Background The Govan SHIP project was implemented in Govan Health Centre, a deprived area in the south of Glasgow, involving GPs and social workers sitting within multidisciplinary teams to identify and support vulnerable patients at risk of attending A&E and GP surgeries. This study aimed to measure the impact of SHIP on A&E presentations and GP interactions.

Methods Rate of A&E presentations per 1000 population in 3 participating GP practices in Glasgow City was compared before and after onset of the service with a 3- month phase-in period, using segmented linear regression with 9-month (April 2014- December 2014) pre- and 36-month (April 2015-March 2018) post- intervention periods. Rate of GP interactions were similarly compared before and after using 21-month pre- and 36-month post- intervention periods. Rates of A&E presentations and GP interactions for practice population in Drumchapel- an area within NHS Greater Glasgow & Clyde with similar rates of deprivation, but with no such service in place - were used as a comparison group. Autoregressive and moving average terms and a fourier term to adjust for seasonality were included in the models.

Results Govan practices had a lower rate of A&E presentations than Drumchapel practices across all time points. A&E presentations did not change significantly over time between April and December 2014 in either area. At April 2015, SHIP onset, a level change of -4.34 (-7.44, -1.24) A&E presentations was observed in both areas, but no significant change in trend over time comparing pre and post SHIP in either area. Onset of SHIP was therefore not associated with a reduction in level or trend in A&E presentations. Rate of interactions with GP was greater for the Govan practices than those of Drumchapel at all time points, prior to January 2015, increasing over time in both areas. After April 2015, there was a significant level change of 33.78 (19.57, 47.99) per 1000 in both areas. GP interactions in Govan however saw a significant reduction of -1.48 (-2.87, -0.09) per 1000 per month, following onset of SHIP, between April 2015 and March 2018. This is equivalent to SHIP being associated with an absolute reduction of 37 GP interactions per thousand and a relative reduction of 7.2% by March 2018.

Conclusion The Govan SHIP initiative was associated with no significant change in A&E presentations and a small reduction in GP interactions. A cost effectiveness analysis is recommended.