Headache headaches! A podcast from the PHARM
Hi all
I got involved in a twitter debate about the work up of headaches in ED, why is it that GP “miss” serious intracranial pathology from time to time and what is the current ‘best practice’ when it comes to the acute headache.
There have been some really big, game-changing papers out in the last few years which might change the way you approach the headache in ED or in GP.
Minh le Cong over at the PHARM podcast recorded this PODCAST with Dr Seth Trueger (@MDaware) and myself last week.
I was hoping to go in to bat for the much maligned GPs who are working with a much bigger haystack when it comes to finding needles. Now let me know – what is your threshold, what features on history or exam will make you jump for the CT or referral papers?
There is not any recognised decision algorithm that I am aware of – so lets hear yours.
Casey
Killer Keeweedoc
Clinical Case 024: Livin’ the ketamine dream
Clinical Case 048: We Ask an Epidural Expert
About The Author
Casey Parker
I am a GP working in Broome, NW of Western Australia. I work as a hospital DMO (District Med Officer) doing Emergency, Anaesthestics, some Obstetrics and a lot of miscellaneous primary care. Also on the web as @broomedocs | + Casey Parker | Contact
This podcast — almost impossibly hard to hear Dr Seth, unfortunately, so I gave up at 9’30 — seems to be about the management of headache in ED, a totally different population to the population of people presenting with headache to their GP. I suspect most ED physicians have no concept of the magnitude of the number and diversity of headache presentations to the GP, of which I suspect at least 98% never get referred on, so I’d love to see some ED physicians put into a GP clinic and told to get on with it without CT and LP. I wonder what the onward referral rates would be (and / or the nervous breakdown rates of the physicians!)?
As a GP, you’re looking for novel presentations in that patient in the wider context of their life and regular presentation pattern. Obviously, sudden onset new headache, any associated features such as LOC, neck stiffness, neuro signs etc will occasion referral, but in 15 years as a GP in the UK I don’t think I ever sent in more than one patient a year (maybe two?) with headache as the primary symptom and I do not have a long backlog of negligence cases stretching behind me for missed subarachnoid or other pathology (or at least they haven’t materialised so far).
The two populations are so totally different as not to be worth comparing in the same article or podcast. If I am a GP and I send in one patient a year with headache then it is probably worth CT and LP, irrespective of the history and exam, because I’ve filtered that patient out for you from thousands, So don’t slag me off if you think I’ve sent you crap because if every single one I send you has serious pathology then my specificity is way too high and I’m leaving a lot of people out there who I should have sent in.
thanks for the honest feedback David! Yes the audio of Seth is low volume and you can sort of hear it ok with headphones but I agree,it would have been better to redo his part or edit it out!
I wanted to ask as was not sure what you meant . How are patients presenting with severe headache to GP or ED different?
Hi Minh,
They are two totally different populations because patients presenting to ED have been screened for severity, either by their GP or by the patients themselves by calling an ambulance or electing to attend the ED rather than to book a routine appointment at their GP. That means the ED population will have a much, much higher rate of organic pathology.
Cheers,
David
I think I can see both arguments here. GP vs. ED.
I agree that on a population-basis the GPs are seeing and screening the higher risk patients – and this is a tough job. Knowing when to pull the trigger on a headache and refer to ED is tough – if you are right 100% of the time then you have not referred enough patients! However, it is impractical to send in everyone with a headache worse than X….
Hence the suggestions around wait and see in some cases
However, in the context of a discussion around pre-test probability relating to HX and Exam – there really is no difference for the individual patient as their symptoms / signs are the same regardless of location. IF you have any signs or symptoms that are concerning, no matter which department, then we need to be conservative and image / work-up in the face of serious badness. I have made my thoughts on the topic clear at: http://broomedocs.com/2012/07/on-evidence-education-errors-ego-and-expert-intuition/
The fact that screening out the needles is a really tough job is not the individual patients problem – it is the poor GPs.
Casey
Casey,
You’re coming at this like a hospital doctor, which is fine for the ED location, but the discussion was billed as being for GPs too.
If a patient presents with a headache across the back of the head, as if they have been hit by a cricket bat, photophobia and vomiting, then it is a pretty easy call in both locations that the next steps in management are likely to include a CT and then probably LP. However, what we’re talking about are the fuzzy situations, the situations where it isn’t clear and which we all face on a daily basis. When you’re standing there, trying to decide whether to investigate further in the ED, you are dealing with a very high risk, selected population, as compared to the ones the GP is seeing. If GPs sent up every fuzzy headache then the ED would be swamped.
The patient population from which your particular patient is selected is extremely relevant, at least if you’re a betting man (or woman) and I’d say all we doctors are.
Cheers,
David
I appreciate what you are getting at David but respectfully disagree.
Whether I am working an ED shift or consulting in GP setting, the patient with severe headache requires a standard level of assessment and workup.
The Linn et al study of GP patients with headache(severe) cited in this UK position paper http://secure.collemergencymed.ac.uk/asp/document.asp?ID=5074
found 37 out of 148 patients had proven SAH as cause. Thats a lot!
This review article by a UK neurologist suggests GP patients presenting with severe headache have 1 in 4 to 1 in 8 odds of an ICH/SAH
http://jnnp.bmj.com/content/72/suppl_2/ii33.full
The key is an adequate history and exam as Casey mentioned. This is supported by this retrospective study on ED evaluation of severe headache
http://qjmed.oxfordjournals.org/content/101/6/435.full
And in that study they cite that the ED miss rate for SAH was as high as 5.4% or more! here is the full article reference for that figure
http://stroke.ahajournals.org/content/38/4/1216.full.pdf
Those authors of that paper claim it is the lower acuity patient ( like what we see in GP land) that are at higher risk of misdiagnosis in the ED. I think the same caution can be translated to GP presentations!
We miss SAH in ED as well as GP presentations. Just because a patient is not as sick and turn up in your GP office does not mean they have a lower risk of having a life changing condition. Sure play the odds but dont rationalise it by location.
I would argue, in the US, they haven’t been all been screened for severity. I had a woman take an ambulance ride the other day for a sore throat. The line is blurring because the “automatic” increase in severity we are taught exists in the ED is slowly being eroded by those using the ED because they don’t have insurance. I’m not sure what percentage of my practice is primary care, but it is a decent chunk.
And please don’t get off on the ‘well you use ct and lp and fast diagnostic testing’ argument… It implies EPs have no clinical acumen and are dependent on testing.
I respect my GP colleagues, I ask the same from them.
Sub topic but this might help you. Its got what we all like -good boxed summaries!
http://www.epocrates.com/dacc/1301/ThunderclapHeadacheBMJ1301.pdf
Minh,
This is a fascinating discussion and I am finding it extremely educational, so thank-you.
The very first case cited in the review in the Journal of Neurology, Neurosurgery and Psychiatry which you refer to ( http://jnnp.bmj.com/content/72/suppl_2/ii33.full ) illustrates exactly the point I am making. The review is written by a neurologist and this is the first paragraph of the case study:
<>
Come on, the initial presentation of this headache, as described above, in someone with known migraine, is a standard presentation in general practice. You see it every week. It could easily have been a further episode of her known migraine, but a bit worse. It wasn’t, but it could have been. If GPs were to send in every patient with migraine who was having a headache that was worse than their usual one (even their worst one ever — and how many times will people tell you ‘Oh this is the worst one ever, Doc’? Some people — drama queens — will tell you this every single time!) then the EDs would be swamped and the ED docs would be even more scathing about GPs than they already are. Hospital doctors have no idea of the range of ‘fuzzy’ presentations in general practice and how hard it is to pick out the wheat from the chaff when faced with a deluge of symptoms and presentations.
On Day 1 I think that GP 1 made a reasonable, or at least understandable, call, based on the presentation as described and the population of people he usually sees — people with migraine and many other causes of headache — out in the community. The specificity of an admission at that point (i.e. the number of people he admitted in this situation who turned out to have serious pathology) would be vanishingly low. By Day 6, the headache was continuing, another GP reviewed the patient and decided to send them in. That is a totally different situation and that patient has now entered a smaller and more select group of patients, filtered out by their GP, as potentially having serious pathology. The specificity now is much, much higher for serious pathology detected on an admission.
And, by the way, the neurologist’s language is sloppy and symptomatic of the perennial misunderstanding I am talking about. He says “On day 6 she was seen by a different GP who recognised the significance of her symptoms and referred her urgently.” By this, he sort of implies the GP on Day 1 should have made this call, but neglects the fact that on Day 6 the new GP was seeing a totally different, much more evolved situation. I don’t think it is at all clear that the second GP would have admitted the patient, had he seen her on Day 1. I am always telling med students and junior docs not to immediately jump to conclusions and judge referring doctors harshly because they think the situation when they see the patient is different to the picture they were initially given. Situations change, and change often very rapidly (even the fact of a preceding medical consultation and someone having taken a history changes the situation) and the retrospectoscope is a dangerous instrument, especially in the hands of a hospital physician.
Now, are there lessons to be learned for GP 1? Yes, probably. In retrospect, it is certainly worth that GP revisiting the initial situation and reconsidering their decision in the light of events. Maybe there were more features than have been described here and maybe in retrospect he will remember that actually he did have more of a bad feeling about this one than he usually had with recurrent migraines. Maybe, but maybe not.
Considering everything from the point of the view of the ED physician, and neglecting the reality of the two different populations they see, does nothing to help GPs sort serious pathology from the vast morass of daily symptoms they see, much greater than that of the ED physician.
In summary, it is bloody hard deciding who is ill at the best of times, but I still think it is a damned sight harder in general practice because so many fewer of your patients have serious pathology, the range of symptomatology is so much greater and, critically, even serious pathology is often viewed at an earlier and less differentiated stage of evolution in its life history, so your betting odds have to be different or else you can’t practise medicine.
Cheers!
David
Hi David
Agree that the initial presentation is common in GP. However the “watch and wait” strategy is dependent on timely follow-up.
If you were opting for waiting over referring – in the face of potential badness – then I would imagine a follow-up window of 12 – 18 hours is justified. If the headache is gone, then great – but if it is worse / still laying in bed, you have now crossed the threshold where a CT / LP is warranted I think.
The patient needs to be brought back regardless. the 6 day window is just too long in my opinion.
Discharge instructions from ED or “safety netting” in GP terminology are crucial in headache, chest pain and abdo pain especially.
Casey
The case study first paragraph didn’t get posted for some reason above. Here it is:
Case study
A 42 year old woman became aware of a mild global headache while warming up for her aerobic class. Several minutes later (before the class started), she had sudden exacerbation of her headache, followed by vomiting and photophobia. She was seen by her GP, who diagnosed migraine and gave her intramuscular morphine and prochlorperazine. She spent the next four days in bed with her worst ever headache (she had a previous history of migraine without aura, but this was much worse). On day 6 she was seen by a different GP who recognised the significance of her symptoms and referred her urgently.
Hi Casey,
Totally agree on the importance of adequate safety netting and DOCUMENTING adequate safety netting.
The articles Minh has posted are excellent and I am finding this so valuable — thank-you to you both. There is one interesting reductio conclusion, though, which I think nicely highlights how the difference in incidence and therefore percentage hit rates between the GP and the ED populations is very important.
In this article it cites that 91 patients with acute headache were seen in the Western General Hospital ED in Edinburgh in a 3 month period: http://qjmed.oxfordjournals.org/content/101/6/435.full The Western covers a very large catchment area, so we can presume GPs in its catchment area had many thousands of consultations for headache in that period, some very acute, many being described as ‘the worst the person has ever had’ and others much less acute. However, GPs saw many more times the number of acute headaches than were seen in the ED. Now, in this article, we see that the percentage missed diagnosis for SAH in the ED is 5.4%, in quite a large, convincing study: http://stroke.ahajournals.org/content/38/4/1216.full.pdf Putting the two together, therefore, we have to conclude that had a CT / LP been performed on the 76% (67) of the patients who did not have one at the Western General, the ‘Number Needed to LP’ to find one case of previously unsuspected SAH would have been 13.64. This is a pretty good ratio, an excellent one in fact, and so the reductio conclusion is that all patients presenting to an ED with acute headache should have a CT / LP, irrespective of the degree of clinical suspicion the admitting clinician has. Puncture them all! Which is what I said in my first post on this subject: if a GP sends up a patient with an acute headache to ED then just investigate them because the hit rate will be very high.
Would we come to a similar conclusion when looking at all cases of ‘acute headache’ in the GP population? Absolutely not because the ‘Number Needed to LP’ will be astronomically higher. Different populations, different odds…
The messages I take home are:
1. Have a damned good reason not to investigate any presentation to the limit in the ED setting, if for no other reason than that your patient has been pre-screened for you in one of a variety of ways.
2. The lot of the GP is an onerous one and we (they) are grossly misunderstood souls, labouring away in obfuscation and darkness for the benefit of humanity, at gross risk to ourselves (cumulative risk), but lesser risk to each individual patient we see (point risk).
Cheers,
David
HI David
Yeah its not always straightforward in GP land thats for sure.
But if you want to look at some scientific research rather than anecdote, into GP presentations of severe headache, Linn et al as I quoted above shows that with severe headache GP presentations in that cohort of 148 GP patients, 37 ended up having proven SAH. Be careful believing low acuity patients are of lower risk.
I think a GP is more than reasonable if after an adequate history and exam, an alternative diagnosis to SAH is made. Appropriate followup is the key, regardless of ED or GP presentation. You dont have to order the CT head straight away if you think its not a SAH but you sure need to followup the next day and document your followup. You know as well as I do that many GPs order analgesia, parenteral even, over the phone if they think the person has a migraine, often without even examining the patient. Yes we play the odds
The Case 1 you cite, the learning I get from that is the history of sudden headache is a red flag not to ignore. And followup would have helped. If the first GP had arranged a followup plan..even a phone call to check, then it is likely the diagnosis would have been made by them on followup.
Sure its hard deciding who has got the life threatening condition but thats why we did medicine!
Hi Minh,
Yep, follow up and safety net is key and documenting it. In all situations.
Thanks again for a great discussion. Certainly thought-provoking for me and raised my awareness of all these issues greatly.
Cheers,
David
Thanks for your feedback David !
This UK discussion forum might interest both of you
Just saw it on twitter
Nice to know our patients support the GP care we provide ! … Mostly …
Which forum is that, Minh? I don’t think you posted a link.