This site is intended for health professionals only

Assessing the evidence for merging and closing hospitals

Assessing the evidence for merging and closing hospitals
By Carolyn Wickware
8 June 2017

It is hoped that changes in the configuration of acute hospitals will improve clinical outcomes, address workforce shortages and control expenditure. Sustainability and transformation plan (STP) footprints in England are now being asked to set out their medium-term plans, but these can often be very difficult to make, and their impact is hard to estimate.

It is hoped that changes in the configuration of acute hospitals will improve clinical outcomes, address workforce shortages and control expenditure. Sustainability and transformation plan (STP) footprints in England are now being asked to set out their medium-term plans, but these can often be very difficult to make, and their impact is hard to estimate.

Many people have taken the time to share their learning across the think-tanks and regulators in this massive, messy, topic. In Suffolk and north east Essex, the stakeholders were particularly keen to consider the evidence carefully, and established a review group to do this. There is no way this can be a fully comprehensive review, but executive and programme leads felt we should try to examine the key points from what has been published.

Closing or radically reducing a hospital

There are studies that compare market concentration (number of hospitals) to outcomes and costs. But these tend to be conducted at a certain period in time, and we found few longer-term studies examining the effects of a significant closure or downgrade over a longer period.

We examined the case of the Martin Luther King hospital in Los Angeles County, which closed because of problems with funding and accreditation. The three main negative effects noted by local primary care clinicians were:

  • A drop in local access to specialist consultations, direct hospital admissions and timely emergency department evaluation.
  • Poor patient understanding of the system, which led to more delays in care and worse health outcomes.
  • Loss of colleagues and opportunities to teach residents and medical students.

There were also relatively few well-documented British examples. Queen Mary’s hospital in Sidcup, Kent, reduced bed numbers from 328 to 44, focusing on an urgent care centre, outpatients and planned inpatient care. An extensive King’s Fund report noted that there had been inadequate costings for community-based prevention and inadequate plans for reducing demand, as opposed to redistributing it. The absence of a major tertiary provider as a core partner in the transformation process also limited the scope for clinical pathway transformation in south-east London.

Closure of A&E departments

A frequently quoted cohort study analysed data on 10,315 conveyances to acute hospitals by four English ambulance trusts between 1997 and 2001. After adjustment for age, sex, clinical category and illness severity, it suggested that a 10km increase in straight-line distance is associated with around a 1% absolute increase in mortality. The effect was particularly marked in respiratory patients, in keeping with two other studies. The study was transparent in describing potential limitations; the Rapid Emergency Medicine Score (age, Glasgow Coma Score, oxygen saturation, pulse, blood pressure and respiratory rate) was used to risk-adjust the cases but was available in only 37.6% of cases.

The study also noted previous studies of road traffic accidents in Norfolk, all serious trauma in Scotland, and ruptured abdominal aortic aneurysms in West Sussex. These studies did not find any relationship between time to hospital and mortality.

In terms of activity and costs, we also found very few published studies of the effect of emergency department closure on the emergency departments themselves. One example was from Sheffield4 in the late 1990s, where three A&Es were streamlined into a specialised unit for adults and another for children. Taking into account slight increases of attendance in neighbouring areas such as Rotherham and Chesterfield, there was overall a 2% reduction in demand. Some of the effects were unexpected, such as time to clinician increasing, time to admission decreasing and costs rising by an inflation-adjusted 3%.

Mixed results seen in past hospital mergers

The group found a very extensive study conducted by the University of Bristol in 2012, looking at control and merged hospitals from one-year pre-merger to four years post-merger.5 Merger, considered overall, had no or stable effects on length of stay, or outcomes for acute myocardial infarction or proximal neck of femur fracture. There was an increase in mean waiting times and worse outcomes for some stroke indicators.

In terms of resources, admissions, staff and beds costs fell by 11-12%, which was maintained for each year following the merger; this was mirrored by falls in expenditure, though at lower proportions. Spend on managers and temporary staff increased as a proportion of expenditure.

Successful service redesign at merged trusts

With the above information on mergers, we did not feel that the simple act of merging would alone be enough to improve services or finance. We had to look deeper into how services at different sites worked together; whether it was through centralisation, different sites concentrating on different aspects of acute care, or adoption of common pathways.

Centralisation has been attempted in a number of services in recent NHS history, most notably in relation to hyperacute stroke units in London and Manchester, and across most areas in relation to surgery for abdominal aortic aneurysm.

There have been a number of systematic reviews examining the relationship between volume and outcome; for example, an Italian review6 found a positive relationship not only for abdominal aortic aneurysm, but also for two other vascular procedures (lower extremity bypass surgery, carotid endarterectomy), two orthopaedic procedures (hip fracture repair, knee arthroplasty), four cardiology situations (myocardial infarction in general, coronary angioplasty, coronary artery bypass, paediatric heart surgery), two neurosurgical conditions (brain aneurysm and subarachnoid haemorrhage), cancer surgery, cholecystectomy and neonatal intensive care.

Many of the above procedures (such as those relating to vascular surgery) have already been centralised in the East of England. The work stream did not underestimate the challenges of further centralising services in specialties with a significant non-elective element. For example, if a patient had acute cholecystitis and turned up at a site that did not have general surgery, they would still need to be assessed by a surgical clinician. Second, it was difficult to imagine how a cholecystectomy at a more distant hospital could be much cheaper than a local hospital unless there were major reductions in fixed costs of rotas, theatre and so forth.

We therefore examined more subtle instances of differentiation between acute sites in a detailed Monitor Report7 from 2014 on six mergers from Sussex, Hampshire, Nottingham, North East London, Trafford, with Manchester and Birmingham. The last of these related to a community service provider but overall these case studies managed to save 1-3% in corporate overheads, generally within the first six to 12 months. In all cases except Trafford they generally had clinical directors covering service areas across several sites. Hampshire and Birmingham community mergers had detailed plans they were able to enact on the day the merger was completed.

In Hampshire, where Basingstoke and North Hampshire Foundation Trust merged with Winchester and Eastleigh NHS Trust, out-of-hours work was shared across the merged trust and reducing locums required to maintain rota. They were also able to recruit to subspecialty posts in cardiology. In London, Barts Health had a similar story, filling six A&E consultant vacancies previously serviced by locums at Whipps Cross. Barts was also able to reduce the length of stay (20 to 14 days) for stroke through a common care process.

More radical changes occurred with the Western Sussex merger in 2012; in this instance, inpatient services were moved from Southlands to Worthing. Southlands Hospital concentrated on outpatient and day-case services, and hip and knee surgery was moved from Worthing Hospital to St Richard’s Hospital. This led to a reduced hospital standardised mortality ratio (HSMR) for fractured neck of femur and length of stay for hip and knee conditions.

An older example was in Nottingham from 2006, where emergency, urgent care and trauma services were concentrated at Queen’s Medical Centre and planned services (such as elective orthopaedics) were concentrated at City Hospital; this resulted in fewer cancellations for orthopaedic surgery.

Hospital chains

The group studied four main reports on this topic, by the King’s Fund, the Dalton Report, Nuffield Trust report9 and a report by PriceWaterhouseCoopers.10 We found an excellent discussion of the legal necessities and the success factors that underpin the formation of hospital chains, such as management capacity for the transformation, an aligned view of aims and a standard operating model. For the purposes of this discussion, we considered hospital chains to consist of central management and procurement function that sets the standards and parameters that site management teams use to make decisions and discharge their responsibilities.

The King’s Fund report11 from Professor Chris Ham gave the most detailed description of outcomes, finding that in the US consolidation had increased prices, not produced any significant cost savings unless sites had been consolidated, and made mixed impact on quality. The German experience was more positive, citing the Helios Medical Advisory Board, which took a lead in measuring, benchmarking and investigating quality, and also in medical integration of new hospitals into the Helios Group.

Hospitals as accountable care organisations (ACOs)

The group noted that there were a variety of definitions for this term, which are discussed in a King’s Fund paper on the topic.12 Regardless of the exact form, it considered key factors to be risk stratification, case management, information sharing, investment in technology and patient engagement, supported by overarching specific objectives, alliances and networks, and a lead role for the commissioner in contracting and payment. For this discussion, we adopted a definition of the hospital as the provider of an integrated delivery system to the population.

In terms of finances, Health Affairs has analysed 92 ACOs in the US that did save money, and 241 that did not. As expected, ACOs with lower baseline spends found it harder to achieve savings; but in addition more savings were found in those that were more successful in a number of primary care and community care indicators, such as controlling blood pressure and lipid levels in diabetics, and mammography screening. The group noted the strong fit between the suggested priorities of the ACO with the aims of the findings of Suffolk’s strategic review in 2015 and its prevention strategy.

Moving forward across the STP footprint

The acute reconfiguration workstream and the STP programme board acknowledged that evidence was by no means perfect in any of the areas; however it felt that the best local solution was service redesign at merged trusts (Ipswich and Colchester) and the possibility of creating a West Suffolk ACO. The work has helped bring the CCGs and providers together in a relationship focused on redesign and transformation and less on organisational boundaries.

Other contributors to this piece are Dr Shane Gordon, director of integration at Colchester Hospital, Pam Green, director of transformation at North East Essex CCG, and Nicky Leach, director of NGL Associates working on behalf of Ipswich Hospitals

Want news like this straight to your inbox?

Related articles