Written evidence submitted by Ms Molly McCarthy, Dr Jason McIntyre, Professor Rajan Nathan, Dr Emma Ashworth and Dr Pooja Saini
(Liverpool John Moores University)
This report is submitted by members of the Suicide and Self-Harm Research Group (SSHRG) at Liverpool John Moores University: Molly McCarthy, Dr Pooja Saini, Dr Jason McIntyre, Dr Emma Ashworth and Professor Rajan Nathan (Cheshire & Wirral Partnership NHS Foundation Trust). The group has expertise in suicide prevention and intervention design and implementation as well as conducting research within emergency care settings and utilising large hospital datasets. The group has conducted research into emergency department (ED) access and coding practices across six National Health Service (NHS) sites for suicidal crisis and self-harm. Key findings from our research include:
Accurate identification of suicidal crisis and self-harm presentations to emergency departments (EDs) can lead to more timely mental health support, improve clinical pathways, patient experience, and support impact evaluation of suicide prevention initiatives. We conducted a mixed-method study with the aim of identifying (1) the most common codes used for suicidal crisis presentations and the extent of missing data, (2) why coding practices for suicidal crisis are inaccurate and inconsistent across EDs and (3) the factors that contribute to patients receiving an incorrect code or no code.
We collected data across six EDs in Merseyside and Cheshire from 2019 to 2021. The total sample for this dataset across the six participating EDs included 14,984 presentations for suicidal crisis, self-harm or suicide attempt. Further to this, qualitative semi-structured interviews were conducted with 23 administrative, medical and mental health staff to better understand the reasons behind inaccurate and inconsistent coding for suicidal crisis.
Results from our study highlighted the inconsistent and inaccurate coding practices across EDs for suicidal crisis. Across all six EDs examined, 3,358 suicidal crisis attendances were coded as ‘depressive disorder’ (21.8%) and 581 as ‘anxiety disorder’ (3.8%), despite the existence of a ‘suicidal crisis’ code. This means that attendances are consistently being under-estimated from ED databases. This has important implications for patients accessing emergency care, as well as ensuring appropriate mental health follow-up care is provided.
Interviews with ED staff elaborated on some of the reasons behind why the primary diagnosis codes of ‘anxiety disorder’ and ‘depressive disorder’ are consistently being used across sites in relation to suicidal presentations. One Consultant in Emergency Medicine stated:
“We're limited with the coding that we can have. So, I think there's depressive disorder, there's anxiety disorder, psychotic disorder, and then a couple of others. Now, what you can do is, code formally the depressive disorder, but then in the actual discharge letter, you can then put an anxiety disorder with suicidal ideation.” (02: Consultant in Emergency Medicine).
This quote illustrates the way codes can conceal the complexity of the real nature of the presentation and can misreprest suicidal presentations to EDs.
The primary diagnosis code of ‘no abnormality detected’ was used across all EDs, with 23.6% of all suicidal crisis attendances receiving the code. In interviews, 12/23 staff reported that they use the code ‘no abnormality detected’ in relation to suicidal presentations. Staff elaborated on the reasons behind this and often discussed the absence of an appropriate or relevant suicidal crisis code, leaving staff with no other option.
Across all six EDs, the primary diagnosis code was left blank on 18.4% of occasions. During the interviews, staff often noted the contextual factors that impede accurate coding of suicidal presentations, resulting in a high level of missing data. In the fast-paced environment, ED staff faced additional pressures and challenges which impacted negatively on their ability to accurately code and make good decisions (McCarthy et al. 2023). Limited time, competing demands and inadequate training in suicide documentation often resulted in staff leaving the code blank. This means suicidal presentations are not being recognised on ED databases, which has implications for patients receiving appropriate care.
Implications for Access to Urgent and Emergency Care:
Poor ED coding practices hinders the ability to accurately identify individuals at risk of suicidal ideation and self-harm (McCarthy et al. 2021). This can have significant implications for patient treatment and support, such as ensuring referrals to mental health services and appropriate follow-up care is received. For example, if self-harm presentations are recorded as ‘laceration’, it is more difficult to identify future mental health support needs.
Furthermore, inaccurate coding within EDs results in significant underestimation of suicidal crisis presentations. Inaccurate or biased data collection can lead to ineffective allocation of funding and resources. Implementing a standardised code for suicidal crisis is urgently needed to ensure accurate and timely data collection. This could be done by piloting the new code within a small number of EDs to ensure the code is being implemented appropriately. It would also be beneficial to develop clear and standardised coding guidelines to support staff responsible for coding presentations and introduce regular and ongoing training for all ED staff. Improved detection and recording of suicidal crises will then support service developments such as the crisis care concordat (Gibson et al. 2016), which aims to provide better access to mental health support in England. Better data could then be used to inform policy to tackle the implementation gap between policy and services and also provide better data links between primary, secondary, tertiary, and community care.
Implementing efficient, clinically relevant and user-friendly electronic health record systems that facilitate accurate and complete documentation of patient encounters is crucial to developing coding practices within EDs. It is therefore essential to recognise how these systems are used and the time pressures ED staff face. Providing focused coding training for administrative, medical and mental health staff that is supported by ongoing clinical supervision would ensure consistency and accuracy, leading to better data quality. EDs would also benefit from a continuous feedback loop that integrates research and practice. Collaborative work with academics, researchers, ED staff, patients and wider stakeholders would help identify issues with coding, address them promptly and implement ongoing training and process enhancements.
Crisis lines are the first mental health service contact point for many people, making them a vital community and public health intervention. Our dataset captured calls to the Cheshire & Wirral Partnership NHS Foundation Trust (CWP) crisis line between August 2020 and August 2021 (Saini et al. 2022). Calls were examined if self-harm, risk to self, or overdose were reported by the caller. This research aimed to examine (1) the referral pathways for people calling the CWP crisis line for self-harm, risk to self, and/or overdose and (2) the socioeconomic factors reported by these callers.
There were 25,106 calls recorded between August 2020 to August 2021. The majority of calls handled by CWP Crisis Line were from women (57.0%), whilst very few callers reported being non-binary (0.1%). The highest proportion of calls were from people aged 50 – 59 years (23.3%), followed by people aged 0 – 18 years (15.9%). The overwhelming majority of calls were from people who indicated they were from white British ethnic backgrounds (74%). Risk to self was noted as the reason for calling for 10.2% of calls (N=2,561). Self-harm made up 7.9% of calls (N=1,983) and overdose was recorded for 435 callers (1.7%).
We found that call handlers were significantly more likely to call 999 or hand over to a practitioner, and less likely to provide advice and guidance, if self-harm, risk to self or overdose was reported. Social issues were found to be significantly associated with all three outcomes: self-harm, risk to self and overdose. The majority of socioeconomic factors explored were significantly associated with risk to self. For example, callers to the CWP crisis line who indicated issues with alcohol, bereavement, COVID, finances, housing, medication, physical health and social factors were significantly more likely to report risk to self.
Implications for Access to Urgent and Emergency Care:
Our findings highlight a number of socioeconomic factors that are significantly associated with callers reporting self-harm, risk to self or overdose. This evidence highlights that issues with alcohol, bereavement, COVID, finances, housing, medication, physical health and social factors may be precipitating factors to self-harm calls, and so these need to be targeted through early identification strategies to improve access and outcomes for people experiencing suicide-related thoughts and behaviours.
Furthermore, our findings provide an understanding of the common referral pathways after contact with the mental health crisis line. We found the majority of callers (63.5%) were given advice and/or guidance and one in eleven callers were handed over to a practitioner. Again, there was also a high level of missing data for type of assistance provided (12.0%). This can be valuable information for service commissioners and funding efforts to ensure accurate and effective targeting of resources.
Since the crisis line service was implemented twelve months ahead of schedule in response to the COVID-19 pandemic, important clinical implications can also be gained in terms of developing the data collected from the service. Improving the data routinely collected by call handlers can benefit both services and research into suicide prevention. Mental health crisis coding has been identified as an issue in NHS settings (McCarthy et al., 2021) and the current study highlights that more specific codes could be implemented to better understand where callers are being signposted following contact with the crisis line. More data captured on the context of advice given by call handlers, and follow-up data on user outcomes, would aid an understanding what support and signposting services are effective for those reporting self-harm. Services may also benefit from additional questions related to suicide and self-harm because suicide prevention is a key focus of the service. For example, there was no data available related to whether the caller was experiencing suicidal thoughts. Since suicidal ideation is a key risk factor for self-harm and suicidal behaviour (Kienhorst, 1995), services and research would benefit from better understanding the symptom characteristics, caller characteristics and referral pathways for those in suicidal crisis.
To summarise, our findings highlight problems with current ED and crisis line coding practices and data collection. Coding practices across EDs for suicidal crisis are ineffective. This means people experiencing suicide-related thoughts and behaviours cannot be tracked effectively and it is not possible to ensure that they are receiving the care they need. This evidence adds substantially to the need to improve access for urgent and emergency care. There is a need to implement a standardised code for suicidal crisis to ensure accurate and timely data collection and improve access to emergency care and develop clear and standardised coding guidelines to support staff responsible for coding presentations is needed, as well as the introduction of regular and ongoing training for all ED staff.
Gibson S., Hamilton, S., & James, K. (2016). Evaluation of the crisis care concordat implementation. McPin Foundation, 2016. https://mcpin.org/wp-content/uploads/CCC-Evaluation_Report.pdf
Kienhorst, I. C. (1995). Crisis intervention and a suicidal crisis in adolescents. Crisis, 16(4), 154-156.
McCarthy, M., McIntyre, J., Nathan, R., & Saini, P. (2023). Factors influencing emergency department staff decision-making for people attending in suicidal crisis: a systematic review. Archives of suicide research, 1-15.
McCarthy, M., Saini, P., Nathan, R., & McIntyre, J. (2021). Improve coding practices for patients in suicidal crisis. BMJ, 375. DOI: https://doi.org/10.1136/bmj.n2480
Saini, P., McCarthy, M., & McIntyre, J. (2022). Crisis Line Evaluation: Cheshire & Wirral Partnership, Mersey Care and Mid-Mersey. June 2022. https://researchonline.ljmu.ac.uk/id/eprint/17106/3/Crisis%20Line%20Evaluation%20Cheshire%20&%20Wirral%20Partnership,%20Mersey%20Care%20and%20Mid-Mersey..pdf