Why new shoes give you blisters & why protective equipment does not fit: Demographic Design Bias

Blog image

Why new shoes give you blisters & why protective equipment does not fit: Demographic Design Bias

Main blog image


Most of us have at some point bought a new pair of shoes; you go to a store, have your feet measured, try on 10+ pairs of shoes and run/walk up and down the walkways. However after all of this, your new shoes will still manage to give you blisters.

The same issue can be found in the healthcare and construction sectors, with as many as 97% of minority ethnic healthcare professionals suffering some form of skin damage from respiratory protective equipment, or RPE. [1] 

Furthermore, In some cases as few as 31% of people are correctly protected by RPE at any given time due to leakage resulting from poor fit. [2]

Ultimately, the reason why your new shoes do not fit is the same as why these respiratory masks fail to protect healthcare, construction and industrial professionals globally; there is Demographic Design Bias in the data used to design and manufacture them.

All of this is because of a disproportionate representation of one or a few demographic groups (typically middle aged white men)! [3]

Hunting for the perfect database.

PolyMetrix started as the Mensura Mask Initiative at Imperial College London, where we aimed to deliver fully bespoke FFP3 masks to the general public over the COVID-19 Pandemic.

However, even when designing fully bespoke masks, we were struggling to be sure our automated system worked equally well for all demographic groups. After searching a range of free and paid databases, our team could still not get fair representations of many demographics.

It was from this hunt that the idea for PolyMetrix was born; by providing digital fit tests, we would be able to build a robust system, tested on a much wider demographic spread, improving our systems and enabling us to build an accurate model for fitting facial data.

Digital fit testing

PolyMetrix is currently focusing on digital fit testing for the construction industry; in the US, UK, Australia and parts of the EU the issue of Demographic Design Bias and poor fit has paved the way for regulations requiring workers to undergo fit tests. [4]

But these tests are time consuming, manual and expensive, costing the construction industry in these regions alone roughly £2.7 Billion a year, and the equivalent of one person working for 65,000 years. 

As such, our BioTwin system has been designed to help and inform fit testers and construction firms about just how protected their employees are, helping them to save time and money, whilst also properly protecting their staff.

How we plan to solve Demographic Design Bias completely

PolyMetrix’s BioTwin pipeline has been tested extensively with various facial databases, and in partnership other UK Universities and health institutions and in all cases our systems performed equally well for almost all demographic groups. 

The system also does not just work for faces and we have also tested it on hands and feet, as well as non-human geometris such as buddha statues, rubber ducks and university crests.

Right now we are working on fit testing, but PolyMetrix has plans to go much further; whether it be fully bespoke CPAP masks for sleep apnea patients, a new pair of prescription glasses perfect for your face, better ways of modelling facial shape for the metaverse or just a new jacket. Our vision is to remove Demographic Design Bias, ensuring all receive equal fit, performance or comfort. 

If you want to find out more or work with us please get in touch via info@polymetrix.org!


[1] Lan, J., Song, Z., Miao, X., Li, H., Li, Y., Dong, L., Yang, J., An, X., Zhang, Y., Yang, L., Zhou, N., Yang, L., Li, J., Cao, J., Wang, J. and Tao, J., 2020. Skin damage among health care workers managing coronavirus disease-2019. Journal of the American Academy of Dermatology, 82(5), pp.1215-1216.

[2] Coffey, C.C., Lawrence, R.B., Campbell, D.L., Zhuang, Z., Calvert, C.A. and Jensen, P.A., 2004. Fitting characteristics of eighteen N95 filtering-facepiece respirators. Journal of occupational and environmental hygiene, 1(4), pp.262-271.

[3] Wilkinson IJ, Pisaniello D, Ahmad J, et al., 2010, Evaluation of a Large-scale Quantitative Respirator-fit Testing Program for Healthcare Workers: Survey Results. Infect Control Hosp Epidemiol, 31:918-25.

[4] https://www.hse.gov.uk/pubns/indg479.pdf