911½ñÈÕºÚÁÏ

Over the past year, George has shared his findings on the links between household size and child health at national conferences while also working closely with families to shape and interpret the research. In this blog, he offers a behind‑the‑scenes look at the evolution of his project, from conference presentations to community conversations with families at the heart of the research.

Working at the intersection of data science and public health, George Melrose is a Research Assistant at the Mohn Centre investigating how family and household structures influence child health outcomes. Here, he reflects on presenting his work to diverse audiences and the invaluable insights gained through patient and public involvement.

-

I am a biostatistician at the Mohn Centre. My daily work involves writing computer code to process and analyse complicated medical data, running statistical models, and discussing the results of this work with paediatricians. I really enjoy my work as there’s always a new puzzle to solve, some interesting work to do, that will result in better outcomes for our patients.

At the start of September, I had the pleasure of going up to Edinburgh to present my poster for the project 'Exploring the impact of household family size on child health and social care outcomes in Northwest London: Analysis to inform service provision and national welfare policy' at the Royal Statistical Society 2025 conference. I got to meet and explain my research to a kaleidoscopic range of stats professionals from software engineers at the Czech software company 'JetBrains' to former Glasgow University Hares and Hounds Running Club colleagues!

Interest abounded for how different the associations were for the number of children in a household (household size) for different outcomes e.g. negative for Emergency Department visits, more children --> lower odds of visiting A&E even after taking into account and adjusting for other variables like ethnicity and gender. More children --> higher odds for under 12 tooth extractions, even after adjusting for other variables, with a clear gradient for households of 3 or more.

Several professionals gave great advice on how to take the project forward. Iain Bell, a PhD candidate at the University of Glasgow, gave his recommendation on dealing with the problem of children living in the same household and so being influenced by things going on in that household, i.e. having the same diet. To account for this, you could include a random effect for the household - a way of telling the model children from the same home are linked. An academic at Lund University pointed out that for outcomes which happen many times, like ED visits, using Poisson or negative binomial regression would be better. As well as presenting my poster, I was able to attend fascinating talks by fellow researchers. Just one such example was that of Professor Oyarzun of the University of Edinburgh, who has used machine learning to find potential novel senolytic compounds - drugs that destroy senescent (zombie) cells, that are one of the hallmarks of ageing!

Later in September, we hosted participants, mothers from the Wealdstone Baby Bank (WBB) at the Mohn Centre for a Patient and Public Involvement and Engagement (PPIE) session. The Wealdstone Baby Bank, coordinated by Families in Action Together, in North London, is run by mothers and local community members to support one another as parents, for example by providing essential items for infants. During the session, I presented an abridged research talk, where I highlighted the stark differences between different household sizes for anti-depressant prescriptions, A&E visits, and under-12 in-hospital tooth extractions. After this presentation, Dr Francesca Neale (Paediatrician Researcher), Esta Orchard (PPIE Lead), Yusuf Ibrahim (Involvement and Engagement Officer), Aghileh Marbini (WBB organiser), and I discussed the different health outcomes in my research. For example, we discussed whether there are any disparities in dental care between different Northwest London boroughs and what are the barriers to good mental health in one's children. These discussions were noted on large reams of A3 paper as well as being beautifully captured by a visiting graphic illustrator, Anna Geyer.

In December, Francesca and I visited Wealdstone Methodist Church in Harrow, for another WBB PPIE session with Aghileh, Esta, and the participating WBB mothers. There were two halves to this session. The first focused on mental health resilience for parents and how to improve it in the face of all of life's various economic and social challenges. I gave the example of my love of cross-country running, an anchor, a touchstone of any workday I have good or bad, where I can forget about any worries I have, lace up, and get some muddy miles in. During the second half of the session, I presented some new data analysis I had generated of the cumulative incidence of under 12 tooth extractions in a 2012 birth cohort. In this analysis, the same previous pattern was borne out - children living in larger households consistently had higher incidence of non-cosmetic, non-orthodontic tooth extractions throughout their childhood. This project has grown and evolved over 2025.

A long-form, 15-minute presentation of this research was presented at the NIHR SPHR Children, Young People & Families Symposium at the University of Birmingham in November alongside colleagues from WBB, and we are writing up a paper focused on the under 12 tooth extractions analysis. We are also planning a presentation for the 'Society of Social Medicine' conference in September, so watch this space!

As a researcher working primarily from home, it can be all too easy to start to feel distant from your research. You understand it will help patients but between complex statistics, lengthy journal publication waiting times, and the plethora of pre-existing publications, the end goals of your research can feel out of reach. Meeting and speaking candidly with patients involved in my research, discussing their lived experiences, has been invaluable. The candid reflections of different health challenges and opinions voiced could not be sourced from the routinely collected data I analyse.

Many thanks to our colleagues Prof Hargreaves and Dr Hope (911½ñÈÕºÚÁÏ), Prof Taylor-Robinson and Dr Eleojo Abubakar (University of Liverpool) , Dr Lateef Akanni (University of Glasgow),and Prof Fairbrother and Dr Nicholas Woodrow (University of Sheffield).