3% (n = 596) agreed that life expectancy of less than 10 years was a reasonable criterion for stopping cancer screening, including 81.3% (n = 488 of 600) of primary care physicians and 56.5% (n = 108 of 191) of gynecologists (P < .001). In multivariable logistic regression, gynecology specialty (odds ratio [OR], 0.41; 95% CI, 0.24-0.69), older physician age (OR, 0.77; 95% CI, 0.66-0.89 per 10 years), and Black race (OR, 0.41; 95% CI, 0.21-0.79) were associated with lower odds of supporting using life expectancy (Table).
A total of 64.4% physicians (n = 509 of 790) agreed that reducing overscreening is part of good patient care, but only 38.8% (n = 300 of 774) perceived overscreening as a substantial problem in older adults (Figure, A). Among participants who supported life expectancy as a criterion for stopping cancer screening, 45.4% (n = 269 of 593) believed using life expectancy may introduce bias against racial and ethnic minority individuals and 48.4% (n = 287 of 593) believed it may introduce bias against those with low socioeconomic status; only 35.1% (n = 209) of these participants believed that life expectancy prediction algorithms were accurate for making cancer screening decisions (Figure, B).
For the full article, click here.
For a link to the abstract, click here.