We thank Subbe et al for engaging with our paper ‘Understanding the enablers and barriers to implementing a patient-led escalation system: a qualitative study’ in their editorial. The editorial highlights the importance of having a system-led response to patient-led escalation. The editorial also draws attention to the patient and family level trauma and distress that can result from having concerns about a deteriorating condition dismissed by staff, as evidenced by the personal experience provided by one of the co-authors, Alison Phillips. Indeed, in one of our earlier publications from the same study, we similarly highlight the potential for epistemic injustice to arise when patients’ concerns are ignored or dismissed, particularly when in a condition of extreme vulnerability.(1) Our paper and the linked editorial by Subbe et al share the aim of supporting patients in being heard and receiving timely responses when their condition is deteriorating.
Despite the increased interest, policy activity and enthusiasm for patient-led escalation, there are still significant gaps in the evidence base around implementation of such systems, particularly with regard to how they work in relation to wider socio-cultural rescue systems and across different clinical settings.(2) The premise of our paper was not to undermine the importance of patient-led escalation systems but rather to explore and take learning from the factors that shaped implementation of one particular pilot sy...
We thank Subbe et al for engaging with our paper ‘Understanding the enablers and barriers to implementing a patient-led escalation system: a qualitative study’ in their editorial. The editorial highlights the importance of having a system-led response to patient-led escalation. The editorial also draws attention to the patient and family level trauma and distress that can result from having concerns about a deteriorating condition dismissed by staff, as evidenced by the personal experience provided by one of the co-authors, Alison Phillips. Indeed, in one of our earlier publications from the same study, we similarly highlight the potential for epistemic injustice to arise when patients’ concerns are ignored or dismissed, particularly when in a condition of extreme vulnerability.(1) Our paper and the linked editorial by Subbe et al share the aim of supporting patients in being heard and receiving timely responses when their condition is deteriorating.
Despite the increased interest, policy activity and enthusiasm for patient-led escalation, there are still significant gaps in the evidence base around implementation of such systems, particularly with regard to how they work in relation to wider socio-cultural rescue systems and across different clinical settings.(2) The premise of our paper was not to undermine the importance of patient-led escalation systems but rather to explore and take learning from the factors that shaped implementation of one particular pilot system. As noted in our paper, the Early 3S system is only one of four intervention strands of the RESPOND study, which as a whole, tests the hypothesis that a bundle of Human Factors interventions, designed to address issues across the rescue response process, can decrease mortality from complications after emergency abdominal surgery. Inclusion of Early 3S within this bundle signifies the value attributed by the RESPOND team to the role of patient and family involvement in rescue.
Subbe et al suggest limited insights can be drawn from our paper, given both the design of Early 3S and the process evaluation. We would argue that our use of ethnographic methods enabled us to learn not only from successful implementation but also from situations where implementation was difficult or unsuccessful. For instance, our paper draws attention to the challenges of implementing a system without a critical care outreach team, where a responder is limited to within the ward team. Similarly, our data highlight complexities around the meaning of patient-led activations and understanding what a low or high rate signifies.
Subbe et al acknowledge that, rather than being down to individual behaviours, there are typically systems and cultural issues at play that prevent a timely and compassionate response. This reaffirms the contribution of our work, which examines and illuminates the socio-technical nature of patient-led escalation systems. Through our use of ethnographic methods, we were able to highlight some of the wider system issues, displacement effects, and interactional factors that require consideration to make these patient-led escalation systems work in practice, given other competing quality and safety improvement efforts. Our paper cautions against the oversimplification of the implementation of patient-held checklists and shared decision-making tools, which in reality require adaptive and complex labour to make them work.(3)
To conclude, we share Subbe et al’s belief that patient-led escalation systems should be considered as part of a compassionate and effective healthcare system. We hope that our paper’s qualitative insights into the socio-technical nature of such systems, and the complexity of relationships involved, can help contribute to this emerging field, particularly given the roll-out of Martha’s Rule within the NHS from April 2024.
References
1. Sutton, E., et al. (2022). Am I safe? An Interpretative Phenomenological Analysis of Vulnerability as Experienced by Patients With Complications Following Surgery, Qualitative Health Research. 32(14), 2078-2089.
2. McKinney, A., et al. (2021). Patient and family involvement in escalating concerns about clinical deterioration in acute adult wards: A qualitative systematic review, Nursing in Critical Care. 26(5), 352-362.
3. Bosk, C.L., Dixon-Woods, M., Goeschel, C.A. and Pronovost, P.J., 2009. Reality check for checklists. The Lancet, 374(9688), pp.444-445.
We were interested to read the recent article on patient safety in remote primary care encounters by Payne et al. We have been reviewing the use of remote consultations specifically for paediatric asthma patients and would like to thank the authors for their work.
Firstly, we agree with the authors’ findings that a remote environment may exacerbate existing inequalities such as economic and language barriers. We would add that an additional factor that must be assessed is a patient’s ability to use technology. Pinnock et al. highlight the risk of virtual consultations to those who lack “e-literacy (or digital healthcare literacy)” (2). These patients must be identified and offered additional support or alternative methods of consultation to maintain the utmost level of care.
Payne et al. highlight the need for a more definitive approach to escalating care rather than a “rule of thumb” or “if in doubt, put it down as urgent” approach (1). We would echo that there need to be clear guidelines and more specific thresholds for escalating care from remote to in-person visits. One suggestion by Galway et al. is having a lower threshold for seeing younger children face to face (3). Galway et al. also suggest having alternative “red-flag” signs that are unique to the remote setting. For example, multiple calls from a patient may indicate the need to escalate their care to face-to-face. More of these red-flag signs unique to this setting need to be...
We were interested to read the recent article on patient safety in remote primary care encounters by Payne et al. We have been reviewing the use of remote consultations specifically for paediatric asthma patients and would like to thank the authors for their work.
Firstly, we agree with the authors’ findings that a remote environment may exacerbate existing inequalities such as economic and language barriers. We would add that an additional factor that must be assessed is a patient’s ability to use technology. Pinnock et al. highlight the risk of virtual consultations to those who lack “e-literacy (or digital healthcare literacy)” (2). These patients must be identified and offered additional support or alternative methods of consultation to maintain the utmost level of care.
Payne et al. highlight the need for a more definitive approach to escalating care rather than a “rule of thumb” or “if in doubt, put it down as urgent” approach (1). We would echo that there need to be clear guidelines and more specific thresholds for escalating care from remote to in-person visits. One suggestion by Galway et al. is having a lower threshold for seeing younger children face to face (3). Galway et al. also suggest having alternative “red-flag” signs that are unique to the remote setting. For example, multiple calls from a patient may indicate the need to escalate their care to face-to-face. More of these red-flag signs unique to this setting need to be identified and then relayed to professionals during specific training on consulting remotely.
Payne et al. highlight that one major challenge to remote consults is the inability to perform physical examination of patients. We agree that the inability to examine a patient means that remote consults are not always appropriate. With particular regards to asthma patients, Eguiluz-Gracia et al. suggested that this limited or inaccurate clinical information can lead to an increased risk of corticosteroid over-prescription (5). As Payne et al. rightly point out, a preliminary rather than definitive diagnosis is perhaps the way forward. Payne et al. also described the limited opportunities for demonstrating or checking patients’ technique, however Agerskov et al. concluded that spirometry can still be effectively carried out remotely (4). Our own recommendation and clinician practice is not to use remote consultations for first diagnostic consultation but rather in a follow-up setting, and instead arrange in-person consultations for an initial diagnosis.
Lastly, we would like to highlight some risks. Payne et al. identify the dangers of relying on information from a third party due to it potentially being inaccurate. In addition to this, another potential danger that requires research is whether remote consultations still enable doctors to safeguard effectively. Southgate et al. also highlight that there is a lack of literature on safeguarding in remote consults (6). We feel that resolving this is paramount to the future of remote consultations and maintaining patient safety.
References:
1. Payne R, Clarke A, Swann N, et al Patient safety in remote primary care encounters: multimethod qualitative study combining Safety I and Safety II analysis BMJ Quality & Safety Published Online First: 28 November 2023. doi: 10.1136/bmjqs-2023-016674
2. Pinnock H, Noble M, Lo D, et al. Personalised management and supporting individuals to live with their asthma in a primary care setting. Expert Rev Respir Med. 2023;17(7):577-596. doi:10.1080/17476348.2023.2241357
3. Galway N, Stewart G, Maskery J, et al. Fifteen-minute consultation: A practical approach to remote consultations for paediatric patients during the COVID-19 pandemic. Arch Dis Child Educ Pract Ed. 2021;106(4):206-209. doi:10.1136/archdischild-2020-320000
4. Agerskov N, Coughlin S, Parrott H, et al. Quality of Unsupervised Home Spirometry in Children with Asthma. B9. B009 PATIENT, FAMILY, AND PROVIDER EXPERIENCES IN CRITICAL CARE AND CHRONIC PATIENT MANAGEMENT. May 1, 2021, A1088-A1088
5. Eguiluz-Gracia I, van den Berge M, Boccabella C, et al. Real-life impact of COVID-19 pandemic lockdown on the management of pediatric and adult asthma: A survey by the EAACI Asthma Section. Allergy. 2021;76(9):2776-2784. doi:10.1111/all.14831
6. Southgate G, Yassaee AA, Harmer MJ, et al. Use of Telemedicine in Pediatric Services for 4 Representative Clinical Conditions: Scoping Review. J Med Internet Res. 2022;24(10):e38267. Published 2022 Oct 26. doi:10.2196/38267
Smith et al (2022) present an excellent review of the current state of safety netting theory and practice in Primary Care. In relation to the education of paramedics working in primary care, the General, Logistical, Red Flag and Risk approach has been suggested as a means of implementing the theory of safety netting into a clinician’s practice (Mallinson, 2023). It reminds clinicians to always give a General worsening statement, meeting Smith et al’s Recommendation No. 1. The second point is to provide clear Logistical advice to patients on how to seek help; the specifics of what phone number to call to seek or summon help. This perhaps aligns to Recommendation 9; “including a specific safety-netting plan”. Red Flag safety netting reinforces the importance of patient education in relation to possible serious deterioration which aligns to Recommendation 4 in terms of “specific situations that should be cause for concern”. The final component of Risk based safety netting relates to shared decision making and specifically situations where there is disagreement and a patient is not following medical advice. The Risk component seeks to ensure that patients are fully aware of the potential risks inherent in their chosen course to action. This is vital in relation to ensuring patients are making informed decisions about their care. There is definitely more work to be done on improving education around Safety Netting, and Smith et al’s work provides us with clear recommendati...
Smith et al (2022) present an excellent review of the current state of safety netting theory and practice in Primary Care. In relation to the education of paramedics working in primary care, the General, Logistical, Red Flag and Risk approach has been suggested as a means of implementing the theory of safety netting into a clinician’s practice (Mallinson, 2023). It reminds clinicians to always give a General worsening statement, meeting Smith et al’s Recommendation No. 1. The second point is to provide clear Logistical advice to patients on how to seek help; the specifics of what phone number to call to seek or summon help. This perhaps aligns to Recommendation 9; “including a specific safety-netting plan”. Red Flag safety netting reinforces the importance of patient education in relation to possible serious deterioration which aligns to Recommendation 4 in terms of “specific situations that should be cause for concern”. The final component of Risk based safety netting relates to shared decision making and specifically situations where there is disagreement and a patient is not following medical advice. The Risk component seeks to ensure that patients are fully aware of the potential risks inherent in their chosen course to action. This is vital in relation to ensuring patients are making informed decisions about their care. There is definitely more work to be done on improving education around Safety Netting, and Smith et al’s work provides us with clear recommendations in this regard moving forwards.
References
Smith C.F., Lunn, H., Wong, G. and Nicholson, B.D.(2022). Optimising GPs’ communication of advice to facilitate patients’ self-care and prompt follow-up when the diagnosis is uncertain: a realist review of ‘safety-netting’ in primary care. BMJ Quality & Safety. 31:541-554.
Mallinson, T. Consultation Skills and the Diagnostic Process. In. Eaton, G., Proctor, A. and St Leger-Francis, J. (2023). Primary Care for Paramedics. Class Professional Publishing. Bridgwater, UK.
I work in a hospital at home (H@H) service and have found the AFN website a very useful learning resource and regularly recommend it to my Clinical practitioners and nurses. Read the paper with interest and tend to agree with the conclusion reached by the authors.
Given that the flow of patients to hospital are from the community the way we deal with an acute frailty crisis in the community needs to be looked at. The atypical presentation of acute illness in the frail older person coupled with the move to virtual consultations ( due to work pressure on GPs) has led to a delay in the diagnosis and treatment of acute illness in this group of patients. We often come across the scenario of patients being prescribed multiple courses of antibiotics when the underlying diagnosis is not an infection. The consequence is that the patients become deconditioned even before they enter an Acute frailty unit (AFU) making them less responsive to all the interventions prescribed. I am sure a proportion of patients on an AFU do not need to be there if their acute illness was dealt with promptly in the community.
A H@H service is well placed to deal with acute frailty crisis in the community but needs to be able to respond in a timely way to the high risk frail older population which are care home residents, the housebound older patient and the frail older person on the ambulance stack waiting a paramedic response. If this service is well resourced it will enable an AFU to...
I work in a hospital at home (H@H) service and have found the AFN website a very useful learning resource and regularly recommend it to my Clinical practitioners and nurses. Read the paper with interest and tend to agree with the conclusion reached by the authors.
Given that the flow of patients to hospital are from the community the way we deal with an acute frailty crisis in the community needs to be looked at. The atypical presentation of acute illness in the frail older person coupled with the move to virtual consultations ( due to work pressure on GPs) has led to a delay in the diagnosis and treatment of acute illness in this group of patients. We often come across the scenario of patients being prescribed multiple courses of antibiotics when the underlying diagnosis is not an infection. The consequence is that the patients become deconditioned even before they enter an Acute frailty unit (AFU) making them less responsive to all the interventions prescribed. I am sure a proportion of patients on an AFU do not need to be there if their acute illness was dealt with promptly in the community.
A H@H service is well placed to deal with acute frailty crisis in the community but needs to be able to respond in a timely way to the high risk frail older population which are care home residents, the housebound older patient and the frail older person on the ambulance stack waiting a paramedic response. If this service is well resourced it will enable an AFU to see the right patients at the right time. Both these services cannot work in isolation of each other.
Having perused the article titled "Diagnostic error among vulnerable populations presenting to the emergency department with cardiovascular and cerebrovascular or neurological symptoms: a systematic review," I found it captivating and of great significance. The notion behind this study is quite innovative, as it tackles the concerns of policymakers who worry about the potential to erroneously misdiagnose emergency patients, who indeed are in the most need of care. I firmly believe that this article will provide invaluable insights into a topic that greatly interests a wide audience.
Given my keen interest in this study, and to enhance its quality and the reliability of the final findings, I would like to offer a few suggestions.
I find that the authors have stated that they dropped the chance for a quantitative meta-analysis as they found substantial heterogeneity. I agree with them on decreased reliability of a pooled estimate with high heterogeneity. However, I believe that a quantitative estimate, even accompanied by considerable heterogeneity, is still much more convenient for readers to infer and relate. In fact, having a high heterogeneity is a good chance for authors to investigate the factors and covariates, providing a more precise insight into the complex relationships, and substantially improving the quality of the study. Therefore, I suggest an appendix that provides such data. Providing the limitation...
Having perused the article titled "Diagnostic error among vulnerable populations presenting to the emergency department with cardiovascular and cerebrovascular or neurological symptoms: a systematic review," I found it captivating and of great significance. The notion behind this study is quite innovative, as it tackles the concerns of policymakers who worry about the potential to erroneously misdiagnose emergency patients, who indeed are in the most need of care. I firmly believe that this article will provide invaluable insights into a topic that greatly interests a wide audience.
Given my keen interest in this study, and to enhance its quality and the reliability of the final findings, I would like to offer a few suggestions.
I find that the authors have stated that they dropped the chance for a quantitative meta-analysis as they found substantial heterogeneity. I agree with them on decreased reliability of a pooled estimate with high heterogeneity. However, I believe that a quantitative estimate, even accompanied by considerable heterogeneity, is still much more convenient for readers to infer and relate. In fact, having a high heterogeneity is a good chance for authors to investigate the factors and covariates, providing a more precise insight into the complex relationships, and substantially improving the quality of the study. Therefore, I suggest an appendix that provides such data. Providing the limitations and defects in the quantitative synthesis of the evidence, together with a narrative review of the findings, will let readers draw conclusions at their discretion.
In the editorial authored by G.D. Schiff, B. L. Lambert, and A. Wright, the concept of "indication-based prescribing" is explored. This involves clearly documenting the reason, or indication, for prescribing a medication and linking it to the prescription itself. Despite recommendations and evidence supporting its potential to enhance medication safety and patient comprehension, this essential piece of information is frequently absent from current practices.
The authors advocate for a drastic reimagining of the prescription process. Rather than treating the indication as a supplementary detail, it should serve as the inception point. Under this proposed model, the prescriber would initially enter the medical condition to be addressed. The electronic prescribing system would then recommend the most appropriate and evidence-based medication for the patient. This suggestion, although raising issues about autonomy and trust, is posited to elevate prescription safety, patient education, medication reconciliation, deprescribing, and efficiency in prior authorization processes.
A further innovative proposal places the prescriber in full control. We have developed clinical decision support software enabling the prescriber to begin with either a) an indication, b) a medication, or c) the administration route. The software then filters the remaining pertinent options. As the combination of these three elements, along with the patient's specific context,...
In the editorial authored by G.D. Schiff, B. L. Lambert, and A. Wright, the concept of "indication-based prescribing" is explored. This involves clearly documenting the reason, or indication, for prescribing a medication and linking it to the prescription itself. Despite recommendations and evidence supporting its potential to enhance medication safety and patient comprehension, this essential piece of information is frequently absent from current practices.
The authors advocate for a drastic reimagining of the prescription process. Rather than treating the indication as a supplementary detail, it should serve as the inception point. Under this proposed model, the prescriber would initially enter the medical condition to be addressed. The electronic prescribing system would then recommend the most appropriate and evidence-based medication for the patient. This suggestion, although raising issues about autonomy and trust, is posited to elevate prescription safety, patient education, medication reconciliation, deprescribing, and efficiency in prior authorization processes.
A further innovative proposal places the prescriber in full control. We have developed clinical decision support software enabling the prescriber to begin with either a) an indication, b) a medication, or c) the administration route. The software then filters the remaining pertinent options. As the combination of these three elements, along with the patient's specific context, can precisely shape the resultant medication order, this system ensures a safe and efficient computerized order entry.
Furthermore, through an innovative computational approach, the order goes beyond simply determining the prescription; it also calculates the preparation and administration of the medication. This allows the same order to be viewed from the perspectives of various stakeholders: the prescriber, pharmacist, nursing staff, and the patient.
The true test of the system lies in its practical application. We've made this system available online for trials. Currently, it incorporates a Dutch dataset of products, dosing rules, and solution rules. However, its principles are universally applicable.
We are writing in response to Abraham et al.’s recent review of systematic reviews (SR) targeting the impact of computerised provider order entry (CPOE) on clinical and safety outcomes [1]. We commend the authors’ inclusion of medication errors and adverse drug events (ADE) among the outcomes assessed. This is particularly timely given the World Health Organisation’s 2017 announcement of the third Global Patient Safety Challenge to motivate actions to reduce medication errors causing actual patient harm by half in five years [2]. Abraham et al. concluded that, based on the evidence reported by three SR of inpatient populations, pooled studies showed significant reduction in ADEs with CPOE use, with considerable variation in the magnitude of relative risk reduction [1]. However, there are significant limitations to the studies on which this conclusion is based, and we believe a more cautious approach should be taken when assessing the current evidence.
Firstly, as the authors acknowledged, there was variation in the definitions of ADE across the three SR and the 18 studies they included. We agree that these are significant limitations when trying to summarise the impact of CPOE on ADE. To be clear, the included studies assessed preventable ADEs (10 studies) and/or potential ADEs (15 studies), and three studies did not specify the type of ADE. An ADE can be preventable, non-preventable, or potential [3]. A preventable ADE refers to a medication error which reached the...
We are writing in response to Abraham et al.’s recent review of systematic reviews (SR) targeting the impact of computerised provider order entry (CPOE) on clinical and safety outcomes [1]. We commend the authors’ inclusion of medication errors and adverse drug events (ADE) among the outcomes assessed. This is particularly timely given the World Health Organisation’s 2017 announcement of the third Global Patient Safety Challenge to motivate actions to reduce medication errors causing actual patient harm by half in five years [2]. Abraham et al. concluded that, based on the evidence reported by three SR of inpatient populations, pooled studies showed significant reduction in ADEs with CPOE use, with considerable variation in the magnitude of relative risk reduction [1]. However, there are significant limitations to the studies on which this conclusion is based, and we believe a more cautious approach should be taken when assessing the current evidence.
Firstly, as the authors acknowledged, there was variation in the definitions of ADE across the three SR and the 18 studies they included. We agree that these are significant limitations when trying to summarise the impact of CPOE on ADE. To be clear, the included studies assessed preventable ADEs (10 studies) and/or potential ADEs (15 studies), and three studies did not specify the type of ADE. An ADE can be preventable, non-preventable, or potential [3]. A preventable ADE refers to a medication error which reached the patient and resulted in actual patient harm. A non-preventable ADE, or adverse drug reaction, is defined as harm resulting from an unexpected reaction to a drug where there was no medication error. A potential ADE, or near-miss, is typically defined as a medication error with the potential for harm but in which no harm occurred due to patient circumstance or an intervention [3]. In the wider medication safety literature, many studies which assess ADEs do not delineate whether this harm was, the much more frequently occurring, potential harm or the relatively rare actual harm [4].
Further limitations of the studies of ADEs and CPOEs not discussed were the data collection methods. There is significant variation in the methods undertaken by the included studies to identify ADEs. Firstly, there were eight studies using voluntary incident reports as the only source for identifying ADEs. Incident reports are known to result in significant under-estimation of errors and ADEs [5]. Eight studies used the more rigorous chart and record review. Finally, the study designs do not allow for drawing causal conclusions. That is, there were no randomised controlled trials. Except for one cohort study, all the studies used before-after designs, only one of which included a control.
In conclusion, only half the studies (n=10) included assessed the impact of CPOE on actual patient harm resulting from a medication error. Of these studies, only three reported a significant reduction in patient harm, with the most recent data collection completed in 2010. Five studies reported no significant effect and two studies reported an increase. Of the six studies identifying preventable ADE by chart review, three studies reported a significant reduction (a pooled total of 157 events compared to 139 events) and three studies reported no significant effect.
Considering the very significant limitations across studies, we believe there is not yet convincing evidence of the direct impact of CPOE on actual patient harm. This conclusion remains consistent with two earlier umbrella reviews [6, 7].
References:
1. Abraham, J., et al., Effects of CPOE-based medication ordering on outcomes: an overview of systematic reviews. BMJ Quality & Safety, 2020: p. bmjqs-2019-010436.
2. Donaldson, L.J., et al., Medication without harm: WHO's Third Global Patient Safety Challenge. The Lancet, 2017. 389(10080): p. 1680-1681.
3. Bates, D.W., et al., Relationship between medication errors and adverse drug events. J Gen Intern Med, 1995. 10(4): p. 199-205.
4. Gates, P.J., et al., Standardising the Classification of Harm Associated with Medication Errors: The Harm Associated with Medication Error Classification (HAMEC). Drug Safety, 2019. 42(8): p. 931-939.
5. Westbrook, J.I., et al., What are incident reports telling us? A comparative study at two Australian hospitals of medication errors identified at audit, detected by staff and reported to an incident system. Int J Qual Health Care, 2015. 27(1): p. 1-9.
6. Ranji, S.R., S. Rennke, and R.M. Wachter, Computerised provider order entry combined with clinical decision support systems to improve medication safety: a narrative review. BMJ Qual Saf, 2014. 23(9): p. 773-80.
7. Weir, C.R., N. Staggers, and T. Laukert, Reviewing the impact of computerized provider order entry on clinical outcomes: The quality of systematic reviews. International Journal of Medical Informatics, 2012. 81(4): p. 219-231.
We feel that this article and accompanying press release have failed to fully acknowledge some significant limitations of the study. We feel these limitations are important when making the conclusion that following guidelines by earlier referral would be associated with earlier cancer diagnosis.
1. There is no recognition that the cancer diagnosed in the year following index consultation may not have any association with the index consultation. For example a non-urgent referral for breast lump who developed bladder carcinoma in the following year would be included as someone who could have benefitted from earlier referral.
2. There is no attempt to acknowledge screening cancer diagnoses. Again these would be included despite them being unrelated to any previous “red flag” symptoms.
3. Most significantly, there is no acknowledgement that not “following guidelines” is often an important part of shared decision-making that prevents morbidity related to diagnostic processes and treatment. Although the article explains that co-morbidities and age greater than 85 are associated with lower referral rates; it fails to recognise that any delay in cancer diagnosis in this group would often not be considered a “missed opportunity”,. There is sometimes no clinical benefit to the patient of earlier diagnosis. In relation to this It also fails to recognise that many local 2ww guidelines include severe frailty as an exclusion criteria for an urgent or 2w...
We feel that this article and accompanying press release have failed to fully acknowledge some significant limitations of the study. We feel these limitations are important when making the conclusion that following guidelines by earlier referral would be associated with earlier cancer diagnosis.
1. There is no recognition that the cancer diagnosed in the year following index consultation may not have any association with the index consultation. For example a non-urgent referral for breast lump who developed bladder carcinoma in the following year would be included as someone who could have benefitted from earlier referral.
2. There is no attempt to acknowledge screening cancer diagnoses. Again these would be included despite them being unrelated to any previous “red flag” symptoms.
3. Most significantly, there is no acknowledgement that not “following guidelines” is often an important part of shared decision-making that prevents morbidity related to diagnostic processes and treatment. Although the article explains that co-morbidities and age greater than 85 are associated with lower referral rates; it fails to recognise that any delay in cancer diagnosis in this group would often not be considered a “missed opportunity”,. There is sometimes no clinical benefit to the patient of earlier diagnosis. In relation to this It also fails to recognise that many local 2ww guidelines include severe frailty as an exclusion criteria for an urgent or 2ww referral.
The authors of this paper describe GP decision-making as ‘good but not good enough’. Their paper also highlights considerable variation among clinicians with the highest referrers being more than two times more likely to refer patients with red flag symptoms for cancer investigation than the lowest referrers. When we consider this wide disparity among individual doctors, we must look at how we can objectively and consistently reduce this. How can we make the diagnostic process in primary care ‘reliable, not heroic’ [1]?
Previous papers have considered interventions that can reduce the global burden of diagnostic errors in primary care [2], including the use of information technology tools. Perhaps we can consider that better use of technology such as Clinical Decision Support Systems (CDSS) could improve clinical diagnostics especially in those areas with clear guideline-based practice such as in the decision to refer urgently for suspected cancer [3]. These tools can be used to provide appropriate suggestions for differentials at any point in the clinical consultation or offer ‘alerts’ at the end of the consultation if there are important diagnoses that haven’t been considered [4].
Uptake of CDSS has been poor [5] and clinician response to these tools when they have been implemented are mixed [6] despite them showing that they increase physician’s diagnostic accuracy [7, 8]. At present, clinical work mostly only uses inconsistently implemented ‘alerts’. Some...
The authors of this paper describe GP decision-making as ‘good but not good enough’. Their paper also highlights considerable variation among clinicians with the highest referrers being more than two times more likely to refer patients with red flag symptoms for cancer investigation than the lowest referrers. When we consider this wide disparity among individual doctors, we must look at how we can objectively and consistently reduce this. How can we make the diagnostic process in primary care ‘reliable, not heroic’ [1]?
Previous papers have considered interventions that can reduce the global burden of diagnostic errors in primary care [2], including the use of information technology tools. Perhaps we can consider that better use of technology such as Clinical Decision Support Systems (CDSS) could improve clinical diagnostics especially in those areas with clear guideline-based practice such as in the decision to refer urgently for suspected cancer [3]. These tools can be used to provide appropriate suggestions for differentials at any point in the clinical consultation or offer ‘alerts’ at the end of the consultation if there are important diagnoses that haven’t been considered [4].
Uptake of CDSS has been poor [5] and clinician response to these tools when they have been implemented are mixed [6] despite them showing that they increase physician’s diagnostic accuracy [7, 8]. At present, clinical work mostly only uses inconsistently implemented ‘alerts’. Sometimes, these alerts can reduce inappropriate prescribing and improve patient health [9] although poorly implemented alerts can add to the doctor’s burden [10]. Only small studies have previously looked at the acceptability of these alerts [11] which highlights an area rich for further action and research to improve patient outcomes.
Investment in better and more effective technologies in the primary care consulting room has the potential to reduce the disparity between different doctors’ referral rates and provide a more consistent adherence to the cancer referral guidelines potentially increasing our detection rate.
References
1 Schiff GD, Bates DW. Can Electronic Clinical Documentation Help Prevent Diagnostic Errors? New England Journal of Medicine 2010;362:1066–9. doi:10.1056/NEJMp0911734
2 Singh H, Schiff GD, Graber ML, et al. The global burden of diagnostic errors in primary care. BMJ Quality & Safety 2017;26:484–94. doi:10.1136/bmjqs-2016-005401
3 Graber ML. Reaching 95%: Decision support tools are the surest way to improve diagnosis now. BMJ Quality & Safety Published Online First: October 2021. doi:10.1136/bmjqs-2021-014033
4 Nurek M, Kostopoulou O, Delaney BC, et al. Reducing diagnostic errors in primary care. A systematic meta-review of computerized diagnostic decision support systems by the LINNEAUS collaboration on patient safety in primary care. European Journal of General Practice 2015;21:8–13. doi:10.3109/13814788.2015.1043123
5 Cheraghi-Sohi S, Alam R, Hann M, et al. Assessing the utility of a differential diagnostic generator in UK general practice: A feasibility study. Diagnosis 2021;8:91–9. doi:10.1515/dx-2019-0033
6 Ford E, Edelman N, Somers L, et al. Barriers and facilitators to the adoption of electronic clinical decision support systems: A qualitative interview study with UK general practitioners. BMC Medical Informatics and Decision Making 2021;21. doi:10.1186/s12911-021-01557-z
7 Kostopoulou O, Rosen A, Round T, et al. Early diagnostic suggestions improve accuracy of GPs: A randomised controlled trial using computer-simulated patients. British Journal of General Practice 2015;65:e49–54. doi:10.3399/bjgp15X683161
8 Riches N, Panagioti M, Alam R, et al. The Effectiveness of Electronic Differential Diagnoses (DDX) Generators: A Systematic Review and Meta-Analysis. PLOS ONE 2016;11:e0148991. doi:10.1371/journal.pone.0148991
9 McKibben S, Bush A, Thomas M, et al. An electronic alert to reduce excessive prescribing of short-acting Beta2-agonists for people with asthma in East London: A retrospective case-control study using routine primary care data. In: M-Health/e-health. European Respiratory Society 2020. 1364. doi:10.1183/13993003.congress-2020.1364
10 Cecil E, Dewa LH, Ma R, et al. General practitioner and nurse practitioner attitudes towards electronic reminders in primary care: A qualitative analysis. BMJ Open 2021;11:e045050. doi:10.1136/bmjopen-2020-045050
11 Kelsey EA, Njeru JW, Chaudhry R, et al. Understanding User Acceptance of Clinical Decision Support Systems to Promote Increased Cancer Screening Rates in a Primary Care Practice. Journal of Primary Care & Community Health 2020;11:215013272095883. doi:10.1177/2150132720958832
Conflict of Interest
None declared
2.
Title: “Inadequate limitations”
Author: William P, Hart, Katarzyna, Grellier
In response to the published article Concordance with urgent referral guidelines in patients presenting with any of six ‘alarm’ features of possible cancer: a retrospective cohort study using linked primary care records"
Rapid response:
Other Contributors:
October 13, 2021
We feel that this article and accompanying press release have failed to fully acknowledge some significant limitations of the study. We feel these limitations are important when making the conclusion that by following guidelines, earlier referral would be associated with earlier cancer diagnosis.
1. There is no recognition that the cancer diagnosed in the year following index consultation may not have any association with the index consultation. For example a non-urgent referral for breast lump who developed bladder carcinoma in the following year would be included as someone who could have benefitted from earlier referral.
2. There is no attempt to acknowledge screening cancer diagnoses. Again these would be included despite them being unrelated to any previous “red flag” symptoms.
3. Most significantly, there is no acknowledgement that not “following guidelines” is often an important part of shared decision-making that prevents morbidity related to diagnostic processes and treatment. Although the article explains that co-morbidities and age greater than 85 are associated with lower referral rates; it fails to recognise that any delay in cancer diagnosis in this group would often not be considered a “missed opportunity”,. There is sometimes no clinical benefit to the patient of earlier diagnosis. In relation to this It also fails to recognise that many local two-week wait guidelines include severe frailty as an exclusion criteria for an urgent or two-week wait referral.
Conflict of Interest
None declared
The recent article by Muskens et al "Overuse of diagnostic testing in healthcare: a systematic review" published online on May 10, 2021, admirably describes the prevalence and overuse of low-value diagnostic tests in multiple healthcare settings, but does not offer possible explanations or solutions. Some salient reasons are (1): 1. Physicians do not consider healthcare costs during a diagnostic workup; 2. Malpractice concerns; 3. An excessive fee-for-service mentality--more tests and consultations generate more income; 4. Lack of trust in clinical skills versus high technology testing; 5. Believing that diagnostic workups should eliminate all conceivable etiologies; and 6. Patients may experience greater satisfaction with more testing.
Underlying these last three possibilities is the failure to appreciate that in science, especially after Darwin, Einstein and Heisenberg, absolute certainty is considered unachievable. Facts and theories, and clinical diagnoses, are now rated on the probability that they are true. A diagnosis is more probable than not, or more probable than another, based on degrees of evidence. Hence recent interest in Bayesianism, which analyses evidence in terms of probability. When determining the efficacy of additional tests, Bayesianism requires a pre-test (a priori) probability estimate of the diagnosis. The more probable a diagnosis is clinically (a priori) and the greater the diagnostician's clinical confidence, the less nec...
The recent article by Muskens et al "Overuse of diagnostic testing in healthcare: a systematic review" published online on May 10, 2021, admirably describes the prevalence and overuse of low-value diagnostic tests in multiple healthcare settings, but does not offer possible explanations or solutions. Some salient reasons are (1): 1. Physicians do not consider healthcare costs during a diagnostic workup; 2. Malpractice concerns; 3. An excessive fee-for-service mentality--more tests and consultations generate more income; 4. Lack of trust in clinical skills versus high technology testing; 5. Believing that diagnostic workups should eliminate all conceivable etiologies; and 6. Patients may experience greater satisfaction with more testing.
Underlying these last three possibilities is the failure to appreciate that in science, especially after Darwin, Einstein and Heisenberg, absolute certainty is considered unachievable. Facts and theories, and clinical diagnoses, are now rated on the probability that they are true. A diagnosis is more probable than not, or more probable than another, based on degrees of evidence. Hence recent interest in Bayesianism, which analyses evidence in terms of probability. When determining the efficacy of additional tests, Bayesianism requires a pre-test (a priori) probability estimate of the diagnosis. The more probable a diagnosis is clinically (a priori) and the greater the diagnostician's clinical confidence, the less necessary is additional testing. Since it is impossible to prove absence of disease, testing can only support or not support a specific clinical diagnosis.
But there is a quid pro quo: Tests can be harmful and costs uncontrolled. Given the subjective nature of characterizing truth and justifying beliefs, the medical profession might consider differentiating levels of evidence or burdens of proof regarding diagnostic accuracy much as the legal profession does: in less stringent civil cases the burden of proof is preponderant, more likely than not, >50%. In more severe criminal cases the burden of proof is beyond reasonable doubt, indubitable. While this legal dichotomization is probably too simplistic for all clinical issues, it can serve as a guide. Graded burdens of proof for differing clinical presentations could alter physician behavior, enhance patient satisfaction, reduce malpractice concerns, and contribute to lower health care costs without compromising individual patient care.
Parenthetically, after decades of teaching medical students and house staff officers of the importance of clinical skills to minimize excessive testing, I joined a multispecialty healthcare group from which I was eventually dismissed for not ordering enough tests and generating sufficient testing income. My philosophy, which I strongly believe in, came back to bite me.
1. Horowitz SH. Diagnostic decision-making, burdens of proof, and a $6000 per hour memory lapse. JAMA Int Med; 2013;176:1493-1494.
We thank Subbe et al for engaging with our paper ‘Understanding the enablers and barriers to implementing a patient-led escalation system: a qualitative study’ in their editorial. The editorial highlights the importance of having a system-led response to patient-led escalation. The editorial also draws attention to the patient and family level trauma and distress that can result from having concerns about a deteriorating condition dismissed by staff, as evidenced by the personal experience provided by one of the co-authors, Alison Phillips. Indeed, in one of our earlier publications from the same study, we similarly highlight the potential for epistemic injustice to arise when patients’ concerns are ignored or dismissed, particularly when in a condition of extreme vulnerability.(1) Our paper and the linked editorial by Subbe et al share the aim of supporting patients in being heard and receiving timely responses when their condition is deteriorating.
Despite the increased interest, policy activity and enthusiasm for patient-led escalation, there are still significant gaps in the evidence base around implementation of such systems, particularly with regard to how they work in relation to wider socio-cultural rescue systems and across different clinical settings.(2) The premise of our paper was not to undermine the importance of patient-led escalation systems but rather to explore and take learning from the factors that shaped implementation of one particular pilot sy...
Show MoreDear Editor,
We were interested to read the recent article on patient safety in remote primary care encounters by Payne et al. We have been reviewing the use of remote consultations specifically for paediatric asthma patients and would like to thank the authors for their work.
Firstly, we agree with the authors’ findings that a remote environment may exacerbate existing inequalities such as economic and language barriers. We would add that an additional factor that must be assessed is a patient’s ability to use technology. Pinnock et al. highlight the risk of virtual consultations to those who lack “e-literacy (or digital healthcare literacy)” (2). These patients must be identified and offered additional support or alternative methods of consultation to maintain the utmost level of care.
Payne et al. highlight the need for a more definitive approach to escalating care rather than a “rule of thumb” or “if in doubt, put it down as urgent” approach (1). We would echo that there need to be clear guidelines and more specific thresholds for escalating care from remote to in-person visits. One suggestion by Galway et al. is having a lower threshold for seeing younger children face to face (3). Galway et al. also suggest having alternative “red-flag” signs that are unique to the remote setting. For example, multiple calls from a patient may indicate the need to escalate their care to face-to-face. More of these red-flag signs unique to this setting need to be...
Show MoreSmith et al (2022) present an excellent review of the current state of safety netting theory and practice in Primary Care. In relation to the education of paramedics working in primary care, the General, Logistical, Red Flag and Risk approach has been suggested as a means of implementing the theory of safety netting into a clinician’s practice (Mallinson, 2023). It reminds clinicians to always give a General worsening statement, meeting Smith et al’s Recommendation No. 1. The second point is to provide clear Logistical advice to patients on how to seek help; the specifics of what phone number to call to seek or summon help. This perhaps aligns to Recommendation 9; “including a specific safety-netting plan”. Red Flag safety netting reinforces the importance of patient education in relation to possible serious deterioration which aligns to Recommendation 4 in terms of “specific situations that should be cause for concern”. The final component of Risk based safety netting relates to shared decision making and specifically situations where there is disagreement and a patient is not following medical advice. The Risk component seeks to ensure that patients are fully aware of the potential risks inherent in their chosen course to action. This is vital in relation to ensuring patients are making informed decisions about their care. There is definitely more work to be done on improving education around Safety Netting, and Smith et al’s work provides us with clear recommendati...
Show MoreI work in a hospital at home (H@H) service and have found the AFN website a very useful learning resource and regularly recommend it to my Clinical practitioners and nurses. Read the paper with interest and tend to agree with the conclusion reached by the authors.
Show MoreGiven that the flow of patients to hospital are from the community the way we deal with an acute frailty crisis in the community needs to be looked at. The atypical presentation of acute illness in the frail older person coupled with the move to virtual consultations ( due to work pressure on GPs) has led to a delay in the diagnosis and treatment of acute illness in this group of patients. We often come across the scenario of patients being prescribed multiple courses of antibiotics when the underlying diagnosis is not an infection. The consequence is that the patients become deconditioned even before they enter an Acute frailty unit (AFU) making them less responsive to all the interventions prescribed. I am sure a proportion of patients on an AFU do not need to be there if their acute illness was dealt with promptly in the community.
A H@H service is well placed to deal with acute frailty crisis in the community but needs to be able to respond in a timely way to the high risk frail older population which are care home residents, the housebound older patient and the frail older person on the ambulance stack waiting a paramedic response. If this service is well resourced it will enable an AFU to...
Dear BMJ Quality and Safety,
Having perused the article titled "Diagnostic error among vulnerable populations presenting to the emergency department with cardiovascular and cerebrovascular or neurological symptoms: a systematic review," I found it captivating and of great significance. The notion behind this study is quite innovative, as it tackles the concerns of policymakers who worry about the potential to erroneously misdiagnose emergency patients, who indeed are in the most need of care. I firmly believe that this article will provide invaluable insights into a topic that greatly interests a wide audience.
Given my keen interest in this study, and to enhance its quality and the reliability of the final findings, I would like to offer a few suggestions.
I find that the authors have stated that they dropped the chance for a quantitative meta-analysis as they found substantial heterogeneity. I agree with them on decreased reliability of a pooled estimate with high heterogeneity. However, I believe that a quantitative estimate, even accompanied by considerable heterogeneity, is still much more convenient for readers to infer and relate. In fact, having a high heterogeneity is a good chance for authors to investigate the factors and covariates, providing a more precise insight into the complex relationships, and substantially improving the quality of the study. Therefore, I suggest an appendix that provides such data. Providing the limitation...
Show MoreIn the editorial authored by G.D. Schiff, B. L. Lambert, and A. Wright, the concept of "indication-based prescribing" is explored. This involves clearly documenting the reason, or indication, for prescribing a medication and linking it to the prescription itself. Despite recommendations and evidence supporting its potential to enhance medication safety and patient comprehension, this essential piece of information is frequently absent from current practices.
The authors advocate for a drastic reimagining of the prescription process. Rather than treating the indication as a supplementary detail, it should serve as the inception point. Under this proposed model, the prescriber would initially enter the medical condition to be addressed. The electronic prescribing system would then recommend the most appropriate and evidence-based medication for the patient. This suggestion, although raising issues about autonomy and trust, is posited to elevate prescription safety, patient education, medication reconciliation, deprescribing, and efficiency in prior authorization processes.
A further innovative proposal places the prescriber in full control. We have developed clinical decision support software enabling the prescriber to begin with either a) an indication, b) a medication, or c) the administration route. The software then filters the remaining pertinent options. As the combination of these three elements, along with the patient's specific context,...
Show MoreWe are writing in response to Abraham et al.’s recent review of systematic reviews (SR) targeting the impact of computerised provider order entry (CPOE) on clinical and safety outcomes [1]. We commend the authors’ inclusion of medication errors and adverse drug events (ADE) among the outcomes assessed. This is particularly timely given the World Health Organisation’s 2017 announcement of the third Global Patient Safety Challenge to motivate actions to reduce medication errors causing actual patient harm by half in five years [2]. Abraham et al. concluded that, based on the evidence reported by three SR of inpatient populations, pooled studies showed significant reduction in ADEs with CPOE use, with considerable variation in the magnitude of relative risk reduction [1]. However, there are significant limitations to the studies on which this conclusion is based, and we believe a more cautious approach should be taken when assessing the current evidence.
Firstly, as the authors acknowledged, there was variation in the definitions of ADE across the three SR and the 18 studies they included. We agree that these are significant limitations when trying to summarise the impact of CPOE on ADE. To be clear, the included studies assessed preventable ADEs (10 studies) and/or potential ADEs (15 studies), and three studies did not specify the type of ADE. An ADE can be preventable, non-preventable, or potential [3]. A preventable ADE refers to a medication error which reached the...
Show MoreWe feel that this article and accompanying press release have failed to fully acknowledge some significant limitations of the study. We feel these limitations are important when making the conclusion that following guidelines by earlier referral would be associated with earlier cancer diagnosis.
Show More1. There is no recognition that the cancer diagnosed in the year following index consultation may not have any association with the index consultation. For example a non-urgent referral for breast lump who developed bladder carcinoma in the following year would be included as someone who could have benefitted from earlier referral.
2. There is no attempt to acknowledge screening cancer diagnoses. Again these would be included despite them being unrelated to any previous “red flag” symptoms.
3. Most significantly, there is no acknowledgement that not “following guidelines” is often an important part of shared decision-making that prevents morbidity related to diagnostic processes and treatment. Although the article explains that co-morbidities and age greater than 85 are associated with lower referral rates; it fails to recognise that any delay in cancer diagnosis in this group would often not be considered a “missed opportunity”,. There is sometimes no clinical benefit to the patient of earlier diagnosis. In relation to this It also fails to recognise that many local 2ww guidelines include severe frailty as an exclusion criteria for an urgent or 2w...
The authors of this paper describe GP decision-making as ‘good but not good enough’. Their paper also highlights considerable variation among clinicians with the highest referrers being more than two times more likely to refer patients with red flag symptoms for cancer investigation than the lowest referrers. When we consider this wide disparity among individual doctors, we must look at how we can objectively and consistently reduce this. How can we make the diagnostic process in primary care ‘reliable, not heroic’ [1]?
Previous papers have considered interventions that can reduce the global burden of diagnostic errors in primary care [2], including the use of information technology tools. Perhaps we can consider that better use of technology such as Clinical Decision Support Systems (CDSS) could improve clinical diagnostics especially in those areas with clear guideline-based practice such as in the decision to refer urgently for suspected cancer [3]. These tools can be used to provide appropriate suggestions for differentials at any point in the clinical consultation or offer ‘alerts’ at the end of the consultation if there are important diagnoses that haven’t been considered [4].
Uptake of CDSS has been poor [5] and clinician response to these tools when they have been implemented are mixed [6] despite them showing that they increase physician’s diagnostic accuracy [7, 8]. At present, clinical work mostly only uses inconsistently implemented ‘alerts’. Some...
Show MoreThe recent article by Muskens et al "Overuse of diagnostic testing in healthcare: a systematic review" published online on May 10, 2021, admirably describes the prevalence and overuse of low-value diagnostic tests in multiple healthcare settings, but does not offer possible explanations or solutions. Some salient reasons are (1): 1. Physicians do not consider healthcare costs during a diagnostic workup; 2. Malpractice concerns; 3. An excessive fee-for-service mentality--more tests and consultations generate more income; 4. Lack of trust in clinical skills versus high technology testing; 5. Believing that diagnostic workups should eliminate all conceivable etiologies; and 6. Patients may experience greater satisfaction with more testing.
Underlying these last three possibilities is the failure to appreciate that in science, especially after Darwin, Einstein and Heisenberg, absolute certainty is considered unachievable. Facts and theories, and clinical diagnoses, are now rated on the probability that they are true. A diagnosis is more probable than not, or more probable than another, based on degrees of evidence. Hence recent interest in Bayesianism, which analyses evidence in terms of probability. When determining the efficacy of additional tests, Bayesianism requires a pre-test (a priori) probability estimate of the diagnosis. The more probable a diagnosis is clinically (a priori) and the greater the diagnostician's clinical confidence, the less nec...
Show MorePages