Clinical decision support tools aim to improve healthcare delivery by enhancing decision making through the provision of targeted health information. The evolution of clinical decision support tools has been boosted by the introduction of new technologies, and their evaluation as useful resources continues to be an area of high interest for health professionals.
This Collection brings together NIHR research on the use of different clinical decision support tools in various areas of healthcare delivery. We asked a number of healthcare professionals and service users to comment on selected Alerts that are relevant and important to them. Their commentary provides important context to the research findings and highlights what we can learn from the studies summarised in the Alerts.
The Alerts included in this collection are:
- A simple test may predict the risk of hospitalisation for flare-up in patients with COPD, a common lung disease
- Decision aids quickly and accurately rule out heart attack for almost half of all patients tested
- Computerised decision support can improve antibiotic prescribing in hospitals
- Tools for GPs can help reduce unnecessary antibiotic prescribing
- Interactive dashboard identifies patients at risk of unsafe prescribing in a flexible and sustainable way
- ICU admission decision support tool showed promise but was rarely used
- Early warning scores for detecting deterioration in adult hospital patients: systematic review and critical appraisal of methodology
- New tool for assessing the severity of type 2 diabetes could help personalise treatment and improve outcomes
A simple test may predict the risk of hospitalisation for flare-up in patients with COPD, a common lung disease
Dr. Sarah Elkin - Consultant Respiratory Physician
“Whilst an excellent assessment tool, the 6 min walk test is often hard to implement as it requires suitable estate in a busy clinic environment and training staff to teach individuals correctly.
“This research which shows that the one minute sit to stand test (STST) can predict hospital admission and length of stay, is exciting as the STST is an outcome measure that could be linked to care planning in primary or community care. For example, if a patient with Chronic Obstructive Lung Disease has a reduced STST, a more intensive follow up by a specialist COPD team may be indicated. It may also identify and motivate those that would benefit from a pulmonary rehabilitation program where their STST may improve.
“I was particularly interested to read this alert as many clinicians have been utilising the STST as part of a desaturation assessment in acutely unwell COVID-19 patients in both primary care and emergency departments. The test was found useful in picking up oxygen desaturation which led to a more timely hospital review or admission.”
Dr. James Larcombe - GP
“The sit to stand test sounds useful. Why? It is simple to administer so could easily be incorporated into routine assessment, just as measuring O2 saturations have recently become a standard examination tool.
“My only doubt is regarding the specificity of this test. I suspect it will pick up general frailty as much as risk of COPD. However, I’ll ignore such a flaw (if it exists) as the utility of the test is great. If it reliably predicts high risk of exacerbations, deterioration, and hospital admissions, it matters little whether this results from a combination of COPD and frailty. GPs and practice nurses with good continuity of care could no doubt predict these risks from good knowledge of patients, but in our current world this situation doesn’t always exist. The tool looks suitable for use in annual review templates. Further research should assess this test for all patients suspected of having frailty and in other comorbidities.”
Decision aids quickly and accurately rule out heart attack for almost half of all patients tested
Read the Alert on decision aids for heart attacks
Dr. Ranjit More - Consultant Cardiologist
“Chest pain presentations comprise over 5% of all visits to Emergency Departments (ED). Most EDs currently use clinical symptoms, patient characteristics, and repeated hs troponin levels 2 hours apart to rule out suspected heart attacks in patients presenting with chest pain. However, the delay involved in this process adds further to the pressures faced by currently stressed EDs and often result in unnecessary hospital admission.
“This study suggests that using an algorithm combining ECG, BP, chest pain characterisation and associated symptoms together with a single hs troponin has a high enough negative predictive accuracy (>99%) to discharge around 40% of chest pain patients early from the ED.
“From the practical point of view the algorithm would in theory be straightforward to implement in most ED departments. One important point however is this study did use a hs troponin T assay (Roche), whereas many other Hospital Trusts in the UK use hs troponin I assays and therefore would not be able to use this approach until a similar validation study has been carried out using hs troponin I assays.
“I look forward to seeing the results of the follow on PRESTO study which utilises this decision making approach earlier in the patient journey i.e. when the patients are seen by paramedics.”
Dr. James Larcombe - GP
“The T-MACS decision aid ruled out heart attacks in 46.5% but missed the diagnosis of only 0.08% (1/132) of patients in the study. Thus it identifies more true negatives without any change in false negatives compared to high-sensitivity troponin I. What isn’t clear is how many patients were excluded before consent as they had obvious non-coronary causes for chest pain or other heart attack variants (they received ‘normal’ treatment).
“Ambulance service testing is planned but will 12-lead ECGs be available universally? Near-patient testing is also usually less reliable than laboratory based tests. We await the outcome of the NIHR- funded study!
“As acute chest pain of possible cardiac origin is a ‘999’ call, this study isn’t relevant to General Practice. It is now rare that we would encounter this situation face-to-face and it would be inappropriate to maintain a near-patient testing facility in every surgery.”
Computerised decision support can improve antibiotic prescribing in hospitals
Read the Alert on decision support for antibiotic prescribing
Tricia Williams - Critical Care Research Nurse
“Although I have been a qualified nurse for a number of years now, I only qualified as an independent prescriber last year. From my experience, I feel that having computerised decision support when considering antibiotic prescribing would be invaluable.
“Our Trust has antibiotic guidelines for most diseases/conditions/infections but these can be time consuming to access. Many antibiotics prescribed are then changed when microbiology results are received. Having a computerised decision support tool to identify the most likely appropriate antibiotic in the first instance would also save financially and prevent the misuse of antibiotics.
“I feel a lot of clinicians would use such a tool as long as there has been involvement from pharmacy and microbiology to establish a strict criteria to work within.”
Dr. Kabir Manchanda - Clinical Pharmacist
“Antibiotics are vital for treating bacterial infections, however, due their irrational use, antimicrobial resistance is on the rise.
“This research indicates that a computerised decision support (CDS) system can help bring about rational use of antibiotics and improve adherence to guidelines among GPs and other prescribers. The use of a CDS can help reduce the volume of antibiotics prescribed as well as the number of serious adverse drug reactions (sADRS) experienced by patients, thus making treatment cost-effective and safer for patients.
“Although CDS presents a promising future for optimizing antibiotic use and improving patient care, its implementation is a major barrier in low- and middle-income countries due to its high cost. It is also necessary to understand the prescriber’s perspectives to ensure that the selected tools meet their needs. Implementation based on a one-size-fits-all approach can lead to prescribers rejecting them.”
Tools for GPs can help reduce unnecessary antibiotic prescribing
Read the Alert on a tool for GPs to reduce antibiotic prescribing
Dr. James Larcombe - GP
“Antibiotics are over-prescribed for minor self-limiting respiratory infections, and may even harm individual patients, as well as contributing to global resistance. This 2018 qualitative meta-analysis is an update of a 2011 review. It didn’t try to determine the best tool to reduce antibiotic prescribing. Instead, it looked at barriers and facilitators of implementation. The 2011 review concluded that interventions should reflect and educate on prescribing habits, facilitate patient-centred care, decrease uncertainty in clinical management, and be beneficial to implement. The 2018 review, however, tells us that a ‘one-size-fits-all’ approach does not work. A negotiation tool that supports patient interaction without damaging patient relationships or interfering with clinical judgement is most effective. It is hoped that such a tool provides ‘golden moments’ in consultations, educating and empowering patients. It is also recognised that interventions can be a compromise to break consultation ‘deadlocks’ and experienced GPs were more likely to feel that tools were unnecessary. Success requires staff engagement and positivity (about the tool), and contextual use.
“The Alert also reports that the NHS committed to halving inappropriate prescribing by this year. I wonder whether this might happen through circumstance? Whilst antibiotic prescribing surged in the initial phase of COVID- 19, possibly through a fear of missing pneumonia, levels of antibiotic prescribing have since reduced perhaps through greater self-management of minor illness. If this shift is maintained during subsequent waves and after the pandemic is over, might it more quickly ensure the success of the NHS aspiration than implementing the findings of this study?”
Interactive dashboard identifies patients at risk of unsafe prescribing in a flexible and sustainable way
Read the Alert on the interactive dashboard
Dr. Kabir Manchanda - Clinical Pharmacist
“Patient Safety is vital to well-functioning health systems. With the fast-growing telemedicine industry and widespread use of electronic health records (EHR), there is an urgent need to develop and implement multi-sectoral strategies to combat the growing medication safety concerns on a global level.
“The Salford Medication Safety daSHboard (SMASH) intervention interrogates the (EHR) databases of general practices and provides feedback to general practitioners (GPs) in a pharmacist-led intervention to curb hazardous prescribing. It involves verifying the information in the dashboard, reviewing patient clinical records and making decisions collaboratively with physicians for enhancing medication safety.
“Although the study helps with the understanding of engagement and collaborative decision approaches of healthcare staff concerning medication safety in general practice, for decisions to be made, the use of the dashboard needs to be clinically minded. Also, the interaction has to be made by the whole healthcare team and not just the pharmacist. Moreover, a separate time should be allotted for the integration of this dashboard into the system.
“Presumably, SMASH presents a promising future for a pharmacist and a clinician to work together to decide upon the ‘best course of action’ for a patient’s safety. We should be utilizing more of such dashboard interventions to rationalize medication use and maximize patient safety measures.”
Dr. James Larcombe - GP
“Potentially hazardous prescribing occurs in over 5% of patients in primary care, and omission of monitoring tests in around 10%.
“This system produces a dashboard of patients at risk. It was developed to address the problems of computerised decision aids, which may interrupt correct decisions and provide too many alerts leading to ‘alert fatigue’ and subsequent non-use.
“The target audience seems primarily to be practice-based pharmacists: in 56% of the study practices they were the sole users. Employment of practice-based pharmacists is widely encouraged and often supported financially, so the study is well-timed.
“It warns us that we can expect increased workload at first, so will practices with current robust systems be prepared to put in extra work for less benefit, or do these practices generate fewer alerts? The study doesn’t help us here.
“I was reassured by the list of hazardous prescribing in the study, but does it consider other prescribing situations where the balance of risks and benefits is more finely balanced, and shared decisions have been made?
“As long as the dashboard is used as a basis for discussion then it will enhance pharmacist/ GP/ patient interactions. It looked unlikely, however, that alerts were generated at the time of prescribing, which reduces its effectiveness but I wonder if there was no other way to avoid over-sensitivity of the system and ‘alert fatigue’.”
ICU admission decision support tool showed promise but was rarely used
Read the Alert on the ICU decision support tool
Professor Paul Dark - Chair of Critical Care Medicine at the University of Manchester
“Previous work with NIHR’s James Lind Alliance has revealed that when patients, carers and healthcare professionals come together in Priority Setting Partnership, they agree that there is a high priority to progress research on how to identify patients who may benefit from intensive care.
“This research provides important multi-source evidence highlighting the considerable challenges for ICU healthcare teams in determining and communicating decisions about care for critically ill patients who often lack mental capacity to take part in these decisions.
“The decision support tools developed by the researchers, as a first step, should help identify and document these complex decisions better and will encourage improved communication of care uncertainties with patients, families, and other healthcare professionals.”
Tricia Williams - Critical Care Research Nurse
“We are a relatively small ICU in comparison to other units which means empty beds are often in short supply. Although not always the case this can be caused by previous admissions that have not always been appropriate. At present in our trust we do not use a ‘tool’ to assess a patient's suitability to receive the somewhat intense therapies and treatments ICU offers. The decision to admit takes into account the already existing guidelines that consider if a patient’s condition is reversible. Often this decision will be difficult to make, and therefore to allow more time to consider the limitations of treatments a patient is to receive and possible discussions with family members, patients are admitted to ICU while that decision is made. Sometimes this results in more difficult decisions and discussions with family having to take place as a result of an inappropriate admission.
“To prevent these admissions happening and to give guidance on admission criteria a Decision Support Tool would be valuable. I do feel it needs to be easily accessible and should be integrated into our existing systems or even form part of the admission paperwork.
“I feel Junior Doctors and Critical Care Outreach team members would gain confidence and justification for their decisions and would also use it as a referral tool. It does need to be adapted to ensure senior clinicians agree with the assessment categories used.
“It would be interesting to know the size of hospitals it was trialled in as this may impact on its success. Our senior clinicians are very proactive and supportive of the junior team and as we are a small DGH I feel they would possibly have the time and resources to adapt such a tool and encourage its use. The impact of a Decision Support Tool, both financially and on staffing an ICU unit safely, is potentially huge, therefore any resource that could ensure appropriate admissions should be considered and implemented.”
Early warning scores used in hospitals must be based on sound science
Read the Alert on early warning scores
Professor Paul Dark - Chair of Critical Care Medicine at the University of Manchester
“Early warning scores (EWS), based on readily measurable vital signs, have been widely adopted in acute healthcare settings. They were developed and implemented to screen hospital patients to identify acute physiological deterioration and to assist the multidisciplinary healthcare team prioritise care. It has also become evident that EWS are associated with poorer patient outcomes (e.g. in-hospital death). This new systematic review identifies important quality gaps for the design and reporting of EWS studies in acute healthcare. However, most of the reported studies were focused on predicting patient outcomes rather than whether the scores were effective in leading to care that improved patient outcomes. As opportunities are realised to progress automated innovations within electronic patient records, with the potential to rapidly embrace a greater range of patient data, this study highlights the need for the highest quality research to assure the best evidence for patient care.
“Through NIHR James Lind Priority Setting Partnerships, patients and relatives have helped highlight the need for acute healthcare systems to identify who is most likely to benefit from acute healthcare interventions (e.g. admission to ICU), ideally in advance of vital sign deterioration: an urgent challenge not addressed by this research to date.”
Dr. James Larcombe - GP
“I’ve always been somewhat sceptical of the way these tools have been implemented so that clinical judgement is downgraded. They were never validated in the community, and our context is often very different to hospital- based care. More worryingly, this review questions the quality of the evidence underpinning their use even in secondary care. It suggests that even small inaccuracies in the scores could cause excess work and reduced predictive value. However, NEWS2 can still be an adjunct to assessment in primary care, until better models are developed, but its use as a threshold for specific actions, to the exclusion of clinical judgement, should not be supported.
“On a positive note, the authors suggest that we can expect more nuanced and personalised risk tools in the future. This might better suit General Practice with our wealth of personalised data – but the studies must be high quality and undertaken in a community population!”
New tool for assessing the severity of type 2 diabetes could help personalise treatment and improve outcomes
Read the Alert on the tool for type 2 diabetes
Jan Davies - Patient and Public Involvement (PPI) volunteer and member of the All Wales Diabetes Patient Reference Group
“This Alert may change outcomes for diabetic related problems which include amputations. This may also help in reducing unscheduled care.
“My experience of introducing an algorithm into GP surgeries has not always been successful. GP’s have commented on the time taken to access and sometimes need to employ extra staff to facilitate. Therefore, convincing patients and GP surgeries to take this algorithm onboard will need to show considerable advantages.
“The significant wording in this study such as COULD and MAY, makes the study less convincing. More in depth research on this Alert is needed for all the issues to be addressed.”
Dr. James Larcombe - GP
“Diabetes and its management affects ever more people and consumes ever more budgets.
“The diabetes severity score (DISSCO) better predicts adverse outcomes, especially cardiovascular disease, compared to HbA1c levels. It is based on 34 indicators of disease severity said to be recorded routinely in electronic health records. To be fully utilised, coding must be accurate and the scoring template populated electronically. The Alert suggests that a paper version might be the first step, but I can’t see that catching on!
“The study admits that it has only shown its worth as a risk predictor, but further studies are needed to show that this can translate to better interventions and patient outcomes.
“HbA1c may not be a good comparator: the study’s data itself showed that higher risk scores were not associated with the highest HbA1c levels. Patients already suffering from complications will have a higher DISSCO score, but subsequent intensive treatment and monitoring will have lowered their HbA1c levels: a case of ‘the horse has bolted’.
“DISSCO may only prove its worth if it can predict adverse outcomes early in a patients’ diabetic career. Even then there is debate about whether provision of risk information when complications are more likely but haven’t yet happened leads to changes in behaviour.”
Dr. Kabir Manchanda - Clinical Pharmacist
“India has the second-highest number of diabetes patients aged 20-79 as of 2019. International Diabetes Federation Diabetes Atlas makes it clear that India needs to pause and re-evaluate its strategy to combat diabetes.
“Diabetes, being a lifestyle disorder with multi-component risk factors, demands a multidimensional treatment approach. At present, our healthcare system is shaped for acute care as opposed to chronic care. India requires an efficient diabetes prevention program. Moreover, the country also contains a massive burden of pre-diabetics. If targeted with evidence-based information on the correct lifestyle options to keep blood-glucose, lipids and blood pressure in check, we will possibly prevent at least a third of people from developing diabetes.
“There is an urgent need to develop and implement multi-sectoral strategies to combat the growing epidemic. As of now, a patient’s glycated haemoglobin (HbA1c) levels provide an indication of their average blood glucose levels that is measured routinely over weeks or months. Doctors use this measure in drawing up a disease management plan for that individual. A validated score that can accurately assess who is most at risk of complications will help focus interventions and increase cost effectiveness.
“The new Diabetes Severity Score (DISSCO) shows promise to outperform the standard HbA1c blood test, which is unidirectional in nature, and thus, a poor indicator of adverse outcomes. With the fast-growing telemedicine industry in India, this new tool could support doctors in providing more tailored interventions for their patients and spare-unnecessary interventions, hence preventing patients from future complications.”
Learn more about our contributors
Dr. Sarah ElkinDr Sarah Elkin is a Consultant Respiratory Physician working both in the acute trust (Imperial College NHS trust) managing respiratory & medical emergencies and in the community leading a large integrated multidisciplinary respiratory team. She works closely with GPs in NW London, as chair of the clinical reference group, running and developing respiratory pathways from early diagnosis to advanced care, running community based clinics, advice and guidance and supporting the pulmonary rehabilitation, oxygen and supported discharge programs. She is Joint-Clinical Director of the Respiratory network - NHSE/I (London Region) and an Honorary senior lecturer at imperial college London.
Her Clinical & research Interests include: Models of care delivery, integrated care, Quality improvement, COPD, & Pulmonary rehabilitation
Dr. James LarcombeJames Larcombe is a locum GP currently working for NHS 111 Covid Assessment Service. He was a partner in Sedgefield, Co Durham for 31 years. He has a clinical and research interest in infections and has helped NIHR in various guises for many years. James is also an RCGP examiner and practice development support, post-graduate tutor, BNFC formulary committee member, and he co-moderates the NICE GP forum.
Dr. Ranjit MoreRanjit is a Consultant Cardiologist at the Lancashire Cardiac Centre Blackpool, co-lead for Cardiovascular Disease for the Northwest Coast Clinical Research Network and a staff governor for the local Acute Trust. He undertook his cardiology training in King’s College Hospital, Leicester and St Mary’s London. Ranjit’s specialist interests are in coronary artery disease and coronary intervention, structural heart disease, heart failure and arrhythmias.
Tricia WilliamsTricia’s background is 3 years as a staff nurse in Emergency Medicine followed by 8 years as a staff nurse in Intensive Care including transferring the critically ill via ambulance to specialist units in the South West of England. She is now in her third year of working as the Critical Care Research Nurse for the Dorset County Hospital NHS Foundation Trust, setting up and running all clinical trials involving Intensive Care and Anaesthetics. Last year Tricia completed the Independent Prescriber Course at Bournemouth University. Currently, she is running The Recovery Trial and the Siren Trial driven by Public Health England and she is Joining the Vaccine Hub at Bournemouth Hospital this week to begin the Vaccine Trial from Oxford University.
Dr. Kabir ManchandaKabir Manchanda (qualified Pharm.D Scholar) is a clinical pharmacist pursuing an internship at Shree Mahant Inderesh Hospital, Uttarakhand, India. He has done his clinical rotations within a range of healthcare settings ranging from internal medicine, infectious diseases, nephrology, psychiatry, neurology, and endocrinology departments. His expertise is in prescription auditing, patient counseling, ADR monitoring & reporting, and drug utilization evaluation (DUE) studies. His research interests are in infectious diseases, depression, and diabetes.
Professor Paul DarkPaul is an academic clinician at the Northern Care Alliance NHS Group (Salford Royal), Professor of Critical Care Medicine at the University of Manchester and sub-theme lead at NIHR Manchester Biomedical Research Centre. He is also NIHR's National Speciality Lead in Critical Care based at King's College London.
Jan DaviesJan has been involved in Patient and Public Involvement (PPI) for many years. She has been a co-applicant in a number of studies and published papers.
Jan is a consumer panel member of SAIL (Secured Anonymised Information Linkage) database and the IGRP (Information Governance Review Panel) at Swansea University. She is also a member with PRIME (Primary and Secondary Research) based at Cardiff University; All Wales diabetic reference group; Operation Working Group, Health and Care Research Wales.
Jan has more than one co-morbidity and therefore her interest in health research is wide. Jan’s main objective of being part of PPI is to make a difference.
