Case studies featuring research projects enabled using data services supplied by the Lothian Research Safe Haven

Case studies featuring research projects enabled using data services supplied by the Lothian Research Safe Haven

Evaluating the impact of the PACT Pilot Service on unplanned hospital admissions and associated costs.

A large proportion of acute hospital admissions in the UK are due to a relatively small number of high demand ambulatory care patients. A service that systematically identifies these high demand patients and implements a proactive clinical case management framework can help reduce the high costs associated with unplanned hospital admissions.

In 2014 NHS Lothian initiated the Patient Experience and Anticipatory Care Team (PACT) Pilot Service. PACT identifies high demand patients and proactively pursues them with the aim of implementing Anticipatory Care Plans (ACPs) for these patients.


This project aims to evaluate the costs and cost savings of PACT, using linked administrative health data among comparable populations. Primarily, it will focus on the impact of PACT on the number and total cost of unplanned hospital admissions. Secondarily, where possible, it will assess whether ACPs help to reduce the healthcare costs for high demand patients.

Professor Paul McNamee


Temporal trends of molecular portraits of breast cancer in Scotland

Breast cancer has been clearly shown to be a heterogeneous disease with distinct receptor-specific prognosis and prediction of responsiveness to treatments. Using information collected on receptor status and molecular profiling data from tumour tissue linked to clinical data a unique linked dataset for Scotland would be created that provides an important model of the potential value of biomarkers in the tracking of cancer to better understand: 1) the burden in the incidence of different molecular subtypes of breast cancer over time; 2)whether demographic subgroups of the population differ in the incidence by subtypes 3) whether improvement in mortality gains by molecular subtype (e.g. ER and ‘omics defined subtypes) are realised across different segments of the population (e.g. age, deprivation index, screening) and 4) estimating how many women might benefit from new ‘omics defined subtypes to provide more targeted treatments (Precision Medicine). Cumulatively, these data would inform the planning of future breast cancer prevention and treatment programs in Scotland, elsewhere in the UK and internationally.

The activity described here is to perform a proof of principle study based on tissue data in Lothian in the first instance with the idea of then expanding this work to the rest of Scotland.

Dr Jonine Figueroa


Developing the Lothian IBD Registry

As part of routine clinical care of patients with Inflammatory Bowel Disease (IBD) we have identified a need to audit our standards of care. We have therefore undertaken an exercise in identifying patients with IBD in Lothian from review of health records. It is hoped that we can then audit our practice for quality indicators of care (such as blood tests, stool tests, endoscopy, radiology etc.) to ensure we meet national guidelines.

To ensure we have identified patients appropriately we have asked LRSH to identify patients with IBD from their own search strategy and wish this information to be added to our registry (identified from in-patient admission coding and primary care prescribing of IBD drugs).

Dr Gareth-Rhys Jones


Patient Flow in Hospital: Using machine learning to predict areas of strain in an inpatient setting

At times of high workload in hospitals, areas of strain can develop where bed shortages and staffing issues become critical. This can impact on patient care, for example reducing nursing availability to patients, increasing bed waiting times in emergency areas, and resulting in cancellations to planned procedures when a patient cannot be accommodated afterwards. We aim to use data about patients admitted and the hospital bed state to design machine learning software that will be able to predict how long patients will stay and what their movements between wards is likely to be. This would create a hospital “weather forecast”, allowing staff to be moved between wards in advance, extra staff brought in, or if necessary procedures to be delayed before a patient has to fast and come into the hospital. This pilot study aims to develop simpler software to assess feasibility, with the future aim of improving detail and accuracy.

Dr J Kenneth Baillie


Blood transfusion triggers in patients undergoing major cancer surgery

Anaemia at the time of surgery is common and may require doctors to prescribe blood transfusions. Possible benefits of transfusion include prevention of heart attacks, or other complications. However, blood transfusions can have other side effects such as infections after surgery or increased risk of cancer recurring.

Some research has been done in this area, but doctors are still unsure when to prescribe blood to these patients or how low the blood count should be before a blood transfusion is ordered.

We wish to look at cancer operations at the Royal Infirmary of Edinburgh to find out:

1.         The rate of blood transfusions in patient having surgery for cancer

2.         How low the haemoglobin was when a transfusion was ordered.

3.         If this has change over the last 10 years.

4.         What the rate of complications, including heart attacks is in these patients.

This information will help us design a larger study of transfusion in the patients

Dr Michael Gillies


Understanding and Managing Polypharmacy

This study seeks to better understand polypharmacy – where patients are taking multiple (5+) medications – with its potential risks – and how these may be mitigated by reviewing prescribing data.

Miss Maram Zahraa, Professor Robin Williams, Professor Stuart Anderson


The urinary microbiome and recurrent urinary tract infections

Is there a difference in urinary bacterial communities between patients with a history of recurrent urinary tract infections and those for whom this is their first urinary tract infection?

Community prescribing to support annotation of surplus clinical diagnostic samples as part of a NHS Lothian BioResource study


Cost-effectiveness of patient-matched pre- and on-treatment biomarkers in cancer therapy response prediction

BACKGROUND The development of new biomarkers allows for the accurate prediction of cancer therapy response in the neoadjuvant setting. The implementation of these diagnostic and prognostic tools in the routine NHS care could have the potential for better clinical outcomes and a more cost-effective use of resources. AIMS To develop a decision analytic model for the early economic evaluation of a biomarker-based predictive tool (currently under development at Edinburgh Cancer Research Centre), to compare the clinical effectiveness and cost-effectiveness in a potential routine NHS implementation. METHODS. The economic evaluation of this technology will use of a novel approach that combines molecular biomarker data with clinical events, clinical outcomes and healthcare costs data within a Bayesian Generalized Evidence Synthesis framework, as a basis for constructing patient simulation models. The simulation will be further developed for the incorporation of NHS data. EXPECTED RESULTS The economic appraisal of this biomarker-based technology will provide clinical, budget impact and cost-effectiveness estimates of its potential implementation on the routine NHS clinical pathways. The project will directly inform upcoming technology adoption decisions required by NHS Lothian and the Edinburgh Cancer Centre.

Dr Peter Hall and Giovanni Tramonti


Antibiotic Review Kit for Hospitals (ARK-Hospital) - Electronic Health Record Data Assessment

To assess whether a package of interventions, designed carefully following theories of health psychology and behaviour change, increase healthcare worker uptake of the Department of Health recommended ‘review and revise’ strategy to support the prudent use of antibiotics in hospitals, leading to reductions in antibiotic use in patients admitted to acute/general medicine without any change in 30-day mortality post-admission.

Professor Clifford Leen


Long term mortality after internal fixation of calcaneal fractures: a retrospective study

A research project looking into mortality for patients who required operative fixation of their fractured calcaneus. Causes of death for patients who died during this study will significantly add to the quality of the discussion and hopefully improve the care of future patients with such injuries.

Dr Oliver Brewster


Diabetes telemonitoring trial follow-up study: diabetes management, hospital admissions and deaths

The purpose of this study is to find if there are differences in HbA1c and BP and long term health events in the form of hospital admissions and deaths in 177 patients with Type 2 diabetes, of whom 88 patients were randomised to telemonitoring and 89 to usual care, (between 2011 and 2013) in Lothian.

Dr Janet Hanley


Data Intelligence for the evaluation pf new treatments for breast cancer.

New drugs for cancer are improving outcomes for individual patients but placing a financial strain on NHS Health Boards in Scotland. The Scottish Medicines Consortium uses cost-effectiveness analysis assess cost effectiveness, but Health Boards do not follow a robust process for assessing the potential impact on their budget. This project builds on established methods for cost-effectiveness modelling to incorporate real NHS data in the forecasting of budget impact of new cancer medicines, to allow robust planning of post-adoption service provision.

Robert Nagy, Dr Peter Hall


Primary care detected acute kidney injury, age, socioeconomic deprivation and associated outcomes: an observational cohort study.

Acute kidney injury (AKI) is a sudden or recent reduction in renal function. In the UK up to 100,000 deaths each year in hospital are associated with AKI (NCEPOD,2009). As around 65% of AKI originates in the community (Selby, et al. 2012), this project aims to investigate and characterise cases of AKI identified in primary care according to age. This information will contribute towards robust and evidence-based early detection and management strategies, with the aim of reducing avoidable harm and death and improve patient safety.

Dr Clare MacRae