Fig. 1

(a)-(b): Smoothened time series of the data from 73 GPs across two different boroughs (blue lines) and best fit fractional polynomial to the data (maroon solid and dashed lines) with equation and specific parameters outlined in the supplementary material. The graphs show the daily referral rate (\( \frac{number\ of\ referrals}{GP\ size}\ast \mathrm{10,000}\Big) \)) over the period for which we have data in borough B (in (a)) and borough C (in (b)). Boroughs B and C had a disruption of IRIS service for respectively six and three months (time period (T₃ − T₂) in (a)-(b)). The dashed lines in (a) and (b) illustrate the temporal trajectory of the fitted polynomials in the scenario where “no disruption of IRIS service” would have occurred in these boroughs b and c