A Collaborative Multimodal Pain Management Pathway Reduces Opiate Consumption After Total Hip and Knee Arthroplasty

Abstract

Background

A collaborative effort with pharmacy, nursing and surgeons developed a multimodal pain management protocol for total hip and knee arthroplasty. This investigation aims to evaluate the effectiveness in reducing opiate consumption.

Methods

Retrospective cohort study comparing 455 total hip and knee arthroplasty patients. We compared the first 206 multimodal pain management protocol and 249 standard-of-care pathway patients from January 2017 to March 2020. Primary outcome was postoperative morphine milligram equivalents (MMEs) per 24 hours. Secondary outcomes included total MME at discharge and total MME in the first 6 weeks following discharge.

Results

Postoperative MME per 24 hours was lower in the multimodal pain management protocol group 57.0 (30.5, 94.3) median (interquartile range), compared with the standard of care group 84.7 (50.4, 139.6), P < .001. MME prescribed at discharge was reduced in the multimodal pain management group 300 (224, 450) compared to the standard of care group 525 (400, 750), P < .001. MME prescribed 6 weeks following discharge was reduced in the multimodal pain management group 150 (0, 590.6) compared to the standard of care group 400 (0, 1050), P = .01. Additionally, the length of stay was reduced in the multimodal group, P = .01, with similar postoperative pain scores across groups.

Conclusions

Our collaborative multimodal pain management pathway reduced opiate consumption for patients undergoing total hip and knee arthroplasty with a reduction in LOS, no difference in pain scores and was effective for patients with various levels of opiate requirements.

Level of Evidence

Level III Therapeutic Study: retrospective comparative study

Introduction

Total hip (THA) and knee arthroplasty (TKA) are common orthopaedic procedures that traditionally relied heavily on opiate pain medications for postoperative pain control [1]. These medications have the potential for adverse effects such as confusion, nausea, constipation, or respiratory depression [2]. After THA or TKA, a number of patients may become chronic abusers of opiate medications [3,4]. As the incidence of opiate abuse in the United States increases [5], it is important for orthopaedic surgeons to consider ways to shift to alternative pain management protocols that do not rely on opiate medications.

The use of evidence-based clinical care pathways have been shown to improve efficiency and improve utilization of resources after THA and TKA [1,6]. Prior investigations have focused on the effectiveness of single interventions such as regional blocks [7,8], local infiltration of analgesia [9], shifting to nonopioid medications [10], or combinations of these interventions [11,12]. These are important investigations but often fail to address challenges associated with protocol development including pre-existing work cultures, bureaucratic road blocks, and departmental silo mentalities [13]. The importance of an interdisciplinary team to drive development and implementation of clinical care pathways for THA and TKA has been shown in previous investigations [13,14]. This team-based approach should include consistent and coordinated approaches to pain management and should consist of key stakeholders including surgeons, anesthesiologists, nursing, pharmacy, and hospital administrators [15].

Preoperative opioid use has been shown to lead to increased opioid consumption and complications after THA and TKA [16]. While important to optimize patients and attempt to reduce preoperative opioid use [17], there is a paucity of evidence on developing and implementing multimodal pain control pathways in these populations. The primary aim of this investigation was to evaluate the effectiveness of our multimodal pain control pathway in terms of opioid consumption during the immediate postoperative period after THA and TKA. Our secondary aim was to evaluate the effectiveness of our multimodal pain control pathway in terms of opioid consumption during the postoperative recovery following THA or TKA. Our tertiary aim was to investigate if the multimodal pain control pathway was effective in terms of opioid consumption reduction for patients with and without prior opioid dependence.

Material and methods

After receiving approval from our institutional review board, we performed a retrospective cohort study to compare 455 patients who underwent hip or knee arthroplasty by 1 of 3 fellowship-trained adult reconstruction surgeons at a single academic medical center. The 3 surgeons engaged without our institutional quality improvement team to identify key stakeholders to be involved with creating our novel multimodal pain management protocol for THA and TKA. We identified leaders in our orthopaedic nursing staff, pharmacy, anesthesia, and hospital administrators to take part in the development of the protocol. This multidisciplinary team then held meetings and took input from all involved to establish goals and identify areas for improvement. Before launching the new protocol, we created specific educational materials for all teams and communicated broadly to all involved in the care of our THA and TKA patients the rationale, key changes, and timing for the launch of the protocol.

Data were collected on the first 206 patients in this multimodal pain management pathway from January 2019 to March 2020 and on 249 patients with the previous standard of care pathway from January 2017 to December 2018. The previous standard of care pathway was developed by the surgeons based on their expertise and involved the following nonstandardized combination of interventions for perioperative and postoperative pain control. Before surgery patients would receive oral acetaminophen 1000 mg, with a preference for a spinal anesthetic. TKA patients would receive a single-shot adductor canal block with our regional anesthesia team. Both hip and knee arthroplasty patients would receive an intraoperative total joints periarticular injection with a combination of weight-based ropivacaine, 30-mg ketorolac, 0.5-mg epinephrine, and 80-mcg clonidine. Postoperatively patients would receive oral acetaminophen 1000 mg every 8 hours as needed, oral oxycodone 5-15 mg every 3 hours as need, with 15-mg intravenous ketorolac every 8 hours as needed and intravenous hydromorphone as needed.

Our multimodal pain management pathway added and modified to this protocol. In the multimodal pain management pathway preoperative medications given before surgery included acetaminophen 1000 mg with the addition of 150-mg pregabalin and 400-mg celecoxib. In TKA patients' preferences, spinal anesthesia, single-shot adductor canal blocks (with bupivacaine 0.5% typically 15 mL and dexmedetomidine 0.5-1 mcg/kg), and our intraoperative periarticular injection did not change. Postoperatively, patients would receive 1000-mg oral acetaminophen and this was changed to a scheduled dosing regime, 15-mg intravenous toradol was also scheduled to be given every 8 hours for the first 24 hours postoperatively. After the first 24 hours, 500-mg oral naproxen was scheduled. We created pathways for both ages, <75 years old and >75 years old. For patients aged <75 years, they received 50-100 mg tramadol every 6 hours as needed as a first-line oral option for moderate pain, with 5-10 mg oxycodone every 6 hours as needed as second-line oral as needed for severe pain. For patients age >75 years, they received 2.5-5 mg oxycodone every 4 hours as needed, and tramadol was avoided due to the potential risks in this population. For patients with chronic kidney disease, the use of scheduled naproxen was at the provider's discretion and held in select cases. Intravenous hydromorphone was removed from the protocol and only given in cases where severe pain persisted after all other as needed medications were attempted and only after examination by the treating surgeon, nurse practitioner or physician assistant with a comprehensive provider bedside evaluation of pain. On discharge in hospital medication use reviewed by the providers, and enough as needed opioid medication (oxycodone or tramadol) was provided until the first postoperative clinic visit 14-21 days postdischarge. Scheduled acetaminophen and naproxen were continued on discharge.

The primary outcome measured postoperative morphine milligram equivalents (MMEs) per 24 hours. Secondary outcomes included total MME provided at discharge and total MME prescribed in the first 6 weeks following discharge. Two patients were excluded due to inability to accurately record MME in the medical record. Demographic characteristics including patient age (years), sex (male or female), body mass index, and American Society of Anesthesiologists classification were collected. In addition, we collected patient characteristics data including preoperative diagnosis of depression or fibromyalgia, total number of allergies, any documented preoperative opioid use, history of chronic pain or narcotic dependence and active opioid use at day of surgery. Perioperative data including surgery type, length of stay, and postoperative visual analog pain scores, were collected.

A power analysis anticipating a 20% reduction in MME at 80% power and alpha = 0.5 indicated a required sample size of approximately 200 patients per treatment group. Demographic and perioperative patient characteristics were compared between the standard of care and multimodal pain management groups with chi-square and Fisher exact test used for categorical variables and independent samples t-tests and Mann–Whitney tests were used for continuous variables (Table 1). Multivariable quantile regression analyses were used to investigate whether patients who received the multimodal pain management intervention consumed fewer postoperative MME per 24 hours and were prescribed fewer MME at discharge, compared with patients in the standard of care group. Significance was assessed at the P < .05 level and all analyses were performed on commercially available statistical software (Stata v14.2, College Station, TX).

Table 1.

Preoperative, intraoperative, and postoperative interventions or medications comparing the standard-of-care cohort to our multimodal pain management pathway.

	Standard of care cohort	Multimodal pain management pathway
Preoperative medications	-Oral acetaminophen 1000 mg	-Oral acetaminophen 1000 mg -Pregabalin 150 mg -Celebrex 400 mg
Intraoperative interventions	-TKA patients receive single-shot adductor canal blocks -Intraoperative periarticular injection (weight-based ropivacaine, ketorolac 30 mg, epinephrine 0.5 mg, clonidine 80 mcg) -Preference for spinal anesthesia	-TKA patients receive single-shot adductor canal blocks -Intraoperative periarticular injection (weight-based ropivacaine, ketorolac 30 mg, epinephrine 0.5 mg, clonidine 80 mcg) -Preference for spinal anesthesia
Postoperative medications	-Oral acetaminophen 1000 mg every 8 h as needed -Oral oxycodone 5-15 mg every 3 h as needed -Intravenous ketorolac 15 mg every 8 h as needed -Intravenous as needed hydromorphone	-Oral acetaminophen 1000 mg every 8 h SCHEDULED -Age <75 y tramadol 50-100 mg every 6 h as needed as a first-line oral option for moderate pain, with oxycodone 5-10 mg every 6 h as needed as second line oral as needed for severe pain. -Age >75 y oxycodone 2.5-5 mg every 4 h as needed -Intravenous ketorolac 15 mg SCHEDULED every 8 h for the first 24 h postoperatively. -After the first 24 h, naproxen 500 mg oral was scheduled -Intravenous hydromorphone REMOVED and only given where severe pain persisted after all other as needed medications were attempted and only after examination by the treating surgeon, nurse practitioner or physician assistant with a comprehensive provider bedside evaluation of pain.

Results

We found no between group differences in patient characteristics (Table 2). Perioperative characteristics showed more knee arthroplasty in the multimodal pain management group (111 of 236, 53.9%) compared to the standard of care group (102 of 249, 41%), P = .03. The length of stay was reduced in the multimodal pain management group (median 29.9 hours, interquartile range (IQR) 27.150.8), compared to the standard of care group (median 33.5, IQR 27.8-51.8), P = .01. Similar postoperative pain scores were noted between the multimodal pain management group (4.5, IQR 3.3-5.7) and the standard of care group (4.7, IQR 3.5-5.8), P = .50 (Table 2).

Table 2.

Demographic and perioperative characteristics.

Variable	Overall (n = 455)	Standard of care (n = 249)	Multimodal pain strategy (n = 206)	P valuea
Age (y), mean (SD)	63.8 (11.5)	64.0 (11.6)	63.7 (11.4)	.78
Male gender	172 (37.8%)	94 (37.8%)	78 (37.9%)	1.00
Body mass index (kg/m2), mean (SD)	31.2 (6.1)	31.5 (6.3)	31.0 (5.8)	.39
Visual analog pain score, mean (SD)	4.0 (2.8)	4.0 (2.9)	4.1 (2.8)	.82
ASA Classification				.33
I	9 (2.0%)	5 (2%)	4 (1.9%)
II	227 (49.9%)	131 (52.6%)	96 (46.6%)
III	217 (47.7%)	111 (44.6%)	106 (51.5%)
IV	2 (0.4%)	2 (0.8%)	0 (0%)
Diagnosis of depression	160 (35.2%)	96 (38.6%)	64 (31.1%)	.12
Diagnosis of fibromyalgia	30 (6.6%)	17 (6.8%)	13 (6.3%)	.98
Number of allergies, mean (SD)	2.2 (2.9)	2.2 (3.1)	2.3 (2.6)	.75
Surgery type				.03
Conversion THA	15 (3.3%)	11 (4.4%)	4 (1.9%)
THA	220 (48.4%)	132 (53%)	88 (42.7%)
TKA	213 (46.8%)	102 (41%)	111 (53.9%)
Unicompartmental knee arthroplasty	7 (1.5%)	4 (1.6%)	3 (1.5%)
Average postoperative pain score, median (IQR)	4.7 (3.3, 5.7)	4.7 (3.5, 5.8)	4.5 (3.3, 5.7)	.50
Length of stay (h), median (IQR)	31.3 (27.6, 51.7)	33.5 (27.8, 51.8)	29.9 (27.1, 50.9)	.01
Any documented preoperative opioid use or history of chronic pain or narcotic dependence	344 (75.6%)	185 (74.3%)	159 (77.2%)	.56
Active opioid use at day of service	105 (23.1%)	55 (22.1%)	50 (24.3%)	.66

ASA, American Society of Anesthesiologists; SD, standard deviation.

^aChi-square tests or Fisher's exact tests used for categorical variables and independent samples t-tests or Mann–Whitney U test used for continuous variables.

Patients in the multimodal pain management group had reduced MME per 24 hours postoperatively (median 57.0, IQR 30.5-94.3) compared to the standard of care group (median 84.7, IQR 50.4-750), P < .001. MME prescribed at discharge was reduced in the multimodal pain management group (median 300, IQR 225-450) compared to the standard of care group (median 525, IQR 400-750), P < .001. MME prescribed in the first 6 weeks following discharge was reduced in the multimodal pain management group (median 150, IQR 0-590.6) compared to the standard of care group (median 400-IQR 0, 1050), P = .01 (Table 3).

Table 3.

Primary and secondary outcomes, bivariate analyses.

Variable	Overall (n = 455)	Standard of care (n = 249)	Multimodal pain strategy (n = 206)	P valuea
Postoperative MME per 24 h	70.9 (40.4, 117.3)	84.7 (50.4, 139.6)	57.0 (30.5, 94.3)	<.001
Total MME prescribed at discharge	450 (275, 616)	525 (400, 750)	300 (225, 450)	<.001
Total MME prescribed in the first 6 wk following discharge	256.3 (0, 881.25)	400 (0, 1050)	150 (0, 590.6)	.01

^aP values from Mann–Whitney U tests.

Multivariable-adjusted quantile regression analysis revealed significant differences and reductions in MME for the multimodal pain management group across the 0.25, 0.50, 0.75, and 0.90 quantiles in postoperative MME consumption per 24 hours (Fig. 1), with percentage differences ranging from 24.1% (95% confidence interval (CI), 9.9%-36.1%) at the 0.90 quantile to 38.6% (95% CI, 26.3%-48.8%) at the 0.25 quantile (Table 4, Fig. 1). The 0.10 quantile showed a reduction of 14% for multimodal pain management patients; however, the difference was not statistically significant. Similarly, the multimodal pain management group received significantly lower levels of MME at discharge across all measured quantiles of MME prescribed, with percentage differences ranging from 35.6% (95% CI, 27.9%-42.5%) at the 0.25 quantile to 43.9% (95% CI, 35.5%-51.2%) (Fig. 2).

Figure 1.

Percentage difference in MME between the multimodal and standard of care groups by quantile. Adjusted for age, active opioid use at the time of service, average postoperative pain score, any documented opioid use, depression diagnosis, fibromyalgia diagnosis, surgery site, and whether tramadol was used as a first-line opioid treatment postoperatively.

Table 4.

Multivariable analysis.

	Model 1: Postoperative MME per 24 h		Model 2: Total MME prescribed, discharge to 6-wk postoperative
	Standard of care	Multimodal pain strategy	Standard of care	Multimodal pain strategy
Overall median (IQR)	84.7 (50.4, 140.0)	57.0 (30.5, 94.3)	900 (503, 1725)	525 (300, 992)

Multivariable analysis results
Quantile	% Difference (95% CI)ab		% Difference (95% CI)ab
0.10	Ref	−31.9 (−49.4 to −8.47)	Ref	−44.4 (−59.0 to −24.7)
0.25	Ref	−39.5 (−49.6 to −27.4)	Ref	−47.6 (−54.6 to −39.6)
0.50	Ref	−39.0 (−46.6 to −30.4)	Ref	−47.8 (−55.7 to −38.5)
0.75	Ref	−31.1 (−38.4 to −23.0)	Ref	−47.9 (−55.3 to −39.0)
0.90	Ref	−27.6 (−38.9 to −14.3)	Ref	−53.6 (−62.8 to −42.2)

Ref, reference group.

^aAdjusted for age, active opioid use at the time of service, average postoperative pain score, any documented opioid use, depression diagnosis, fibromyalgia diagnosis, surgery site, and whether tramadol was used as a first-line opioid treatment postoperatively.

^bPercent (%) = 100∗(exponentiated regression coefficient −1), which is interpreted as % difference in MME consumption in the Multimodal Pain Strategy group vs the Standard of Care group at the specified quantile of MME consumed (Model 1) or prescribed (Model 2).

Figure 2.

Percentage difference in total MME prescribed at discharge between the multimodal and standard of care groups by quantile. Adjusted for age, active opioid use at the time of service, average postoperative pain score, any documented opioid use, depression diagnosis, fibromyalgia diagnosis, surgery site, and whether tramadol was used as a first-line opioid treatment postoperatively.

Discussion

It is important for orthopaedic surgeons to consider ways to shift away from and reduce opioid consumption after THA and TKA. Our interdisciplinary team created a multimodal pain management protocol that reduced MME per 24 hours postoperatively, on discharge, and prescribed in the first 6 weeks after surgery. This was achieved without an increase in postoperative pain scores with a reduction in LOS. Additionally, we found that this multimodal pain management protocol was effective for patients across a spectrum of opioid consumption, suggesting this protocol lowered opioid consumption even for those patients with preoperative narcotic dependence.

Our multimodal pain management protocol was evidence based as we identified areas where prior research supported the opportunity for potential improvement. Before our multimodal pain management protocol postoperative IV opioid prescribing was utilized per surgeon discretion. Lampolt et al. in a 2014 randomized control trial showed that compared to IV opioid standardization of multimodal pain management protocols can decrease opioid use, improve pain scores, and enhance early recovery [18]. The American Society of Anesthesiologists Task Force on Acute Pain Management recommends incorporation of nonsteroidal anti-inflammatory drugs when appropriate [19]. We incorporated preoperative celecoxib and postoperative ketorolac into our protocol. Straube et al. conducted a systematic review of randomized studies and concluded that preoperative coxibs showed clear benefits in terms of reduced postoperative pain, analgesic consumption and patient satisfaction compared with placebo [20]. Another addition to our protocol was incorporation of selective use of tramadol for patients less than 75 years of age. Tramadol use has been shown to decrease MME in patients undergoing TKA [21]. We limited the use of this medication to patients less than 75 years old due to concerns for safety events associated with tramadol use among older adults with osteoarthritis [22].

Utilizing a team approach and engaging stakeholders from across an institution can be critical for surgical process improvement. We started by identifying an interdisciplinary team that included the surgeons, anesthesiologists, pharmacists, and nursing leadership representatives from the operating room and ward. The Six sigma methodology was used to define, measure, analyze, improve, and control following a process defined by Cima et al. [14]. This methodology has been used by Attarian et al. [13] to create high-efficiency operating rooms for total joint arthroplasty in an academic setting. To our knowledge this is the first investigation to specifically use this methodology for opioid reduction in total joint arthroplasty. While our process map was focused on opioid reduction rather than operative room efficiency this methodology can be used for a variety of quality improvement. Overcoming preconceived notions and silo mentalities was critical, some stakeholders expressed concerns for increased pain scores or delayed discharges due to the changing protocols. These concerns helped us define our secondary outcome measures, length of stay, and patient-reported pain scores. Our finding of decreased length of stay with no increase in pain scores was critical to achieve buy in from some in nursing leadership and underscores the importance of the methodology used and reporting clearly defined measures for process improvement.

A key finding of this investigation is that our multimodal pain protocol was effective in reducing MME across a spectrum of patients including those who require higher dosages of opioids after THA and TKA. Specifically, patients who have a history of preoperative narcotic use are at higher risk of poor rehabilitation progression and complications such as nausea, oversedation, urinary retention, and gastrointestinal complications [23,24]. These patients are also more likely to experience in hospital complications, increased length of stay, and require oral narcotics at 3 months postoperatively [25]. Our multivariable-adjusted quantile regression analysis showed differences and reductions in MME across the quantiles, suggesting that regardless of the narcotic requirements of patients even those who require higher dose narcotics will benefit from the comprehensive multimodal pathway of interventions.

This investigation has several potential limitations. This was a retrospective investigation and subject to selection bias. Despite this, we included all consecutive THA and TKA patients over the study period to limit this concern. As there is growing focus and attention on the opioid crisis and reducing opioids, it is possible that over time we could have seen a reduction in our use opioids as societal and other pressures would create a natural drift. An additional potential limitation is that we did not perform subset analyses comparing THA or TKA patients and it is possible that our intervention would be more effective in one group vs another. Patient-reported outcome measures were not included in this investigation and that limits our ability to identify this important patient-centric outcome measure between cohorts. Finally, it is impossible to point out 1 specific intervention or weigh one change over another. While that limits the ability to point to any one of our protocol changes, it also supports the utilization of a team approach to engage multidisciplinary stakeholders. However, we are unable to prove causation of the multidisciplinary pathway to the success at reducing MME. It is possible that just by engaging stakeholders and creating educational materials on the importance of opioid reduction even without formal medication changes improvements can be made. We are limited in this investigation to determine the relative importance of the education and multidisciplinary team approach.

Conclusions

The use of an intradisciplinary team approach and use of defined process improvement methodologies supported the development of a collaborative multimodal pain management protocol for use with patients undergoing THA and TKA. We were able to reduce opioid consumption for patients in the immediate postoperative recovery period and reduce opioid prescribing on discharge. This was achieved with a reduction in length of stay and no difference in pain scores. It was also effective for patients with various levels of opioid requirements. This is only preliminary work, and future investigations are required to validate the multidisciplinary approach and compare to other methods for quality improvement. We hope that other institutions consider both the interventions and intradisciplinary approach we undertook for process improvement and use these to support opioid reduction.

Conflicts of interest

Thomas Huff is a paid consultant for Smith & Nephew, Zimmer Biomet, Medacta, and Exactech; receives research support as a principal investigator from Smith & Nephew and Zimmer Biomet; receives fellowship or educational support from Smith & Nephew and Zimmer Biomet. Kathryn Schabel receives research support from Zimmer Biomet and is a board member/committee appointments for Western Orthopaedic Association, Oregon Association of Orthopaedic Surgeons, and American Academy of Orthopaedic Surgeons. Ryland Kagan is a Paid consultant for OrthAlign, Smith & Nephew, and Enovis; receives Research support from Smith & Nephew, OrthoDevelopment, and 3M; has medical or orthopaedic publications or is in the editorial or governing board of Journal of Arthroplasty as Elite Reviewer; and is a board member/committee appointments for the AAOS Hip Program Committee, the AAHKS Evidence-Based Medicine Committee, and the AJRR Research Committee. The other authors declare there are no conflicts of interest.

For full disclosure statements refer to https://doi.org/10.1016/j.artd.2025.101716.

CRediT authorship contribution statement

Mahmoud Khellah: Writing – review & editing, Writing – original draft, Validation, Methodology, Investigation, Data curation, Conceptualization. Thomas Huff: Writing – review & editing, Writing – original draft, Validation, Supervision, Resources, Methodology, Investigation, Data curation, Conceptualization. Kathryn Schabel: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Resources, Methodology, Investigation, Data curation, Conceptualization. Jessica Foerster: Writing – review & editing, Writing – original draft, Visualization, Validation, Resources, Methodology, Investigation, Data curation, Conceptualization. Leah Esposito: Writing – review & editing, Writing – original draft, Validation, Resources, Methodology, Investigation, Conceptualization. Megan Rushkin: Writing – review & editing, Writing – original draft, Visualization, Validation, Software, Resources, Project administration, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Ryland Kagan: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Software, Resources, Project administration, Methodology, Investigation, Formal analysis, Data curation, Conceptualization.