Assessing the 5×-Multiplier Calculation to Reduce Discharge Opioid Prescription Volumes After Inpatient Surgery

Abstract

This case-control study examines the rates of refills under 2 separate strategies for opioid prescribing after operations in a hepatopancreatobiliary surgery department.

State-specific limits on total days and procedure-specific recommendations of discharge opioid volumes have had mixed success in mitigating postoperative opioid dissemination.^1,2 Most prescribers still expose their clinician-specific bias in writing round numbers of opioid doses (eg, 30-50 pills). In the theme of patient-centered care, this study analyzed oncologic surgery discharge opioid prescriptions and 30-day refills when a novel, patient-centered prescription calculation was implemented.

Methods

Based on a US Centers for Disease Control and Prevention report³ showing a salient inflection point in the association of initial opioid prescription volume with long-term opioid dependence at 5 days’ supply, we created a simple 5× multiplier, wherein patients received 5 times the amount of their opioid use in the last 24 inpatient hours (eg, 2 pills × 5 = 10 pills; 0 pills × 5 = 0 pills). Beginning with clinician education sessions in August 2018, this patient-centered calculation was voluntarily and gradually adopted by an increasing proportion of hepatopancreatobiliary (HPB) clinicians.⁴ This was a nonrandomized cohort study of consecutive HPB operations at MD Anderson Cancer Center in Houston, Texas, conducted from September 2018 through June 2019. It was approved by the MD Anderson Cancer Center institutional review board, with a waiver of informed consent because the study activities were part of usual patient care.

Two prescribing patterns emerged: patients in the usual care (UC) group, who received clinician-specific opioid pill volumes, vs patients who received 5× multiplier–compliant volumes. Actual oral morphine equivalents (OMEs) were converted using institution-approved values. Refills were trackable in the electronic medical record. Continuous variables were compared using Mann-Whitney U test, and categorical variables were compared using χ² test or Fisher exact test, as appropriate. The threshold for statistical significance was P < .05, with all tests 2-sided. All tests were performed with SPSS version 22 (IBM).

Results

There were 278 consecutive elective HPB tumor resections by 9 faculty surgeons in 10 months. Operations included 152 liver resections (54.7%) and 126 pancreas resections (45.3%), with 125 patients (45.0%) receiving 5×-multiplier dosing. Other than slight differences in regional nerve block use (72 [47.1%] vs 75 [60.0%]; P = .03), both groups had similar clinical characteristics, including length of hospital stay and use of opioid-weaning nonopioid bundles (Table). Median OMEs consumed in the last 24 inpatient hours were 10 mg for both groups (interquartile ranges [IQRs], UC group: 5-20 mg; 5× multiplier, 0-20 mg). Median (IQR) discharge prescription OMEs were 150 (100-150) mg for UC vs 50 (0-100) mg in the 5×-multiplier group (P < .001). The cumulative volume curve of OMEs disseminated was notably flattened in the 5×-multiplier group (Figure).

Table. Comparison of Patient Characteristics Managed in Usual Care and 5×-Multiplier Pathways.

Characteristic	Patients, No. (%)		P value
	Usual care (n = 153)	5× Multiplier (n = 125)
Age, median (IQR), y	58.3 (48.4-68.6)	59.2 (49.6-68.5)	.62a
BMI	25.8 (22.8-29.6)	26.5 (22.8-29.9)	.64a
Female	74 (48.4)	59 (47.2)	.85
Preoperative opioids	29 (22.6)	12 (18.4)	.03
Liver operation	88 (57.5)	64 (51.2)	.29
Laparoscopic or robotic operation	24 (15.6)	22 (17.6)	.67
Combination operationb	14 (9.2)	19 (15.2)	.12
Regional nerve block	72 (47.1)	75 (60.0)	.03
Epidural	46 (30.1)	27 (21.6)	.11
Major complication	28 (18.3)	15 (12.0)	.15
Readmission	28 (18.3)	18 (14.4)	.38
≥2 Nonopioid medications bundled with opioidsc	122 (79.7)	97 (77.6)	.66
Length of stay, median (IQR), d	4 (3-6)	5 (3-6)	.52a
Refills within 30 d	32 (20.9)	21 (16.8)	.39
OMEs in last 24 h, median (IQR), mg	10 (5-20)	10 (0-20)	.36a
Initial discharge OMEs, median (IQR), mg	150 (100-150)	50 (0-100)	<.001a

Abbreviations: BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); IQR, interquartile range; OME, oral morphine equivalents.

^aMann-Whitney U test.

^bA combination operation consists of a second organ resected or a concomitant ventral hernia repair.

^cNonopioid medications include acetaminophen, nonsteroidal anti-inflammatory drugs, muscle relaxers, and gabapentinoids.

Figure. Cumulative Opioids Prescribed by Usual Care vs 5×-Multiplier Groups.

The curves represent the difference in opioid dissemination to the community that can be avoided by using the 5×-multiplier dosage strategy. The curves would likely separate more as time goes on. OME indicates oral morphine equivalents.

Thirty-day opioid refill rates were similar between groups (UC group, 32 [20.9%] vs 5×-multiplier group, 21 [16.8%]; P = .39). Sixty-nine patients (24.8%) were already weaned to no opioid use in the last 24 hours of hospitalization, but 31 (45%) were still discharged with opioids (UC). Among the 69 patients who had used no opioids in their last 24 inpatient hours, none subsequently received an outpatient opioid prescription. Among the 15 patients who received at least 1 opioid pill but an amount less than the 5×-multiplier dosage, 8 (53%) required a refill.

Discussion

In patients undergoing curative-intent oncologic surgery, a novel, patient-centered 5× multiplier of last 24 hours of inpatient opioid use reduced discharge opioid prescriptions by 67% over UC, with no increase in 30-day refill rates. No patients who used 0 mg of opioids in the 24 hours prior to discharge received refills of opioid pain medications. Patients already weaned down to 0 mg of opioids by the last 24 hours before discharge could be safely assumed to not need outpatient opioids.

Even as more clinicians acknowledge the desire to abandon the antiquated model of clinician-specific volume calculations and try to adopt a procedure-specific model, the discharging clinician still has the ability to supersede any recommended range based on clinical instinct, fear of refill logistics, opioid education bias, and/or perceived empathy for the patient.^1,4,5,6 Writing discharge opioid prescriptions based on each patient’s actual opioid use overcomes clinician bias, is patient centered, and markedly decreases the trajectory of cumulative opioid dissemination to the community (Figure).

Study limitations include the nonrandomized design and the voluntary rollout of the 5× multiplier by advanced practice clinicians and fellows in the study period. Despite these study limitations, the major strength of the 5× multiplier is that it does not require any special online or smartphone calculator or pain team consultation. It complies with all current state laws and US Centers for Disease Control and Prevention guidelines. Future prospective studies expanding to other types of surgery and comparing the patient-centered 5× multiplier with procedure-specific and state-specific guidelines are needed.