Opioids are routinely prescribed to patients after a surgery—whether they use them or not. A new study, led by LDI Senior Fellow Anish Agarwal will help tailor prescribing so patients get the right amount of opioid medications; not too little and not too much. 

Unused opioids are a particular problem. Opioid misuse can start when extra tablets end up with a friend, or relative, or in the community. Agarwal and colleagues are working on individualizing decision-making about post-surgery opioids so they more closely match patients’ pain-management needs.

About five years ago, Agarwal and colleagues created a system that automatically sends text messages to Penn Medicine patients for three weeks after surgery, asking about their pain level and opioid use. The system prospectively collects patient-reported data that Agarwal and his team investigate to inform clinical care. In a 2021 study, they found that 61% of post-surgery opioid tablets were not used.  

The main function of the patient text-messaging system, however, is to give surgeons feedback on their patients so they can give them the right amount of medication to manage their pain. Once the surgeons saw the results from the text-messaging system, though, they had another request. Agarwal explains more in this Q&A.

What did surgeons request from the messaging system?

Agarwal: We gave surgeons information about their patients’ pain and opioid use to guide them in managing existing patients. They said, “This is great work, but this could also help us before we operate by telling us who might need a little more in their opioid prescription and who might need less.”Surgeons want to be more tailored in their opioid prescribing so patients manage anticipated pain following common surgeries with minimal exposure to opioids. They want to practice good opioid stewardship and “right-size” their approach to their patients. Our team, including LDI Statistical Analyst Ruiying Xiong, Jeffrey Ebert, Evan Spencer, and Senior Fellows Daniel Lee, Zarina Ali, and M. Kit Delgado, used data from the text-messaging system on 3,603 surgical patients. We looked at characteristics such as age, gender, race, chronic conditions, and previous exposure to prescription opioids. We applied different analytic models to find associations between patient-level factors and opioid use.

Which patients tend to use fewer opioids?

Agarwal: Three factors were consistently linked to lower opioid use in all the models. One was 35 to 54 years of age compared to 55 years of age or older. The other factors were no previous use of opioids, and a low patient-reported pain score at the time of discharge. When post-surgical patients leave the hospital, they’re asked to describe their pain level, from 1 (low) to 10 (high). A lower pain score was associated with no, or lower, opioid use.

How did you account for the effects of race on opioid prescribing?

Agarwal: Studies by LDI Fellows and others show that unfortunately, structural racism and biases affect opioid prescribing. Black patients are potentially undertreated, receiving fewer opioids than white patients, who are potentially overtreated. We also know that opioid use is associated with the amount we prescribe–studies have shown that if we prescribe more, patients take more.

Thus, our analyses attempted to adjust and account for these effects. We specifically investigated patient-reported use relative to the amount prescribed and prescriptions within the recommended guidelines. These findings did not show a significant association between race and opioid use. We hope our work contributes to lowering variation in opioid prescribing for everyone, regardless of race.

How could these results change surgical pain management?

Agarwal: Most surgical pain management guidelines recommend a standard number of opioid tablets. We’re working to help move toward a personalized strategy so patients get pain management that works best for them.  The standard question that patients are asked at discharge about their pain level can help surgical teams determine the right number of opioid tablets for that person. In our study, patients reporting lower pain levels tended to use no, or low, amounts of opioids. Younger age was also linked to lower opioid use. And of the patients who hadn’t used opioids before, more than half used zero, or fewer than five, of their prescribed tablets.

Interestingly, all this information is in electronic health records, so wouldn’t it be great if clinicians had a built-in tool that helped them tailor an opioid prescription up or down based on the patient? We envision these types of decision-support and shared decision-making tools to help clinicians get information as they are prescribing opioids for an individual. We also hope our results promote more personalized prescribing guidelines.

The study, “Identifying Patient Characteristics Associated With Opioid Use to Inform Surgical Pain Management,” was published in December 2023 in Annals of Surgery Open. Authors include Anish Agarwal, Ruiying Xiong, Jeffrey Ebert, Fran Shofer, Evan Spencer, Daniel Lee, Zarina Ali, and M. Kit Delgado


More on the Opioid Epidemic