The Impact of Opioid-Related Disorder on Healthcare Resource Utilization Following Distal Radius Fracture Repair Surgery

Abstract

Purpose

This study assesses the impact of opioid-related disorders (ORDs) on healthcare utilization in patients following distal radius fractures (DRF) repair surgery—one of the most common fracture surgeries performed. The study hypothesis was that patients with ORD would experience higher rates of emergency department (ED) admissions, hospitalizations, and physical therapy (PT) evaluations compared to patients without ORD (NORD).

Methods

The TriNetX US Collaborative Database was queried to identify patients who underwent DRF repair between the ages of 18 and 49 years. The ORD cohort was propensity matched in a 1:1 ratio to NORD patients based on age, sex, race, ethnicity, and other comorbidities to limit confounder influence of results. Postoperative rates of healthcare resource utilization outcomes, including hospital readmissions, ED visits, and PT utilization, were compared within 90 days of surgery and prescriptions were compared between 30 days and 1 year postoperatively.

Results

A total of 789 patients were included in each cohort after propensity matching. Within 90 days postoperatively, a significantly greater percentage of ORD patients were admitted to the ED (21.8% vs 11.2% of NORD; P < .0001), hospitalized (13.6% vs 8.5% of NORD; P = .0013), and evaluated by PT (18.8% vs 12.0% of NORD; P = .0002). There was no significant difference in outpatient visits between groups (P = .9534). Within 30 days and 1 year postoperatively, patients in the ORD group received significantly more opioid prescriptions per patient (8.3 ± 17.4) compared to NORD patients (4.7 ± 8.8) (P = .002).

Conclusions

ORD patients received significantly more opioid prescriptions and were more likely to be admitted to the ED, be hospitalized, or be evaluated by PT after DRF repair compared to those without ORD. Surgeons and healthcare teams should prepare for greater support of patients with ORD and consider deliberate strategies to minimize perioperative complications, while optimizing opioid stewardship.

Level of evidence

Level 3 (retrospective cohort study).

Opioid analgesics have long been a cornerstone of postoperative pain management across surgical specialties, including orthopedics and hand surgery. In the United States, national opioid prescribing rates have declined in recent years following recognition of the opioid crisis as a public health emergency by the Centers for Disease Control and the Department of Health and Human Services.¹ Despite this encouraging trend, orthopedic surgeons continue to account for a substantial proportion of opioid prescribing. Recent data demonstrate a 38.7% reduction in opioid dispensing rates among orthopedic surgeons between January 2016 and December 2022; however, patients continue to receive opioid prescriptions in quantities that often exceed their actual postoperative needs.²

As the risks associated with opioid use have become more clearly defined, prescribing practices and surgeon attitudes are evolving. Survey data among hand surgeons in the United States suggest increased selectivity in postoperative opioid prescribing, with 75% reporting comfort in refilling opioid prescriptions following fracture surgery compared to only 13% after minor elective procedures.³ However, there remains a need for more comprehensive evaluation of patient-specific risk factors for opioid misuse. Notably, only 40% of surveyed hand surgeons reported consistently assessing a history of opioid abuse or dependence before scheduling surgery, highlighting a critical gap in preoperative risk stratification and an opportunity for more standardized screening practices.³

Distal radius fractures (DRF) account for about 25% of pediatric fractures and 18% of fractures in the elderly population, making them one of the most common orthopedic injuries.⁴ Given their high incidence, optimizing perioperative care and understanding factors that influence postoperative outcomes is of value. This study aims to evaluate the impact of a prior diagnosis of opioid-related disorder (ORD) on postoperative healthcare resource utilization as well as quantify the rates of opioid and antidepressant prescriptions after DRF repair. This investigation hypothesizes that patients with a history of ORD will be associated with increased healthcare utilization, opioid prescriptions, and antidepressant prescriptions in the postoperative period.

Materials and Methods

Patients

The TriNetX US Collaborative Database was retrospectively queried using International Classification of Diseases, 10th Revision (ICD-10), and Current Procedural Terminology (CPT) codes to identify patients. TriNetX is a global research network encompassing data from over 170 healthcare organizations and more than 400 million patients.⁵ The TriNetX database does not involve patient-identifiable information and is therefore exempt from Institutional Review Board review and approval. TriNetX data is derived from electronic medical records of participating institutions and includes demographics, diagnoses, procedures, medications, and outcomes. Patients who underwent DRF repair (CPT codes: 25607, 25608, 25609) between the ages of 18 and 49 years were eligible. Pathologic and neoplastic fractures were excluded from analysis. Patients were then divided into two groups: ORD and no opioid-related disorder (NORD). ORD was classified as any patient with a diagnosis of an OPD (ICD-10 F11) before the date of the index event: DRF repair. NORD patients did not have a prior diagnosis of OPD, and ORD was listed as an exclusion criterion in this cohort. The study period extended from January 1, 2015, to January 5, 2026 (date of database query). A complete list of codes used to identify patients can be found in Appendix A (available on the Journal’s website at www.jhsgo.org).

Methods of assessment

Patient outcomes were analyzed within 90 days and 1 year postoperatively. Postoperative rates of healthcare resource utilization were compared within 90 days of DRF repair. Healthcare resource utilization included emergency department (ED) visits, hospitalizations, outpatient visits, and physical therapy evaluations. ED visits and physical therapy evaluations were identified using CPT codes for those services. Inpatient hospitalizations were identified as new admissions beginning one day after surgery or patients being admitted and discharged on the same day beginning 1 day after surgery. Outpatient visits included established patients or postoperative visits included in the patient’s surgical package. Opioid and antidepressant prescriptions were identified using the Veterans Administration drug classification system. These prescriptions were recorded from 30 days to 1 year postoperatively in order to accurately capture prolonged opioid prescribing outside of initial postoperative pain management. ICD-10, CPT, and Veterans Administration codes used for patient and outcome identification can also be found in Appendix A.

Statistical analysis

All statistical analyses were performed using TriNetX Analytics. A 1:1 nearest-neighbor matching algorithm was employed for propensity score matching of cohorts to reduce selection bias and balance comorbidities between groups. The ORD cohort was matched to the NORD cohort based on age, sex, race, ethnicity, alcohol-related disorders, major depression, anxiety disorders, antidepressant prescriptions, diabetes, heart disease, hypertension, and metabolic disorders. These comorbidities were identified based on known risk factors for opioid-use disorder, as well as postoperative complications and poor bone healing.6, 7, 8, 9, 10, 11, 12, 13 Descriptive statistics were collected before and after propensity matching was performed.

Outcomes were analyzed based on the percentage of patients with the outcome (rate). Odds ratios with 95% confidence intervals were calculated, and t tests were performed for continuous variables. Statistical significance was set to P < .05. Risk ratios (RRs) with 95% confidence intervals were calculated using the TriNetX analytics platform measure of association analysis. This analysis compares the proportion of patients experiencing a given outcome between cohorts. Z-tests were used to assess differences between proportions, and corresponding P-values and RRs with 95% confidence intervals were generated. Statistical significance was defined as a two-tailed P < .05. Risk ratios were considered statistically significant when the corresponding 95% confidence interval did not include 1.00.

In all analyses, ORD was used as the reference group. Accordingly, RRs greater than 1 indicate a higher risk associated with ORD relative to NORD, whereas RRs less than 1 indicate a higher risk associated with NORD.

Results

Patient demographics

Using the inclusion and exclusion criteria to query the TriNetX database, 21,923 patients were identified, of which 792 were identified as ORD patients and 21,131 were identified as NORD patients. Before matching, patients in the ORD cohort were significantly older than the NORD cohort (35.1 ± 6.7 years and 31.4 ± 8.0 years, respectively). There were also significant differences between the cohorts in regard to comorbidities before propensity matching. After matching, the two cohorts were not significantly different based on the selected matching criteria (P < .05, Table 1).

Table 1.

Patient Demographics of ORD and NORD Cohorts Before and After Propensity Matching^∗

Comorbidity	ICD-10 Code	Before Propensity Matching			After Propensity Matching
		ORD	NORD	P	ORD	NORD	P
Number of patients		792	21,131		789	789
Age at index, y		35.1 ± 6.66	31.4 ± 7.96	<.0001	35.1 ± 6.67	35.2 ± 7.17	.7741
Female sex		342 (43.2)	9,024 (42.7)	.7900	342 (43.3)	345 (43.7)	.8789
Anxiety disorders	F40-F48	435 (54.9)	3,478 (16.5)	<.0001	432 (54.8)	443 (56.1)	.5774
Alcohol-related disorders	F10	246 (31.1)	1,393 (6.6)	<.0001	243 (30.8)	223 (28.3)	.2697
Major depressive disorder	F33	114 (14.4)	637 (3.0)	<.0001	113 (14.3)	98 (12.4)	.2672
Diabetes	E08-E13	46 (5.8)	633 (3.0)	<.0001	44 (5.6)	41 (5.2)	.7380
Metabolic disorders	E70-E88	303 (38.3)	3,205 (15.2)	<.0001	300 (38.0)	306 (38.8)	.7561
Hypertension	I10	169 (21.3)	1,845 (8.7)	<.0001	167 (21.2)	160 (20.3)	.6637
Heart disease	I30-I5A	196 (24.7)	1,323 (6.3)	<.0001	193 (24.5)	186 (23.6)	.6800
Medications	VA Code
Antidepressant prescription	CN600	377 (47.6)	3,387 (16.0)	<.0001	374 (47.4)	378 (47.0)	.8402

ICD-10, International Classification of Diseases, 10th Revision; ORD, opioid-related disorder; NORD, no opioid-related disorders; VA, Veteran’s Administration Drug Classification System.

^∗Values are presented as n (%) or mean ± SD. Bold indicates statistical significance (P < .05).

Postoperative healthcare resource utilization

Healthcare resource utilization between the cohorts was analyzed within 90 days postoperatively (Table 2). Patients with ORD were significantly more likely than NORD patients to visit the ED (21.9% vs 11.2%), be hospitalized (13.6% vs 8.5%) and undergo an evaluation by physical therapy (18.8% vs 12.0%). Outpatient visit rates did not differ significantly between cohorts. Utilization was also analyzed on a per-patient basis by examining the number of visits for each service type. ORD patients had a significantly higher number of ED visits per patient than NORD patients. There were no significant differences between cohorts in the number of hospitalizations, physical therapy evaluations, or outpatient visits per patient.

Table 2.

Ninety-Day Postoperative Healthcare Resource Utilization of ORD and NORD Patients Undergoing Distal Radius Fracture Repair After Propensity Matching^∗

Characteristic	ORD	NORD	Odds Ratio	95% CI	P
ED visits
Rate, n (%)	172 (21.8)	88 (11.2)	2.221	1.680−2.934	<.0001
Visits per patient	1.95 ± 1.7	1.42 ± 1.0			.006
Inpatient hospitalizations
Rate, n (%)	107 (13.6)	67 (8.5)	1.691	1.224−2.335	.0013
Admissions per patient	1.58 ± 1.2	1.45 ± 0.9			.4232
Outpatient appointments
Rate, n (%)	594 (75.3)	595 (75.4)	0.993	0.79−1.249	.9534
Visits per patient	3.21 ± 2.3	3.39 ± 2.7			.2217
Physical therapy evaluations
Rate, n (%)	182 (18.8)	95 (12.0)	1.687	1.276−2.23	.0002
Visits per patient	1.29 ± 0.6	1.25 ± 0.8			.6733

ED, emergency department; ORDs, opioid-related disorders; NORD, no opioid-related disorder.

^∗Values are presented as n (%) or mean ± SD. Bold indicates statistical significance (P < .05).

Postoperative prescriptions

Within 30 days and 1 year postoperatively, patients in the ORD group received significantly more opioid prescriptions per patient (8.3 ± 17.4) compared to NORD patients (4.7 ± 8.8) (P = .002). During the same follow-up interval, there was no significant difference in prescription rates for antidepressants between ORD (21.9%) and NORD patients (18.1%). Additionally, there was no significant difference in the number of antidepressant prescriptions per patient between groups. Table 3 presents all results related to prescriptions made for each cohort within 1 year of DRF repair.

Table 3.

Medication Prescriptions for ORD and NORD Patients Undergoing Distal Radius Fracture Repair After Propensity Matching, Prescribed Between 30 days and 1 Year Postoperatively^∗

Characteristic	ORD	NORD	Odds Ratio	95% CI	P
Opioid analgesics
Rate, n (%)	358 (45.4)	261 (33.1)	1.680	1.370−2.061	<.0001
Prescriptions per patient	8.3 ± 17.4	4.7 ± 8.8			.002
Antidepressants
Rate, n (%)	173 (21.9)	143 (18.1)	1.269	0.991−1.625	.059
Prescriptions per patient	7.3 ± 18.8	4.5 ± 7.3			.090

ORD, opioid-related disorder; NORD, no opioid-related disorder.

^∗Values are presented as n (%) or mean ± SD. Bold indicates statistical significance (P < .05).

Discussion

This retrospective database study was conducted to analyze healthcare utilization and postoperative opioid and antidepressant prescription rates in patients with ORD who underwent DRF repair surgery. The study hypothesis was upheld, finding that patients with ORD had a higher risk of increased healthcare utilization as well as increased opioid prescriptions. Patients with a prior diagnosis of ORD were more likely to present to the ED, require hospitalization or undergo physical therapy evaluation within 90 days following DRF repair. However, there was no significant difference in outpatient postoperative visits between groups. Although ORD patients demonstrated a higher frequency of ED visits on a per-patient basis, there were no significant differences between cohorts in the number of hospital admissions, physical therapy evaluations, or outpatient visits per patient.

While the underlying mechanisms remain incompletely understood, prior literature has demonstrated an association between increased opioid use and worse postoperative outcomes across orthopedic surgeries, including DRF repair. In a retrospective review, Runge et al⁹ found that the presence of surgical complications, ED utilization, and hospital readmission were significantly increased in patients with prolonged (>6 months postoperatively) opioid use after DRF repair. A dose-dependent risk relationship between preoperative opioid use and postoperative complications—including impaired wound healing, increased length of stay and higher revision rates—has been demonstrated in arthroplasty populations.¹⁴ Consistent with these findings, the present study demonstrates that patients with a history of ORD are more likely to require ED evaluation and hospitalization following DRF.

Within 30 days and 1 year postoperatively, ORD patients received more opioid prescriptions compared to patients without ORDs. The present study employed a 30-day washout period after the index event when assessing opioid prescriptions after DRF. This was done in an effort to exclude routine postoperative opioid prescriptions following DRF fixation. Survey data collected by Bhashyam et. al indicated that almost 1 in 5 orthopedic surgery residents would prescribe more than 7 days of opioids after DRF repair despite state laws limiting the duration of prescriptions.¹⁵ A cohort study conducted by Scully et al¹⁶ found that the optimal length of opioid prescription after musculoskeletal procedures lies between 6 and 15 days.¹⁶ This washout interval of 30 days was determined by the studies listed and the use of 30 days insured exclusion of longer postoperative prescriptions to highlight solely prolonged opioid prescribing.

The patient groups were propensity matched for major depressive disorder, anxiety disorders, and prior antidepressant prescriptions, so this difference could not be attributed to these concomitant psychiatric diagnoses. In a TriNetX study conducted by Monahan et al,¹⁷ patients with ORD received more postoperative opioid and antidepressant prescriptions after primary hip arthroscopy. Prior studies have identified several patient-specific risk factors for prolonged postoperative opioid use, including depression, substance use, diabetes mellitus, preoperative opioid exposure, obesity, and male sex.6, 7, 8, 9, 10, 11 In the context of DRF repair, Newton et al¹¹ reported that patients with a history of substance use were at increased risk of hardware complications, persistent pain, and greater postoperative opioid consumption. Building upon this foundation, the findings of the present study suggest that patients with a diagnosis of ORD represent a particularly high-risk subgroup, demonstrating increased healthcare utilization in the postoperative period, including higher rates of ED visits and hospitalization. Together, these findings reinforce the role of substance use and opioid-related pathology as important drivers of both clinical outcomes and resource utilization following DRF.

These study results align with the current literature and mirror trends in opioid use and postoperative outcomes across several orthopedic procedures.^18,19 Therefore, there is a need for more thorough preoperative screening for ORDs and other patient-specific risk factors for improved risk stratification. Additionally, patients undergoing nonoperative management of DRF are also at risk of prolonged opioid use. In a Prescription Drug Monitoring Program study conducted by Ramtin et al,²⁰ over 80% of opioid prescriptions after DRF came from nonorthopedic providers, highlighting the need for multidisciplinary coordination in opioid stewardship.²⁰

There are a number of limitations involved in using a deidentified database for a retrospective study. First, the possibility of coding errors or incomplete coding should be considered because patients were identified using CPT and ICD-10 codes. The analysis is dependent on the accuracy and completeness of diagnostic and procedural coding within the TriNetX platform. For example, it is possible that patients included in the NORD group may have a history of ORD that was not documented or was not diagnosed correctly. Furthermore, the nature of the database prevents identification of reasons for outpatient visits, the reasons for outpatient visits, hospitalizations, ED admissions, or PT evaluations. Although we were able to temporally filter encounters in relation to the index procedure, there is no way to confirm the reason for the visit. These limitations are inherent in large database studies and have been previously described in studies evaluating the TriNetX platform.²¹ However, this large multi-institutional propensity-matched database study allows for the identification of patterns of postoperative healthcare resource utilization in patients with ORD undergoing DRF. Despite its limitations, the results of this study indicate that there is a need for improved risk stratification and thorough screening for ORDs in the preoperative setting.

This large propensity-matched database study allows for the identification of consistent patterns of increased healthcare utilization, prescription opioid, and analgesic requirements seen in patients with a history of ORD who undergo DRF fixation. Our results, along with the current literature, underscore a need for better risk stratification before surgery for patients with ORDs to improve outcomes, reduce chronic postoperative opioid use, and financial burden on patients related to increased healthcare resource utilization.

Conflicts of Interest

Asif M. Ilyas, MD, MBA, has an association and/or financial involvement with the following entities: Stryker (Other Professional Activities), Globus Medical, Inc (Other Professional Activities), SurgiColl (Fiduciary Officer), AxoGen (Other Professional Activities), Pacira Pharmaceuticals Incorporated (IP Royalties), Arthrex (Other Professional Activities), and Walant Surgical Solutions (Fiduciary Officer). The other 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.