A protocol to improve analgesia use in the accident and emergency department

Last month I pointed out Prehospital use of analgesia for suspected extremity fractures. The study I am covering now looks at a similar group of patients in A&E (Accident & Emergency - the British version of ED - Emergency Department).

We aimed to assess the use of analgesia within our department and identify shortcomings. Having done this we devised a protocol for intervention and assessed whether this was successful. We chose acute skeletal injuries as a well defined group of unequivocally painful injuries that could readily be assessed.[1]

The authors don't even doubt that they have a problem. They already have a plan to fix the problem. They claim that they are looking at a well defined group of unequivocally painful injuries that could readily be assessed. Not all fractures cause moderate to severe pain. Using opioids for minor pain, even if from a fracture, is not good medicine. Even their results do not agree. The doctors in this study do seem to have been equivocal, when it came to the pain of their patients.

One hundred consecutive referrals to the fracture clinic and 100 orthopaedic admissions were analysed for the following data: (1) patient's sex and age; (2) injury sustained; (3) analgesia given in A&E department: drug used, dosage, route of administration.[1]

Using consecutive referrals is a good way to randomize to avoid selection bias. This is one of the things that Ex-Dr. Andrew Wakefield did that was a reason for eventually retracting his autistic enterocolitis paper, although it was just one of many problems with that paper.

The results of this audit were presented at a staff meeting. An analgesic protocol (figure) was then introduced and circulated through the department to all staff.[1]

Here is the protocol they came up with.

The Notes at the bottom are interesting.

1. Act on clinical suspicion - do not wait for x rays.[1]

This is nice and logical.

An x ray does not tell you anything about the pain the patient is experiencing. Waiting for an x ray does nothing to help the patient. If anything, trying to position a painful injury for x ray, will probably increase the patient's pain and is likely to make it more difficult to obtain a clear x ray.

Waiting for an x ray makes as much sense as delaying treatment of abdominal pain until after the patient is assessed by a surgeon. There is no evidence to support either excuse for delaying pain management.

2. Use Entonox during assessment.[1]

Entonox is a brand of nitrous oxide/oxygen mixture. This can decrease the amount of opioid needed to appropriately manage the patient's pain. Since the goal is pain management, not opioid dose, this is a good thing.

One huge confounder is the complete lack of mention of nitrous oxide anywhere else in the study.

Did all patients receive nitrous oxide?

What about before the study?

If they have been using nitrous oxide and using it aggressively, then their pain management is better than it appears, but why is there no other mention of nitrous oxide. They mention paracetamol, which is the Commonwealth name for acetaminophen, (commonly sold under the brand name of Tylenol), but ignore any mention of nitrous oxide? Nitrous oxide can be an excellent pain medication. Nitrous oxide is not used enough. In this study, we don't know if it was used at all, even though we are presented with a protocol stating Use Entonox during assessment.

3. All opiates must be given with an IV anti-emetic in adult patients.[1]

This could generate a couple of blog posts on its own.

The preferred route of administration for titratable medication is IV (IntraVenous). The use of IM (IntraMuscular) opiate is not a good idea. IM use may be based on a combination of ignorance and being too lazy to start an IV, although the authors suggest that this is a vascular access problem. Maybe.

Why would someone use IV anti-emetic medication in combination with IM pain medication? Given IM, the effect is delayed. Given IM, the ability to titrate is compromised. Given IM, much more medication is required to produce the same effect.

There is rarely a need for anti-emetic medication when treating patients with opioid medications - even with large doses of opioids. It is important to not give the medication as a fast push. Almost all of the adverse effects/side effects are rate/dose related. Either the rate of administration increases the frequency of the side effect, or the total increases the frequency of the side effect, or both. The rate may be most likely to induce side effects during the first dose of a medication, while the total dose may be more likely to produce side effects with the later doses.

An anti-emetic generally produces sedative effects. If we want to sedate the patient, midazolam may be a more appropriate choice.

If nausea is present, I find that morphine or fentanyl will do a great job of decreasing the nausea by decreasing the pain that appears to be causing the nausea.

In one of the Star Wars movies, Yoda explain a problem with fear. However, Yoda does not point out the obvious. Fear also has an origin. Fear is often due to ignorance.
“Fear is the path to the dark side. Fear leads to anger. Anger leads to hate. Hate leads to suffering.”
Yoda also stops with suffering, but even suffering is not the worst of it. I would write it this way -

Ignorance is the path to abuse. Ignorance leads to fear. Fear leads to denial. Denial leads to suffering. Suffering leads to self-delusion. Self-delusion leads to complacency.

When we see the suffering we have increased, we delude ourselves about the actual cause of that suffering. We pretend that we are really acting in the best interest of the patient, but aren't we really acting to protect our ignorance?

Self-delusion is what allows us to make up all sorts of excuses for abuse. Once we start deluding ourselves, where do we stop?

Once we start deluding ourselves, do we maintain any reason to examine what we do?

Their pain isn't that bad.

I know a faker when I see one.

Pain builds character.

We don't want to turn them into addicts.

It's a slippery slope.

Once we start treating one patient's pain, we will be invaded by hordes of drug seekers faking hip fractures, tib/fib fractures, humerus fractures, third degree burns, et cetera.

That doesn't hurt.

These drugs are dangerous.

Fentanyl is much too potent.

Respiratory depression will sneak up on the patient and carry the patient off to the morgue before anyone can notice.

It's for your own good.

Man up.

What if . . . ?

The use of these phrases indicates ignorance about the appropriate use of opioid medications.

We should not let the Jenny McCarthys of pain management guide our treatment decisions. We should be educating the ignorant, so that they better understand pain management.

Ignorance is only a problem if it is maintained. We are all ignorant, just on different subjects.

How we deal with our ignorance determines how we treat our patients.

Act on clinical suspicion - do not wait for x rays.[1]

There is a big problem with this study and the previous study I wrote about. They only look at fractures. They do not look at pain. They do not measure pain. They use the fracture as a surrogate for pain. Fractures often are painful.

What does an x ray tell you about the patient's pain?

Not a thing.

The x ray will tell you something about what appears to be causing the pain, but opioids are not a part of the treatment of fractures. Opioids are a part of the treatment of pain.

If the pain is not due to a fracture, but is due to a sprain, or a strain, or a tumor, or an insect bite, or a poisoning, or an embolus, or a burn, or ischemia, or any of the many other causes of pain - should we ignore that pain just because there is no fracture?


We should not ignore pain until after an x ray.

We should not ignore pain until after arrival at the hospital.

We should not ignore pain until after transport.

We should not ignore pain until after immobilization.

We should treat the pain as soon as is appropriate.

If I have a patient with an apparent hip fracture - painful deformity to a leg that appears shortened and externally rotated - I only need to know a little bit of information before initiating treatment. A quick assessment (including circulation, sensation, and movement distal to the injury, vital signs, and pain rating), a list of allergies and medications, and a history. This takes a couple of minutes. I also need IV access. This may take another couple of minutes. Then it is all a matter of give a bit of fentanyl and reassess, repeating every 5 minutes, until the pain is reduced to a level, where it is appropriate to move the patient.

If it takes 10 minutes to manage the pain, that is not a problem.

If it takes half an hour to manage the pain, that is not a problem.

If it takes over an hour to manage the pain, that is not a problem.

These are only problems if we are more worried about our hypothetical next patient than we are worried about our current patient. I am able to provide a significant benefit to this patient, but some people are more interested in some other patient who might benefit more. If we are going to be treating patients, we need to act as if we have the maturity to pay attention to the current patient, not dreaming about other patients with potentially greater life threats.

What if . . . ?

What if we actually provide competent care, rather than worry about the things that are not happening?

What if we really do have other ambulances and mutual aid agreements to deal with more than one call at a time?

What if we have exaggerated the importance of response times beyond what is remotely reasonable?

Over the following one month a further 100 consecutive referrals to fracture clinic and 100 orthopaedic admissions were then assessed using the same criteria.[1]

One nice thing about this is that it makes the percentages of the total easy to work with. 1 = 1%. There were 100 patients in each of the original groups, so it is easy to compare. That is a total of 400 patients, but it would be better with a lot more patients.



These were divided into four groups: forearm, lower leg/ankle, hand/foot, and others. The numbers of fractures assessed in the initial and repeat audit are shown in table 1. The analgesia given is detailed in table 2. Fracture clinic referrals receiving unsatisfactory analgesia were reduced from 91 % to 69%, a difference of 22% (95% confidence interval 10.9% to 33.1%, P < 0.001).[1]

Let's look at that a little bit differently. If 91% of these patients are receiving unsatisfactory analgesia, then we might conclude that 9% of patients are receiving satisfactory analgesia. After the protocol, this increases to 31%. They describe this as a 22% improvement.

If they were trying to increase sales of a drug, would they be reporting these numbers this way? If a competing drug showed an effect in 3% of patients, and the drug being studied showed an effect in 6% of patients, would the advertisements claim 3% more effective than our competitor? No, they would not. The drug would have doubled the effect. Whether that doubling is relevant depends on a lot of things. The ads would probably claim 100% more effective than our competitor or, the same information expressed for a different effect, 200% as effective as our competitor.

The increase in patients receiving satisfactory analgesia increases from 9% to 31% an increase of 22% of the total number of patients, but an increase of 244% above the original 9%. 22 is almost 2 1/2 times 9.

There are several things to mention about this.

91% unsatisfactory is pretty bad. The authors acknowledged that they were not happy with that. They sought to improve things. The improvement was dramatic.

Imagine if you are back in grade school and you get a score of 9% on a test. We will assume that this was not a grade on a multiple choice test, because unless there were an average of eleven choices for each question, you would be expected to do better by just guessing.

If you have a 9% score on your first test, but get 31% on the next test, are you going to be bragging to your parents? No, you probably are not. It is more likely that you will be offering to get the mail every day for a while. You may be engaging in a bit of triage of mail from the school, or hacking the parents' email to redirect school emails to you.

65% is the minimum score to be considered passing in many places. If you took 31% and doubled it, you still would not receive a passing score. Can you claim that you have learned 244% more between Test 1 and Test 2? That may depend on how little you knew at the time of Test 1.

There is no Limitations section, so I created a sample of what that might include.

What is not reported in this study:

Was there any change in availability of opioids?

Was there any change in supplier of opioids?

Was there any change in requirements for documentation?

Had there been any event before the first part of this study that had staff especially worried about pain management, a worry that might have eased with time?

Were there any incentives offered with the introduction of the new protocol?

Was there anything done to track longer term changes in treatment?

Was there any change in the staffing (different hours, new hires, vacation coverage, et cetera)?

Were there any other changes to the way that the patients were treated?

As you can see, there are many things that are not described as being controlled for in the study. Therefore, we should not presume that any of them have been controlled for. What is the confidence interval based on? Just statistics, or is there some understanding of the potential variables? Since this is an early pain management study, I do not expect that there is the kind of awareness of the relevant variables that the same authors would have today.

How sad is it that a study of pain management published in 1996 is considered early?

What did we base our treatment on before then? Unfortunately, we based our treatment on a lot of misinformation. We still do this, but we are getting better at pointing out the misinformation and correcting those spreading misinformation.

That was the Fracture Clinic. Initially, 91% of patients did not receive any medication for their fractures. None of those patients declined pain medicine.

Since I am writing mostly about opioid medication for pain management, I should mention that in the first part, the number of patients receiving opioids is zero IM and zero IV. 4 patients received paracetamol. A whopping 4 patients received this aggressive treatment. They could have done just as well by going to their medicine cabinet and taking their own over-the-counter medicine.

In the second part, those numbers zoom all the way up to IM opioids = 3 and IV opioids = 3. This is not impressive. This is just something is better than nothing.

For those in America, reading this and saying to yourself, I'm just glad I am in America.

Do you really think that pain management is/was better in America?

If you do think pain management is/was better in America, what do you base that on?

These were also divided into four groups: neck of femur, forearm, lower leg/ankle, and others. The numbers of fractures assessed are given in table 3 and the analgesics used in table 4. Orthopaedic admissions receiving unsatisfactory analgesia were reduced from 39% to 22%, a fall of 17% (95% confidence interval 4.2% to 29.8%, P = 0.009). The number of orthopaedic admissions receiving intravenous opiates increased by 28%, from 9% to 37% (95% confidence interval 16.3% to 39.7%, P < 0.001). The most appreciable improvements in analgesia used were obtained treating patients with fractured neck of femur (table 5).[1]

Orthopaedic admissions receiving unsatisfactory analgesia were reduced from 39% to 22% It is great that we are dealing with numbers that are much higher in this group. Still, this implies that only 61% rated their pain management as acceptable. That improved to 78% after the introduction of their protocol.

Intravenous analgesia is superior to intramuscular analgesia for reasons of speed of onset, reliability of uptake, and the ability to titrate doses to response.4 6 The number of orthopaedic admissions receiving intravenous opiates increased by 28% following the introduction of the protocol.[1]

I have no argument with the reasoning, but they should have already been aware of that before the start of the study. Just as they already should have been aware of this little tidbit - Pain Hurts.

We were particularly successful in improving analgesia for patients with fractured neck of femur. Perhaps the value of intravenous titration of doses is better appreciated in the elderly population who typically suffer this injury.[1]

I would reword that, but I don't think that I am really changing their meaning.

Perhaps the doctors in this study are more comfortable using intravenous titration of doses to better manage the pain in the elderly population who typically suffer this injury (hip fracture).

And by better I mean more safely and with fewer side effects and with faster onset and with a more accurate end point.

This was from the dark ages of pain management, which was only 14 years ago where this study was done. Other places may not have progressed beyond this point, even now. It is good that the authors looked at their practices. I expect that they are much more aggressive today than what is shown in this study.

And for the punchline we have this -

Setting- University teaching hospital.[1]

These are supposed to be doctors with the most education, the most resources, the most experience, et cetera. These are supposed to be the doctors who are educating the doctors of the next generation. Even in an academic medical center, they appear to have been held back by the mythology of pain medicine, but they are changing that mythology.

The management of pain in acute trauma is often neglected.1 2 Patients arriving at A&E departments with acute trauma are unlikely to have received sufficient analgesia,3 so responsibility lies with the attending doctor.[1]

Sometimes the best way to encourage the doctors to treat pain appropriately is to authorize EMS to treat pain appropriately. Doctors are capable of both compassion and learning, but sometimes they do seem to need a push. The authors of this study did provide a bit of a push. Did things continue to progress, or did they revert to whatever was the mean at this A&E?


^ 1 A protocol to improve analgesia use in the accident and emergency department.
Goodacre SW, Roden RK.
J Accid Emerg Med. 1996 May;13(3):177-9.
PMID: 8733653 [PubMed - indexed for MEDLINE]

Free Full Text from PubMed Central

Free Full Text PDF from PubMed Central



Anonymous said...

Greate post.
You should lookt at ketamin for severe pain management, greate drug when combined with midazolam or low dose propofol. I with you on fentanyl. In Sweden we use it in many organizations with good result. Ketamin, fentanyl, morphine and one Nsaid drug is in our amb. All puched Iv.

Do you use ketamin in the US for pain management or just for RSI in hypotensive pt?

Thanks for a create blog.

Rogue Medic said...


Thank you.

We do not use ketamine in Pennsylvania. I believe that there are some EMS agencies that do use ketamine. From what I have read, I would expect that it is very useful.

Expect that propofol will be very hard to find for a while.

Anesthesiology News writes -

Teva Exits Propofol Market
Move, which leaves two U.S. suppliers, seen exacerbating ongoing shortage

You may have to register. Registration is free. If you do not want to register, email me and I will send you a copy of the article.


Here is a link to the FDA Drug Shortage page information on propofol.

Teva Pharmaceuticals has discontinued their propofol injection.

Hospira Inc. recalled specific lots because some containers may contain particulate matter. Hospira has implemented improvements to its manufacturing process and is coordinating with the FDA to begin distribution of propofol manufactured under a new process as soon as possible.

APP has increased production in response to the increased demand. Please see information on availability in the Related Information section.