a b s t r a c t

Objectives: After initial emergency department (ED) management of acute renal colic, recurrent or ongoing se- vere pain is the usual pathway to ED revisits, hospitalizations and rescue interventions. If index visit Pain severity is associated with stone size or with subsequent failure of conservative management, then it might be useful in identifying patients who would benefit from early definitive imaging or intervention. Our objectives were to de- termine whether pain severity correlates with stone size, and to evaluate its utility in predicting important out- comes.

Methods: We used administrative data and structured chart review to study all ED patients with CT proven renal colic at six hospitals in two cities over one-year. Triage nurses recorded arrival Numeric rating scale (NRS) pain scores. We excluded patients with missing pain assessments and stratified eligible patients into severe (NRS 8-10) and less-severe pain groups. Stone parameters were abstracted from imaging reports, while hospitaliza- tions and interventions were identified in hospital databases. We determined the classification accuracy of pain severity for stones >5mm and used multivariable regression to determine the association of pain severity with 60-day treatment failure, defined by hospitalization or rescue intervention.

Results: We studied 2206 patients, 68% male, with a mean age of 49 years. Severe pain was 52.0% sensitive and 45.3% specific for larger stones >5mm. After multivariable adjustment, we found a weak negative association (adjusted OR =0.96) between pain severity and stone width. For each unit of increasing pain, the odds of a larger stone fell by 4%. Index visit pain severity was not associated with the Need for hospitalization or rescue interven- tion within 60-days.

Conclusions: Pain severity is not helpful in predicting stone size or renal colic outcomes. More severe pain does not indicate a larger stone or a worse prognosis.

(C) 2021

1. Introduction

Most patients with acute ureteral colic will pass their stones sponta- neously but some fail and require urologic intervention [1,2]. Early in- tervention improves outcomes for high-risk patients [3,4], but may cause Iatrogenic injury [4,5] and increase morbidity for those with smaller, more distal stones [6,7]. Clinical findings and urinalysis are often sufficient to suggest a renal colic diagnosis, and Decision tools like the STONE score, along with hydronephrosis on bedside ultrasound, increase diagnostic certainty. [8] However, these modalities rarely pro- vide specific stone size and location data, and do not predict passage success. [9-11] Without the use of computed tomography (CT) imaging, physicians may have little concept of prognosis, which is important to guide patient discussions and referral decisions.

* Corresponding author.

E-mail address: [email protected] (G. Innes).

CT reliably defines stone size and location, which are the primary de- terminants of spontaneous passage; consequently, most patients now undergo CT imaging during a ureteral colic episode [9,12,13]. But CT is costly and carries radiation risks that are multiplied in stone formers who experience recurrent stone episodes and undergo repeated imag- ing. [14] Growing concerns about radiation, as well as evidence that CT may not improve patient outcomes, have led to recommendations that physicians avoid their use if possible. [15] If physicians are expected to reduce their dependence on CT, it would be helpful to find prognostic clinical parameters that identify high-risk patients more likely to fail spontaneous passage, who may require definitive imaging and early referral for intervention.

Severe pain is the cardinal manifestation of renal colic. The experi- ence and expression of pain are highly individual and multifactorial. Intolerable pain, defined by the patient, is the main criterion for “treat- ment failure” in renal colic, [10,11] and the most common proximate cause of return visits that lead to hospitalization or rescue intervention.

https://doi.org/10.1016/j.ajem.2021.02.049

0735-6757/(C) 2021

[6] Individual differences in pain experience and pain tolerance may therefore be important outcome determinants. Our hypothesis was that patient-reported pain severity has independent prognostic value and might identify patients more likely to experience treatment failure after index visit discharge. Our objectives were to evaluate the associa- tion of ED pain scores with larger stones and with the occurrence of clin- ically important outcomes.

1. Methods
   1. Study design

We used administrative data and structured chart review to perform a retrospective cohort study of all ED patients with acute renal colic seen at six hospitals in two xxx cities during 2014. Research Ethics Boards at the University of xxx and University of xxx approved this study.

• 1. Setting and patients

The two cities have similar demographics and health infrastructure. Each has a unified health delivery system with regional program over- sight, Quality management, and common health information systems that register all ED visits, hospitalizations and interventions occurring in the region. This enabled us to conduct a population-based study with a low likelihood of missing outcome events. All patients with an ED diagnosis of renal colic based on the ICD-10 codes N200, N201, N202, N132, N23 and N209 were evaluated for inclusion. [1] Eligible pa- tients required CT imaging describing stone size, location and hydronephrosis severity. We excluded patients with out-of-region postal codes whose outcomes might not be captured, and those with a preceding ureteral colic ED visit within 30-days, to focus on incident cases and avoid patients already failing outpatient management. We also excluded patients who did not have a Numeric rating scale pain score documented at triage. Based on the Canadian Triage and Acu- ity Scale (CTAS), NRS scores of 8-10 are considered severe pain, while lower scores were considered moderate pain [16].

• 1. Study protocol

We obtained patient demographics, NRS pain score, Arrival mode (ambulance, self), triage acuity, index diagnosis, disposition, hospitali- zations and procedures from regional hospital databases. To assure in- clusion of all relevant procedures, we searched the xxx diagnostic imaging database to identify outpatient lithotripsy procedures as well as inpatient or outpatient surgical procedures not coded in hospital dis- charge abstracts. urological interventions in xxx occur at a single site and are reliably captured in administrative data, but we verified these by auditing a 20% sample of electronic patient charts. To assess the reli- ability of our record review for stone characteristics, two research assis- tants independently abstracted 100 consecutive imaging reports, and we calculated Cohen’s kappa to determine inter-observer agreement. Based on eligibility criteria, all patients had complete data for arrival pain scores and main outcomes (stone size and hospital-based care ep- isodes); however, we used multiple imputation for missing covariate data-notably, heart rate, blood pressure, creatinine level and white blood cell count.

• 1. Measures

Our primary explanatory variable was arrival pain severity. Main out- comesincludedsubsequentconfirmationofalargestone>5 mm, and 60- day treatment failure defined by post-discharge hospitalization or rescue intervention. The latter are morbidity markers and proxies for patients in severe distress. [7,17-19] We chose a 60-day window because prior re- search showed that 80% of these outcomes occurred within 30 days, 15% between 30 and 60 days, and few later than this [7]. As a secondary

objective, we explored the association between pain severity, as a dependentvariable, andthefollowingcharacteristicsthatmightbecorre- lates or determinants of pain severity: patient age, sex, arrival mode, stone size, stone location, hydronephrosis (moderate or severe) and peri-renal stranding.

• 1. Data analysis

We summarized descriptive statistics for pain severity groups using 95% confidence intervals (CI) around group differences. We determined the classification accuracy of self-described severe pain (NRS = 8-10) as an indicator of larger stones >5 mm, using standard formulae to cal- culate sensitivity, specificity, predictive values and likelihood ratios. We then developed Multivariable logistic regression models to evaluate the association of arrival pain, our primary explanatory variable, with the subsequent confirmation of a large >5 mm stone, and with post-index treatment failure (hospitalization or rescue intervention within 60- days). These variables were part of standard data collection at the study sites or were available in diagnostic imaging reports. Adjustment variables for logistic models included age, sex, triage Acuity level, arrival day and time, arrival mode (ambulance, self), heart rate, blood pressure, hydronephrosis, white blood cell count and serum creatinine level. By definition, all eligible patients had complete data for main ex- planatory and outcome variables, but we performed multiple imputa- tion for missing covariate data. Because our goal was to identify prognostic factors that physicians can consider without performing CT, we did not include stone size or location in these regression models. We incorporated hydronephrosis severity as a potential predictor be- cause this can be ascertained using beside or formal ultrasound.

To evaluate possible determinants of pain severity, we developed multivariable Linear regression models using NRS pain score as the dependent variable, with candidate explanatory variables including age, sex, stone size, stone location, triage acuity level, arrival day and time, arrival mode, heart rate, blood pressure, hydronephrosis, creatinine and white blood cell count . Based on previous re- search [6] we assumed that the rate of 60-day admission or interven- tion would be ~20%. We determined that, with alpha set at 0.05, 1023 patients per group would provide 80% power to detect a 5% absolute difference, from 18% to 23%, between patients with severe vs. less severe pain.

All statistical analyses were performed using the R statistical pack- age (R Foundation for Statistical Computing; http://www.R-project. org/), and the Amelia II package was used for multiple imputation [20]. We adhered to the STROBE checklist for observational studies.

1. Results

We studied 2206 patients, 68% male, with a mean age of 48.9 years (Fig. 1). imaging parameters were reliably captured, with kappa values of 0.97 for stone length, 0.92 for width, 0.95 for location, and 0.90 for hydronephrosis severity. Unadjusted comparisons (Table 1) showed that patients with severe pain were younger, more likely to present at

Study Sample (N=2206)

-No CT imaging performed (N=1118)

-Out of region (N=112)

-Missing stone size/location/hydronephrosis data (N=22)

-Absent triage pain assessment (N=397)

Index emergency department visit with renal colic diagnosis (N=3855)

Fig. 1. Study flow sheet.

Table 1

Baseline characteristics and outcomes by pain severity

Table 3

Multivariable predictors of stone size and renal colic outcomes (without CT).

1. Presence of large stone

Pain severity	Less severe	Severe (NRS	Difference	95% CI		Outcome	Predictor	aOR (95%CI)
	(NRS < 8)	8-10)

Demographics Male: N (%)	N = 1027 712 (69.3)	N = 1179 792 (67.2)	-2.1%	(>=5 mm) unit) -6.1, 1.8 EMS arrival 0.74 (0
Mean age (SD)	50.1 (14.9)	47.8 (13.8)	-2.3	-1.1, -3.5 Male sex 0.82 (0
Mean NRS (SD)	5.12 (1.93)	9.04 (0.86)	3.9	3.8, 4.0 Hydronephrosis* 2.7 (2.3
Mean heart rate (SD)	79.3 (19.2)	79.1 (15.4)	-0.2	-1.7, 1.2 Age (per year) 1.01 (1
Mean systolic BP (SD) Arrival data: N (%)	142 (29.9)	144 (22.5)	2	-0.4, 4.1 1.015)

NRS pain score (per

0.96 (0.93, 0.99)

.58, 0.94)

.68, 0.99)

, 3.3)

.003,

White blood cell count 9.9 (3.3) 10.2 (3.1) 0.3 -0.1, 0.6

Night arrival	125 (12.2)	255 (21.6)	9.4%	6.3, 12.6	2. 60-day treatment failure:	NRS pain score (per	1.01 (0.95, 1.07)
Weekend/Holiday	753 (73.3)	831 (70.5)	-2.8%	-6.7, 1.0	hospitalization or rescue	unit)
EMS arrival	246 (24.0)	118 (10.0)	-14.0%	-10.7,	intervention	Hydronephrosis*	1.55 (1.11, 2.13)
				-17.2		Male sex	1.45 (1.03, 2.01)
Investigations: Mean (SD)						Age (per year)	1.01 (1.00, 1.022)

Creatinine (mg/dL) 93.7 (40.1) 91.3 (25.5) -2.4 -5.3, 0.5

Stone length: mm 5.2 (3.1) 4.9 (3.0) -0.3 -0.1, -0.5

Imaging results: N (%)

Distal ureteral stone 571 (55.6) 684 (58.0) 2.4% -1.8, 6.7

Middle Stone 86 (8.4) 114 (9.7) 1.3% -1.2, 3.8

Proximal stone 228 (22.2) 267 (22.6) 0.4% -3.1, 4.0

renal stone 142 (13.8) 114 (9.7) -4.1% -1.4, -7.0

Large stone >5 mm 507 (49.4) 550 (46.6) -2.8% -7.0, 1.6

Hydronephrosis^a 340 (33.1) 399 (33.8) 0.7% -3.3, 4.8

Stranding (edema) 504 (49.1) 655 (55.6) 6.5% 2.2, 10.7

Outcomes: N (%)

Index admission 422 (41.1) 543 (46.1) 5.0% 0.7, 9.2

Index intervention 384 (37.4) 506 (42.9) 5.5% 1.3, 9.7

60-day ED revisit 276 (26.9) 342 (29.0) 2.1% -1.7, 6.0

60-day admission 118 (11.5) 165 (14.0) 2.5% -0.37, 5.4

60-day rescue procedure 282 (27.5) 306 (26.0) -1.5% -5.3, 2.3

Outcome 1 is based on the entire study population (N = 2206). Sixty-day outcomes are based on patients who were discharged from the index visit without intervention (N = 1315). aOR = adjusted odds ratio. *Hydronephrosis refers to moderate or severe hydronephrosis. No models explained more than 10% of total outcome variability.

positively associated (aOR = 2.7 and 1.01 per year, respectively). None of the candidate predictors, including pain severity, predicted 60-day ED revisits. NRS pain score did not predict any post-ED outcomes; how- ever, in models that excluded stone size and location, hydronephrosis, male sex and increasing age were significantly associated with 60-day hospitalization and rescue intervention.

In evaluating possible determinants of severe pain, multivariable

60-day admit or procedure

296 (28.8) 330 (28.0) -0.8% -4.7, 3.0

models confirmed a weak negative association between stone size and pain severity (Table 4). For each 1 mm increase in stone width, the like-

^a Hydronephrosis refers to moderate to severe hydronephrosis. Bold values indicate that the 95% confidence interval for the difference between groups did not cross zero, therefore suggested group imbalance.

night, less likely to arrive by EMS, and more often had smaller stones, ureteric (versus renal) stones, and Perinephric stranding on CT. Patients with severe pain were more likely to be admitted and to have an inter- vention during the index hospital encounter; however, if they were discharged for a trial of spontaneous passage, they were no more likely to require subsequent 60-day ED revisits or hospitalizations.

Self-reported severe pain was 52.0% sensitive and 45.3% specific for the presence of a large stone (Table 2), and severe pain was more com- mon among patients with smaller stones. Positive and negative likeli- hood ratios for predicting a large stone were 0.95 and 1.06 but confidenceintervalsfor bothestimatesincorporatedzero. Thesefindings show that pain severity is of no value in predicting stone size. Table 3, which summarizes relationships between renal colic outcomes and pre- dictorsavailablepriorto CTimaging, confirms a very weak negativeasso- ciation between NRS pain score and the presence of a large stone (adjusted OR = 0.96), meaning that, for each unit increase in pain score, the odds of a large stone diminished by 4%. Male sex and EMS ar- rival were also negatively associated with stone size (aOR = 0.82 and 0.74, respectively), while hydronephrosis and increasing age were

lihood of severe pain decreased, with an adjusted odds ratio (aOR) of

0.97 (95%CI = 0.93 to 1.0). Multivariable models also confirmed that Advancing age and renal (versus ureteral) location were associated with less severe pain. Perinephric stranding on CT was most strongly as- sociated with severe pain, while hydronephrosis, white blood cell count, and creatinine level were not significantly associated with pain severity. A sensitivity analysis excluding variables with imputed values showed similar adjusted odds ratios for male sex (0.87; CI, 0.73-1.05), age (0.987; CI,0.981-0.993), stone width (0.96; CI, 0.93-1.0), renal location (0.71; CI, 0.54-0.94), and stranding (1.35; CI, 1.12-1.62). This analysis did not significantly change any of the relationships seen.

1. Discussion

We found a negative association between pain severity and stone size, whereby severe pain predicted smaller, not larger stones; however, this relationship was too weak to be helpful in clinical decision-making. We also found that arrival pain severity did not predict any 60-day renal colic outcomes; therefore, our hypothesis was not supported. However, while pain severity had minimal prognostic value in terms of stone size or outcome, physicians and patients can be reassured that more severe pain does not increase the probability of a larger stone, of significant ureteric obstruction, or of adverse outcomes. Physicians should also be

Table 2

Classification accuracy of severe pain for larger stones.

NRS pain	Large stone >= 5 mm	Smaller stone	Total
Severe (8-10)	550	629	1179
Less severe (<8)	507	520	1027
	1057	1149	2206
Sensitivity (CI)	52.0% (49-55%)	Positive predictive value (CI)	46.6% (44-50%)
Specificity (CI)	45.3% (42-48%)	Negative predictive value (CI)	50.6% (48-54%)
Positive likelihood ratio: 0.95 (0.88, 1.03)		Negative likelihood ratio:	1.06 (0.97, 1.16)

CI = 95% confidence interval.

Table 4

Variables associated with the presence of severe pain

Predictor Univariable associations

	OR	95%CI		aOR	95%CI
Male sex	0.91	0.76, 1.08		0.84	0.70, 1.01
Age in years*	0.99	0.983, 0.995		0.985	0.98, 0.99
Stone size Width (mm)*	0.96	0.92, 0.99		0.97	0.93, 1.0
Length (mm)*	0.97	0.95, 1.0

Stone location

Distal ureter 1.0 Reference

Middle ureter 1.1 0.82, 1.5

Proximal ureter	0.98	0.79, 1.21
Renal	0.67	0.51, 0.88	0.71	0.54, 0.94
Hydronephrosis^	1.03	0.87, 1.23
Stranding	1.3	1.1, 1.5	1.33	1.11, 1.59
WBC*^	1.03	0.99, 1.05
Creatinine (mg/dL)*	1	0.99, 1.01
Systolic BP (mmHg)*	1	0.99, 1.01	1.005	1.001, 1.009

Multivariable associations

captured and our exposure variable, pain severity, was recorded for all eligible subjects. To assure we were studying accurate stone parame- ters, we excluded patients who did not have CT imaging. We are there- fore confident that we had complete and reliable data for our primary explanatory and outcome variables, but some covariate data was miss- ing. Specifically, we imputed WBC and creatinine values for 9.2% of pa- tients who did not have bloodwork drawn, and blood pressure values for 1.1% of patients who lacked triage vital signs. In addition, our eligibil- ity criteria may have skewed our study sample. The exclusion of patients without CT imaging might have led us to disproportionately study males and patients with more severe illness. Conversely, the require- ment for a triage NRS score might have excluded patients with extreme pain who could not participate in a triage interview. We measured pain at a single time point using NRS scores. This is a unidimensional mea- sure that captures none of the qualitative or emotional aspects of the pain experience. We believe that our paradoxical finding of lower pain scores among EMS patients reflects analgesia provided by EMS en route, thus compromising the observed associations of pain severity in

OR represents the odds a patient will/won’t have severe pain if they have the specified predictor. aOR = adjusted odds ratio. *Odds ratio per unit rise. ^moderate or severe hydronephrosis. *^per 1000 cell increase. Variables that improved model fit are listed under ‘multivariable associations.

aware that, in the absence of CT stone size and location data, moderate or severe hydronephrosis and increasing age are associated with increasing likelihood of passage failure and the need for rescue intervention.

In evaluating potential determinants of severe pain, we found that ureteral stones were more painful than renal stones but that stone loca- tion within the ureter (e.g. proximal vs. distal) did not affect pain severity. Thestrongestpredictorofseverepainwasperi-renalstranding, whichisa marker of renal inflammation or acute obstruction. Other factors includ- ing sex, creatinine level, white blood cell count and hydronephrosis did not correlate with pain severity, but increasing age correlated with less severe pain. Although we identified several intriguing associations, the predictors studied explained relatively little of the variability in self- reported pain severity, suggesting that pain is complex and not easily ex- plained by the measurable clinical parameters we studied.

We found no prior studies assessing the prognostic value of arrival pain, and related research is conflicting. Portis reported that pain scores after ED discharge did not correlate with stone burden [21]; however, a 2005 study showed that ED discharge pain scores did predict subse- quent rescue intervention. [10] We believe this finding, contrary to ours, is because arrival and discharge pain reflect distinct phenomena

-the former a marker of disease severity without medical intervention and the latter a marker of treatment responsiveness, which logically could predict treatment failure if patients are discharged with appropri- ate analgesics and alpha blockers. This suggests that a favorable re- sponse to ED treatment is a more important prognostic indicator than the patient’s pre-treatment pain severity.

In our dataset, male sex emerged as an unexpected predictor of ad- verse outcome. While previous studies [22,23] have reported that sex has little Prognostic importance, we found that males required more downstream rescue interventions (Table 3). Inapost-hocanalysis, wede- rived a credible explanation. We found that, after controlling for stone sizeandlocation, malesweresubstantiallylesslikelytoundergoindexin- tervention (aOR = 0.54; 95%CI, 0.43 to 0.68). This means more males were discharged with significant stones and explains a higher rescue in- terventionrate. Wethereforebelievethatthesex-relatedoutcomediffer- ences seen resulted from differences in index management and not from intrinsic sex-related differences in prognosis.

Our use of administrative data from six hospitals in two cities allowed us to enrol a large diverse sample which enhances external va- lidity; however, administrative data is often incomplete. Our outcome measures (stone size, hospitalizations, interventions) were reliably

this subset. We determined that, in the absence of CT findings, hydronephrosis has significant prognostic utility; however, when CT stone size and location data were incorporated in regression models, these proved to be the only significant outcome predictors, and other variables, including hydronephrosis, lost their independent prognostic value. In addition, our hydronephrosis estimates were based on CT find- ings that may not translate exactly to bedside ultrasound. Previous au- thors have shown that ED ultrasound and CT provide comparable hydronephrosis grading [24-26], but if these two modalities differ sig- nificantly, our conclusions about hydronephrosis may be uncertain.

1. Conclusions

Index pain severity is not helpful in predicting stone size or renal colic outcomes. In the absence of stone size and location data, hydronephrosis and increasing age increase the likelihood of passage failure. Physicians can reassure patients that more severe pain does not indicate a larger stone or a worse 60-day prognosis.

Prior presentations

N/a.

Funding sources/Disclosures

This study was unfunded and the authors declare no conflicts.

Author contributions

Katie Gourlay and Graeme Splinter proposed the study and devel- oped the first draft of the study protocol. Grant Innes and Jake Hayward revised the study design. Grant Innes wrote the ethics submission, over- saw data acquisition and performed some aspects of the analysis. Jake Hayward performed some aspects of the analysis. KG and GS developed the first draft of the manuscript. All authors contributed to multiple manuscript revisions.

Acknowledgments

We thank Dongmei Wang for her assistance in data acquisition for this study.

References

1. Ordon M, Andonian S, Blew B, et al. Canadian Urological Association guideline: man- agement of ureteral calculi. Can Urol Assoc J. 2015;9:837-51.
2. Pearle MS, Nadler R, Bercowsky E, et al. Prospective randomized trial comparing shock wave lithotripsy and ureteroscopy for management of distal ureteral calculi. J Urol. 2001:1255-60.
3. Frassetto L, Kohlstadt I. Treatment and prevention of Kidney stones: an update. Am Fam Physician. 2011;84:1234-42.
4. Tonolini M, Villa F, Ippolito S, et al. Cross-sectional imaging of iatrogenic complica- tions after extracorporeal and endourological treatment of urolithiasis. Insights Im- aging. 2014;5:677-89.
5. Summerton DJ, Kitrey N, Lumen N, et al. EAU guidelines on iatrogenic trauma. Eur Urol. 2012;62:628-39.
6. Innes G, McRae AD, Grafstein E, et al. Variability of renal colic management and out- comes in two Canadian cities. Can J Emerg Med. 2018;20:702-12.
7. Innes GD, Scheuermeyer FX, McRae AD, et al. Which patients should have early sur- gical intervention for acute ureteric colic. J Urol. 2020;204. https://doi.org/10.1097/ JU.0000000000001318 In Press.
8. Mirfazaelian H, Doosti-Irani A, Jalili M, et al. Application of decision rules on diagno- sis and prognosis of renal colic: systematic review and meta-analysis. Eur J Emerg Med. 2020;27:87-93.
9. Jendeberg J, Geijer H, Alshamari M, et al. The width and location of a ureteral stone accurately predict the chance of spontaneous passage. Eur Radiol. 2017;27:4775-85.
10. Papa L, Stiell IG, Wells GA, et al. Predicting intervention in renal colic patients after emergency department evaluation. Can J Emerg Med. 2005;7:78-86.
11. Daniels B, Schoenfeld EM, Taylor A, et al. Predictors of hospital admission and uro- logical intervention in adult emergency department patients with computerized to- mography confirmed ureteral stones. J Urol. 2017;198:1359-66.
12. Schoenfeld EM, Pekow PS, Shieh M, et al. The diagnosis and management of patients with renal colic across a sample of US hospitals: high CT utilization despite low rates of admission and inpatient urologic intervention. PLoS One. 2017;12:1-15. https:// doi.org/10.1371/journal.pone.0169160.
13. Bajaj M, Yuan L, Holmes LC, et al. Predictors of surgical intervention following initial surveillance for acute ureteric colic. World J Urol. 2018;36:1477-83.
14. Brisbane W, Bailey MR, Sorensen MD. An overview of kidney stone imaging tech- niques. Nat Rev Urol. 2016;13:654-62.
15. Moore CL, Bomann S, Daniels B, et al. Derivation and validation of a clinical predic- tion rule for uncomplicated ureteral stone-The STONE score: retrospective and prospective Observational cohort studies. BMJ. 2014;348:g2191.
16. Murray M, Bullard M, Grafstein E. Revisions to the Canadian emergency department triage and acuity scale implementation guidelines. Can J Emerg Med. 2004;6:421-7.
17. Ordon M, Urbach D, Mamdani M, et al. The surgical management of kidney stone disease: a population based Time series analysis. J Urol. 2014;192:1450-6.
18. Dauw C, Kaufman SR, Hollenbeck B, et al. Expulsive therapy versus early endoscopic stone removal in patients with acute renal colic: comparison of indirect costs. J Urol. 2014;191:673-7.
19. Scales CD, Saigal CS, Hanley JM, et al. The impact of unplanned postprocedure visits in the management of patients with Urinary stones. Surgery. 2014;155:769-75.
20. Honaker J, King G, Blackwell M. Amelia II: a program for missing data. J Stat Softw. 2011;45:1-47.
21. Portis JL, Neises SM, Portis AJ. Pain is independent of stone burden and predicts surgical intervention in patients with ureteral stones. J Urol. 2018;200:597-603.
22. Innes GD, Scheuermeyer FX, Law MR, et al. Sex related differences in emergency department renal colic management: females have fewer computed tomography cans but similar outcomes. Acad Emerg Med. 2016;23:1153-60.
23. Schoenfeld EM, Shieh M, Pekow PS, et al. Association of patient and visit character- istics with rate and timing of urologic procedures for patients discharged from the emergency department with renal colic. JAMA Netw Open. 2019;2(12):e1916454.
24. Leo M, Langlois B, Pare J, et al. Ultrasound vs. computed tomography for severity of hydronephrosis and its importance in renal colic. West J Emerg Med. 2017;18: 559-68.
25. Riddell J, Case A, Wopat R, et al. Sensitivity of bedside ultrasound to detect hydronephrosis in patients with computed tomography-proven stones. West J Emerg Med. 2014;15:96-100.
26. Gaspari RJ, Horst K. Emergency ultrasound and urinalysis in the evaluation of flank pain. Acad Emerg Med. 2005;12:1180-4. https://doi.org/10.1197/j.aem.2005.06.023.