Thread by @veitchemma

A thread about infant readmissions to hospital in the UK. Prompted by a recent paper in @bmcseries #bmcpediatrics. I’ll talk thru the data & some wider issues that are the focus of debate in #infantfeeding - 1/n

bmcpediatr.biomedcentral.com/articles/10.11…

Why is this important? Well, there is a lot of discussion going on, mainly in mass media and social media, about possible relationships between #infantfeeding policy & practice and sufficiency of #breastfeeding in the early days/weeks of life 2/n

My personal interest in this area: while having my own kids, I trained & volunteered to provide #breastfeeding support locally and continue to follow the research and evolving discussions 3/n

Back to the paper: I saw this as a nice example of descriptive #epidemiology: the unsexy sibling to aetiological epidemiology (of the “coffee/chocolate/bacon cures/causes cancer” headlines…) 4/n

The background context: a lot of ongoing discussion esp stories in news & #socialmedia re infant readmissions to hospital due to feeding complications, which can include excessive weight loss, jaundice, hypoglycaemia, and hypernatremia 5/n

However, there aren’t many studies explicitly addressing these issues for the UK, even just to generate representative population-based statistics (eg, how many readmissions are we talking about?) as a starting point for discussion and further evaluation 6/n

OK, so as soon as I spotted this paper, definitely knew worth a careful look. Re personal COI, neither of my babies were readmitted for feeding (or any other) issues, but I have a few friends whose babies were, and I know it was a substantial source of anxiety and concern 7/n

This paper used 6yr data from Hospital Episode Statistics, from 1 Apr 2008 to 31 March 2014, to assess how many infants in England are being readmitted to hospital in the first year of life. HES data covers the whole of the English NHS – so it’s astonishingly complete 8/n

The readmissions data is combined with denominators on all live births for England from @ons, so that incidence rates for particular types of readmission can be calculated 9/n

The authors set up an advisory panel to produce a pre-specified set of ICD-10 codes to define “potentially avoidable” readmissions; things that could have been dealt with by good care in the community and prevented 10/n

This work was done, and prespecified prior to data analysis (this is good!). Their coding framework includes physiological jaundice, feeding difficulties, and gastroenteritis 11/n

The dataset also included info on gender, ethnicity, region of admission, and deprivation (estimating IMD based on LSOA data, below for the nerds – like me) 12/n

The analysis is basically descriptive: stats geeks, look away now. I love this type of stats section: let’s just describe the data. How many babies get readmitted every year? What for and at what ages? How are trends changing over time? 13/n

Okay, let’s see some results. The big picture trend is that readmissions are creeping up, in terms of incidence rates. Around 334/1000 babies were readmitted (all causes) in 2008/9 but 354/1000 in 2013/14 (14/n)

This is most obvious for the youngest babies: 0-6 days and 7-28 days of life (15/n)

However, the increase in readmissions is most striking for the avoidable conditions, in the first week of life 16/n

Just to let that sink in, in 2013/14, nearly 2% of babies born alive in England were admitted to hospital for things that could be avoided, in their first week (GIF) 17/n

Looking at a cause breakdown: whilst 16.3/1000 babies were readmitted for jaundice in 2008/9, in 2013/14, this was 22.35/1000 (just over 2% of babies), a really substantial rise 18/n

Unusually, there is a rare patterning by SES in this dataset: readmissions for jaundice and feeding difficulties are LOWEST in the most deprived quintiles, but HIGHEST in the most affluent (this is not true for gastroenteritis, which is patterned in the opposite way) 19/n

In total, their data show 14,806 admissions for jaundice and 8694 for feeding difficulties in English babies in 2013/14, ie 23,500 in total. Collectively this is around 10% of the total infant admissions in the English NHS (234,900) *but* heavily stacked to 1st weeks 20/n

SO, WHAT DOES THIS ALL MEAN? 21/n

Firstly, this is really crucial data that gives a clearer picture both of infant readmissions overall and time trends, enabling policy makers to start trying to address these problems 22/n

I think it’s useful to read this paper alongside another HES analysis by @Richard_GP @SoniaKSaxena et al, which also maps patterns of infectious (and other) admissions in babies in England – I’ll come back to this later 23/n
bmcmedicine.biomedcentral.com/articles/10.11…

Overall, we can say that it is not rare, or unusual, for a baby to be admitted for feeding problems in the UK (you sometimes see assertions that this is #fakenews, particularly on social media). 24/n

Importantly, and sadly, trends are going in the wrong direction, at least for the period 2008-2014 25/n

Currently, #infantfeeding policy in the UK is very much focussed on the importance of raising population #breastfeeding rates, and not specifically on preventing complications linked with feeding issues in babies 26/n

However: cf pop'n rates for outcomes in THIS paper with infectious admissions outcomes in @richard_gp paper. Admissions for feeding difficulties of same order of magnitude (10s thousands) as admissions for URTI/LRTI ( partly preventable thru #breastfeeding) 27/n

bmcmedicine.biomedcentral.com/articles/10.11…
28/n

Jones dataset, 2013/14: 22.35/1000 jaundice admissions, plus 13.12/1000 feeding diff admissions. Cecil data (2000-2013), 26.6 admissions/1000 for URTI, 26.1/1000 for LRTI. 29/n

My POV: start thinking about *preventing* admissions for feeding problems as seriously as concern for increasing #breastfeeding rates (in order to reduce things like URTI/LRTI admissions) 30/n

If infant admissions for respiratory infections in the UK are a priority for public health action, then admissions for jaundice and poor feeding are, too. We need to be consistent. 31/n

Thankfully, this does seem to be receiving some attention – see recent presentation @rcpchtweets specifically focussing on trying to avoid term admissions with feeding problems adc.bmj.com/content/104/Su… 32/n

But what can we say about why these rates are rising, and *how* policy and care need to be changed to reverse these trends? IMHO, this paper doesn’t tell us, unfortunately 33/n

In this paper, rates aren’t compared by feeding type or other variables, so we have no info internal to this dataset to shed light. But there are insights from elsewhere 34/n

In recent data from the US, from @ianpaulmd et al, it’s clear that babies discharged exclusively #breastfeeding have a higher risk of readmission (probably due to inadequate intake) than other babies ncbi.nlm.nih.gov/pubmed/29191700 35/n

Currently, infant feeding policy promotes exclusively #breastfeeding, & emphasis on ensuring adequacy of intake on “conservative” measures, ie the breastfeeding technique – not on when (and why) you might supplement (ie, add formula, or donor milk if it’s available) 36/n

Guidelines vary between trusts on the issue of management of feeding problems. This paper suggests to “consider” supplementing (with donor milk or formula) at 12.5% infant weight loss 37/n

ncbi.nlm.nih.gov/pubmed/29848502

@nicecomms CG37 doesn’t say when supplementation should be considered to ensure baby’s full nutritional/hydration needs are met in order to avoid feeding related complications 38/n

nice.org.uk/guidance/cg37/…

One poss implication from the Jones paper is that fragmentation of postnatal care, poss combined with rise in early post-birth discharge, is behind these trends. That’s plausible, and would highlight that closer postnatal care from community #midwifery teams is important 39/n

40/n

Reductions in quality & depth other support services may also be linked- some readmissions poss preventable w/better #breastfeeding management. In England, provision community #breastfeeding support with #publichealth budgets held by local authorities, undergone cuts 41/n

However, IMHO improvement in services has to be combined with attention to raising awareness eg for HCP, around potential for feeding problems post-birth and at what point supplementation should be recommended to avoid complications 42/n

The evidence base reviewed by @WHO in 2017 clear that early supplementation not likely to harm #breastfeeding success & no evidence from HIC settings that this poses risk for health outcomes of itself (providing safely prepared appropriate supplement – eg formula/DHM). 43/n

Conclusion: data underscores need for focus on potential for complications linked with #infantfeeding. Need better understanding driving factors and clearer policy and implementation at national level #qualityimprovement #patientsafety 44/n

My bias/COI on this issue. I did ebf both kids, for (not quite) the recc 6 months and then extended to 18 months for each. I was lucky in that although it wasn’t easy, the sufficiency was there and no feeding complications. Know many with readmissions & difficulties 45/n FIN