Time to pain relief: A randomized controlled trial in the emergency department during vaso-occlusive episodes in sickle cell disease

INTRODUCTION

Individuals with sickle cell disease (SCD) often experience severe painful, vaso-occlusive episodes (VOE’s) requiring an emergency department visit for treatment of severe pain with opioids. In 2014, the National Heart, Lung and Blood Institute (NHLBI) published evidence based recommendations treatment of VOE including the use of individual pain protocols when possible and standard pain protocols specifically for SCD when individual pain protocols are not possible.¹ In 2020, the American Society of Hematology (ASH) recommended tailored opioid dosing to treat VOE which should consider baseline opioid therapy for chronic pain, and doses that have been effective in the past.² Both the NHLBI and ASH recommendations suggest treatment for pain within 60 minutes of arrival.

In an effort to further test the effectiveness of these protocols on pain relief, our team conducted, and previously reported, results from a randomized controlled trial that compared changes in pain scores from placement in an Emergency Department (ED) treatment space to ED discharge, between patients randomized to receive an individualized opioid protocol developed by the patient’s hematologist or SCD provider, or an opioid weight based protocol that served as the “standard” protocol.³ Patients randomized to treatment with an individualized pain protocol experienced a 16.6 mm greater reduction (clinically and statistically significant) in pain scores compared to patients treated with the weight-based protocol.

While these results are important because they provide a higher level of evidence than previously reported, EDs remain under intense pressure to rapidly treat and provide a disposition for all patients. In 2018, the Centers for Medicare and Medicaid Services implemented an outpatient quality reporting program measure for the ED which requires reporting of ED throughput, (OP-18). The measure mandates hospitals to report the median time from arrival to discharge for patients discharged home.⁴ The NHLBI and ASH guidelines, including time to first dose of 60 minutes, are based on the premise that rapid, aggressive pain control will result in quicker resolution of pain, and ideally, lower hospital admission rates. Due to the pressure to reduce ED length of stay, and the clinical goal to provide timely pain relief, it is also important to determine which protocol led to more timely pain relief. Our clinical trial included every 30-minute pain measurements obtained from the patient by the research staff. The aim of this analysis was to compare time to pain relief during an ED visit among the patients treated with either an individualized pain protocol or weight-based pain protocol. This supplemental analysis of the trial data examined two time-to-event outcomes: (1) time to a minimum of 13 mm reduction from initial pain score and (b) time to a minimum of a 30% reduction in the initial pain score.

METHODS

Study Design and Setting -

We previously reported the design, methods, and primary results for our two site randomized control trial.³ Institutional review board approval was obtained at Duke University and both study sites. Patients signed an informed written consent prior to randomization. The trial was registered in ClinTrials.gov (NCT02222246).

Data collection occurred at two academic medical center affiliated EDs (clinical sites). Setting criteria included willingness of the site hematologist to randomize their patients to either protocol, a robust ED research staffing infra-structure, a strong informatics team and willingness to place protocols in the EHR that would allow the protocols to be visible when the ED provider attempted to order analgesics for an enrolled patient, and a minimum of two ED visits/day for the treatment of VOC.

Selection of Participants –

Patients age 21 and older with the one of following Sickle Cell genotypes verified in the EHR were eligible for inclusion: hemoglobin SS, SB⁰, SC and SB⁺. Patients were ineligible if they were allergic to both morphine sulfate and hydromorphone, had an explicit care plan stating no admittance to the hospital for only pain control, were non-English speaking, had greater than 24 ED visits in the prior 12 months, or presented with acute organ dysfunction that could affect their opioid tolerance.

Patients were recruited and enrolled either in clinic, during an in-patient hospitalization, or at the end of an ED visit for data collection during a future ED visit, should one occur. Patients were allowed to contribute data for a maximum of five different ED visits during the study. Patients received up to $20 compensation per ED visit and data provided.

Randomization–

After consent was provided, patients were randomized to either a patient-specific (individualized) or standard weight-based pain management protocol. Study staff were un-blinded. A stratified randomization method with clinical site as the stratifying variable, permuted block sizes of four, and a 1:1 treatment allocation applied by the study statistician using a computerized random number generator. The study was un-blinded to study staff and patients.

Intervention - Pain plans and placement in the electronic health record (EHR) –

Both protocols were based upon the NHLBI recommendations for the treatment of vaso-occlusive episodes (VOE’s). All patients were assigned to either morphine sulfate or hydromorphone (determined by the hematologists or SCD provider), route was intravenous unless access was difficult upon which the sub-cutaneous route was recommended, and dosing intervals for both protocols were every 20–30 minutes until pain was controlled. The primary difference between protocols was the initial opioid dose. The process used by the hematologist to determine both the weight based and patient specific doses has been previously reported.³ After the protocols were written, they were placed in the EHR and a pop-up or best practice alert was developed to alert the ordering ED physician that the patient had a patient specific or weight based opioid protocol. ED physicians were aware of the trial and encouraged to use the protocol. Adherence to the protocols was very good and previously published.³

Data Collection Procedures –

Patients completed a brief demographic survey at enrollment. Research staff monitored the ED for the arrival of enrolled subjects and placement of the patient in an ED treatment space. Upon entering the ED, research staff obtained a pain score and re-assessed pain every 30 minutes after placement in a treatment space. A 10-minute window was set to obtain pain scores. If research staff could not obtain a pain score within the 10-minute window, the pain score was missing for that 30-minute interval. Study endpoint was: 1) decision to admit to the hospital for pain or transfer to observation unit, 2) discharge home from the ED, or 3) six hours after placement in an ED treatment space. These criteria were determined to indicate either successful or unsuccessful control of pain.

Measurements - Pain outcomes –

Pain level was measured using a 0–100 mm visual analog scale (VAS), with higher scores indicated greater pain severity. Pain was assessed at the time of placement in an ED treatment space (baseline, T0) and every 30 minutes over the subsequent 6-hours (360 minutes, T360) or until discharge to home from the ED or a decision to admit to the hospital for pain occurred within the 6-hour period. Pain scores obtained at time of discharge to home or time of admission to hospital for pain management that occurred within 6-hour period were reviewed to determine the 30-minute interval to which the score most aligned. If the pain score at the assessment point to which this score aligned was missing, the pain score collected at discharge/admission was imputed at that time point.

Time-to-event outcomes –

Two event outcomes were examined: (1) time to a first minimum of 13 mm reduction in pain scores from T0 and (b) time to a first minimum of 30% reduction in pain scores from T0 within a 6-hour period using the VAS pain scores at each assessment point and imputation process detailed above. A 13 mm reduction had been previously determined to represent a clinically meaningful reduction in pain scores ^5,6. In addition we analyzed a 30% reduction in pain which is another measure of clinical significance.⁷ Patients with a 30% reduction described their pain as either much, or very much improved.⁷ A 13mm or greater reduction at each assessment was determined based on a T0 pain score minus assessment pain score, with higher positive scores indicating greater absolute reduction in pain scores. Percent reduction at each assessment point was calculated as the (T0 pain score minus assessment pain score)/T0 pain score) * 100, with higher positive scores representing greater reduction. As a final step, each event outcome at each assessment within the 6-hour period was coded as occurring (1) or not (0). Time-to-event was determined by the first occurrence of the event.

Analysis –

Descriptive statistics were used to describe the sample characteristics and outcomes. The unit of analysis was each ED visit rather than patient. Non-directional statistical tests were performed with significance set at 0.05 per test. Non-parametric Fisher’s Exact Test for categorical measures and Wilcoxon Two-Sample Tests for continuous measures were used to test for protocol differences in patient characteristics. Independent t-tests were employed to compare protocol differences in pain scores at T0 and discharge as well as change in pains score from T0 to discharge. Next, a random coefficients regression model for repeated measurements, a type of multi-level, mixed-effects regression model for longitudinal data, was used to compare protocol differences in the trajectory of pain severity scores across the 6-hour period. Time ranged from T0 to T360 minutes. Fixed effects were treatment, time, their interaction, and site, while random effects were ED visit and ED visit*time. The site-by-protocol and site-by-protocol-time interactions were not statistically significant at the 0.05 level and were omitted from the final model. This multi-level model allowed us to control for nesting effects of patients, which was an important consideration since patients could have up to five ED visits. The best fitting trajectory model indicated a linear pattern of change in both protocols from T0 to T360 minutes. Finally, time-to-event analyses for right censored data were conducted to compare the individualized protocol to the standard protocol with regard two events: namely, time to first: (a) 13 mm or greater pain reduction and (b) 30% or greater pain reduction during the 360 minutes. For each event, the hazard ratio (HR) and its 95% confidence interval (CI) was estimated using a Cox Proportional Hazards model with protocol and site as covariates and a time-to-event plot was generated with Kaplan Meier methods.

RESULTS

We randomized a total of 52 patients and analyzed 126 ED visits due to some patients having multiple ED visits.³ The current analysis included 49 patients and analyzed 122 ED visits. Four of the original 126 EDs visits were omitted due to ED visit having pain scores of 4 or less at each assessment point (n=1) or the ED visit did not have any VAS pain assessments (n=3). Pain scores of less than four are considered to have mild pain and unable to have a significant decrease in pain. Among the four ED visits omitted, three were from patients randomized to the standard weight-based protocol and one was randomized to the individualized protocol. A patient with three ED visits had one of their visits with no pain scores recorded omitted; however, the remaining two visits with pain scores collected were retained. This resulted in an analysis sample with N=122 ED visits, representing 49 patients. Among the 49 patients, 24 were randomized to the standard weight-based control protocol and 25 were randomized to the individualized protocol.

Patient Characteristics.

In Table 1, we report patient characteristics captured at the enrollment visit for the 49 patients. Most were African American/Black (89.8%), with five reporting their race/ethnicity as multiracial or other (10.2%). The median age was 27 years of age (range: 21 to 60). The majority were males (59.2%) and almost all had health insurance (93.9%). Only 26.5% were employed and almost 50% reported an annual income of less $11,000/year. The majority had a sickle cell genotype of SS (71.4%). The number of ED visits/patient ranged from 1 to 5, with a median of 2.0 ED visits/patient reported for each protocol. Individuals randomized to the two protocols did not significantly differ on any patient characteristics.

Table 1.

Patient Characteristics (N=49 Patients)

Characteristic	Total Patients (N=49)	Standard Weight-based Protocol (Standard, N=24)	Patient-Specific Protocol (Individualized, N=25)	p-value
Age, median (25th,75th)	27.0 (23.0, 32.0)	27.0 (23.0, 32.5)	28.0 (23.0, 32.0)	0.77
Male, n (%)	29 (59.2%)	14 (58.3%)	15 (60.0%)	1.00
African American/Black, n (%)	44 (89.8%)	22 (91.7%)	22 (88.0%)	1.00
Hispanic/Latinx, n (%)	7 (14.9%)	2 (9.1%)	5 (20.0%)	0.42
Health insurance, n (%)	46 (93.9%)	21 (87.5%)	25 (100.0%)	0.11
Secondary education, n (%)	26 (53.1%)	16 (66.7%)	10 (40.0%)	0.09
Full employment, n (%)	13 (26.5%)	5 (20.8%)	8 (32.0%)	0.52
Annual income < $11K, n (%)	15 (48.4%)	7 (50.0%)	8 (47.1%)	1.00
Sickle cell genotype, n (%)				1.00
Hemoglobin SS	35 (71.4%)	17 (70.8%)	18 (72.0%)
Hemoglobin SC	11 (22.5%)	6 (25.0%)	5 (20.0%)
Hemoglobin SB+	3 (6.1%)	1 (4.2%)	2 (8.0%)
ED visits, median (25th,75th)	2.0 (1.0, 3.0)	2.0 (1.0, 3.0)	2.0 (1.0, 3.0)	0.89
Clinical sites, n (%)
Number of patients	N=49	N=24	N=25	0.67
Clinical site 1	23 (46.9%)	12 (50.0%)	11 (44.0%)
Clinical site 2	26 (53.1%)	12 (50.0%)	14 (56.0%)
Number of ED visits	N=122	N=61	N=61	0.28
Clinical site 1	64 (52.5%)	35 (57.4%)	29 (47.5%)
Clinical site 2	58 (47.5%)	26 (42.6%)	32 (52.5%)

Wilcoxon Two-Sample test for continuous measures and Fisher’s Exact Test for categorical characteristics. Non-parametric methods applied due to small sample sizes and/or skewness > ± 1.0.

ED Visits:

The 49 patients had a total of 122 ED visits during the 13-month study period, with 64 visits occurring at site 1 and 58 visits at site 2. Of the 122 ED visits, 61 visits were for patients randomized to the standard protocol and 61 visits for patients assigned to the individualized protocol (Table 1). For each ED visit, the total ED analgesic dose delivered was determined. The total dose calculation included doses of morphine sulfate, hydromorphone, fentanyl, oxycodone, and ketorolac prescribed. The median total analgesic dose per visit in IV mg of morphine sulfate equivalent (IVMSE) was 42.3 (25^th, 75^th=26.8, 66.7, range=7.7 to 114.3) for the standard protocol and 46.7 (25^th, 75^th=31.0, 58.0, range=5.0 to 265.3) for the individualized protocol (Wilcoxon p=0.70). Specific dosing results for the original N=126 are detailed in Tanabe et al (2018).³

VAS Pain Scores.

Table 2 presents the unadjusted pain scores for each protocol at time of placement in an ED treatment space (T0), every 30 minutes across the subsequent 6-hours (T360 minutes), and at ED discharge, defined as time of discharge to home or admission to hospital for pain management when it occurred within 360 minutes. The mean pain score at T0 for all visits was 84.9 mm (SD=15.6, range=38.0 to 100.0). The two protocols did not significantly differ at T0 (t=−1.80, df=120, p=0.07). The mean discharge pain score was 50.4 mm (SD=29.5, range=0.0 to 100.0), with the standard protocol having significantly higher pain scores at discharge (M=57.0, SD=27.5) relative to the individualized protocol (M=43.7, SD=30.2, t=2.46, df=113, p=0.02, Cohen’s d = 0.46, Cohen’s d 95% CI=0.09, 0.83). A T0 minus discharge difference score was calculated to compare change in pain from T0 to discharge. The mean change score for all visits was 34.7 mm (SD=28.3, range=−25.0 to 100.0). The standard protocol yielded significantly less improvement in pain (M=26.3, SD=23.4) relative to the individualized protocol (M=43.2, SD=30.4, t=−3.35, df=113,p=0.001, Cohen’s d=0.62, Cohen d 95% CI=0.25, 0.99).

Table 2.

Unadjusted Visual Analogue Scale Pain Scores Across 6-hours (N=122 ED Visits)

	Standard Protocol (N=61)		Individualized Protocol (N=61)
Time (T)	N	Mean ± SD	N	Mean ± SD	p-value
T0	61	82.4 ± 16.1	61	87.4 ± 14.9	0.07
T30	60	84.4 ± 15.2	61	87.9 ± 13.5
T60	60	82.3 ± 16.9	59	79.3 ± 22.1
T90	61	74.0 ± 19.5	59	70.6 ± 25.7
T120	61	71.4 ± 21.1	58	63.1 ± 28.2
T150	61	65.6 ± 23.3	58	59.2 ± 29.3
T180	55	62.3 ± 23.1	55	56.8 ± 26.6
T210	52	58.7 ± 26.0	54	53.2 ± 26.8
T240	49	59.1 ± 27.5	45	53.9 ± 26.6
T270	44	60.8 ± 28.6	38	50.5 ± 30.3
T300	35	60.2 ± 26.6	29	49.0 ± 31.0
T330	29	63.6 ± 27.3	22	45.5 ± 30.2
T360	25	61.2 ± 27.0	18	37.1 ± 29.0
ED Discharge	58	57.0 ± 27.5	57	43.7 ± 30.2	0.02
T0 minus Discharge	58	26.3 ± 23.4	57	43.2 ± 30.4	0.00

N=Data Available; SD = Standard Deviation, P-values for independent t-tests for T0, discharge, and T0 minus discharge pain scores. Columns “N” = Data available, number of patients with non-missing data at the time indicated for each protocol. Data points are missing due to discharge before 6 hours or a missed assessment. SD = Standard Deviation, independent t-tests for T0, discharge, and T0 minus discharge pain scores. Time to discharge, in minutes: Median (25th,75th percentile) was 295 (225, 357) for the Standard protocol and 262 (224, 347) for the individualized protocol (Wilcoxon Two-Samples Test, p=0.312).

Pain Trajectory Analysis.

Figure 1 presents the adjusted mean pain scores across the 360 minutes for the two protocols, adjusting for the fixed, random, and nested effects in the linear RRM model for repeated measurements. The trajectory analysis revealed a non-significant protocol effect (F=0.82, df=1,121, p=0.37) and site effect (F=0.00, df=1,119, p=0.98). However, the time (F=151.7, df=1,99.6, p<.0001) and protocol-by-time interaction (F=5.83, df=1,99.6, p=0.02) were statistically significant, indicating that, although there was an improvement in pain over the 360 minutes in both protocols, the rate of reduction across the minutes was significantly greater in the individualized compared to the standard protocol. A posteriori protocol contrasts of the adjusted pain scores at each time point indicated significant protocol differences at the last four time points, with patients in the individualized protocol reporting significantly lower pain scores compared to the standard protocol at T270 (Standard: M=55.0, SD=27.7; Individualized: M=43.7, SD=30.4, p=0.05), T300 (Standard: M=51.8, SD=30.3; Individualized: M=38.9, SD=33.2, p=0.04), T330 (Standard: M=48.5, SD=32.9; Individualized: M=34.0, SD=36.0, p=0.04), and T360 (Standard: M=45.3, SD=35.6; Individualized: M=29.2, SD=38.8, p=0.03, Cohen d=0.43, Cohen d 95% CI= 0.07, 0.79)

Figure 1.

Mean Adjusted Pain Scores Across 360 Minutes (N=61/protocol)

Patients randomized to receive individualized dosing began to see a more rapid reduction in pain that began at 120 minutes; this gain was sustained throughout the ED stay.

Time-to-event Analyses.

A 13mm pain reduction occurred at or before 360 minutes in 105 (86.1%) of the 122 ED visits. The event incidence during this period was 55 (90.2%) for the 61 individualized protocol visits compared 50 (82.0%) for the 61 standard protocol visits (HR=1.50, 95% CI=1.01, 2.22, p=0.02). The HR for site 1 compared to site 2 was 0.87 (95% CI=0.59, 1.28, p=0.47); thus, site was omitted from subsequent analyses. The hazards model without site further demonstrated that the probability of experiencing a 13 mm or greater pain reduction was significantly higher in the individualized protocol compared to the standard protocol (HR=1.54, 95% CI=1.05, 2.27, p=0.03), with a small effect.⁸ A supplemental log-rank test also indicated that the individualized protocol was significantly more effective than the standard protocol in reducing pain by 13 mm or greater (χ²=6.40, df=1, p=0.01), with an estimated median time to the first event of 120 minutes (95 CI%=90, 120) for the individualized protocol compared to 150 minutes (95% CI=120, 150) for the standard protocol. For the 13 mm reduction, it was estimated that 75% of those receiving the individualized protocol will achieve this outcome within 150 minutes compared to 210 minutes for those receiving the standard protocol. Figure 2a presents the Kaplan Meier plot for the 13 mm pain reduction event across 360 minutes.

Figure 2a and 2b.

Kaplan-Meier Plots for Pain Reduction Events

The reduction in pain for those randomized to the individualized protocol was greater when measured as achieving both a 13 mm reduction and a 30% reduction, which may be more clinically meaningful.

A 30% pain reduction during the 360-minutes occurred in 86 (70.5%) of the 122 ED visits, with 48 (78.7%) reported for the individualized protocol visits compared 38 (62.3%) for the standard protocol visits (HR=1.69, 95% CI=1.10, 2.61, p=0171). Site 1 relative to site 2 did not differ in probability of a 30% pain reduction (HR=0.93, 95% CI=0.61, 1.43, p=0.75; thus, site was also omitted as a covariate in the subsequent analyses for outcome. The final model indicated that the probability of experiencing 30% or greater pain reduction was significantly higher in the individualized protocol compared to the standard protocol (HR=1.71, 95% CI=1.11, 2.63, p=0.01), indicative of a small effect.⁸ A supplemental log-rank test also indicated that the individualized protocol was significantly more effective than the standard protocol in reducing pain by 30% or greater (χ²=7.23, df=1, p=0.007), with an estimated median time to the first event of 150 minutes (95 CI%=120, 210) for the individualized protocol compared to 210 minutes (95% CI=150, 360) for the standard protocol. For a 30% reduction, it was estimated that 75% of those receiving the individualized protocol will achieve this outcome within 300 minutes. However, less than 75% of those receiving the standard protocol reported a 30% reduction at 360 minutes. Figure 2b presents the Kaplan Meier plot for the 30% pain reduction event across 360 minutes.

Statistical Power.

The sample size of 122 ED visits, with 61 visits/protocol, provided 80% power for the hierarchical linear trajectory analysis comparing protocol differences in change in pain scores over the 360 minutes, assuming an average intraclass correlation of 0.70 or higher, medium protocol effects (Cohen’s d equivalent of 0.50), and significance set at 0.05 for a two-sided test. The estimated power for the Cox proportional hazards model for each time-to-event analysis was at least 80%, based on values derived from the final model for HR, r-square, standard deviation, sample size, and event probability along with the assumption of a two-sided test with a significance level of 0.05.

DISCUSSION

We present results from the first randomized controlled trial that determined when patients with SCD received individualized pain doses they received a more rapid reduction in pain scores, when compared with weight based dosing. Until now, there has been limited evidence that providing rapid treatment of VOE with individualized protocols is associated with meaningful outcomes. Kavanagh et al (2015) conducted a quality improvement project in one pediatric ED aimed at decreasing time to initiation of patient controlled analgesia and receipt of opioid pain medications in the ED.⁹ Interventions included use of intra-nasal fentanyl as the first parenteral opioid, a standard time-specific VOE protocol, a SCD pain medication calculator and provider as well as patient and family education. Improvements in time to receiving first and second dose, initiation of PCA, time to admission and discharge decisions, as well as a reduction in the proportion of children that required hospital admission.⁹ Further evidence of the utility of individual pain plans for pediatric patients was found in one hospital that made individual pain plans accessible to the ED providers. From 2002–2008, the authors compared hospital admission and re-admission with four like pediatric US hospitals, without individual pain plans.¹⁰ They demonstrated a greater decline in hospital admission for patients treated with individual pain plans, over time, compared with the four like hospitals.¹⁰ These studies demonstrate rapid treatment of pain, with individualized protocols can result in lower hospital admission rates; however, these are both single site pediatric EDs.

Limited research has been performed in adult centers. The most compelling data thus far comes from a trial comparing ED care with treatment of VOE in a SCD infusion center.¹¹ For every 10 minute increase in time to first dose of pain medication, the relative risk of admission increased by 0.7%.¹¹ In one adult center, using a pre-post design, individualized plans for 221 adults were developed and resulted in both a more rapid time to initial opioid and a shorter ED length of stay, but not a lower hospital admission rate.¹² Our findings take the next step to demonstrate that individual plans resulted in more rapid pain relief, and lower hospital admission rates.³ In our sample, persons randomized to receive individualized relative to weight-based protocols not only experienced greater reductions in pain at discharge time, but they experienced pain relief sooner. This is important to patients, ED providers, and administrators.

The difference for patients is clear; pain relief is achieved sooner when using an individualized protocol. The decrease in pain scores began to differentiate between the two groups at one hour and continued throughout the entire length of stay. Among those who received individualized protocols, the time-to-event estimates suggest that 75% of these patients achieve a 13 mm pain reduction by 2.5 hours and 30% pain reduction by five hours, greater reductions than the weight based group. These analyses represent the most sensitive and clinically meaningful perspective. EDs typically tend to try to administer three doses and then determine a disposition. These data supports taking a little more time to make a disposition, and that by 5 hours most patients who have received rapid aggressive pain management with individualized protocols will be more likely to be able to be discharged home.

Our data is important for ED providers and hospital administrators. In 2019, the American College of Emergency Physicians published a policy statement on crowding (originally approved in 2006).¹³ The statement listed the following, evidence-based, negative effects of crowding including increased morbidity and mortality, increased length of stay for admitted patients, decreased patient satisfaction for hospitalized and ED patients, patients leaving prior to completion of medical treatment and significant delay in evaluation and treatment of emergency patients. Further increasing the stress on ED’s is the Hospital Outpatient Quality Measure, ED-Throughput (OP-18), that requires ED’s to report their median time from ED arrival to ED departure for discharged ED patients, to the Centers for Medicare and Medicaid.⁴ Our results demonstrate that patients who receive individualized pain plans achieve more rapid pain relief. It is clear that rapid, aggressive pain control for treatment of VOE can help ED and hospital administrators be more likely to meet this quality measure.

The major findings we present support a more timely reduction in pain with individual protocols. Our work also lends further support to the ACEP SCD point of care tool published in 2021 which recommends use of individual pain protocols when possible and immediate treatment.¹⁴ In an overcrowded ED, there is a very legitimate concern with how to achieve immediate treatment, or even within 60 minutes of arrival as recommended by both NHLBI and ASH^1,2. In the past, ED’s have had to learn to adapt their processes to deliver ECG’s in 15 minutes and facilitate rapid assessment of patients with possible stroke and sepsis. Improvement in clinical outcomes for individuals with sepsis, chest pain and stroke has resulted from these efforts and are clear examples that changes in processes can be made.^15,16 At this time, there are no “monitoring bodies” for SCD, however evidence based guidelines exist on how to provide optimal treatment of VOE. EDs are encouraged to implement ED specific quality indicators, from the existing guidelines including accurate triage as ESI level 2, time to first dose in 60 minutes, re-assessment and re-administration of opioid every 15–30 minutes until pain is controlled, and use of individualized protocols. Compared with implementing the need to obtain a rapid ECG for patients with chest pain, there are typically far fewer patients with SCD making it easier to implement processes to facilitate rapid pain treatment.

The feasibility of developing individual pain control will require close collaboration with hematologists or sickle cell experts. With the advent of sharing electronic health system data, it is now possible to share pain protocols from a sickle cell center, with hospital systems that do not have sickle cell experts. Finally, this analysis approach may facilitate the design of future clinical trials of investigational approaches to pain reduction in emergency departments, where time to pain relief is of critical relevance to both patients and ED providers and administrators.

There are several limitations that have been previously reported and include only two EDs, the number of unique patients for this analysis (n=49), and un-blinded to patients and ED providers.³ These factors may influence our findings in either direction. Further research is needed to validate findings with additional patients and EDs. In addition data was imputed for missing data. However, much rigor was used to impute values and was described in the methods.

What is new/central finding/specific clinical relevance?

We report the findings from the first trajectory analysis ever conducted examining improvements in pain scores in the acute emergency department phase of care treating VOE. Over a maximum of six hours in the ED, patients randomized to receive an individualized opioid protocol vs. a standard weight based protocol experienced a more rapid reduction in pain. As EDs and A&E’s globally suffer from overcrowding, it is critically important to provide rapid reduction in pain scores to patients experiencing VOE.