In this before and after implementation study in a UK Emergency Department, the authors retrospectively evaluated the effect of an analgesic protocol. In this protocol, treating physicians should administer oral analgesia or IV morphine to patients with fracture pain, depending on injury severity and fracture type.
The analgesic protocol was presented at a staff meeting at the same time as the baseline results of the primary audit were presented as well.
In this protocol pain scores were not mentioned and analgesic choices were made depending on the type of fracture. For example, patients with significant head injury were supposed to be given IM diclofenac (something our neurosurgeons would certainly disagree with) or IV morphine,
Although significantly more patients received analgesia after implementation, it should be noted that - as in many other pain studies - pain treatment was the worst before implementation, but was still bad after implementation of the protocol.
In patients with minor fractures, nearly 70% still got no pain medication offered and in admitted patients this number was 22%.
However, to end with a positive note, the IV route of administration was recognized as a superior route and was used in nearly 40% after implementation.
It is certainly nice to read that - already back then in 1996 - nerve blocks were explicitly mentioned in the pain protocol and this local anesthetic technique was mentioned before the administration of systemic opioids!
The authors included all consecutive patients during 2 periods (before and after implementation).
It was a retrospective study, therefore prone to potential (unknown) confounding factors, for which no correction methods were applied. It appeared as if patients were interviewed in a later stage and recall bias might have influenced results. Methodology and data collection were not clearly described. No formal sample size calculation was performed.
It was not clear whether the protocol was intended for use by physicians only, but it probably was. Post-implementation data were collected in the month following implementation, therefore could have been falsely positive due to a 'honeymoon' period.
Implementation of a pain protocol, not based on pain assessment, but on injury type increased the use of IV pain medication in adult patients with fractures who are discharged from the Emergency Department as well as patients who were admitted to the hospital.
Level of evidence 4
Decosterd I, Hugli O, Tamchès E, Blanc C, Mouhsine E, Givel JC, Yersin B, Buclin T
Oligoanalgesia in the emergency department: short-term beneficial effects of an education program on acute pain
The authors of this Swiss prospective pre- and postimplementation cohort study evaluated the effect of implementing a structured, VAS-based pain protocol.
The pain protocol included:
VAS <4: acetaminophen and/or NSAID (ibuprofen 600mg or diclofenac 75mg)
VAS >4: morphine 0.1mg/kg IV (titrated to max 0.15mg/kg) AND acetaminophen or NSAID
No opioid administration in headache patients.
Although the pain protocol analgesics were to be prescribed by the physician, the protocol stated explicitly that this was a joint-responsibility of both physicians and nurses, as was the goal to administer analgesia within 15 minutes and to perform reassessments every 15 minutes.
Implementation was combined with educational sessions that covered evaluation and treatment of acute pain. Pain guidelines were widely distributed (in pocket-form and as posters) among physicians and nurses and discussed during all formal staff meetings, rounds and shift-changes. Additionally, there was a hotline installed that was staffed by an experienced ED physician (for urgent questions during office hours).
As a result, they found that pain assessments increased, as did reassessment and reporting of pain severity, however still half of patients did not have their pain severity documented after implementation (26% by physicians and 45% in nurses).
Administration of pain meds increased to 63%, as well did the administration of IV morphine in higher dosages (mean dose from 2.4 to 4.6mg). This all resulted in more effective pain management as VAS pain scores decreased significantly more after implementation.
This was a prospective cohort study.
Patients were recruited by a dedicated research assistant, therefore confounding such as crowding and staff-related factors were minimized.
A patient flow chart was included reporting excluded patients including reasons of exclusion.
Patients were not recruited consecutively (however, bias was minimized by recruiting during different 8-hour shifts so that admissions during 24 hours were represented in the data as much as possible).
No sample size calculation was performed.
Due to the prospective, unblinded and observational design, the Hawthorne effect could not have been excluded as a confounder.
Implementation of a clinical pain management guideline improved pain assessments and reassessments, as well as analgesic administration and efficacy and patient satisfaction in patients presenting with pain of any cause.
Level of evidence 2b
The authors of this Israelian study prospectively evaluated implementation of a structured pain protocol combined with an educational intervention in patients with minor-to-moderate traumatic injuries and patients with renal colic.
Before implementation available analgesics were acetaminophen, dipyrone, acetylsalicylic acid, ibuprofen, naproxen, buprenorphine, parenteral diclofenac, meperidine, morphine and fentanyl. The structured pain protocol consisted of:
VAS 1-3: dipyrone, aspirine, acetaminophen, ibuprofen or naproxen
VAS 4-6: acetaminophen with codeine, buprenorphine or dypirone with ibuprofen
VAS 7-10: IV morphine, fentanyl, ketamine or nerve blocks
Educational sessions were 2-hours didactic small-group lectures and a 1-hour case-based discussion involving physicians and nurses. Moreover, during the 3-months after educational sessions, patient charts spot-checks were performed and personal feedback to treating staff provided.
The study had a single-blinded design as Emergency Physicians and nurses were unaware of the purpose of the study.
Data was collected by a dedicated physician, not involved in care for recruited patients.
Both pre- and after-intervention data were collected prospectively
Patients were only included in case the research physician was present (non-consecutively).
No sample size calculation was performed beforehand.
Although the administration of analgesia improved impressively, it is remarkable that there was no difference in pain scores (analgesic efficacy). This might have been caused by an unknown confounder or factor that had influenced patient selection.
The implementation of a VAS-based pain protocol in combination with educational sessions for physicians and nurses including personal feedback improved analgesic administration and time to analgesia in patients with minor/moderate traumatic injuries or renal colic significantly.
Level of evidence 2b
Gawthorne J, Welch S, Robertson F, McDonell K, Finckh A
Implementation of a guideline to improve prescription of analgesia for adult trauma patients in an emergency department
The authors of this retrospective observational cohort study in Australia evaluated the effect of implementation of a pain management guideline in trauma patients (intubated and non-intubated), identified from a trauma database.
The pain management guideline consisted of recommendations for intubated as well as for non-intubated patients.
Intubated patients: morphine and midazolam infusions (unless neurological injury or age >70 years, then morphine and propofol were used).
Non-intubated patients: NRS-based;
NRS 5 or lower: acetaminophen/codeine PO AND ibuprofen or indomethacine PO (in case no contraindications for NSAIDs. Reassessment at 30-60 minutes NRS >5: oxycodone 5 or 10mg
NRS >5: IV morphine 2.5mg titrated up to 0.1mg/kg AND acetaminophen 1000mg PO. NSAID added in case no contraindications. If longer opioid pain treatment anticipated, initially subcutaneously morphine was utilized and after 24 hours PCA was considered.
After development of this guideline, educational sessions were held in 2 separate months for both medical and nursing staff, including simulation training and feedback from audit data. Additionally, large posters of the guideline were displayed in resuscitation bays and throughout the ED and were made available electronically. PCA accreditation was required for nursing staff.
The authors reported increased analgesic administration in both intubated (32 to 68%) as well as non-intubated patients (86 to 100%). Median time to analgesia in non-intubated patients decreased from 38 to 14 minutes. Documentation of pain scores increased non-significantly.
The pain management guideline was specifically developed for major trauma patients (in whom pain management usually is neglected due to other high treatment priorities), including intubated patients.
Along with implementing the guideline, extensive educational sessions were held during 2 separate periods with audit feedback and simulation training.
Retrospective chart review and therefore, unknown confounders might have influenced study results.
No formal sample size calculation was performed.
It is not clear who selected the patients.
The control group was selected during a period of 5-6 months and the post-implementation group during a period of 2-3 months.
It was not explicitly described how long after implementation of the guideline, the post-intervention cohort was included in the study.
Implementation of a clinical pain management guideline in combination with extensive educational sessions for Emergency Department treating staff increased analgesic administration in intubated and non-intubated major trauma patients and pain assessment (in the latter). Time to analgesia decreased significantly.
Level of evidence 4
Steinberg PL, Nangia AK, Curtis K
A standardized pain management protocol improves timeliness of analgesia among emergency department patients with renal colic
The authors of this before and after observational cohort study compared effectiveness of a NSAID/opioid pain protocol in patients with renal colic in an Emergency Department in the USA.
The protocol prescribed ketorolac 30mg IV AND morphine 0.05-0.1mg/kg IV (and allowed for a 2nd dose of morphine if pain persisted). It was not explicitly described whether nurses were allowed to administer these drugs autonomously, therefore, it was probably physician-driven analgesia.
Physicians and nurses were educated about the study and its background before implementation of the pain protocol.
The authors found that implementation of this pain protocol resulted in a 35-minute decrease in time to effective analgesia.
A power calculation was performed in order to detect a time reduction of 50% (time to analgesia)
Patient flow chart was shown, including reasons why patients were excluded.
The control group was a historical cohort and data were collected retrospectively and therefore prone to confounding factors.
The calculated sample was 45 patients in each group, however, in the after-implementation group 44 patients were analysed. Moreover, 1 patient was excluded from the prospective group as time to analgesia was 446 minutes and this patient was considered a statistical outlier. The analysis including this patient was not reported.
The patient flow chart only shows the prospective arm, it is not known what patients were excluded in the historical cohort group, therefore selection bias might have occurred. However, in order to minimize this, a table 1 was reported (baseline parameters) showing no differences between the groups.
In the prospective after-implementation group, a Hawthorne effect might have influenced study results.
Implementation of a pain protocol utilizing NSAIDs and opioids decreases time to effective analgesia in patients with renal colic.
Level of evidence 4
The authors of this USA study evaluated the effect of implementation of a so-called ‘Pain Management 30 Policy’, that emphasized administering analgesics as soon as possible (30 minutes target from triage) to patients presenting with severe pain, instead of awaiting complete patient evaluation and diagnostics. The pain protocol was physician-driven.
No specific analgesics were included in this policy and acetaminophen (IV and PO), ketorolac parenteral and opiates were utilized most frequently.
Paradoxically, the authors found that after implementation of this 30 minutes analgesia target, median time to analgesia increased and proportionally less patients with severe pain received pain meds within 30 minutes. As a possible explanation, increased patient volume from 973 to 1058 patients per week (8.7% increase) was reported, as the authors found increased time was needed to allocate patients to a bed from triage.
By starting recruitment 5 months after implementation, the authors aimed to minimize bias by the Hawthorne effect (although, by the retrospective design of this study this would probably have been of little significance) and, additionally, the honeymoon-effect.
A sample size calculation was performed in order to detect a 20% improvement in proportion of patients who received analgesics within the 30 minutes target time.
The authors evaluated all eligible patients and explicitly reported the patients who were excluded from the study, minimizing selection bias.
This was a retrospective study and was vulnerable to potential confounding factors, such as Emergency Department crowding and other department- or staff-related factors. Moreover, information bias might have influenced results.
Baseline groups were not identical in sex, as significantly more female patients represented the before-implementation study group. This might have influenced the study results.
Implementing a pain management policy that required analgesic administration within 30 minutes in patients with severe pain with selected painful conditions paradoxically resulted in delayed time to analgesia as compared to a historical control group. Emergency Department crowding might have played a role.
Level of evidence 4
This was a prospective before- and after-implementation cohort study performed in India. The authors tested the hypothesis that implementation of their pain protocol would reduce time to analgesia and provide more pain relief.
The pain protocol that was implemented was both NRS based as well as based on origin of pain. Specific pharmacological recommendations were made when patients presented with abdominal pain; renal colic; fractures and dislocations; acute musculoskeletal injuries; back pain; dental pain; cardiac pain; burn injuries and migraine headache.
Proposed analgesics included IV morphine, fentanyl and acetaminophen; IM diclofenac; nerve blocks and cardiac ischemia medication as well as metoclopramide and haloperidol in migraine. Besides implementation, all emergency physicians were trained in its use, including education regarding pain and pain assessment.
They found that time to analgesia significantly improved and there was greater pain relief after implementation of the protocol in adult patients with mainly abdominal pain, acute musculoskeletal injuries and fractures / dislocations.
The pain protocol that was implemented consisted of both NRS pain scores as well as condition specific analgesic choices, utilizing as much IV administration as possible. Nerve blocks were specifically mentioned.
All patients were recruited prospectively, decreasing information bias.
A formal sample size was performed. Data were collected systematically using standard case report forms.
It is not known how long the pain protocol had been in place at the moment the post-implementation patient group was recruited. A possible 'honeymoon' period could have been present.
Patient satisfaction was measured utilizing the NASS scale. This instrument has been used in measuring pain and disability in patients with spinal conditions. It was probably not validated for use in the acute setting in patients with various causes of pain.
Additional information regarding the E5 scale could not be retrieved.
It is not clearly described who collected the data and how these persons were trained or whether they were involved in direct patient care. Therefore, socially desired answers and a Hawthorne effect could not have been ruled out.
Crowding and other departmental factors were not accounted for.
Implementation of a physician's directed pain protocol based on both NRS pain scores and different painful conditions resulted in faster administration of analgesia and superior pain relief.
Level of evidence 2b
Van Zanden JE, Wagenaar S, ter Maaten JM, Ter Maaten JC, Ligtenberg JJM.
Pain score, desire for pain treatment and effect on pain satisfaction in the emergency department: a prospective, observational study
This was a prospective, observational study from The Netherlands. In this study the authors evaluated the effectiveness of a pain protocol that was already in place and investigated decrease in pain scores, patients' satisfaction and explored desires to receive analgesics when presenting to the Emergency Department with pain from various causes.
The pain protocol was NRS based: NRS <4: acetaminophen 1000mg orally; NRS 4-7: acetaminophen 1000mg orally AND naproxen 500mg orally. In case of contra-indications for NSAIDs oxycodone 5-10mg orally was used. NRS >7: acetaminophen 1000mg IV AND diclofenac 75mg IV AND piritramide 2mg IV OR fentanyl 50ug IV (if necessary in combination with haloperidol 1mg IV). The goal of pain treatment was to achieve an NRS decrease of 2-4 NRS points.
They found that pain clinically significantly decreased when protocolized analgesics were administered (decrease 2.41 NRS points). Moreover, it seemed that the initial pain score was the only significant predictor of desiring analgesics in a multivariable analysis.
In case patients denied offered analgesia, this was mainly due to 'bearable pain' and other reasons were that they wanted to know the diagnosis first of had already used pain medication before coming to the Emergency Department. Overall, patients were more satisfied in case they received analgesics compares to patients who did not receive analgesics.
Data were collected prospectively in consecutive patients during a period of 10 days.
Both a univariate (to identify potentially significant factors) and subsequently, a multivariate logistic regression analysis were utilized to identify factors that predicted desire to receive analgesia.
Patient factors were explored by using standardized questionnaires.
This prospective study did not have a formal sample size calculation and it is unknown whether it was powered to detect clinically relevant differences in NRS pain scores.
It is not known how long the pain protocol had been in place and whether a possible 'honeymoon' period could have been present.
Patients were included in case they had pain < 3 months, which seems as a long time, it is not clear how long pain was actually present among participants. Therefore, external validity is questionable.
Research assistants utilizing prospective questionnaires could lead to socially desired answers (from patients' perspective) and a Hawthorne effect (health care staff's perspective).
Emergency Department conditions such a patient crowding were not evaluated.
Patient satisfaction was not specifically measured in pain-related questions and other factors, such as waiting time or communication might have influenced satisfaction.
The study objective was to evaluate effectiveness of the pain protocol, but medication, such as nitroglycerine and proton pump inhibitors were included as well, while not described in the protocol.
The pain protocol that was NRS based seemed effective and resulted in clinically significant decreases in pain scores and increased patient satisfaction. It is important to evaluate patient factors when analgesics are refused, as in this study nearly half of all patients did so.
Level of evidence 2b
Minhaj FS, Hoang-Nguyen M, Tenney A, Bragg A, Zhang W, Foster J, Rotoli J, Acquisto NM
Evaluation of opioid requirements in the management of renal colic after guideline implementation in the emergency department
Online ahead of print
In this retrospective cohort study from the USA, the authors evaluated the effect of a multimodal analgesia pain treatment guideline in adult patients with renal colic in the Emergency Department.
This step-wise treatment guideline that was implemented was physician-driven and included:
1. Oral acetaminophen AND IV ketorolac AND 1000mL normal saline fluid bolus
2. IV lidocaine 1.5mg/kg (max 200mg) diluted in 100mL normal saline in 20 minutes
3. Rescue medication IV opioids
Contraindications lidocaine: heart rate <60/min; heart block; arrhythmia; liver failure; allergy to amide anesthetics
The authors mention that patients who were not identified by ICD 10 codes, were identified by IV lidocaine administration as a cross-check.
Data were retrieved by a data collection team (pharmacists and Emergency Medicine residents) and the extraction methods were validated. Very meticulously data collection methods that were tested, optimized and standardized.
The before-implementation period was 1 year and the after-implementation period 9 months, not clear why these were different and whether this difference could have influenced results. No formal sample size calculation was performed.
Stone size was collected from CT or from ultrasound and it was not described which method was used in which patients. It must be noted that these methods are not equally sensitive for detecting stones (as small kidney stones and those close to the cortico-medullary junction can be quite hard to find and to measure using ultrasonography).
In the after-implementation group there were significantly more chronic opioid users. This difference might directly have influenced results as treating physicians might not prescribe opioids to opioid users as liberal as to non-chronic opioid users.
One endpoint was incidence of re-presentation. It might be that revisits to other hospitals were missed.
The use of a multimodal analgesia treatment guideline probably decreases opioid requirements in adult patients with renal colic in the Emergency Department and after discharge and the use of lidocaine in such a protocol seems safe. Protocol adherence might be a barrier though.
Level of evidence 2b