When treatment algorithms fail: A response to the development of a nomogram to determine the frequency of elevated risk for non-medical opioid use in cancer patients

We read the February 2021 article by Yennurajalingam et al. (2021) with great interest given the growing recognition that individuals with cancer remain at risk for opioid misuse and overdose (Tan et al., 2015). Caring for this patient population requires carefully balancing risks to manage cancer-related pain while keeping patients safe from the risk of potential overdose (Alyssa et al., 2020). The authors developed a nomogram to help identify patients at-risk for non-medical opioid use (NMOU) and include a patient’s race as a risk factor (Yennurajalingam et al., 2021).

Recent publications caution against the use of race in such algorithms (Vyas et al., 2020). The nomogram in this article finds that a Black patient matched with a White patient for all other categories would be at elevated risk for NMOU (Yennurajalingam et al., 2021). Disappointingly, the authors neither reflect on why this may be true in their clinic setting nor mention structural or other forms of racism. While the authors correctly conceptualize race as a social variable with marital status and financial distress, we believe that inclusion of race is more likely to promote race-based disparities in opioid and pain care than it is to improve care. Furthermore, the inclusion of race and recommendations for future research of genetic factors for NMOU leaves room for the reader to incorrectly interpret race as a biologic factor rather than a sociopolitical one (Boyd et al., 2020).

Black Americans are less likely than White Americans to have their pain managed, both because clinicians are less likely to believe their reports of pain and because Black Americans face significant barriers to access pain medications (Meghani et al., 2012; Trawalter et al., 2012; Jefferson et al., 2019). Black patients may experience a high burden of chronic cancer pain, which is associated with higher unemployment, lower socioeconomic status, and inadequate insurance (Meghani and Chittams, 2015; Jiang et al., 2019). Additionally, Black patients are subjected to more burdensome opioid care such as more frequent required office visits and restricted refills, despite similar rates of opioid use disorder between Black and White patients (Becker et al., 2011). Finally, Black patients are less likely to be offered medical treatment for opioid use disorder and are more likely to die from an overdose (James and Jordan, 2018; Lagisetty et al., 2019). A treatment algorithm that includes race is likely to reinforce known inequities in pain and opioid care due to structural racism, provider implicit bias, and the false idea about biological differences in pain perception (Hoffman et al., 2016; Hirsh et al., 2020; Meghani et al., 2020).

The authors concede limitations to their nomogram, but not the likely harm that it perpetuates. Our field should prioritize creating decision-making and treatment algorithms that reduce bias, promote equity, and ultimately disrupt racism through antiracist action (Obermeyer et al., 2019). In the end, it is the journals themselves that bear responsibility of gatekeeping and vigilant adherence to the current standards for publishing on racial health inequities (Boyd et al., 2020).