Martin Bardsley

Article Information Article first published online: July 16, 2020 Ben Goldacre1, Martin Bardsley2, Tim Benson3, Kate Cheema4, Roger Chinn5, Ellen Coughlan6, Sarah Dougan7, Marc Farr8, Loraine Hawkins9, Adrian Jonas10, Andy Kinnear11, Morag Mcinnes12, Mohammed Amin Mohammed13, Caroline Morton1, Rahul Pasumarthy14, Neil Pettinger15, Ben Rowland16, Neil Sebire17, Paul Stroner18, Jeni Tennison19, Samantha Warnakula20, Oliver Watson21, Emma Wright22, Hamish Young12, Jessica Morley1 1DataLab,...

Data have been widely hailed as ‘the raw material of the 21st century’1 and ‘better use of data’ is a central feature of the NHS Long Term Plan.2 Yet, data alone does not produce insights. To capitalise on opportunities to improve health and care, we need the data and outstanding data analysis. However, policymakers and academia have almost exclusively focused on pure academic research around the aetiology of disease; the field of practical coalface analytics has been largely neglected.

We compared a sample of patients enrolled in a large telehealth programme in a predominately rural area in England to a well-balanced matched control group using person-level data. We found that, contrary to the aims of the programme, telehealth patients had a higher likelihood of an emergency admission, and accident and emergency department attendance, than matched controls.