Highlights
-
•
Implementation of a calculator for post-operative opioid prescribing following gynecologic surgery is feasible.
-
•
Reduced opioid prescribing did not worsen post-operative pain control nor impact pain-related admissions or outpatient encounters.
-
•
Calculator platforms outside of the electronic health record may limit end-user usability.
Keywords: Post-operative pain management, Opioid use disorder, Quality improvement, Predictive modeling
Abstract
Objectives
The Gynecologic Oncology Postoperative Opioid use Predictive (GO-POP) calculator is a validated tool to provide evidence-based guidance on post-operative opioid prescribing. The objective of this study was to evaluate the impact of the implementation of GO-POP within an academic Gynecologic Oncology division.
Methods
Two cohorts of patients (pre-implementation and post-implementation) who underwent surgery were compared with reference to GO-POP calculator implementation. All patients were included in the post-implementation group, regardless of GO-POP calculator use. An additional expanded-implementation cohort was used to compare pain control between GO-POP users and non-GO-POP users prospectively. Wilcoxon rank sum tests or ANOVA for continuous variables and Chi-square or Fisher’s exact tests were used to categorical variables.
Results
The median number of pills prescribed post-operatively decreased from 15 pills (Q1: 10, Q3: 20) to 10 pills (Q1: 8, Q3: 14.8) after implementation (p < 0.001). In the expanded-implementation cohort (293 patients), 41% patients were prescribed opioids using the GO-POP calculator. An overall median of 10 pills were prescribed with no difference by GO-POP calculator use (p = 0.26). Within the expanded-implementation cohort, refill requests (5% vs 9.2%; p = 0.26), clinician visits (0.8% vs 0.6%, p = 1), ED or urgent care visits (0% vs 2.3%, p = 0.15) and readmissions (0% vs 1.7%, p = 0.27) for pain did not differ between those prescribed opioids with and without the GO-POP calculator.
Conclusions
A 33% reduction in post-operative opioid pills prescribed was seen following implementation of the GO-POP calculator into the Gynecologic Oncology division without increasing post-operative pain metrics or encounters for refill requests.
1. Introduction
Prescription of opioids following surgery is often central to pain management in combination with acetaminophen and non-steroidal anti-inflammatory drugs. It is well documented, however, that misuse and diversion of prescription opioids postoperatively contributes to the epidemic of opioid use disorder (Kennedy-Hendricks et al., 2016). Approximately 6–8% of opioid-naïve patients develop chronic use following surgery (Lee et al., 2017), and this risk increases with the duration of therapy (Scully et al., 2018, Shah et al., 2017). During the COVID-19 pandemic, this problem has only grown, and deaths attributed to opioids increased from 53,099 in the 12-month period ending in March of 2020 to 80,772 in the 12-month period ending in March of 2022 (Drug Overdose Deaths, 2022, Products - Vital Statistics Rapid Release - Provisional Drug Overdose Data, 2022). While most physicians are aware that over-prescription of pain medication contributes to the epidemic of opioid use disorder, surgeons, including gynecologists, often over prescribe postoperatively (Lamvu et al., 2018, As-Sanie et al., 2017). Many factors contribute to over-prescription, including a desire to provide adequate pain control, attempts to avoid refill requests and a lack of prescribing guidelines and/or formal training on opioid management.(Chiu et al., 2018) Therefore, the post-operative period serves as a critical opportunity to limit opioid exposure.
The Gynecologic Oncology Postoperative Opioid use Predictive (GO-POP) calculator was created to address this opportunity (Davidson et al., 2021). This predictive model was developed at Duke University from 2018 and 2019 and generates an individualized estimate of predicted outpatient opioid use following gynecologic surgical procedures. The 7 predictors are: pre-operative patient-reported anxiety, anticipated opioid need, age, operating time, pregabalin use, smoking status, and education level. Following internal validation, a quality improvement initiative was initiated within the Gynecologic Oncology division to implement the GO-POP calculator into clinical practice.
The aim of this study was to evaluate the implementation of the GO-POP calculator within our academic institution. The primary outcome was the effect of GO-POP implementation on the number of opioid pills prescribed. Secondary outcomes included the effect of GO-POP implementation on post-operative pain control, as measured by surrogate markers, such as refill requests and reengagement with the healthcare system, as well as direct measures from validated patient surveys.
2. Materials and methods
2.1. Application development and implementation
Our institution’s internal review board approved this study as exempt (Pro00109585 and Pro00087827) prior to any study activities. The university’s Algorithm-Based Clinical Decision Support (ABCDS) committee participated in the approval process prior to implementation of GO-POP into clinical practice.
The GO-POP model uses seven predictive variables of post-operative opioid needs: patient’s pre-operative anticipated need for post-operative pain medication [below average, average or above average], patient’s pre-operative anxiety regarding surgery [score 0–100], age, total operative time [incision to closure], pre-operative pregabalin administration, highest education attainment, and smoking history (Davidson et al., 2021). Electronic health record templates were created to facilitate pre-operative documentation of patient-reported metrics as well as post-operative use of the GO-POP calculator. Feedback was solicited to ensure template functionality and ease of use.
A Tableau (Seattle, Washington) mobile application was then developed in partnership with the Duke Institute for Health Innovation (DIHI), which serves as a facilitative platform for technology-enabled healthcare innovation. The application extracts relevant and available data from the electronic health record (EHR) and generates a single recommended number of opioid pills as well as the probability of needing 5, 10, and 15 pills following hospital discharge. The clinician can then use the recommendation to inform their prescribing practice on an individual basis and adjust based on their clinical judgement.
Prior to release of the templates and GO-POP calculator, education was provided to nursing staff, Obstetrics and Gynecology trainees, advanced practice providers (APPs), and attending Gynecologic Oncology physicians on the templates, mobile application use, and implementation into routine post-operative care.
2.2. Patient selection and follow-up
Patients 18 years of age undergoing surgery by one of eight Gynecologic Oncology faculty at a single academic institution were included in the prospective GO-POP implementation and evaluation from November 1, 2021 to March 31, 2022. This study included patients with both benign and malignant indications for surgery as well as those with pre-operative opioid use, which is consistent with the historic cohorts. This study excluded patients who did not have an outpatient pre-operative visit, were not discharged by the Gynecology Oncology service, or died during the follow up time period (Fig. 1).
Fig. 1.
CONSORT diagram of eligible patients for GO-POP use and post-operative follow-up.
Demographic and clinical data were abstracted from the EHR and managed using REDCap electronic data capture tools hosted at our institution (Harris et al., 2009). Documentation of pre-operative patient-reported responses regarding anxiety and expected opioid used was recorded in the pre-operative history & physical. Utilization of the GO-POP calculator in post-operative opioid prescribing was documented in the discharge summary template. Number and type of opioid pills prescribed were collected retrospectively.
In line with the historic cohorts, patients were attempted to be contacted, based on provider and patient availability, post-operatively on a weekly basis for up to 6 weeks or until they reported no longer using opioids to control post-operative pain or, for chronic opioid users, when they returned to their pre-operative baseline use. During these calls, satisfaction with pain control was assessed using the validated Patient-Reported Outcomes Measurement Information System (PROMIS) Short Form v1.1- Pain Interference 4a survey (Chen et al., 2018). Responses to the PROMIS survey questions were compared using a t-score distribution where 50 is the mean of the reference population and 10 is the standard deviation. Patients were also asked to report their average pain score (scale of 1–10: 1 being least severe and 10 being most severe) over the last seven days, as well as their pain score at the time of survey contact. Utilization of post-operative opioids and over-the-counter pain medications was recorded. Patients were not contacted if they were still admitted to the hospital; a subsequent call was placed at the next week interval. Patients who were readmitted during the follow-up period were excluded from the follow-up phone call analysis, as readmission would introduce additional analgesic regimens, as well as further medical or surgical management that could confound assessment of post-operative pain following initial surgery.
Opioid refill requests, number of additional opioid pills prescribed, and number of hospital or clinical encounters for pain management were also recorded. Upon completion of post-operative follow up, the total number of opioid pills used was documented.
2.3. Pre- and post-implementation comparison
To evaluate the effect of GO-POP implementation on opioid prescription patterns and pain control, patients who underwent laparoscopic, robotic, or open surgery from November 2021 to March 2022 (Post-I) with a Gynecologic Oncologist were compared to those undergoing the same surgical approaches from March 2018 to June 2019 (Pre-I). This Pre-I cohort was prospectively accrued during the development and internal validation portion of the study.
2.4. Expanded-implementation comparison
An expanded post-implementation cohort (Expand-I) that included all gynecologic surgeries (Expand-I; including vulvar, vaginal and hysteroscopic approaches which were excluded from the pre-implementation portion of the study) was prospectively collected from November 2021-March 2022 regardless of whether or not the GO-POP calculator was utilized to guide post-operative opioid prescribing. This Expand-I cohort was utilized to compare prescribing outcomes and pain management between those prescribed opioids with and without the GO-POP calculator. Use of the GO-POP calculator at discharge was influenced by provider preference, knowledge of how to utilize the mobile application and availability of the patient input variables.
2.5. Statistical analysis
Descriptive statistics were used to assess baseline patient and clinical demographics. Post-operative pain was assessed via surrogate metrics of refill requests, provider visits, emergency department (ED) or urgent care visits, and hospital readmissions for pain. Wilcoxon rank sum tests or ANOVA for continuous variables and Chi-square or Fisher’s exact tests were used to categorical variables. A p-value <0.05 was considered significant. Analyses were conducted in R 4.1.0 (R Core Team 2021) (R: The R Project for Statistical Computing, 2022).
3. Results
3.1. Cohort demographics
A total of 426 patients were identified in the Pre-I cohort (Table 1). The median patient age was 57.6 years (Q1, Q3: 46, 66.6). Chronic pre-operative opioid use was noted in a small proportion (6.9%) of patients. Most patients self-identified as White (70.7%) and non-Hispanic (97.9%). The majority (53.1%) had an ECOG performance status of 1, indicating a relatively high level of functioning pre-operatively. This cohort was compared to 242 women in the Post-I cohort. Patients in the Post-I group had a different racial (p < 0.001) and ethnic (p < 0.001) breakdown compared to the Pre-I cohort, with fewer White (64.0% vs 70.7%), more Hispanic (7.0% vs 1.6%) and more patients selecting Other (4.1% vs 0.5%) or Prefer not to answer (5.4% vs 0.2%) for their racial and ethnic identification. Patients in the Post-I period were also more likely to have pre-operative opioid use (13.6% vs. 6.9%; p = 0.005) and an ECOG score of 0 (66.1% vs. 40.6%).
Table 1.
Patient demographics of pre and post-implementation periods.
| Characteristics | Pre-implementation n (%) | Post-implementation n (%) | Total n (%) | P-value |
|---|---|---|---|---|
| 426 | 242 | 668 | ||
| Age at surgery (years) | 0.11 | |||
| Median (IQR) | 57.6 [46, 66.6] | 60 [48, 66.9] | ||
| Race | <0.001 | |||
| American Indian or Alaska Native | 6 (1.4) | 3 (1.2) | 9 (1.3) | |
| Asian | 12 (2.8) | 8 (3.3) | 20 (3.0) | |
| Black or African American | 99 (23.2) | 52 (21.5) | 151 (22.6) | |
| Native Hawaiian or Other Pacific Islander | 0 (0.0) | 0 (0.0) | 0 (0.0) | |
| White | 301 (70.7) | 155 (64.0) | 456 (68.3) | |
| More than one race | 3 (0.7) | 1 (0.4) | 4 (0.6) | |
| Other | 2 (0.5) | 10 (4.1) | 12 (1.8) | |
| Prefer not to answer | 1 (0.2) | 13 (5.4) | 14 (2.1) | |
| Missing | 2 (0.5) | 0 (0.0) | 2 (0.3) | |
| Hispanic | <0.001 | |||
| Yes | 7 (1.6) | 17 (7.0) | 24 (3.6) | |
| No | 417 (97.9) | 213 (88.0) | 630 (94.3) | |
| Not reported | 2 (0.5) | 12 (5.0) | 14 (2.1) | |
| Pre-operative opioid use | 0.005 | |||
| Yes | 29 (6.9) | 33 (13.6) | 62 (9.3) | |
| No | 393 (93.1) | 209 (86.4) | 602 (90.7) | |
| Missing | 4 | – | 4 | |
| ECOG score at pre-op | <0.001 | |||
| 0 | 173 (40.6) | 160 (66.1) | 333 (49.9) | |
| 1 | 226 (53.1) | 69 (28.5) | 295 (44.2) | |
| 2 | 7 (1.6) | 11 (4.5) | 18 (2.7) | |
| 3 | 0 (0.0) | 1 (0.4) | 1 (0.1) | |
| Unknown | 5 (1.2) | 1 (0.4) | 6 (0.9) | |
| Missing | 15 (3.5) | 0 (0.0) | 15 (2.2) | |
Abbreviations used: ECOG: Eastern Cooperative Oncology Group; IQR: interquartile range; SD: standard deviation.
A total of 293 surgeries were performed in 283 unique patients during the Expand-I period. Patient and surgical characteristics are shown in Tables 2 and 3, respectively. The median patient age was 59.9 years (Q1, Q3: 24.4, 96.2). Chronic pre-operative opioid use was noted in 13.7% of patients. Most patients were White (63.5%) and non-Hispanic (87.0%). The majority (65.2%) had an ECOG performance status of 0. Just under half (45.7%) of the patients had a pre-operative diagnosis of malignancy. Minimally invasive approaches (laparoscopic, robotic, and minimally invasive) were used most frequently (60.4%) followed by open surgery (22.9%). Most patients (90.1%) were prescribed an opioid post-operatively, with oxycodone as the most frequently prescribed opioid (90.9%).
Table 2.
Patient demographics of expanded implementation period.
| Characteristics | Go-POP not used n (%) | GO-POP used n (%) | Total n (%) |
|---|---|---|---|
| 173 | 120 | 293 | |
| Age at surgery (years) | |||
| Median (IQR) | 60.1 [49.2, 68.2] | 59.7 [46.6, 67.1] | 59.9 [47.9, 67.5] |
| Race | |||
| American Indian or Alaska Native | 4 (2.3) | 1 (0.8) | 5 (1.7) |
| Asian | 4 (2.3) | 4 (3.3) | 8 (2.7) |
| Black or African American | 36 (20.8) | 26 (21.7) | 62 (21.2) |
| Native Hawaiian or Other Pacific Islander | 0 (0.0) | 0 (0.0) | 0 (0.0) |
| White | 108 (62.4) | 78 (65.0) | 186 (63.5) |
| More than one race | 1 (0.6) | 0 (0.0) | 1 (0.3) |
| Other | 10 (5.8) | 3 (2.5) | 13 (4.4) |
| Prefer not to answer | 10 (5.8) | 8 (6.7) | 18 (6.1) |
| Hispanic | |||
| Yes | 16 (9.2) | 7 (5.8) | 23 (7.8) |
| No | 147 (85.0) | 108 (90.0) | 255 (87.0) |
| Not reported | 10 (5.8) | 5 (4.2) | 15 (5.1) |
| Pre-operative opioid use | |||
| Yes | 30 (17.3) | 10 (8.3) | 40 (13.7) |
| No | 143 (82.7) | 110 (91.7) | 253 (86.3) |
| ECOG score at pre-op | |||
| 0 | 111 (64.2) | 80 (66.7) | 191 (65.2) |
| 1 | 46 (26.6) | 33 (27.5) | 79 (27.0) |
| 2 | 14 (8.1) | 6 (5.0) | 20 (6.8) |
| 3 | 2 (1.2) | 0 (0.0) | 2 (0.7) |
| Unknown | 0 (0.0) | 1 (0.8) | 1 (0.3) |
Abbreviations used: ECOG: Eastern Cooperative Oncology Group; IQR: interquartile range; SD: standard deviation.
Table 3.
Surgical demographics of expanded implementation period.
| Characteristics | Go-POP not usedn (%) | GO-POP usedn (%) | Totaln (%) |
|---|---|---|---|
| 173 | 120 | 293 | |
| Pre-operative diagnosis of malignancy | |||
| Yes | 79 (45.7) | 55 (45.8) | 134 (45.7) |
| No | 94 (54.3) | 65 (54.2) | 159 (54.3) |
| Surgical approach | |||
| Laparoscopic | 70 (40.5) | 71 (59.2) | 141 (48.1) |
| Robotic | 20 (11.6) | 14 (11.7) | 34 (11.6) |
| Open | 30 (17.3) | 15 (12.5) | 45 (15.4) |
| MIS converted to Open | 12 (6.9) | 10 (8.3) | 22 (7.5) |
| Hysteroscopy | 16 (9.2) | 0 (0.0) | 16 (5.5) |
| Vaginal | 5 (2.9) | 2 (1.7) | 7 (2.4) |
| Vulvar | 11 (6.4) | 2 (1.7) | 13 (4.4) |
| Cervical | 6 (3.5) | 5 (4.2) | 11 (3.8) |
| Inguinal incision | 1 (0.6) | 1 (0.8) | 2 (0.7) |
| Minimally invasive | 2 (1.2) | 0 (0.0) | 2 (0.7) |
| Days until Discharge | |||
| Median (IQR) | 1 [0, 2] | 0.5 [0, 1] | 1 [0, 2] |
Abbreviations used: IQR: interquartile range; MIS: minimally invasive surgery; TAP: Transversus abdominus plane.
3.2. Pre- and post-implementation comparison
There was a reduction in median number of opioids prescribed from 15 (Q1: 10, Q3: 20) to 10 (Q1: 8, Q3: 14.8) pills (p < 0.001) (Table 4) following GO-POP implementation (Pre-I vs Post-I). Despite the decrease in the number of opioid pills prescribed, surrogate markers of post-operative pain control did not significantly differ between the two groups. For example, rates of requesting at least one opioid refill (Pre-I 6.6% vs. Post-I 8.3%; p = 0.44), needing at least one pain-related post-op appointment (Pre-I 0.7% vs Post-I 0.8%, p = 1.00), having an ED or urgent care encounter (Pre-I 2.1% vs Post-I 1.7%, p = 0.78) or requiring admission (Pre-I 0.2% vs Post-I 1.7%, p = 0.14) were similar across the two cohorts.
Table 4.
Pre-implementation vs post-implementation group.
| Characteristics | Pre-I n (%) | Post-I n (%) | Total n (%) | P-value |
|---|---|---|---|---|
| 426 | 242 | 668 | ||
| Number of pills prescribed | <0.001 | |||
| Median (IQR) | 15(Davidson et al., 2021, Straubhar et al., 2021) | 10 [8, 14.8] | 15(Davidson et al., 2021, Straubhar et al., 2021) | |
| Refills | 0.44 | |||
| No refill | 381 (89.4) | 222 (91.7) | 603 (90.3) | |
| At least one refill | 28 (6.6) | 20 (8.3) | 48 (7.2) | |
| Missing | 17 (4.0) | – | 17 (2.5) | |
| Provider visits for pain | 1 | |||
| No visit | 418 (98.1) | 240 (99.2) | 658 (98.5) | |
| At least one visit | 3 (0.7) | 2 (0.8) | 5 (0.7) | |
| Missing | 5 (1.2) | – | 5 (0.7) | |
| ED or urgent care visits for pain | 0.78 | |||
| No visit | 412 (96.7) | 238 (98.3) | 650 (97.3) | |
| At least one visit | 9 (2.1) | 4 (1.7) | 13 (1.9) | |
| Missing | 5 (1.2) | – | 5 (0.7) | |
| Readmissions for pain | 0.14 | |||
| No readmission | 420 (98.6) | 238 (98.3) | 658 (98.5) | |
| At least one readmission | 1 (0.2) | 4 (1.7) | 5 (0.7) | |
| Missing | 5 (1.2) | – | 5 (0.7) | |
3.3. Expanded-implementation comparison
Of the 293 surgeries performed during the Expand-I period, GO-POP was used to guide prescribing in 120 (41.0%) of cases. The median number of pills prescribed both with GO-POP and without GO-POP use was 10 (p = 0.26) (Table 5). Overall opioid refill requests were low (7.5%) and did not differ by GO-POP utilization (GO-POP 5% vs. no GO-POP 9.2%; p = 0.26). Similarly, no differences in post-operative outpatient appointments for pain management (GO-POP 0.8% vs. no GO-POP 0.6%; p = 1.00), ED or urgent care visits for pain (GO-POP 0.0% vs. no GO-POP 2.3%; p = 0.15) or pain-related hospital readmission (GO-POP 0.0% vs. no GO-POP 1.7%; p = 0.27) between the two groups.
Table 5.
GO-POP use during expanded implementation period.
| Characteristics | Go-POP not used n (%) | GO-POP used n (%) | Total n (%) | P-value |
|---|---|---|---|---|
| 173 | 120 | 293 | ||
| Number of pills prescribed post-operatively | 0.26 | |||
| Median (IQR) | 10(Products - Vital Statistics Rapid Release - Provisional Drug Overdose Data, 2022, Kaafarani et al., 2019) | 10(Products - Vital Statistics Rapid Release - Provisional Drug Overdose Data, 2022, Chen et al., 2018) | 10(Products - Vital Statistics Rapid Release - Provisional Drug Overdose Data, 2022, R: The R Project for Statistical Computing, 2022) | |
| Refills | 0.26 | |||
| No refill | 157 (90.8) | 114 (95) | 271 (92.5) | |
| At least one refill | 16 (9.2) | 6 (5) | 22 (7.5) | |
| Provider visits for pain | 1 | |||
| No postop visit for pain | 172 (99.4) | 119 (99.2) | 291 (99.3) | |
| At least one postop visit for pain | 1 (0.6) | 1 (0.8) | 2 (0.7) | |
| ED or urgent care visits for pain | 0.15 | |||
| No visit for pain | 169 (97.7) | 120 (1 0 0) | 289 (98.6) | |
| At least one visit for pain | 4 (2.3) | 0 (0) | 4 (1.4) | |
| Readmission for pain | 0.27 | |||
| No readmission for pain | 170 (98.3) | 120 (1 0 0) | 290 (99) | |
| At least one readmission for pain | 3 (1.7) | 0 (0) | 3 (1) | |
Of the total 293 surgical procedures included in the Expand-I cohort, 266 (91%) were successfully contacted at least once to complete the post-operative survey during the follow-up time period. The median time from surgery to first successful contact was 7 (Q1: 7, Q3: 8.8) days. Those who were successfully contacted during the follow-up time period were less likely to have pre-operative opioid use (11.7% vs. 33.3%; p = 0.005) and had shorter length of hospital stays (median 0 (Q1: 0, Q3: 1) vs. 1 (Q1: 0, Q3: 6) days; p = 0.01) compared to those who were unable to be contacted. The median time in days to first successful contact did not differ by GO-POP utilization (no GO-POP 7(Lamvu et al., 2018, Chiu et al., 2018) vs GO-POP 7(Lamvu et al., 2018, As-Sanie et al., 2017); p = 0.92).
Only 54.9% of all patients contacted at follow-up had taken a prescription opioid for post-operative pain control (Table 6). This did not differ by GO-POP utilization (GO-POP 54.4% vs no GO-POP 55.3%; p = 0.99). At time of first successful contact, both the median pain score over the past week (4; p = 0.83) and pain score on day of contact (2; p = 0.43) did not differ between groups. Median t-score on the PROMIS pain interference scale did not differ between those prescribed post-operative opioids with and without GO-POP (58.5 (Q1: 53.9, Q3: 62.5) vs 58.5 (Q1: 50.7, Q3: 61.6); p = 0.35). Distribution of responses to the individual questions asked on the PROMIS survey are displayed in Fig. 2.
Table 6.
Follow up expanded implementation cohort.
| Characteristics | Go-POP not used n (%) | GO-POP used n (%) | Total n (%) | P-value |
|---|---|---|---|---|
| 152 | 114 | 266 | ||
| Days from surgery to first successful contact | 0.92 | |||
| Median (IQR) | 7(Lamvu et al., 2018, Chiu et al., 2018) | 7(Lamvu et al., 2018, As-Sanie et al., 2017) | 7 [7, 8.8] | |
| Days from surgery to first successful contact | 0.02 | |||
| 7–13 | 138 (90.8) | 112 (98.2) | 250 (94.0) | |
| 14–27 | 14 (9.2) | 2 (1.8) | 16 (6.0) | |
| 28+ | 0 (0.0) | 0 (0.0) | 0 (0.0) | |
| Days from discharge to first successful contact | 0.84 | |||
| Median (IQR) | 7(Products - Vital Statistics Rapid Release - Provisional Drug Overdose Data, 2022, As-Sanie et al., 2017) | 7(Products - Vital Statistics Rapid Release - Provisional Drug Overdose Data, 2022, As-Sanie et al., 2017) | 7(Products - Vital Statistics Rapid Release - Provisional Drug Overdose Data, 2022, As-Sanie et al., 2017) | |
| Number of completed post-op surveys | 0.67 | |||
| 1 | 134 (88.2) | 100 (87.7) | 234 (88.0) | |
| 2 | 12 (7.9) | 12 (10.5) | 24 (9.0) | |
| 3 | 4 (2.6) | 2 (1.8) | 6 (2.3) | |
| 4 | 2 (1.3) | 0 (0.0) | 2 (0.8) | |
| Use of prescription pain medication post-op | 0.99 | |||
| Yes | 84 (55.3) | 62 (54.4) | 146 (54.9) | |
| No | 68 (44.7) | 52 (45.6) | 120 (45.1) | |
| Average pain score over the last week | 0.83 | |||
| Median (IQR) | 4 (Lee et al., 2017, Products - Vital Statistics Rapid Release - Provisional Drug Overdose Data, 2022) | 4 (Lee et al., 2017, Products - Vital Statistics Rapid Release - Provisional Drug Overdose Data, 2022) | 4 (Lee et al., 2017, Products - Vital Statistics Rapid Release - Provisional Drug Overdose Data, 2022) | |
| N observed | 151 | 113 | 264 | |
| Missing | 1 (0.66) | 1 (0.88) | 2 (0.75) | |
| Pain score at time of contact | 0.43 | |||
| Median (IQR) | 2 (Kennedy-Hendricks et al., 2016, Scully et al., 2018) | 2 (Kennedy-Hendricks et al., 2016, Shah et al., 2017) | 2 (Kennedy-Hendricks et al., 2016, Shah et al., 2017) | |
| N observed | 151 | 113 | 264 | |
| Missing | 1 (0.66) | 1 (0.88) | 2 (0.75) | |
| PROMIS t-score at first successful contact | 0.35 | |||
| Median (IQR) | 58.5 [50.7, 61.6] | 58.5 [53.9, 62.5] | 58.5 [52, 62.5] | |
| N observed | 144 | 104 | 248 | |
| Missing | 8 (5.26) | 10 (8.77) | 18 (6.77) | |
Fig. 2.
Comparison of the distribution of patient response to questions on the PROMIS pain interference scale between those who were prescribed post-operative opioids with and without GO-POP. (A) How much did pain interfere with day-to-day activities? (B) How much did pain interfere with work around the house? (C) How much did pain interfere with ability to participate in social activities? (D) How much did pain interfere with household chores?
4. Conclusions
In the four months following initial implementation of the mobile Tableau GO-POP application, we saw increasing uptake of GO-POP utilization for post operative opioid prescribing. Compared to the pre-implementation cohort, quantity of opioids prescribed decreased by a third without negative impact on pain control as measured by patient-reported metrics, refill requests, and encounters for pain control. This study demonstrates that the implementation of evidence-based patient-centered tools can mitigate excess opioids in the community without compromising pain control or burdening clinicians.
GO-POP utilization within the expanded-implementation cohort, however, did not impact the number of opioids prescribed, suggesting that increased clinician awareness and education of appropriate opioid prescribing post-operatively throughout this initiative also influences prescribing practices. This is consistent with the existing literature that demonstrates reduced opioid prescribing rates following education interventions for both surgeons and patients (Hill et al., 2018, Kaafarani et al., 2019).
Importantly, patient-reported pain and surrogate markers of pain management were not impacted by the decrement in post-op opioid prescribing, likely due to historical over prescription of opioids. This is consistent with existing literature demonstrating that efforts to decrease over prescribing have not led to significant increases in refill requests (Titan et al., 2021, Stanek et al., 2015). In fact, in some studies, more restrictive opioid prescription policies were associated with fewer refill requests (Zsiros et al., 2023). Implementation of a restrictive opioid prescription protocol, which was limited to three days or fewer following surgery, led to a 45% decrease in prescribed post operative opioids without an increase in refill requests (17.9% vs. 20.9%, p = 0.02) (Zsiros et al., 2023).
This work highlights future research directions to decrease excess opioids in our communities. Specifically, more than 45% of patients did not use a single opioid pill following surgery in this study. This is consistent with previous literature reporting 20.4–47.3% of opioids prescribed following a variety of gynecologic surgeries are not utilized (Straubhar et al., 2021, Nwosu et al., 2022, McEntee et al., 2021). Opportunities to identify patients unlikely to need any opioids for post-operative pain management would allow for further reduction of unnecessary opioid prescribing practices.
We are continuing to expand GO-POP utilization post-operatively at our institution and have solicited feedback from end-users to identify workflow processes that would increase use. In particular, development of an electronic health record workflow that would obviate the need for using an external website or mobile application is the most commonly cited need from users and is actively being developed.
There were several limitations to our study. Given the continuous nature of subject enrollment in the historic cohort, differences exist in patient-level demographics between the pre- and post-implementation cohorts. These differences may have introduced confounding variables that impacted opioid prescribing between the two groups. As mentioned previously, clinician awareness of the importance of judicious opioid prescribing in the implementation phases likely impacted opioid prescribing, particularly for patients where GO-POP was not utilized to guide prescribing practices. Failure of the prescribing clinician to document GO-POP use in the discharge summary could have contributed to the relatively low utilization of the calculator. The calculator is also intended to only be a guide in post-operative prescription and we did not compare the recommended number of pills to what was actually prescribed at discharge. Post-operative phone contact remained a challenge through all phases of this study. While rates of contact improved over time, the variability in post-operative follow-up could have impacted the ability to assess post-operative pain control thoroughly and accurately.
In conclusion, this study found that the implementation of the GO-POP calculator into clinical practice is feasible and helped to reduce the number of opioids prescribed following gynecologic surgeries for both benign and malignant indications. The education on opioid prescribing required implement the calculator into clinical practice likely contributed to the decrease in opioids prescribed pre and post-implementation. We are currently planning an expansion of the GO-POP implementation to all gynecology divisions at our institution with hopes of embedding the calculator in the electronic health record to maximize its usability.
Funding
Funding for this study was received from a grant through the Duke Institute for Health Innovations program at Duke University.
CRediT authorship contribution statement
Nicole C. Zanolli: Investigation, Data curation, Writing – original draft. Stephanie Lim: Methodology, Writing – review & editing. William Knechtle: Conceptualization, Software, Visualization. Kelvin Feng: Formal analysis, Writing – review & editing, Visualization. Tracy Truong: Formal analysis, Writing – review & editing, Visualization. Laura J. Havrileskey: Conceptualization, Methodology, Writing – review & editing. Brittany A. Davidson: Conceptualization, Investigation, Supervision, Writing – review & editing, Funding acquisition.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
References
- As-Sanie S., Till S.R., Mowers E.L., et al. Opioid prescribing patterns, patient use, and postoperative pain after hysterectomy for benign indications. Obstet. Gynecol. 2017;130(6):1261–1268. doi: 10.1097/AOG.0000000000002344. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Chen C.X., Kroenke K., Stump T.E., et al. Estimating minimally important differences for the PROMIS pain interference scales: results from 3 randomized clinical trials. Pain. 2018;159(4):775–782. doi: 10.1097/j.pain.0000000000001121. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Chiu A.S., Healy J.M., DeWane M.P., Longo W.E., Yoo P.S. Trainees as agents of change in the opioid epidemic: optimizing the opioid prescription practices of surgical residents. J. Surg. Educ. 2018;75(1):65–71. doi: 10.1016/j.jsurg.2017.06.020. [DOI] [PubMed] [Google Scholar]
- Davidson B., Jelovsek J.E., Rodriguez I., et al. Development and validation of the Gynecologic Oncology Predictor Of Postoperative opioid use (GO-POP) model. Gynecol Oncol. 2021;162:S55. doi: 10.1016/S0090-8258(21)00748-4. [DOI] [Google Scholar]
- Drug Overdose Deaths in the U.S. Top 100,000 Annually. <https://www.cdc.gov/nchs/pressroom/nchs_press_releases/2021/20211117.htm> (Accessed May 23, 2022).
- Harris P.A., Taylor R., Thielke R., Payne J., Gonzalez N., Conde J.G. Research electronic data capture (REDCap)–a metadata-driven methodology and workflow process for providing translational research informatics support. J. Biomed. Inform. 2009;42(2):377–381. doi: 10.1016/j.jbi.2008.08.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hill M.V., Stucke R.S., McMahon M.L., Beeman J.L., Barth R.J. An educational intervention decreases opioid prescribing after general surgical operations. Ann. Surg. 2018;267(3):468–472. doi: 10.1097/SLA.0000000000002198. [DOI] [PubMed] [Google Scholar]
- Kaafarani H.M.A., Eid A.I., Antonelli D.M., et al. Description and impact of a comprehensive multispecialty multidisciplinary intervention to decrease opioid prescribing in surgery. Ann. Surg. 2019;270(3):452–462. doi: 10.1097/SLA.0000000000003462. [DOI] [PubMed] [Google Scholar]
- Kennedy-Hendricks A., Gielen A., McDonald E., McGinty E.E., Shields W., Barry C.L. Medication sharing, storage, and disposal practices for opioid medications among US adults. JAMA Intern. Med. 2016;176(7):1027–1029. doi: 10.1001/jamainternmed.2016.2543. [DOI] [PubMed] [Google Scholar]
- Lamvu G., Feranec J., Blanton E. Perioperative pain management: an update for obstetrician-gynecologists. Am. J. Obstet. Gynecol. 2018;218(2):193–199. doi: 10.1016/j.ajog.2017.06.021. [DOI] [PubMed] [Google Scholar]
- Lee J.-S.-J., Hu H.M., Edelman A.L., et al. New persistent opioid use among patients with cancer after curative-intent surgery. J. Clin. Oncol. 2017;35(36):4042–4049. doi: 10.1200/JCO.2017.74.1363. [DOI] [PMC free article] [PubMed] [Google Scholar]
- McEntee K.M., Crawford K.D., Wilson M.D., et al. Postoperative opioid prescribing and consumption after hysterectomy: a prospective cohort study. J. Min. Invas. Gynecol. 2021;28(5):1013–1021. doi: 10.1016/j.jmig.2020.10.023. [DOI] [PubMed] [Google Scholar]
- Nwosu I., Bougie O., Pudwell J. Evaluation of opioid prescribing patterns and use following hysterectomy for benign indication: a prospective cohort study. J. Obstet. Gynaecol. Can. 2022;44(3):240–246.e1. doi: 10.1016/j.jogc.2021.09.019. [DOI] [PubMed] [Google Scholar]
- Products - Vital Statistics Rapid Release - Provisional Drug Overdose Data. <https://www.cdc.gov/nchs/nvss/vsrr/drug-overdose-data.htm> (Accessed May 23, 2022).
- R: The R Project for Statistical Computing. https://www.r-project.org/. Accessed December 19, 2022.
- Scully R.E., Schoenfeld A.J., Jiang W., et al. Defining optimal length of opioid pain medication prescription after common surgical procedures. JAMA Surg. 2018;153(1):37–43. doi: 10.1001/jamasurg.2017.3132. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Shah A., Hayes C.J., Martin B.C. Characteristics of initial prescription episodes and likelihood of long-term opioid use – United States, 2006–2015. MMWR Morb. Mortal. Wkly. Rep. 2017;66(10):265–269. doi: 10.15585/mmwr.mm6610a1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Stanek J.J., Renslow M.A., Kalliainen L.K. The effect of an educational program on opioid prescription patterns in hand surgery: a quality improvement program. J. Hand Surg. Am. 2015;40(2):341–346. doi: 10.1016/j.jhsa.2014.10.054. [DOI] [PubMed] [Google Scholar]
- Straubhar A.M., Dalton L., Rolston A., et al. Optimization of postoperative opioid prescriptions in gynecologic oncology: Striking a balance between opioid reduction and pain control. Gynecol Oncol. 2021;162(3):756–762. doi: 10.1016/j.ygyno.2021.06.023. [DOI] [PubMed] [Google Scholar]
- Titan A., Doyle A., Pfaff K., et al. Impact of policy-based and institutional interventions on postoperative opioid prescribing practices. Am J Surg. 2021;222(4):766–772. doi: 10.1016/j.amjsurg.2021.02.004. [DOI] [PubMed] [Google Scholar]
- Zsiros E., Ricciuti J., Gallo S., et al. Postoperative restrictive opioid protocols and durable changes in opioid prescribing and chronic opioid use. JAMA Oncol. 2023 doi: 10.1001/jamaoncol.2022.6278. [DOI] [PMC free article] [PubMed] [Google Scholar]