Globally, approximately 800,000 die by suicide every year (World Health Organization, 2014). Each death is a tragedy that affects many people; with an estimated 135 adults exposed to each suicide death (Cerel et al, 2018). Suicides are preventable and occasions when patients present to healthcare services are a key opportunity to identify and mitigate suicide risk. A small number of studies have previously examined contact with primary care and/or mental health services in the year prior to suicide, finding that contact is predominantly within a primary care setting (Stene-Larsen & Reneflot, 2019). However, limitations of earlier work include not having a population-level of controls to compare with people who died by suicide, a small sample size and/or using data collected from interviewing next-of-kin and healthcare professionals.

In this study, Professor Ann John and colleagues sought to address these limitations by synthesising multiple existing databases with national coverage (in Wales). They sought to identify the nature and timing of contact with primary care, secondary care and emergency department services in the year leading up to a death by suicide (referred to by the authors as cases). The researchers also compared contact by people who died by suicide with controls who were alive at the time of the cases’ death.

Methods

The researchers examined deaths that occurred between 1st January 2001 and 31st December 2017. They also included an index date for each person which was either the date of death, or the first date of contact with healthcare services in which their death occurred, e.g. the date when they were admitted to an emergency department following self-harm. Each case was matched to five unique controls who were alive when the case’s death occurred. Matching was done based on gender and week of birth (in the same birth year, or one year prior or after).

Suicide Information Database-Wales (SID-Cymru) is a database holding information on individuals who have died by suicide in Wales, irrespective of if they had been in contact with mental health services. This is hosted in the Secure Anonymised Information Linkage (SAIL) Databank, which contains health and population datasets. Every individual in the databank has a unique, anonymous identifier that allows linkage of each person’s data across multiple databases.

The researchers synthesised information from six different databases for the entire Welsh population for the study period unless otherwise noted. Information extracted from these databases included the date and reason for healthcare service contact prior to the index data.

Results

Information from 5,130 of the 5,237 individuals (98%) who had died by suicide in Wales between 2001 and 2017 were linked across the databases. Most were male (78%) with 11% aged under 25 years at the time of death and 16% over the age of 64.

Type of contact

Within one year before the index date, compared to general population controls, people who died by suicide had more often:

• Seen their GP (84% vs. 70%) for all examined reasons including non-mental health related (54 vs. 45%), mental health related (29% vs. 6%) and self-harm (7% vs. 0.3%)
• Been in contact with the emergency department (41% vs. 16%). For both groups, contact was most often recorded as relating to ‘injury and poisoning’ (17% in cases and 6% in controls)
• Had a hospital admission (34% vs. 13%). Cases were more likely to be admitted for any of the examined reasons compared to controls (29% vs. 6%). This included for mental health (18% vs. 1%), injury and poisoning (14% vs. 1%), accidental hanging and self-poisoning (2% vs. 0.1%) and self-harm (9% vs. 0.2%).

Timing of contact

• Contact with any service was more common in cases compared to controls in 1 week (31% vs. 16%), 1 month (73% vs. 53%) and 1 year (89% vs. 76%) prior to the index date
• For both cases and controls, females more often had any healthcare contact at each of these three time-points than males
• Contact was most often with the GP for cases and controls in the 1 week (26% cases vs. 14% controls), 1 month (68% vs. 50%) and 12 months prior to the index date (85% vs. 72%)
• A larger proportion of cases were in contact with any of the examined healthcare services at any week in the 12 months before the index date
• Similarly, a larger proportion of cases were in contact for mental health reasons at any week than controls.

Conclusions

Most (73%) people in the fully linked dataset who died by suicide had been in contact with healthcare services in the month prior to the index date. Almost a third (31%) of those who died by suicide had been in contact in the week prior, double that of matched controls (16%). The last contact with healthcare services prior to the index date was most often with a general practice (71% of cases). People who died by suicide were, at any week in the 12 months before the index date, more likely to have contact health care services than general population. This was particularly evident the closer to the index date. The authors comment that

there does appear to be an escalating build-up of help-seeking contacts over time leading to an acute crisis.

Strengths and limitations

Strengths

• The researchers were able to link individual level data on contact with different healthcare services in the 12 months prior to the index date. This longitudinal design provided a detailed picture of the different services people had been in contact with, the reason why they approached the service, and the timing of the contact.
• The availability of general population data used as matched living controls to people who died by suicide, allowed comparison of the frequency and nature of contact between these two groups.
• Linkage of data across databases was undertaken for 98% of the 5,237 people who died by suicide in Wales between the study period of 2001-2017. This large-scale national sample therefore addresses limitations of some earlier studies which used a smaller sample.
• Previous research has shown that there are gender differences in seeking help for mental and physical health issues (e.g. Thompson et al., 2016). The large study sample in this John et al study (2020) allowed examination of differences in males and females in seeking help-seeking from healthcare services.

Limitations

• For the fully linked database, the authors only included individuals where the index date was 2010 or later and general practice had submitted data to SAIL in the year prior to the index date. Therefore, information across the full study period of 2001-2017 was not available as the databases were not linked prior to 2010.
• Routinely collected health data were synthesised across multiple databases for this study. Previously highlighted potential bias from such health data include the codes used to extract the study population, potentially confounding variables which are not routinely collected so their potential impact cannot be assessed, and missing data (Benchimol et al, 2015). John et al (2020) sought to mitigate these with the use of codes validated by expert clinicians.
• The reasons for lack of contact were not identified, although this was outside the scope of this study. Therefore, it is not known why some people who may have needed help from services did not access these. Datasets on contact with third sector services and private healthcare services do not form part of the SAIL databank. These would help to provide an even more comprehensive overview of service contact.

Implications for practice

Findings from this study indicate the frequency and nature of patient contact with healthcare services prior to suicide. In doing so, this sheds light on key opportunities for suicide prevention.

Given that general practice was usually the final service people contacted, the authors emphasise the importance of GPs in identifying patient contact with any other health services. However, they also recognise the demands within the primary care setting, such as training, which need addressing to facilitate this.

Statement of interests

The author of this blog, Su-Gwan Tham, works with one of the authors of this paper, Professor Nav Kapur at the National Confidential Inquiry into Suicide and Safety in Mental Health (NCISH). However, she had no involvement in this study by John et al (2020).

Links

Primary paper

John, A., DelPozo-Banos, M., Gunnell, D., Dennis, M., Scourfield, J., Ford, D., Kapur, N., Lloyd, K. (2020). Contacts with primary and secondary healthcare prior to suicide: Case–control whole-population-based study using person-level linked routine data in Wales, UK, 2000–2017. The British Journal of Psychiatry, 217(6), 717-724.

Other references

Benchimol, E. I., Smeeth, L., Guttmann, A., Harron, K., Moher, D., Petersen, I., Sørensen, H. T., von Elm, E., Langan, S. M., & RECORD Working Committee (2015). The REporting of studies Conducted using Observational Routinely-collected health Data (RECORD) statement. PLOS Medicine, 12(10), e1001885.

Cerel, J., Brown, M.M., Maple, M., Singleton, M., van de Venne, J., Moore, M. and Flaherty, C. (2019). How Many People Are Exposed to Suicide? Not Six. Suicide and Life-Threatening Behavior, 49: 529-534.

Stene-Larsen K, Reneflot A. (2019). Contact with primary and mental health care prior to suicide: A systematic review of the literature from 2000 to 2017. Scandinavian Journal of Public Health. 47(1):9-17.

Thompson, A.E., Anisimowicz, Y., Miedema, B., Hogg, W., Wodchis, W.P. and Aubrey-Bassler, K. (2016). The influence of gender and other patient characteristics on health care-seeking behaviour: a QUALICOPC study. BMC Family Practice, 17, 38.

World Health Organization (2014). Preventing suicide: A global imperative

Photo credits

• Photo by Waldemar Brandt on Unsplash
• Photo by Catrin Ellis on Unsplash
• Photo by Paico Oficial on Unsplash
• Photo by Jason Leung on Unsplash
• Photo by Théo rql on Unsplash
• Photo by National Cancer Institute on Unsplash