When practices become ingrained in practice it can be very difficult to change them, even when new evidence becomes available, or a review of existing evidence points out that the practice is worthless.
An editorial towards the end of last year in JAMA (Powers BW et al De-adopting Low-Value Care: Evidence, Eminence, and Economics. JAMA. 2020) discusses “Evidence, Eminence, and Economics”, 3 factors that are important in whether or not useless therapies or investigations continue to be common or are de-adopted. Briefly, good quality evidence that a procedure or test is ineffective often has little effect on practice patterns. Statements by learned societies tend to have some effect but it is variable and often a small effect. When a system stops paying for a test or procedure the impacts tend to be major and rapid.
The choosing wisely campaign recommended 5 common neonatal procedures that could be abandoned (Ho T, Dukhovny D, Zupancic JA, Goldmann DA, Horbar JD, Pursley DM. Choosing Wisely in Newborn Medicine: Five Opportunities to Increase Value. Pediatrics. 2015;136(2):e482-9).
These are my top 4, there is some overlap with Ho et al.
- Routine day 3 head ultrasound, (followed by day 7-10, day 30, 36 weeks and pre-discharge, near term).
- Routine investigations as a universal screen should be proven to improve outcomes in some way. Clearly, that has never been shown for routine repetitive head ultrasounds, but they have become standard practice and are often performed on multiple occasions.
- Generally speaking, investigations should only be done if they will change the clinical approach. A good rule of thumb is to ask the question, what will I do if the test result is negative, and what will I do if it is positive? If the answers are the same: don’t do the test.
- I think we should do head ultrasounds only if they are going to have an impact on clinical care.
- We should therefore focus on finding treatable conditions, which in this situation means acute post-haemorrhagic ventricular dilatation.
- A single head ultrasound at 5 to 7 days of age in infants who are at risk of that complication, such as babies under 27 weeks and those who have been critically ill, would suffice. Further head ultrasounds are only needed if the initial images show a lesion that could lead to post-haemorrhagic hydrocephalus.
- C-Reactive Protein
- CRPs are frequently performed in many NICUs, but a recent systematic review shows that they are basically useless, with a PPV and NPV no better than tossing a coin. They are neither sensitive nor specific for the diagnosis of true infections, and many babies receive antibiotics to treat an elevated CRP!
- Anti-acid medications
- Many babies with symptoms attributed to reflux receive either H2 blockers or PPIs
- There is no clinical sign that is adequately diagnostic of reflux except for overt regurgitation
- Most reflux in the newborn is non-acid, either neutral or alkaline
- There is no evidence that clinical signs attributed to reflux are due to acid
- Gastric acid is good for you, blocking it changes the intestinal microbiome, increases translocation, and increases infections and necrotising enterocolitis
- Prokinetic/”anti-reflux” medications
- Even if you have proved, by multiluminal impedance, that a baby has reflux which is temporally related to their symptoms, prokinetics don’t work. There is no evidence of beneficial effects of prokinetics on reflux in the newborn, and they are all toxic.
Anyone have other suggestions?
Neonatal Research 
Well my friend, I would still disagree with throwing out CRP, as we had discussed in an earlier blog. CRP has terrible PPV primarily because studies did not exclude patients who currently had antibiotics floating around in their blood stream (in our world, from maternal antibiotic administration). SERIAL low CRPs have a very good NPV, guiding the clinician toward stopping antibiotics, which is how I (and others) use it.
By the way, a new publication from people in my group adds credence to term-adjusted MRIs, one of the five items in Tim Ho’s “Choosing Wisely” paper.
“Comparing head ultrasounds and susceptibility-weighted imaging for the detection of low-grade hemorrhages in preterm infants”
Journal of Perinatology volume 41, pages736–742(2021)
I would recommend looking at the most recent AAP guideline for EOS. Currently blood cultures are selective for specific antibiotics, hence maternal antibiotic exposure is no longer an issue. I would also recommend JB Cantey’s commentary in Pediatrics “Ending the culture of culture-negative sepsis.”
I wrote a letter to the editor about that EOS statement, and am quite aware of Cantey’s paper.
It baffles me that adults can have ‘culture negative sepsis’ with only 35% of ‘sepsis’ in adults having a positive blood culture, but premature babies are so different physiologically that they will always demonstrate a positive blood culture? And, again in adults, blood cultures turn (-) two hours after antibiotic administration in 2/3 of patients (Open Forum Infectious Diseases, Volume 6, Issue 5, May 2019, ofz179) but no one would consider stopping antibiotics in these patients.
CRP – friend or foe? To quote JB Cantey just a hazardous waste.
If not already stopped as a routine; consider stop routine gentamicin trough levels before 2nd or 3rd dose. Trough values have little PPV regarding ototoxicity, and nephrotoxicity is rare and reversible.
https://pediatrics.aappublications.org/content/145/2/e20192373 (private bias)
Hi-flo
As we all know and probably agree on: CRP has a very low PPV and a high NPV for serial measurements. We probably agree as well on the heuristic way of many physicians assessing CRP measurements in a binary way: negative or positive with a fix threshold leading to antibiotic treatments for increased CRPs. Last, we know as well and I think we agree on that as well: Culture-negative sepsis exists, but probably much less frequent than we think. But this is it, isn’t?
For sure, we do not know and we do not agree on what is the best management approach for suspected neonatal sepsis. And that is a pity which needs to be addressed ASAP. Incidences of neonatal sepsis are going down, but many of us are using still the same strategy as decades ago. Therefore I agree with Keith, change is needed. But I am not sure if just throwing away CRP helps. We submitted just a study showing that the PPV of clinical sings regarding culture-positive EOS is lower than the PPV of CRP or leukcytopenia (and the PPV of risk factors is as expected complete useless). CRP is a crutch, and not a good one, therefore we have to develop better strategies … and we have to educate physicians that CRP (and many other measurements) are not just negative or positive, but there is a large grey area in between.
Dear Prof Barrington
I follow your blogs regularly and thoroughly enjoy your critical and candid analysis of the evidence.
Use of diuretics for BPD is another area which we need to de-adopt. Sometimes these premises just need time. Optimising their nutrition to me is more crucial rather than diuretics because they are in tiny amount of oxygen.
I have one question. What would you propose to replace CRP? Pro calcitonin?
We do use CRP on D2 to guide us in stopping Abx.