Search

Clinical prediction models (CPMs) are tools that compute the risk of an outcome given a set of patient characteristics and are routinely used to inform patients, guide treatment decision-making, and resource allocation. Although much hope has been placed on CPMs to mitigate human biases, CPMs may potentially contribute to racial disparities in decision-making and resource allocation. While some policymakers, professional organizations, and scholars have called for eliminating race as a variable from CPMs, others raise concerns that excluding race may exacerbate healthcare disparities and this controversy remains unresolved. The Guidance for Unbiased predictive Information for healthcare Decision-making and Equity (GUIDE) provides expert guidelines for model developers and health system administrators on the transparent use of race in CPMs and mitigation of algorithmic bias across contexts developed through a 5-round, modified Delphi process from a diverse 14-person technical expert panel (TEP). Deliberations affirmed that race is a social construct and that the goals of prediction are distinct from those of causal inference, and emphasized: the importance of decisional context (e.g., shared decision-making versus healthcare rationing); the conflicting nature of different anti-discrimination principles (e.g., anticlassification versus antisubordination principles); and the importance of identifying and balancing trade-offs in achieving equity-related goals with race-aware versus race-unaware CPMs for conditions where racial identity is prognostically informative. The GUIDE, comprising 31 key items in the development and use of CPMs in healthcare, outlines foundational principles, distinguishes between bias and fairness, and offers guidance for examining subgroup invalidity and using race as a variable in CPMs. This GUIDE presents a living document that supports appraisal and reporting of bias in CPMs to support best practice in CPM development and use., (© 2024. The Author(s).)

The machine learning community has become alert to the ways that predictive algorithms can inadvertently introduce unfairness in decision-making. Herein, we discuss how concepts of algorithmic fairness might apply in healthcare, where predictive algorithms are being increasingly used to support decision-making. Central to our discussion is the distinction between algorithmic fairness and algorithmic bias. Fairness concerns apply specifically when algorithms are used to support polar decisions (i.e., where one pole of prediction leads to decisions that are generally more desired than the other), such as when predictions are used to allocate scarce health care resources to a group of patients that could all benefit. We review different fairness criteria and demonstrate their mutual incompatibility. Even when models are used to balance benefits-harms to make optimal decisions for individuals (i.e., for non-polar decisions)-and fairness concerns are not germane-model, data or sampling issues can lead to biased predictions that support decisions that are differentially harmful/beneficial across groups. We review these potential sources of bias, and also discuss ways to diagnose and remedy algorithmic bias. We note that remedies for algorithmic fairness may be more problematic, since we lack agreed upon definitions of fairness. Finally, we propose a provisional framework for the evaluation of clinical prediction models offered for further elaboration and refinement. Given the proliferation of prediction models used to guide clinical decisions, developing consensus for how these concerns can be addressed should be prioritized., Competing Interests: Competing interestsThe authors declare no competing interests., (© The Author(s) 2020.)

Objective: To determine if serum screen analytes identify preeclamptic patients at risk for small-for-gestational age newborns, maternal laboratory abnormalities and preterm delivery (<37 weeks gestation)., Study Design: Using a retrospective cohort of 102 preeclamptic patients, associations between serum screen analytes and newborn birth-weight percentile, gestational age (GA) at delivery and maternal pre-delivery laboratory abnormalities were evaluated using correlation coefficients and local polynomial regression., Result: Inhibin-A and maternal serum alpha fetoprotein were inversely correlated with newborn birth-weight percentile (-0.27, P=0.006; -0.35, P=0.00004) and delivery GA (r=-0.42, P<0.0001; r=-0.26, P=0.008) and positively correlated with pre-delivery aspartate aminotransferase (r=0.22, P=0.03; r=0.21, P=0.04) and lactate dehydrogenase (r=0.33, P=0.0007; r=0.29, P=0.004). A positive correlation was noted between both second-trimester beta human chorionic gonadotropin and estriol and maternal pre-delivery creatinine (0.28, P=0.004; 0.4, P<0.0001, respectively). Hundred percent of patients with ≥ 2 abnormal analytes delivered before 37 weeks gestation., Conclusion: Preeclamptic patients with abnormal serum screen analytes are more likely to have small-for-gestational age newborns, deliver preterm and have pre-delivery laboratory abnormalities.

Objective: Uric acid is known to be elevated in preeclampsia. We sought to determine if uric acid levels on admission correlate with the length of expectant management in preterm patients with preeclampsia., Study Design: A retrospective chart review was conducted on singleton preeclamptic pregnancies delivered between 24 0/7 and 37 0/7 weeks' gestation at Tufts Medical Center between January 2005 and December 2007. Patients with a multiple gestation and those transferred or discharged before delivery were excluded. Data regarding signs and symptoms of preeclampsia, laboratory values, pregnancy complications and outcome were abstracted from the medical records. Correlation between admission uric acid level and days of expectant management was assessed. The relative risk (RR) was used to estimate the effect of uric acid levels on expectant management length >7 days. Mantel-Haenszel χ(2) values were used to construct 95% confidence intervals (CIs) around the RR., Result: Four hundred seventy-one charts were reviewed. Of these, 190 met inclusion criteria. In all, 55 patients (28.9%) were managed expectantly for >1 week. Admission uric acid level correlated with days of expectant management (P<0.0001). Uric acid levels at admission were categorized as ≤4.0 mg dl(-1) (low uric acid level), 4.1 to 6.0 mg dl(-1) (medium) and ≥6.1 mg dl(-1) (high). Relative to women with high uric acid levels at admission, we observed a sevenfold higher rate of extending expectant management for >1 week among women with low uric acid level (7.0; 95% CI: 3.34 to 14.68). Women with medium uric acid levels at admission also had a higher likelihood of prolonging pregnancy relative to women with high uric acid levels (RR: 2.81; 95% CI: 1.32 to 5.96) (P-value for trend <0.0001)., Conclusion: Admission uric acid levels correlate with the length of expectant management in preterm patients with preeclampsia. Pregnancy prolongation for >1 week is significantly more likely in patients with low and medium uric acid levels at the time of admission. Uric acid levels may be helpful in assessing disease severity and counseling preeclamptic patients regarding likelihood of extended expectant management.