4.3 Comparison with existing literature
The need for extra time to conduct a risk calculation and communicate this effectively, listening, informing, explaining and discussing further investigations with the patient, as found in this study, are known to add complexity and costs to the patient consultation.10 Additional time is a scarce resource in the face of increasing practitioner workload, which will affect implementation of innovations like this. Integration of the risk assessment tool within the general practice IT system, with links to existing patient data, and training on how to use the tool were seen as essential to addressing issues of time and complexity of use.
Patients feeling worried or anxious about being referred for cancer investigations was perceived as another barrier to the use of the tool, although this does not correspond with a recent systematic review of randomised controlled trials of cancer risk assessment tools in primary care, which found no increase in cancer worry.17 This contradiction in the evidence could be due to the fact that, the current study is qualitative and based on the views of participants, particularly service users who had not yet experienced the process of using the tool with a clinician but were expressing what they felt could happen, if the cancer risk information including referral, is not properly communicated to them.9,18 On the other hand, the participants in studies included in the systematic review of trials might have benefited from interventions including clinician input, which could have reduced their anxiety levels. Nevertheless, the fact that these views were expressed in this study suggests that some patients could indeed experience worry and anxiety, if the patient care, from investigations through to diagnosis and treatment of cancer is not properly planned and carried out to meet the needs of individual patients.
Another barrier identified by participants was over-referral, although some practitioners felt this was unlikely to be a problem because they felt they could use risk assessment tools alongside their professional judgement, only referring patients who needed this. This is supported by current guidance that Cancer Decision Support tools should prompt primary care practitioners to think about the possibility of cancer, and then decide on referral based on their clinical judgment.19 Despite this, the potential for increased rates of referral of people without cancer (false positives) remains a concern, with potential costs needing to be weighed against late referral.
Practitioners in a simulation study conducted in Australia appeared not to trust some risk outputs of the QCancer tool,20 and this accords with scepticism expressed by some clinicians in this study. Practitioners might be sceptical because they perceive the evidence that QCancer or other cancer risk assessment tools improve outcomes is limited. Intervention characteristics can be an important potential facilitator or barrier,10 particularly when evidence of effectiveness in practice is lacking.21
Previous studies have found barriers that needed to be considered for successful implementation of cancer risk assessment tools in primary care including: lack of trust on the part of some GPs in the risk calculation, especially when it conflicted with clinical judgement; variable interpretation of symptoms leading to variations in risk assessment, and the difficulty experienced by some GPs in communicating numerical risk to patients.20 Participants in this study felt that the tools might conflict with existing guidelines, a finding in line with the construct of compatibility,10i.e. whether an intervention fits with existing workflows.
Dikomitis and colleagues found that training and guidance were needed in using cancer risk assessment tools in routine practice22 because of difficulties experienced by practitioners in employing the tools. Similarly, practitioners in this study expressed concerns relating to difficulties in understanding, accessing and using the cancer risk assessment tool, and called for integration of the tool in general practice IT systems and training of practitioners on accessing and using the tool. Hence, lack of integration of the tools in the practice IT system and lack of or inadequate training were perceived as barriers to implementation of the tool. Indeed, to meet the needs of patients, clinicians need to be supported with their learning needs such as sources of information,23,24 about decision aids like cancer risk assessment tools.
Another barrier identified was the perception that patients with symptoms suggestive of cancer would need to be referred for further investigations irrespective of their quantified risk. Indeed, it has been suggested that when using Cancer Decision Support tools, practitioners who suspect a possible cancer diagnosis can refer a patient even if their quantified risk is low or does not meet the referral NICE guidelines.19 Furthermore, Macmillan Cancer Support who have integrated the QCancer and Risk Assessment Tool (RAT) and have called them the electronic Cancer Decision Support (eCDS) tools, have suggested that these tools can complement existing NICE guidelines by flagging an alert on the computer screen about the possibility of cancer. Following this flagging on the computer, the clinician can then decide whether to refer a patient, based on NICE guidelines.19
With reference to facilitators, participants in this study felt that the use of the cancer risk assessment tool could support decision-making especially with patients whose cancer symptoms were unclear, or when cancer was a differential diagnosis, helping to speed up the assessment, diagnosis and treatment of cancer. The findings from this study support findings from a previous study on GPs’ experiences of using diagnostic tools, which found that the RAT helped GPs with lung and colorectal cancer symptom recognition and confirmed their decision about whether to refer.25 In another study, embedding electronic decision-support tools was found to have educational benefits, with GPs reportedly learning about cancer symptoms when using cancer risk assessment tools.22,25 Green and colleagues also found that embedding clinical decision support tools in clinical practice was more likely to be achieved when they were used to support, rather than supersede, the clinical judgement of practitioners,25corresponding with the views expressed by practitioners in this current study.
Another facilitator to the use of the tools found in this study, that the use of the cancer risk assessment tool could help to identify, raise awareness of and promote positive health behaviours in patients, supports findings from recent systematic review of randomised controlled trials of cancer risk assessment tools in primary care, which suggested that health promotion messages within tools may have positive effects on behaviour change.17
The cancer risk assessment tools meant for people presenting with symptoms are based on an individual patient’s risk factors and symptoms; this helps to personalise care. Since the risk generated is for an individual patient, the referral and further investigations as well as the subsequent diagnosis and treatment or cancer care will be specific to that individual, taking into consideration their individual attributes and symptoms. Personalised or person-centred care is about taking into consideration the desires or values, social circumstances and lifestyles of people, while working with people as individuals to develop appropriate solutions.13,26
4.4 Implications for practice and further research
For a successful implementation, barriers to the use of the cancer risk assessment tool need to be addressed. This will include ensuring that the tool is integrated in the general practice IT systems and training practitioners on how to access and use the tool during the patient consultations. It is of note that Macmillan Cancer Support and Cancer Research UK have worked with the major primary care IT providers (EMIS, SystmOne and Vision+) in integrating the eCDS tools in general practice systems, although this integration will need to be enforced as expressed by participants in this study.
Practitioners are likely to refer patients with symptoms suggestive of cancer whatever their quantified risk if these fall within NICE cancer referral guidelines, suggesting that risk assessment tools should be used flexibly to fit with clinical practice.
Quantitative research is needed to examine the effects of the use of the cancer risk assessment tool (QCancer) on rates of referral, investigations and diagnoses, whether positive or negative for cancer. Further research is also needed to quantitatively evaluate whether the cancer risk assessment tool improves patient outcomes compared with current practice.