Clinical Decision Support Tools
Clinical Decision Support Tools Brief
Click here to read the CRUK CDS brief which provides and overview of clinical decision support tools available in primary care
Computer-based algorithm tools
Some CDS tools are underpinned by algorithms which calculate the risk of a patient having an undiagnosed cancer. Algorithms can be based on symptoms and tests alone or consider other factors such as patient age, gender and risk factors. Computer-based algorithms are often integrated into a GPs usual patient management system, meaning they are accessible during consultations.
The Risk Assessment Tools (RATs) were developed by Professor Willie Hamilton, designed to be used in symptomatic populations presenting to primary care and support GPs in deciding which patients require further investigation or referral. The RATs are based on algorithms which estimate the risk of a patient having an undiagnosed cancer, expressed as the positive predictive value (PPV) of their current symptom(s).
The tool includes a series of matrix tables for 15 cancer types, colour-coded to depict PPV ranges:
• <1% = white
• 1-2% = yellow
• 2-5% = orange
• >5% = red
The current 2015 NICE referral guidelines for suspected cancer referral suggest referral at a threshold of 3% and above. Although the colour code and associated risk values may help support GP decision making, these should not replace clinical judgement.
The QCancer tool was developed by Professor Julia Hippisley-Cox and Carol Coupland, based on an algorithm to calculate the absolute risk of a symptomatic or asymptomatic individual having a yet undiagnosed cancer. The risk score considers patient symptoms, risk factors, such as age and sex, smoking status, family history, postcode and certain comorbidities.
For symptomatic populations, the QCancer tool provides absolute risk of all cancers combined, as well as a breakdown of risk for 12 different cancer types individually.
Macmillan have developed a CDS tool which integrates the RATs and QCancer® tools into a GPs IT system, designed to support clinical decision making by displaying the risk of patients having a yet undiagnosed, site-specific cancer. The CDS tool is currently available for all EMIS and INPS Vision users across the UK.
Distinguishing malignant from benign pigmented skin lesions in practice can be challenging. The weighted 7-point checklist (7PCL) is a single-site CDS tool which can be used to support GPs in assessing patients who present with pigmented skin lesions. The 7PCL is recommended by NICE for GPs in the UK and was disseminated by EMIS in 2016.
Lesions scoring 3 or more is suggestive of possible melanoma and patients should be referred urgently. However, any one feature is adequate to prompt urgent referral if melanoma is suspected. If a lesion scores less than 3, and there is a low degree of suspicion that the lesion is melanoma, GPs should advise the patient to monitor the lesion (explaining the features to watch for) and to come back if they notice any changes.
Evaluation of computer-based algorithm tools
There are several studies which have evaluated the use and impact of CDS tools which are based on algorithms to calculate a patient’s risk of having an undiagnosed cancer.
For example, in an evaluation of the 7PCL for melanoma, GPs found the tool useful and easy to use, particularly for ‘borderline decision making’ and to facilitate discussion with patients [1]. However, authors noted that there is variation in the implementation and use of the tool by GPs.
A key challenge to the use of CDS tools is the effective implementation into practice. A 2017 survey disseminated to GPs across the UK assessed the availability and use of QCancer® and Risk Assessment Tools (RATs) [2]. Less than 20% of practices had electronic CDS tools available and more than a third of GPs were unaware of the tools. To better understand the use and impact of CDS tools, more robust evaluation is needed.
References:
- Price, S., et al., Availability and use of cancer decision-support tools: a cross-sectional survey of UK primary care. Br J Gen Pract, 2019.
- Pannebakker, M.M., et al., Understanding implementation and usefulness of electronic clinical decision support (eCDS) for melanoma in English primary care: a qualitative investigation. BJGP open, 2019. 3(1): p. bjgpopen18X101635-bjgpopen18X101635.
Research undertaken as part of the ACE Programme set out to evaluate the QCancer tool, to find out if GP decision making can be influenced by CDS tools [1]. Findings showed that the tool can heighten a GP’s awareness of cancer and help with decision making around complex patients. Additionally, a QCancer risk score can help to legitimise a GP’s referral and reinforce gut feeling. To increase the impact of CDS tools, increased training and promotion of tools is needed, as this had a positive impact on GP uptake and understanding of QCancer. Although the study hoped to clarify an association between a QCancer risk score and resulting cancer diagnosis, this was not successful and so further research in this area is needed. Read the full ACE Programme report.
References:
- Accelerate, Coordinate, Evaluate (ACE) Programme. Using Cancer Decision Support Tools to support the early diagnosis of cancer. 2017.
Cancer Research UK coordinated an independent evaluation of the 2013 Macmillian pilot of their integrated CDS tool which used the BMJ Informatica platform, where GPs from over 400 practices across England had access to the software between March and November 2013. The GPs were divided into two groups, with one group being presented with risk scores from the RATs and the other from QCancer.
Cancer Research UK analysed the data collected and a Policy Research Unit within the Department of Health gathered qualitative feedback from interviews with patients and GPs, to gain insight on the use of the CDS tool.
Overall, the pilot evaluation suggests that the electronic CDS can:
- raise GPs’ awareness of cancer symptoms
- alert and remind GPs when patients are potentially at risk
- influence the decisions GPs make about how to care for patients
In 19% of cases where GPs provided feedback, they reported that had they not used the symptom checker, they would not have referred/investigated the patient.
GPs also expressed some concerns:
- they didn’t always agree with or understand how to interpret the risk scores produced for their patients
- the risk scores produced rely on data recorded in a patient’s medical records – which may not necessarily be 100% accurate. This could flag patients who aren’t actually at risk or miss others who are
- they could end up with ‘prompt fatigue’ because of how regularly the programmes prompt them alongside existing computer prompts
- in a 10-minute consultation it was sometimes difficult to use the symptom checker function properly
- some GPs felt they weren’t focusing on interacting and talking to their patients as they would be looking at their computers more
How can Cancer Decision Support tools be used in practice?
General functionalities that may be included within an electronic CDS tool:
An automated prompt that draws on the information in a patient’s medical record and sends an alert if the calculated score exceeds a set threshold
A symptom checker that provides a cancer risk score based on symptoms entered onto the system
A risk stratification list draws on medical data within patient records and produces a list of all patients within that practice exceeding a set threshold of risk.
Translating the large number of cancer guidelines into practice can be challenging, but there are several visualisation CDS tools which may make guidelines for suspected cancer easier to use.
Cancer Research UK have produced visualisation summary tools for NICE NG12 and the Scottish Referral Guidelines for Suspected Cancer (SRG) in the form of body maps.
CRUK Summary of NICE cancer recognition and referral guidelines (NG12)
CRUK Summary of Scottish referral guidelines for suspected cancer
Cancer Maps:
The interactive Cancer Maps were developed by Dr Ben Noble, designed to help clinicians identify possible cancer by illustrating the NICE NG12 recommendations for reference during consultations. Users can input the patient’s gender, age and symptoms to ascertain whether they appear in the NG12 guidance. If relevant, selected symptoms will flash red on the map(s) where they appear, with the corresponding referral pathway guidance.
Formal evaluation of visualisation tools which illustrate cancer referral guidelines for suspected cancer is limited and more insight into what GPs find useful in practice is needed.
In a pilot evaluation of the Cancer Maps in England, around 9 in 10 GPs felt more confident when making cancer referrals and 97% of GPs would recommend the tool to other clinicians. Although only 31% of GPs stated it gave them new knowledge, 68% felt it consolidated existing knowledge. Additionally, the Cancer Maps helped with patient communication, as 76% of GPs felt more confident when talking to patients about possible cancer indicators. Over half of the GPs in the pilot evaluation made referrals they otherwise wouldn’t have, proving that the maps can influence decision-making in practice.
See more about the pilot evaluation of the Cancer Maps tool.
Clinical Decision Support (CDS) tools are designed to support health professionals to recognise potential cancer signs and symptoms and manage patients appropriately. CDS tools can be broadly categorised into 2 different groups:
- Computer-based algorithm tools
- Electronic and printed visualisations of guidelines
CDS tools do not replace clinical judgement. Instead they provide further guidance to inform patient management decisions. This may be to refer, to safety net, or it may support the decision on which pathway is most appropriate.
At the University of Exeter, Professor Willie Hamilton is leading the Electronic Risk Assessment for Cancer (ERICA) trial to evaluate 6 newly developed electronic RATs. Find out more about how your practice could get involved here.