David Dickerson, MD

Although the medical literature has described which patients undergoing surgery are at increased risk for more significant pain or opioid use postoperatively, it is only recently that mechanisms have been created to automate the identification or patients at risk.

Tools such as the one in Soens and colleagues’ abstract would have been unheard of as recently as 5 to 10 years ago. Their predictive model uses technology that exists in most EMRs and thus, allows us to be more aware of which patients are at risk and who may have higher needs in terms of pain management. Out of 163 factors, Soens and colleagues identified 21 predictors they deemed most influential. Many of these predictors have been well established for 15 years, such as anxiety, depression and preoperative use of opioids. The benefit of Soens and colleagues’ model is that it streamlines the acquisition of the data and enhances clinician awareness.

However, it is important to note that there will not always be a foolproof way of identifying all patients at risk for increased pain or needing more significant pain management, even with machine learning and predictive models. Clinicians still need to be ready to ask: what information am I missing about this patient? And when and how will iteration on a standardized approach be taken to ensure all patients receive timely, effective pain care?

References

• Brummet CM, et al. JAMA Surg. 2017;doi:10.1001/jamasurg.2017.0504.
• Ip HY, et al. Anesthesiology. 2009;doi:10.1097/ALN.0b013e3181aae87a.

Anita Gupta, DO, PharmD, MPP

In the United States, many prescriptions for opioids following surgery serve as an unintended gateway to long-term opioid use. We know that more than 50 million surgical procedures happen in the U.S. annually and an estimated 2.6 million Americans will become persistent users of opioids following initial exposure after surgery in the hospital. Reducing opioids exposure in the postoperative setting is a crucial way to help curb the opioid epidemic.

Surgery plays a dual role in the opioid epidemic. It can be the initial introduction to opioids for many patients and can lead to unused pills becoming available for misuse and abuse. To address the opioid epidemic, much attention has been focused on reducing the widespread use of opioids to treat patients with chronic pain. While there is no question that this population is most prone to dependence and addiction, not enough scrutiny has been given to another significant population at risk of long-term consequences related to opioid overexposure — that is, patients who have just undergone surgery. Prior studies have shown that when fewer opioids are used to treat patients, these patients are less likely to experience adverse events.

Therefore, it is imperative that we begin to use the power of data intelligence and machine learning to address the opioid crisis, as Soens and colleagues do in this research. They identified patients at high risk for experiencing more severe pain and higher opioid requirements after surgery, which guided physicians to choose individually targeted analgesic approaches without the use of time-consuming EMRs. The challenges ahead will be how we validate these approaches for all populations of patients in diverse populations to ensure equity and trust to reduce the risk for chronic opioid use long term across the world.

References

• CDC. National Health Statistics Reports. 2017;102:1-15.
• Jones GH, et al. Cancer. 2018;doi:10.1002/cncr.31713.