4. RISK AND PROTECTIVE FACTORS

Suicide is seldomly explained by one causal factor; often multiple factors contribute to the act. Researchers in the Nordic countries have contributed with a substantial amount of high-quality research on suicide prevention. This has partly been facilitated by using the national administrative registers, which exist in all the Nordic countries (Erlangsen, et al., 2018a), in addition to other data sources.

Evidence regarding risk and protective factors for suicide were identified by conducting a literature search. Due to the abundance of Nordic literature, it was opted to restrict the search to studies on death by suicide, as this is less frequent than suicide thoughts and suicide attempts, thus, generating more conservative estimates and reliable findings. Further, the search for risk and protective factors was limited to studies published after 2000, which were based on data from the Nordic countries and where a comparison group equivalent to the general population had been used. When several studies were identified, those with a higher level of evidence, larger study population, more recent data, and estimates, which had been adjusted for relevant confounders were prioritized. Factors were divided into socio-demographic, mental disorders, physical disorders, stressful live-events, psychological, biological, and healthcare-related factors. Explanations for the elevated risks of suicide for some of these factors are also presented.

Several socio-demographic characteristics of individuals who die by suicide have been identified. Excess suicide rates have been reported among Greenlandic people, and the suicide rates in Greenland belong to the highest in the world (Naghavi, 2019). The Sami people of Finland, Norway and Sweden have also been reported to have high rates of suicide (Young, et al., 2015, Jacobsson, et al., 2020). High suicidality among Indigenous populations has been related to the periods of colonization and rapid modernization causing cultural and social disruption and problems with alcohol and childhood adversities (Bolliger & Gulis, 2018). Further, many members of the Indigenous population have personal experiences of losing someone to suicide, which can increase the risk that suicide is viewed as an option to solve a life crisis (Bolliger & Gulis, 2018, Granheim, et al., 2021). Lack of access to mental health care and availability of firearms are other concerns (Jacobsson, et al., 2020, Bolliger & Gulis, 2018, Granheim, et al., 2021).

In Faroe Islands, rates of suicide have been remarkably lower than in other Nordic countries (Wang & Stórá, 2009). The contributing factors are challenging to demonstrate quantitatively, but social connectedness, a traditionally religious community with low consumption of alcohol have been suggested as contributing factors. In addition, the fact that there are few suicides does not create precedence for new cases, i.e. there are no social role models to follow (Wang & Stórá, 2009, Wang, 2005).

Higher rates of suicide have consistently been found among males in the Nordic countries when compared to females (Øien-Ødegaard, et al., 2021). The difference in suicide rates between the sexes have been attributed to males’ preference for more lethal methods, while females tend to use method with a lower fatality ratio, such as poisoning (Nordentoft, 2007). Relationship problems has been mentioned as an explanatory factor for male suicides in Norway (Knizek & Hjelmeland, 2018).

In Greenland, the highest rates are observed among adolescents and young adults aged 15-29 years (Seidler, et al., 2023a), while younger and middle-aged age groups were accountable for the highest rates in Finland, Iceland, and Norway. In Denmark and Sweden, the highest rates of suicide were seen among the oldest age groups. Different circumstances may contribute to the high suicide rates in specific age groups. For instance, misuse behaviours, violence and exposure to sexual abuse have been mentioned as explanations for the high rates of suicide among Greenlandic youth (Seidler, et al., 2023a, Bolliger & Gulis, 2018). The high occurrence of suicides among older adults has been attributed to undiagnosed depression as well as personality traits, such as being strong-willed and a reluctance towards becoming a burden to others (Pitkälä, et al., 2000, Kjølseth, et al., 2002, Kjolseth, et al., 2009a, Kjolseth, et al., 2009b).

Being married has consistently been found to protect against suicide in the Nordic countries (Øien-Ødegaard, et al., 2021, Qin, et al., 2003). Individuals who have never been married, are widowed, divorced, or separated have higher rates of suicide (Øien-Ødegaard, et al., 2021, Crump, et al., 2014) The highest risks were found among those who were separated, followed by those who were divorced (Øien-Ødegaard, et al., 2021).

Having one or more children was found to be protective of suicide compared to not having children, based on data from Greenland and Norway (Øien-Ødegaard, et al., 2021, Seidler, et al., 2023b). Being the parent of a young child has also been shown to be protective of suicide, especially for moms (Qin, et al., 2003, Gissler, et al., 2005, Lysell, et al., 2018).

Lower educational level has been linked to higher suicide risks than higher educational levels (Crump, et al., 2014, Mäki & Martikainen, 2009). Being unemployed has been associated with suicide (Seidler, et al., 2023b, Mäki & Martikainen, 2012). Further, a significant association was found for long-term unemployment in a Finnish population-based sample of adults in the working ages (Mäki & Martikainen, 2012). Also, having to take sick leave has been linked to a higher rate of suicide when compared to those not on sick leave (Wang, et al., 2014). An inverse relation between income level and risk of suicide has been documented in Denmark and Sweden, in the sense than lower income is associated with higher risks of suicide (Qin, et al., 2003, Crump, et al., 2014, Hiyoshi, et al., 2018). With regard to those in employment, higher rates have been reported for medical doctors, dentists, veterinarians, pharmacists, and nurses when compared to other occupational groups (Dalum, et al., 2022, Hawton, et al., 2011). The excess risk in these professions has been attributed to a high workload and stress burden in addition to knowledge about and access to lethal methods (Lindfors, et al., 2009).

Other factors, for instance, where one is born, has also been linked to suicide although findings are equivocal. Living in the country where one was born has been linked to lower suicide rate in Denmark and Sweden (Qin, et al., 2003, Crump, et al., 2014). However, low rates of suicide were also observed among first and second generation immigrants in Norway when compared to those born in the country (Puzo, et al., 2017).

Rates of suicide have been suggested to be slightly higher among young persons living in semi-rural or rural areas in Sweden versus those living in urban areas (San Sebastián, et al., 2020). In Greenland, large variation in suicide rates were seen across different regions (Seidler, et al., 2023a). Area-level markers, such as social deprivation or unemployment rates, seem to play less of a role for individuals’ suicide risk when compared to individual-level factors (Agerbo, et al., 2006). However, differing from the norm, for instance, being born by foreign-born parents and living in an area where most people do not have foreign-born parents, might be a risk factor for suicide (Borczyskowski, et al., 2006). As described by the minority stress hypothesis, individuals belonging to minority groups might be more likely to experience stigmatisation and discrimination, which have been linked to higher risks of suicide (Erlangsen, et al., 2023a). The same mechanism has been suggested to explain the higher suicide risks observed among individuals who were international adoptees and might have a different appearance than others and when compared to national adoptees (Borczyskowski, et al., 2006). The minority stress hypothesis has also been mentioned in relation to sexual minority groups. Linkage studies from Denmark and Sweden have revealed elevated rates of suicide among individuals who entered same-sex marriages and transgender individuals (Erlangsen, et al., 2023a, Erlangsen, et al., 2020a, Bjorkenstam, et al., 2016). In Iceland, adolescents who self-identified as belonging to sexual minority groups were found to have higher risks of suicide attempt than other adolescents (Arnarsson, et al., 2015).

Mental disorders and conditions have consistently been associated with excess risks of suicide. A cumulative incidence of 4.3% was found for males who had a mental disorder in Denmark; implying that 4.3% of this group had died by suicide (Table 4.1). In comparison, just 0.7% of males with no mental disorder had died by suicide (Nordentoft, et al., 2011). For females, the cumulative incidences were 2.1% and 0.3% for those with and without mental disorders, respectively (Nordentoft, et al., 2011). The excess mortality among males and females with mental disorders quantifies to 1.6 and 0.9 years of shorter life spans due to deaths by suicide alone when compared to those of people with no mental disorders (Erlangsen, et al., 2017b). The population attributable risks for females and males who at some point in their lives had been admitted to psychiatric hospital have been calculated as 53.9 and 32.5, respectively, using Danish data (Qin & Nordentoft, 2005). These figures provide evidence of the magnitude, with which the overall female and male suicide rates would be reduced, i.e. approximately 54% and 32%, respectively, if suicide risks in these groups were reduced to the levels of the general population, under the assumption that mental disorders is the only causal factor determining those deaths. The measure of population attributable risk provides information about the reduction in the overall suicide rate, which one might achieve, if eliminating a causal risk factor. As such, the measure provides useful information for setting priorities regarding target groups for interventions.

With respect to types of mental disorders, cumulative incidences show that between 6-8% of males and 4-5% of females diagnosed with schizophrenia and affective disorders subsequently die by suicide (Nordentoft, et al., 2011, Aaltonen, et al., 2018). Elevated risks of suicide have been documented for many mental disorders and conditions (Table 4.2). Adjustment disorder, bipolar disorders, borderline personality disorders, PTSD, reaction to stress/adjustment disorders, recurrent depression, schizophrenia, and substance use disorders are some of the disorders, which are linked to the highest risks of suicide.

*Cumulative incidence for depression (ICD-9: 2961A-G, 2968A; ICD-10: F32-33) was estimated using data from Finland, while the bipolar (ICD-8: 296.19, 296.39; ICD-10: F30 and F31) and unipolar affective disorders (ICD-8: 296.09, 296.29, 296.89, 296.99, 298.09, 298.19, 300.49, 301.19; ICD-10: F32-F34, F38, F39) were estimated using data from Denmark.

^† Unless otherwise stated, the estimates are for both sexes.

Table 4.1 Cumulative incidence rates for suicide among individuals with mental disorders.

Individuals who had been admitted as an inpatient to a psychiatric ward within the preceding year were found to have a 44-fold higher suicide rate when compared to those who has never been in any type of psychiatric treatment and adjusted for socio-demographics and previous suicide attempt (Hjorthoj, et al., 2014). In comparison, those who attended the emergency department had a 27-fold higher suicide rate, while those in outpatient treatment and those in treatment with psychiatric medication had 8- and 5-fold higher rates, respectively.

Two time points are crucial in relation to psychiatric hospitalisation; when individuals are being admitted and discharged. Rates of suicide were found to be 200-300 fold higher during the first week of admission to a psychiatric ward when compared to those never admitted (Madsen, et al., 2020). Similarly, 200-400 fold higher rates of suicide were recorded during the first week after being discharged (Madsen, et al., 2020). The excess risks of suicide at the time of admission and discharge have been demonstrated across a range of disorders, including schizophrenia, affective disorders, substance abuse disorders, anxiety and stress reactions as well as personality disorders (Aaltonen, et al., 2018, Qin, et al., 2006). Although efforts, such as a risk assessment at the time of discharge, have been implemented, the suicide rates after discharge have not improved over recent decades in Denmark (Madsen, et al., 2020).

Time of the first diagnosis has been shown to be linked to elevated risks of suicidality for individuals with schizophrenia and depression (Aaltonen, et al., 2018, Reutfors, et al., 2009). Multiple admissions, for instance for schizophrenia, have also been linked to excess risk of suicide in Sweden (Reutfors, et al., 2009).

Suicide attempt is another important predictor of suicide. Approximately 3.0% of males and 1.4% of females who had a suicide attempt died by suicide within the following year (Tidemalm, et al., 2015). Being older than 35 years and having used a method with high case fatality in the initial attempt were identified as predictors of later suicide (Fedyszyn, et al., 2016).

Abbreviations: HR, Hazard ratio, IRR, Incidence rate ratio, OR, Odds ratio.
Symbol denotes level of adjusting covariates. ^† Adjusted for sex and age. ^†† Adjusted for sex, age and sociodemographic covariates or other context-specific variables. ^††† Adjusted for sex, age, sociodemographic covariates, and psychiatric comorbidity.

^‡ Unless otherwise stated, the estimates are for both sexes.

Table 4.2 Risk estimates for suicide among individuals with mental disorders.

A range of physical disorders have been linked to suicide (Stenager, et al., 2020). When reviewing evidence from the Nordic countries, a range of physical disorders have been examined in relation to suicide deaths (Table 4.3). Some of the highest suicide rates have been reported for fibromyalgia, amyotrophic lateral sclerosis (ALS), and Huntington disease. When compared to the general population, previous reports found suicide rates to be 30-90% higher among individuals with hospital-treated infections, heart diseases, head injury, stroke, cancer, chronic respiratory diseases, diabetes, and epilepsy. These disorders affect large segments of the population, implying that large numbers of suicide deaths can be related to those disorders (Petersen, et al., 2020, Erlangsen, et al., 2020b, Lund-Sørensen, 2016, Madsen, et al., 2018, Sariaslan, et al., 2022). Although population attributable risks have seemingly not been calculated for most physical disorders, it was found to be 10.1 for infections based on Danish data (Lund-Sørensen, 2016). Thus, underscoring the potential for suicide reductions if directing interventions towards this and other relatively large groups, as proposed by Roses’ theorem of efforts directed towards large groups with a low excess risk may reduce a substantial number of cases (Rose, 1992).

In relation to the recent COVID-19 pandemic, it was found that individuals who had been hospitalized with SARS-CoV-2 infection did not have an elevated rate of suicide attempt when compared to the general population with no such infection. (Erlangsen, et al., 2023b).

The elevated risks of suicide linked to physical disorders have been attributed to the stressful experience of being diagnosed with a severe disorder (Stenager, et al., 2020). Depressions might also be an explanatory factor, as certain disorders, such as stroke and cancer, are linked to an excess risk of depression (Petersen, et al., 2020, Fang, et al., 2012). Level of pain might play a role for diseases, such as ALS (Erlangsen, et al., 2020b). In addition, chronic disorders may limit functional skills and lead to an increased dependency on others, which has been linked to suicide (Stenager, et al., 2020). Some physical disorders, such as traumatic brain injury, have been linked to changes in cognition and emotion regulations (Madsen, et al., 2018). Lastly, severity, as measured through multiple hospital contacts or more complicated stages of a disorder, has been linked to excess rates of suicide for multiple sclerosis and heart diseases (Petersen, et al., 2020, Erlangsen, et al., 2020b).

The time shortly after hospital contact for severe disorder has been associated with the highest increase in risk (Erlangsen, et al., 2020b, Fredrikson, et al., 2003). Thus, supporting the assumption that being diagnosed with a severe disorder can be distressing and might be an important time point for support.

Table 4.3 Risk estimates for suicide among individuals with physical disorders.

Abbreviations: HR, Hazard ratio, IRR, Incidence rate ratio, OR, Odds ratio, SMR, Standardized Mortality Ratio.

Symbol denotes level of adjusting covariates. ^† Adjusted for sex and age. ^†† Adjusted for sex, age and sociodemographic covariates or other context-specific variables. ^††† Adjusted for sex, age, sociodemo­graphic covariates, and psychiatric comorbidity. ^‡ Unless otherwise stated, the estimates are for both sexes.

Stressful life-events may lead to an excess risk of suicide, which has been demonstrated for a large number of events (Table 4.4). Stressful life-events may happen unexpectedly and require adaptation. Albeit suicide deaths are not always preceded by mental disorders, stressful life-events may increase individual’s susceptibility to psychological stress, which may lead to mental disorders, and suicidal behaviour (Erlangsen, et al., 2017a, Howarth, et al., 2020). As mentioned in Sections 4.2 and 4.3, diagnosis of mental and physical disorders may constitute stressful life-events in the own right.

Having been exposed to a range of childhood adversities, including death in the family, financial and residential instability, was found to be associated with higher rates of subsequent suicide when compared to individuals not exposed to such events in Sweden (Björkenstam, et al., 2017). Based on Greenlandic data, childhood adversities, such as growing up in a poor emotional environment or a home with alcohol problems as well as experiences of neglect or sexual abuse, were predictive of later suicidal behaviour (Seidler, et al., 2023b). Other early life factors, such as being born by a mom who was below the age of 25 years at the birth or having been placed outside the home by the authorities, were also significant predictors of suicide (Sujan, et al., 2022, Wall-Wieler, et al., 2018).

Bereavement, in the form of the death of a parent or close relative, has been linked to an excess risk of suicide, especially in the first weeks after the loss (Mogensen, et al., 2016a, Burrell, et al., 2018).

Children, siblings and partners who have been bereaved by suicide have higher rates of suicide themselves when compared to the general population (Erlangsen, et al., 2017a, Ranning, et al., 2022a, Tidemalm, et al., 2011). The risk of suicide seems to be higher among first-degree relatives and partners than among distal relatives, although both groups have elevated risks (Tidemalm, et al., 2011). Further, evidence suggest that children who were exposed to a parental suicide at a young age will themselves have higher rates of suicide than children exposed at older ages (Ranning, et al., 2022b). The familial transmission of suicide may be due to genetic factors, that are likely to be related to mental disorders, but also social role modelling may play a role (Ranning, et al., 2022b, Kendler, et al., 2021). Experiencing a suicide attempt of a parent has also been linked to a higher risk of suicide in children when compared to children who were not exposed (Ranning, et al., 2022a).

Perceived job insecurity and loss of job have been linked to elevated risks of suicide (Blomqvis, et al., 2022). Experiences of financial hardship, such as financial debt and being evicted from one’s home, have been identified as predictors of suicide (Rojas, 2021, Rojas & Stenberg, 2016). In addition, elevated rates of suicide were found among individuals with frequent residential relocation (Qin, et al., 2009).

Suicide rates have been observed to increase during recessions in countries outside the Nordic region, for instance Greece and Ireland (Economou, et al., 2011, Corcoran, et al., 2015). In Iceland, researchers did not find a statistically significant increase of suicides during periods of recession (Ásgeirsdóttir, et al., 2020, Óskarsson, et al., 2019). Also, suicide rates did not increase during a recession in 1993-1996 in Sweden. Still, the suicide rate among unemployed increased over the subsequent years; suggesting a possible late effect. The Nordic social welfare models has been hypothesised to act as a buffer for adverse effects of economic crises (Garcy & Vagerö, 2013).

There are only few examples of how natural disasters may impact risks of suicide from the Nordic literature on suicide prevention. Swedish residents who were evacuated from the 2004 tsunami in Southeast Asia had a higher rate of suicide attempts when compared to other Swedish residents, while accounting for relevant confounders (Arnberg, et al., 2015). Other crises situations are witness by military conscripts; veterans who reported combat exposure during military deployment had higher risks of suicide attempt, yet this association was found to be explained through presence of PTSD and other mental disorders (Vedtofte, et al., 2021). Findings from studies of Norwegian and Swedish peacekeeping forces and other deployed military troops have not shown higher rates of suicide than for the general population (Michel, et al., 2007, Thoresen, et al., 2003).

Personal stressors, include fertility problems, which seemingly are linked to a higher risk of suicide among females (Kjaer, et al., 2011). Although women who underwent an induced abortion had higher rates of suicide than women in the reproductive ages in general, the elevated rate was found to decrease after introduction of better care guidelines in Finland (Gissler, et al., 2015). 

Being bullied have been linked to elevated risks of self-harm (Landstedt & Gillander, 2011). This finding has been confirmed in workplace settings, where bullying was associated with a higher risk of suicide when compared to individuals who did not report bullying (Conway, et al., 2022). In addition, victims of domestic violence, as reported in a Swedish household survey, were found to have higher risks of attempted suicide (Dufort, et al., 2015). Based on Norwegian survey data, adolescents who self-reported having been subjected to sexual or physical abuse had an increased risk of self-harm when compared to peer with no such exposure (Mossige, et al., 2016). This is supported by data from Denmark where female victims of sexual offences were found to have a 30-fold higher suicide rate when compared to age, marital status and income level-matched controls (Gradus, et al., 2012). A systematic review of Greenlandic studies demonstrated evidence that exposure to sexual and physical assaults was associated with suicidal behaviour (Seidler, et al., 2023b).

Refugees are likely to experience traumatizing events both in their home country, during their migration, and after arrival (Khan Amiri, et al., 2021). Nevertheless, refugees were in general not found to have higher suicide rates when compared to the general population in Denmark, Norway and Sweden (Khan Amiri, et al., 2021, Geirsdottir, et al., 2021, Hollander, et al., 2020, Amin, et al., 2021a, Amin, et al., 2021b). An exception is unaccompanied minors who seek asylum; this group was found to have higher suicide rates when compared to youth in Sweden (Mittendorfer-Rutz, et al., 2020). It is possible that cultural differences are responsible for the lower prevalence of suicide among asylum seekers, given that many derive from countries with general lower rates of suicide.

Individuals who at some point experienced homelessness were shown to have elevated risks of suicide, in particular in relation to mental disorders (Nielsen, et al., 2011).

People who were convicted of criminal acts or incarcerated were found to have higher suicide rates than found for the general population in Denmark, Iceland, Norway, and Sweden (Morthorst, et al., 2021a, Stenbacka, et al., 2014). Among males who had been found guilty of a homicide, a particularly high rate of suicide was found within the first years of the verdict (Jokinen, et al., 2009). Also, juvenile delinquency, which led to a verdict, has been associated with elevated risks of later suicide when compared to those with no criminal conviction (Bjorkenstam, et al., 2011). In Sweden, the risk of suicide remained elevated after release from incarceration; thus, underscoring the need to be attentive towards mental health and substance use disorders in this group also after being released (Haglund, et al., 2014). This finding is supported by Danish data, which indicated an excess risk of suicide after release from prison when compared to individuals with no contact to the criminal justice system (Webb, et al., 2011).

Table 4.4 Stressful life-events associated with excess risks of suicidal behaviour.

Numerous psychological factors have been identified in international studies. However, relatively little evidence is available from the Nordic region. Still, it has been found that resilience, measured as the ability to cope with psychological stressful situations among young males through semi-structured interviews with a psychologist at conscription, was protective of suicide attempts later in life when compared to siblings and cousins in Sweden (Lannoy, et al., 2022). Severe symptoms of feeling, down, anger, and trouble falling asleep, as reported by males during conscription, have been linked to later suicide in adjusted analyses (Hogstedt, et al., 2018). In addition to these measures, severe symptoms of headache and being nervous were linked to later suicide attempts (Hogstedt, et al., 2018). Moral values, endorsing authority and loyalty, as assessed in a survey of Icelandic youth, was also found to be associated with a lower probability of suicide attempts (Silver, et al., 2021).

Relatively little is known regarding biological factors for suicide. Genetic liability, mainly mitigated through a genetic disposition for mental disorders whose association with suicide is well-established, has been suggested (Kendler, et al., 2021, Erlangsen, et al., 2018b). Yet, the evidence is meagre.

Suicide has been shown to vary by season, for instance, in Finland, Norway, and Sweden where rates were found to peak during May and the summer months (Holopainen, et al., 2013, Bramness, et al., 2015). Although no clear causal mechanism has been demonstrated, changes in temperature, hours of sunshine, ambient warming and global radiation have been suggested as explanations for this (Hiltunen, et al., 2011).

Although about half of those who die by suicide have not been in touch with psychiatric care prior to dying by suicide, a significant proportion have attended their primary care provider in the months before dying.

In Sweden, 72% of individuals who died by suicide had been in contact with their primary care provider within the last 12 months (Bergqvist, et al., 2022). In Denmark, 83% of those who died by suicide had seen their primary care physician within the last year. This was significantly more than an age- and sex-matched comparison group where 76% had attended their primary care physician. Also, as many as 60% had attended care within the last four months (Pedersen, et al., 2019). In Greenland, one third of individuals who died by suicide were found have had contact to healthcare providers within the last 6 months (Grundsøe & Pedersen, 2019). Norwegian findings revealed that 74% and 86% of Norwegian males and females were in contact with a primary health care provider within 6 months of dying by suicide. In comparison, just 68% of male and 71% of female immigrants were in contact with primary health care providers in the last 6 months before dying by suicide (Øien-Ødegaard, et al., 2019). Using a more recent scale, 66% of those who died by suicide during 2016-2018 in Finland were found to have been in contact with primary or specialized care during the last 30 days. As many as 21% had been in contact on the actual day where they died by suicide (Partonen, et al., 2022).

Concerningly, Swedish researchers detected significant deficiencies in the healthcare, which was provided to more than half of the patients who saw a primary health care provider prior to suicide (Roos Af Hjelmsäter, et al., 2019).