MIC CAVAZZINI: Welcome to Pomegranate, podcast of the Royal Australasian College of Physicians. I’m Mic Cavazzini, and before we get into today’s show I’d like to mention the creation of our Virtual Editorial Group. We’re looking for listeners who have a keen ear for podcasts and who’d like to help us develop topics of interest. The editorial group would also provide feedback on the show’s content before it’s published each month. It’s not a huge commitment of time and would be a fun way of earning a few CPD points. Please check out the Pomegranate website for more information, and or email email@example.com to express your interest.
Today’s episode is about low-value care in general paediatrics. Almost three quarters of physicians surveyed in the U.S. admit to ordering at least one unnecessary test, procedure or treatment every week. EVOLVE is the RACP’s initiative to reduce clinical practices that don’t do any good, according to the current evidence base. We talked about this and the Choosing Wisely campaign back in Episode 10.
EVOLVE has just published a list of the top five practices in general paediatrics that need to be pulled back. The list was boiled down from consultations with Fellows of the Paediatrics and Child Health Division, three of whom I cornered at a recent masterclass:
MIC CAVAZZINI: …and just for fun, why don’t I start by asking each of you to read off one of the items off this list and then introduce yourself.
HARRIET HISCOCK: OK so, “Do not routinely prescribe oral antibiotics to children with fever without an identified specific bacterial infection.”
I'm Harriet Hiscock, I'm a paediatrician from the Royal Children’s Hospital in Melbourne where I'm director of the Health Services Research Unit. And I also run the Australian Paediatric Research Network.
SARAH DALTON: “Do not routinely undertake chest radiography for the diagnosis of bronchiolitis in children or routinely prescribe salbutamol or systemic corticosteroids to treat bronchiolitis in children.”
Hi, my name is Sarah Dalton. I'm a paediatric emergency doctor at the Children’s Hospital at Westmead. I also work at the Agency for Clinical Innovation in New South Wales, and I'm the president for paediatrics at the Royal Australasian College of Physicians.
HAMISH MCCAY: “Do not routinely order test radiography for the diagnosis of asthma in children.”
Kia ora, my name is Hamish McCay, I am a paediatrician and clinical director of paediatrics at Waikato Hospital in Hamilton in New Zealand.
MIC CAVAZZINI: Maybe I’ll jump in with the next one: “Do not routinely treat gastroesophageal reflux disease in infants with acid suppression therapy”.
And finally, “Do not routinely order abdominal radiography for the diagnosis of non-specific abdominal pain in children.”
Now, is there any sense as to whether these low-value practices are more coming from junior staff who are trying too hard to show they're doing something—or perhaps senior staff who are stuck on inferences from an outdated evidence base?
HAMISH MCCAY: I think there’s a degree of both in that, and certainly the trainee feels the pressure of the boss. And no one wants to be the person on the ward round that next morning who gets asked, “So what was the CRP?” and you go, “Oh, I didn't get one.”
Certainly my juniors well know that within our team there are some people who have a particular tendency to want particular tests and others who don’t. And to be honest they do an amazing job of varying their practice day to day, week to week depending on which boss is on. And I think particularly as senior clinicians, we need to be very careful about what we model to our juniors.
HARRIET HISCOCK: Yeah I agree, Hamish. There was a study in the U.S. in adult medicine and they just targeted the junior doctors to reduce unnecessary cardiac testing in veterans. And it was a spectacular failure because the junior doctors were just acting on the orders of their senior consultants. So I think if any hospital or system is trying to change low-value care they need to do some careful ground work first to understand what are the drivers, where’s it coming from, and target it. And that might be different at different sites.
SARAH DALTON: I remember starting in paediatrics and feeling really scared that you had to be a really special, good doctor to do paediatrics. What I learnt was a long list of differential tests, and one way that I could show how expert I was was whenever I was consulted about a patient to say, “Well, have you done the hyperglycaemia protocol?” or, “Have you done the chest x-ray in the new patient in with asthma? Because did you know there's a small chance that they might have leukaemia, and you shouldn’t give steroids unless we've actually excluded that.” So I think we train people to prove their worth by the number of tests they do, and I think we need to change that so they prove their worth by the number of tests they don’t do.
One final thing that I learnt a long time ago from a very imminent man called William Runciman was to bring back HOPE in medicine, which is: History, Observation and Physical Examination—and not to rely so much on tests.
MIC CAVAZZINI: So I think you've already addressed my next question, which was how to change that culture—what kind of conversations need to take place in the hospital about low-value care to make more permissive the diagnostic procedures that you are recommending?
HARRIET HISCOCK: I think it’s almost more than conversations, so you need multiple approaches. And it will be a combination of yes, education, but audit and feedback’s really important. So you become clinician 1, 2, 3, 4, or 5 and you see what your chest x-ray ordering was for the month for asthma, and if you’re an outlier that is one of the strongest things to see it visually in front of your peers. They may not know who you are but you pretty much want to get back into the pack, so that helps. And the other approach if there’s an electronic medical record in the mix, then putting in clinical decision support at the time of care to say “Oop, you don't need to do a chest x-ray or a blood test here.”
HAMISH MCCAY: I think also there’s a bit of an issue for those in provincial centres that they feel a bit of the steely gaze of the tertiary centre and they don't want to, you know, get stuff wrong. And certainly a couple of places I worked in my training there were comments about, “We just want to make sure we don’t miss anything.”
HARRIET HISCOCK: Yeah it was really bought out for me by a study of paediatric hospitals in America. They've got an association called PHIS of 42 hospitals and have shown very clearly that if you’re in a hospital that doesn't do much testing, that’s across the board. If you’re in a culture in a hospital that does do a lot of testing, then you get everything: FBEs, U&Es, liver tests, CRP, chest x-rays, CT heads, everything goes up. So it’s a really cultural problem that people have to tackle at their individual hospital sites.
MIC CAVAZZINI: It’s no one practice, it’s really the—
HARIETT HISCOCK: —it goes across the board in these institutions.
MIC CAVAZZINI: Well let's focus on the top five. The first is of course, “Do not routinely prescribe oral antibiotics to children with fever without an identified bacterial infection.” Because the literature shows that the vast majority of children presenting with fever don't have a bacterial infection, and therefore won’t benefit from antibiotics. Is the persistence of this practice largely driven by patients? Is that your experience?
SARAH DALTON: As a paediatric emergency physician I would have this conversation many, many times every day as most paediatricians would. And the number of times I've sat down and literally said, “There are these things called viruses and there are these things called bacteria and this is how we know what the difference is”—that's probably a ten- or fifteen-minute conversation depending on how worried the parents are. So sometimes I think another driver is, it’s just quicker to give people what they want than it is to spend the time explaining what’s going on.
HARRIET HISCOCK: So I wonder, you know, how do we go about trying to solve this? And I was in clinic yesterday with a very irritable unsettled baby, mum and dad in tears—and I talked to them about temperament and their eyes glazed over. And then I went to a website and I showed them a video of an intense baby and they went, “Oh now I understand.”
I'm wondering if we should be more clever in how we impart information other than just talking at parents and whether there’s apps, visual things, videos that we can use that can be played in emergency department waiting rooms, for example.
SARAH DALTON: We really must talk about antimicrobial stewardship as well. One of the reasons we don’t use antibiotics all the time is because if we use them all the time they won’t work anymore. And that’s actually something I talk to parents about.
MIC CAVAZZINI: Looking at the evidence that went into this was a large cohort study of almost a thousand febrile patients up to the age of two, in which 13 per cent had a serious bacterial infection such as pneumonia or UTI. But comparing those that had bacterial infection and those that didn't there was actually no difference in the mean temperature or their white blood cell count. So does this really go back to your point that there is no rational way of determining who might benefit from the antibiotics, so why give them?
HAMISH MCCAY: I think that we need to be very clear that this is about the prescription of oral antibiotics for children with a high temperature and who don’t look that sick. Now we’re not talking about the kid with, you know, a prolonged capillary refill time, a tachycardia, who you’re about to give the fluid bolus to. That kid needs IV antibiotics and, you know, that’s absolutely appropriate. But we’re talking about the child who’s a bit hot and a bit unhappy and you can’t really figure out what’s going on and it’s a “Hmmm, maybe just in case.”
And I think the cohort study that you’re quoting probably blurs the lines a little bit because it will involve a number of patients with that. And having trained through a meningococcal epidemic in New Zealand, we would have never advised anyone to give those kids oral antibiotics. If you were really concerned it was “give intramuscular antibiotics and send them in.”
MIC CAVAZZINI: Three of the recommendations on this EVOLVE list do mention the use of x-ray imaging, so remind us Harriet again why x-rays are so bad for kids and what a reasonable dosage might be?
HARRIET HISCOCK: So the reason why we don’t want to do X-rays is that they cause radiation and cumulative radiation does pose risk to the child—for later cancers, for example. So it’s not a trivial issue. Certainly abdominal x-rays actually have more radiation than chest x-rays. And these practices were specifically chosen because it’s been shown that these x-rays aren't needed for a diagnosis and/or don’t change management.
But I’m not sure that parents have got that message about radiation, and a couple of days ago in Melbourne we had an asthma thunderstorm with 500 children hit the Royal Children’s Hospital Emergency Department in 24 hours—the largest presentation in history. And I was speaking to a colleague actually at a private hospital in Melbourne and she said the demand for chest x-rays went skyrocketing for both adults and children. And I said, “Well, they shouldn’t be doing chest x-rays.” And she said, “Tell that to the parents who turn up to our private facility and get charged a $500 facility fee to come and then don't get an x-ray.”
So a lot of care for children goes on in Australia in private hospitals and we need to bring them into the conversation and we need to put this information in parents’ hands about unnecessary radiation of their child.
MIC CAVAZZINI: In fact, this unprecedented asthma storm in Melbourne is a good segue to jump to the third item on the list, because as you say for diagnosis of wheezing disorders in children, chest x-rays generally don't provide clinically relevant information. Can you guys recap some of the literature that went into this point?
HAMISH MCCAY: Well I suppose there was a Swedish study in preschool children who were newly diagnosed with asthma, which showed that follow-up x-rays were normal in 50 out of 59 cases, and those who had abnormal x-rays it didn't actually change treatment. And certainly one of the things that we try and drum into our trainees, and I think sometimes have to drum into ourselves as well, is that the purpose of any investigation is not to make a diagnosis—it’s actually to change management. And otherwise it simply becomes a scientific nicety.
It’s very easy to take an x-ray in asthma and go “Oh yeah, there’s a bit of hyperinflation and a bit of oh the lung markers are up and oh yeah, and you can always see something behind the heart.” But honestly is that going to change what you do for that child?
SARAH DALTON: I also think there’s a driver around doing x-rays which is partly that it’s a delay tactic. “I don’t really know what else to do, so I’ll order an x-ray and that will buy me a little bit of time.” And it’s actually easier to do than a blood test and the doctor doesn’t have to hurt the child. So you can send the child to a nearby room, not see what happens and then what happens next for you is you look at this x-ray. So it’s a very easy thing to do that makes us feel good but doesn’t always change practice.
HAMISH MCCAY: Totally, that was so me as a junior doctor—but of course I don’t do that anymore.
SARAH DALTON: No one does, don't worry!
MIC CAVAZZINI: You talk about changing practice—in another U.S. study of children admitted through ED with acute asthma exacerbations, only 18 out of 180 had radiographic results that changed their treatment.
HAMISH MCCAY: I think there’s also a risk that you do find something a bit odd and I think the classic thing in asthma and bronchiolitis is you find a degree of collapse, which will resolve with time but, “Ooh, is it collapse? Is it consolidation? I'm not sure maybe I should start some antibiotics.” So there is a risk of causing harm for our patients.
MIC CAVAZZINI: And according to another study pneumonia is also unlikely to be diagnosed by chest x-ray in children presenting to ED with wheezing. But pneumonia and asthma are treated very differently. So can effective treatment decisions be made based on other clinical assessments apart from an x-ray?
HAMISH MCCAY: Absolutely, and one of the difficulties that some of my junior doctors have is that when they get investigation done and the result isn't what they were expecting, they start to doubt their own clinical judgement. I've had junior doctors who’ve had a child who’s febrile, they’re tachycardic, they have creps easily audible at the left base and they get a chest x-ray to confirm for themselves left lower lobe pneumonia, and the x-ray doesn’t look that unremarkable—which we know happens.
And then they start going, “Oh well, maybe I was wrong,” and you have to go back and say to them, “No you’re not wrong, the x-ray’s just not showing it.” And, you know, pneumonia is not a radiological diagnosis—it’s a clinical diagnosis. So again, what are the potential consequences of doing an investigation that is not only not helpful but distracting?
SARAH DALTON: I think it’s worth stating like with many of these things, there are times when you do need to do a chest x-ray in asthma. The most important thing though is knowing where those times are and specifically we say do not routinely order chest x-ray for children with asthma. But children with severe asthma absolutely should have an x-ray because we’re looking for serious complications like pneumothorax.
MIC CAVAZZINI: Now we will come back to Hamish’s pet subject, “Don’t undertake routine chest radiography for the diagnosis of bronchiolitis,” because the evidence shows that chest x-rays don’t discriminate well between bronchiolitis and other forms of lower respiratory tract infection. Now, a recent Cochrane Review found that use of chest x-rays did not reduce duration of illness or severity of symptoms in patients. But is it hard to shake the intuition among physicians that more information is better?
HAMISH MCCAY: As you say, when we see the hyperinflation we have our diagnosis confirmed to us and we can pat ourselves on the back and say what good clinicians we are and how good our clinical judgement is. And I think there’s a degree of that with all these investigations.
So there is a really nice article from the Journal of Paediatrics which found that when they had infants under 2 presenting with barndoor-typical symptoms of bronchiolitis in fact you needed to scan 133 patients to identify one which might suggest an alternative diagnosis. It’s one of those things—realistically it’s an exercise in futility.
MIC CAVAZZINI: Now let's move on to the children that are diagnosed with bronchiolitis—the EVOLVE list tells us not to routinely prescribe salbutamol or systemic corticosteroids for the treatment. And a 2014 Cochrane Review of 30 different trials found that salbutamol did not improve oxygen saturation, did not improve duration of hospitalisation, and in outpatients there was no effect on resolution of illness and hospital readmission rates. Would these findings come as a surprise to many listeners?
HAMISH MCCAY: Oh look, you know, the theory is good there—that steroids maybe should make a difference, or maybe salbutamol would because it will bronchodilate. But the reality is that even if you’re using good scientific premise but have no evidence or evidence to the contrary, realistically you’re practicing alternative medicine. And I think we just need to be very real with ourselves about, you know, this stuff doesn’t work.
MIC CAVAZZINI: Well, more than not working it has the potential for adverse side effects—what are some of the side effects of bronchodilators?
SARAH DALTON: One of the commonest things I see is tachycardia and vomiting. And the reason that that’s concerning is that I’ll often hear about a child who came in with wheeze and a bit breathless, who has had a little bit of salbutamol to see if it works, but they're looking much worse now and they've got a temperature now and their heart rate’s really high and they're vomiting and the mother is really concerned. And then we think, “Well, this child potentially could have sepsis so we’d better give them some IV antibiotics and bring them into hospital for a couple of days until we’re sure that the blood cultures are clear.” But we would never have been there if we hadn't caused the tachycardia to start with.
MIC CAVAZZINI: And there are similar underwhelming data for the benefit of corticosteroids. Again, a Cochrane Review of over two and a half thousand participants showed there was no clinically relevant effect on outpatient admissions, on length of hospitalisation, clinical scores or oxygen saturation.
SARAH DALTON: And in fact there was a study in 2013 that found even in individual patients there was no difference in symptom severity comparing episodes of bronchiolitis that were treated with corticosteroids versus those that were not.
MIC CAVAZZINI: And how widespread is the use of corticosteroids at the moment?
HAMISH MCCAY: I know that some clinicians in my hospital use them. And if they're listening, you know who you are.
HARRIET HISCOCK: I think that’s a good question—we don’t have Australia-wide data on these low-value care practices and one of the challenges for EVOLVE I think is to start to document where is this happening, what is the variation in that practice.
MIC CAVAZZINI: You talk about how to quantify whether these guidelines, these recommendations actually make a difference to practice. What would your dream situation be in that respect?
HARRIET HISCOCK: Oh, my dream situation is that we would all be capturing practice data on every child we see. That we would have diagnosis recorded and that we’d have the tests recorded and the outcomes of the tests to routinely look at whether they are normal or abnormal. And that just doesn’t exist despite our emerging electronic health capacities across our two countries.
There’s a saying in the sort of administrative data: “Garbage in, garbage out.” So it’s all about the quality of the data collected. And routinely we've collected good data for in-patients because that’s what generates money, and very, very poor data in outpatient care where most kids go, and reasonable data in ED presentations.
MIC CAVAZZINI: And other sort of broader MBS records or whatever are just too non-specific?
HARRIET HISCOCK: MBS has no diagnosis attached to it, so you don’t know what the child had the test for.
MIC CAVAZZINI: Now, item the fourth on the EVOLVE general paediatrics list is not to treat gastroesophageal reflux disease in infants with acid suppression therapy—since randomised control trials have shown no effect of PPI drug esomeprazole on GORD symptoms in neonates is the main finding that goes towards this recommendation. Is the persistence of the practice more to do with the compelling logic than it is actually observing positive results?
HARRIET HISCOCK: So I think the issue here is—there's a couple of issues. One is if they have disease where they're actually vomiting up blood or inhaling vomit into their lungs or they’re not putting on weight properly, that’s a little bit more serious. And intuitively you’d think that these medications would help, but they actually haven't been shown to reduce any of those complications or symptoms.
A degree of reflux is normal in many infants, healthy thriving infants, but a lot of children and infants particularly under 1 year of age get misdiagnosed as having gastroesophageal reflux disease. So as soon as you put disease on something then dictates parent expectations of treatment.
We've done some work with maternal child health nurses in Victoria and we’ve shown great variation—some of them are very aware about the evidence about reflux and PPIs and not to use them, others are completely oblivious and still believe in reflux and silent reflux.
MIC CAVAZZINI: And what about the side effects, how widely understood are they?
HAMISH MCCAY: I think particularly esomeprazole is perceived as being a very benign therapy but what we know is that there are significant side effects. It can cause headache, diarrhoea, constipation, nausea, in up to 14 per cent of children. The other thing is that, you know, gastric acid is there as a protective layer and we know that there is an increased risk of gastroenteritis in these children. So they are very real concerns.
HARRIET HISCOCK: There’s also concerns about the effect of these medications on the microbiome of children and young infants and it’s really not well understood, but that might be predisposing them to infections as well. And there’s also concern about later bone health.
MIC CAVAZZINI: And in fact there was a finding also showing that PPIs could increase the risk of food allergies.
And the fifth and final commandment on the list is to not order abdominal x-rays for the diagnosis of non-specific abdominal pain in children. One study in particular of patient records from a non-trauma emergency department found that they might have changed patient treatment in only 4 per cent of cases. But reading the list of possible conditions led me to think that x-rays must help somehow narrow down the options—why isn't this the case, Sarah?
SARAH DALTON: The most common presentation for abdominal pain is most likely to be constipation, and there’s good evidence to say that x-rays don help us in the assessment of constipation. Then the other differential diagnoses that we’re looking at are sometimes important, but not as common. The most important one we would think about would be intussusception, but intussusception in itself is not always seen on abdominal x-ray. So I get very concerned when someone says to me “I thought about intussusception, but I did the x-ray and it’s normal so I'm sending the patient home.” Because to me that’s—we can’t really rule in diagnoses based on x-rays, but we also can’t rule them out.
MIC CAVAZZINI: In fact, one of the findings referred to in the list shows that radiographic evidence of constipation was found just as often in clinically constipated kids as it was in those that didn't have symptoms. How do you make sense of that? Is that what you've just said?
SARAH DALTON: Yeah I think it’s a combination of if you see something that you want to attribute to a diagnosis like constipation then you can easily do that, because it’s a non-specific test. But it also gives false reassurance at times as well. So I think that it leads to over-diagnosis in some cases and misdiagnosis in others.
HAMISH MCCAY: I had a radiologist when I was training who once said, “You know, if you see the poo in the bowel that’s where it’s meant to be. If the poo’s outside the bowel, that’s when you get worried.”
MIC CAVAZZINI: And is it another case of, if the x-ray doesn’t fit with your initial diagnosis it might send you down the wrong rabbit hole?
HARRIET HISCOCK and HAMISH MCCAY: Absolutely.
MIC CAVAZZINI: I like how you chimed in there together.
Another retrospective study from paediatric ED presentations found that those radiographs that were found to be diagnostically useful, could be filtered by selecting only those patients who’d had prior abdominal surgery, who’d had suspected foreign body ingestion, abnormal bowel sounds, or abdominal distension. Would these markers be well-recognised by our physicians?
HARRIET HISCOCK: Yeah I think they absolutely would be, and that comes back to, don’t routinely order it for non-specific abdominal pain, and all those things you’ve mentioned would be more specific abdominal pain markers.
I think the other thing we have to be careful of that some people go, “Oh well x-rays cause radiation, ultrasounds don’t. So I’ll just go and do an ultrasound.” And that’s been seen overseas, that you get a spike in other low-value care practices when you rule out one—so we’ll have to be aware of that. And they’re just as unhelpful in non-specific abdominal pain.
MIC CAVAZZINI: Are there any more definitive imaging modalities or…
HARRIET HISCOCK: No, it’s back down to HOPE.
HAMISH MCCAY: Look, I think we have to be really clear about our top five list here. All of these are useful tools, useful investigations, useful therapies. We’re not wanting to throw the baby out with the bathwater. The important thing is context: when they're used out of the context where they are useful, they can actually become a hindrance and a harm. And that’s what we’re talking about here.
HARRIET HISCOCK: I think we need to get the language around this right. And I know what they say in the U.S. is doing less, safely and that's a good—
HAMISH MCCAY: You wouldn't want to get the emphasis wrong would you?
HARRIET HISCOCK: —but that’s a good mantra to have, you know. We’re not trying to just save costs, we’re just trying to do less, but do it in a safe way.
MIC CAVAZZINI: That was Harriet Hiscock ending this episode of Pomegranate. Many thanks also to Sarah Dalton and Hamish McCay for their generous contributions. The views expressed are their own, and may not represent those of the Royal Australasian College of Physicians.
To find links to the EVOLVE lists and other resources mentioned in the program, visit the Pomegranate website at racp.edu.au/pomcast. Feel free to email us with your feedback or if you’d like more information about joining the podcast’s editorial group. And you can share the discussion round using the Twitter hashtag RACPpod.
I’m Mic Cavazzini. Thanks for listening.