Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
Rumours of the demise of the observational study have been grossly exaggerated
So, maybe I’ve taken some artistic licence with Mark Twain’s original version, but this methodological ‘family’ deserved a fair backing. Take the cohort design: this may have had a difficult start to the 2010s fending off the ‘gold standard’ (an expression that always makes me wary, not to mention suspicious), but has emerged reinvigorated perhaps as a result of new cohort data exploring new angles and helping regain momentum. The principles are, of course unchanged: the starting point exposure or not as it has been since their infancy. The means of assessing outcome, though, have refined the previously rough and ready markers. Examples include modifications to Poisson regression (adapted to deal with chunks of time – think splines, join points and interrupted series among many) and non-independence of outcomes (robust standard errors and multilevel modelling are just two examples). I could go on, but probably don’t need to make the point – better to simply flag the many excellent observational (cohort and other) studies in this edition that show the transitional period is over and that they are as deserving of their champions’ league contender position as much as their vaunted trial counterparts.
I have an over-developed fear of missing an operable subdural haemorrhage: and part of my amelioration is reading sensible arguments for non-irradiation of the low-risk child. Based on the seminal PECARN and CHALICE studies, Sonia Singh and colleagues in Melbourne use per item charges from the Australian publicly funded service to persuasively show that delaying imaging is not only economically but child friendly. Accepting the provisos inherent to all health economic analyses and that safety netting always provides an opportunity to reassess, this inference is hard to reject. This summary says more than 1000 words (figure 1). See page 712
Dietary diversity and cognition
One of the reasons for the re-emergence of observational studies is the refinement of the nuancing of outcomes which has allowed much more subtle dissection of previously slightly blunter ones. Lilia Bliznashka and colleagues in Boston provide an excellent example in their pooled analysis of predictors of developmental trajectories using pooled global demographic health surveillance data. They show among other exposures, the positive association between dietary diversity and child development which Marco Kerouac at the LSHTM, London puts into context while John Wright and colleagues describe how the Born in Bradford cohort can unveil subtle exposures using routinely collected data. See page 719
Getting on with it
If ever there were an area in need of (no, craving) scientific scrutiny, the speed and process of weaning from nasogastric feeding would be a certain medallist. Charlotte Wright and colleagues in Glasgow give this the attention it deserves in their analysis of weight loss and recapture in a heterogeneous group of neurologically impaired long-term tube-dependent babies. Inevitably, there was some weight loss at the start, but overcaution with the rate of reduction of feeds ultimately did not help the children, the time to independence only being delayed. There are gains here for individual families (de-medicalisation particularly) and the children (both enjoyment of food and the re-assessment of the near-inevitable percutaneous gastrostomy trajectory many will have been on). See page 767
Have a great summer!
Patient consent for publication
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Provenance and peer review Commissioned; internally peer reviewed.