January 14, 2022

expert reaction to two SAGE papers looking at early data on omicron and hospitalisations in children

Two papers, published this afternoon in the latest batch of SAGE documents, provide early data on omicron and hospitalisations in children.

This Roundup accompanied an SMC Briefing.

Prof Kevin McConway, Emeritus Professor of Applied Statistics, The Open University, said:

“It looks concerning that the ‘CO-CIN Update January 2022’ report says that there are proportionally more infants (children aged less than 1 year) admitted to hospital, in the period 14 December to 12 January, that in previous periods, and that the difference in proportion of infants compared to earlier times is statistically significant (that is, too large to be explained just by chance variation). Overall, during that latest period, the Omicron variant was becoming dominant, while in the earlier periods, other variants (most recently Delta) dominated. Indeed at one level this is concerning – one never likes to read about very young children being in hospital. But we should ask what the reason is for this change. In particular, is it something directly to do with Omicron having a more serious effect on very young children than was the case for previous variants?

“These are preliminary data so it’s impossible to say for sure. But there are some points that mean that the proportional increase could well be explained in other ways. It could have something to do with vaccinations in older children, and it could possibly have something to do with the variants that infected the children (because they are unlikely to have all been Omicron, and it’s possible that Omicron was less dominant in the children than in older people).

“First, numbers of hospitalisations have been generally increasing, in all age groups, according to data on the Government dashboard. Infection rates have been increasing across all ages, according to the ONS Coronavirus Infection Survey, and are relatively high in the age groups under about 18 (though the survey does not include very young children under 2). If there’s a lot of infection around, and infections in young children are increasing, it’s inevitable that more of them end up in hospital. So the question is really, are the youngest children being disproportionally likely to be hospitalised, not just that the numbers hospitalised have gone up. The report indicates that this is the case, though it’s looking only at hospitalisations at different ages out of all hospitalisations in those aged under 18, not compared to all people of every age. Hospitalisation rates remain much higher for the oldest age groups (late middle age and older).

“But we’ve got to bear in mind that vaccination policy for young people has varied with age. Of those in the update report (aged under 18), those aged 12 to 18 can get two vaccine doses, and some of those aged 16 and 17 can be eligible for a booster. Some children aged 5 to 11 who are at particularly high risk are now being offered vaccines, though that was announced only recently, and few children aged 5 to 11 would have been vaccinated in the period covered by the new data. Though vaccines are less effective against Omicron, they are not totally ineffective. So different levels of vaccination in different age groups amongst the 0-17 year olds could possibly explain some of the proportional differences that the update report mentions.

“However, I’d expect those differences in vaccination rates to make a difference particularly for the 12-17 age group in the update report. If I leave them out of consideration, and just look at the three age groups <1, 1-4 and 5-11, there is still a difference in the proportions of hospitalisations between the latest figures and previous ones, that is too large to explain just by chance, but it’s considerably less marked than if the 12-18 group is included. This remaining difference might possibly have something to do with vaccinations of children at particularly high risk aged 5-11.

“But just comparing the proportion of hospitalisations in infants under 1 with those aged 1-4, the difference in proportions isn’t there anymore, so I don’t think there is really evidence that infants under 1 year old are more likely to be hospitalised than those aged 1 to 4.

“I think also that we have to be careful in saying that the higher rate of hospitalisation in children under 5, compared to previous times, is directly something to do with Omicron. It could be – but a report from ONS* early in the rise of Omicron did provide some evidence that the chance of an infection being Omicron rather than some other variant (usually Delta) was lower in young children than in older age groups. That was an early report, and I haven’t seen evidence on whether or not this age difference has continued or not. But the ONS report does consider infections in some of the period that could have led to hospitalisations in the CO-CIN update report, and maybe it’s possible that the proportionally higher rate of hospitalisation in very young children is connected with the variant that was infecting them.

“I think it’s also important to point out that only one of these hospitalised children died (a girl, aged 5-11), in the most recent period covered by the two reports. Any death of a child is sad, but the chance of illness severe enough to kill in these age groups is extremely low. Also, though the data will still be incomplete, the average length of stay in hospital of the hospitalised children is considerably lower for the most recent period than for earlier periods. That applies to all four of the age groups in the update report, with average length of stay just 1.7 days for the under 1s, 1.6 days for ages 1-4, 2.0 days for 5-11 and 2.5 days for 12-17. Those lengths are all, very roughly, half the average length of stay during the time when Delta predominated, though I don’t know (and the reports don’t really say) how much this might be due to improvements in treatment.”

Dr James Doidge, Senior Statistician, Intensive Care National Audit & Research Centre (ICNARC); and Honorary Associate Professor, London School of Hygiene and Tropical Medicine, said:

“The CO-CIN update for January 2022 indicates some changes in the characteristics of patients aged < 18 years who are in hospital with COVID-19, but there is no direct indication of whether COVID-19 is contributing or incidental to the patients’ reason for admission. A closer look at the data suggests that there are likely to be significant differences in the proportion of patients admitted with COVID-19 as an incidental diagnosis between time periods. While the number of patients aged < 1 year admitted with a diagnosis of COVID-19 has increased from 3.3 patients per day in the period 1 May to 13 December, to 5.7 patients per day in the period 14 December to 12 January (rates calculated from data provided in the table titled “Patient characteristics by epoch”), these patients have shorter hospital stays (1.7 days vs 3.2 days), lower use of oxygen (12% vs 21%) and of the 171 patients aged < 1 year, there has not been a single death recorded (compared with 1-2% in previous epochs).

“These findings indicate that there has either been a substantial increase in patients aged < 1 year admitted with COVID-19 incidental to their primary reason for admission and/or that there has been a substantial reduction in severity of COVID-19 among infants associated with emergence of the Omicron variant. It is also interesting to note that there has been a reduction in the rate of hospital admission with COVID-19 among patients aged 12-17 from 3.7 patients per day during the period 1 May to 13 December, to 2.6 patients per day in the period 14 December to 12 January (rates calculated using data provided in table titled “Patient characteristics by epoch”).”

Dr Michael Absoud, Honorary Reader, Dept of Women & Children’s Health, King’s College London, said:

“The Clinical Information Network (CO-CIN) continues to provide invaluable information on trends of COVID-19 admissions. The data on children is preliminary, and will need further confirmation from longitudinal datasets. The data provides reassuring news for unvaccinated 5-11 year olds, with no increase in relative proportions of admissions compared to other age-groups and relatively low rates. The proportion of admissions appears increased for those <1 year old, compatible with increased hospital admissions usually seen at this time of year due to bronchiolitis and other upper respiratory infections (with relatively lower overall numbers).  It is as yet unclear whether this is due to intrinsic features of Omicron, co-infection with other respiratory viruses, or incidental notifications. There are also no obvious signs of a change in severity or management needed. The UK data linkage studies and population based surveys, should continue to provide important insights this winter.”

Prof Christina Pagel, Professor of Operational Research, UCL, said:

“The absolute number of hospital admissions in children of all ages have been going up steeply since the Omicron wave (dashboard data) but this analysis provides a lot more detail. It is obviously concerning that significantly more infants under 1 are being admitted to hospital than previously. This cannot be explained just by vaccination status or high community prevalence, since the same is not true for instance of 1-4 year olds. . Reassuringly, babies are not needing to stay long in hospital once admitted. It is possible that Omicron’s preference for upper airways is affecting young children more, even while it reduces burden in adults and older children and we urgently need to understand more about what might be causing this increase.

“The other notable finding from this study is that children from the most deprived areas are far more likely to be admitted and this has become significantly more pronounced with the Omicron wave: children from the most deprived fifth of households account for almost half of admissions. This is seen across all ages and is likely reflecting lower vaccination rates in teens from deprived areas, more exposure to infection and higher rates of other health condition making them more vulnerable to severe illness once infected. The disproportionate impact of the pandemic on more deprived communities has been long known but far too little done to address it.”

Dr Alasdair Munro, Clinical Research Fellow in Paediatric Infectious Diseases, University of Southampton, said:

“These preliminary data from CO-CIN appear to provide a picture of shifting proportions of admissions which are likely due to different increases in population immunity. The vast majority of the adult population have now been immunised, and the majority of children over the age of 5 have either been infected previously, have been immunised, or both. The group with the lowest rate of immunity is children under 5, and in particular, children under the age of 1 year. This means we would expect a larger proportion of admissions to be among this population, which is what is demonstrated in this data. Given the extraordinary recent rates of infection in age groups who are likely to be parents to young children, we are seeing higher case numbers in that age group which is resulting in higher rates of hospitalisation. This data would support that the difference is mainly in infants (<1y) rather than pre-school aged children. These data cannot inform us of the severity of omicron in younger children compared to other variants, or any differences in the clinical presentation of disease.

“Rates of high dependency or intensive care admission remain very low among children, although higher in children aged <5 which is consistent with previous analyses. There do not appear to be any deaths in children aged <10 years in this report.

“Analysis will continue to look for other trends or signs of ways in which omicron may differ for younger children compared to previous variants of SARS-CoV-2. The interesting signal is a particular rise in the children <1y compared to other pre-school ages, which is worthy of further investigation.”

‘CO-CIN update January 2022’ and ‘Dynamic CO-CIN report to SAGE and NERVTAG’ were both posted on gov.uk at 13:30 UK time Friday 14^th January.

Declared interests

Prof Kevin McConway: “I am a Trustee of the SMC and a member of its Advisory Committee.  I am also a member of the Public Data Advisory Group, which provides expert advice to the Cabinet Office on aspects of public understanding of data during the pandemic. My quote above is in my capacity as an independent professional statistician.”

Dr James Doidge: “No COIs.”

None received.