This is a prospective cohort study including patients, living in the catchment area of the Hospital Clinic of Barcelona (Catalonia, Spain), with an activated CSRC from July 14th 2014 until November 1st 2017. The Hospital Clinic acts as a community general hospital, which covers a population of 540,000 inhabitants from the south of the city. It also operates as a tertiary and university (University of Barcelona) health care facility for highly complex psychiatric cases, and it is one of the most active research members of the Center of Biomedical Research Network in Mental Health.22

In the present study, we included all the patients aged ≥18 years, with activated CSRC because of suicidal ideation or a suicidal attempt. We collected their clinical data at baseline and during a 12 months follow up. The study was approved by the ethical committee of the Hospital Clinic of Barcelona (approval code number HCB/2017/0211) and all the patients gave their oral consent to be included in the CSRC.

The Catalonia Suicide Risk Code (CSRC) was developed in 2014 in the Catalonia region (ES), in order to prevent suicide attempts. More than 90 primary care services and hospitals in Catalonia have been included. The program has been described in detail elsewhere.2

Briefly, the CSRC includes three phases (Fig. 1): (1) Alert and activation phase: the emergency room staff evaluates the risk of the suicidal behavior and ask the patient to be included in the CSRC. The psychiatrist completes the MINI suicidal module, from MINI Interview,23 a risk factor checklist, and assesses the clinical risk at entrance and discharge. The psychiatrists at the Emergency Room received specific training in completing the CSRC website and its modules before the CSRC was started. Furthermore, audits were conducted and recommendations were provided for the implementation of the CSRC after its application was established. (2) Follow-up: two different pathways are activated. One consists of a follow-up phone call from a regional emergency system after 30 days of discharge, in order to detect alarm signs or the worsening of the clinical symptoms. At the same time, the patient is given an appointment in the outpatient clinic before their 10th day after discharge. In this first follow-up visit the psychiatrist, on the basis of the clinical condition, decides if the patient needs to be followed-up in the Mental Health Service during a one-year period or can be discharged because the risk no longer exists. (3) The general practitioner follows-up the patient in order to improve the prevention of the risk of suicide.

Fig. 1.

The Catalonia Suicide Risk Code (CSRC) phases: (1) alert and activation, (2) proactive telephone and face-to-face follow-up and (3) comprehensive preventive health monitoring.

(0.22MB).

An ad-hoc dataset of the CSRC has been created and each doctor visiting patients in the Emergency Room can access the electronic website with his/her personal data.

Three research members SGC, NV and ES were specifically trained to obtain data for the purpose of the study, through specific designed patients’ charts on baseline and follow-up study points (6 and 12 months). Data extraction was anonymously made from the specific CSRC website (https://salut.gencat.cat/pls/gsa/gsapk030.portal).

The following socio-demographic and clinical information were collected for each patient, when available: sex; birth date; age; Mental Health outpatient facility of the patient; DSM-5 diagnoses (including substance-related disorders); substance abuse during the suicide attempt; type of suicide attempt (i.e. drug overdose, precipitation, hanging, poisoning, drowning, severe cutting) including suicidal ideation without the attempt; clinical risk assessment.

The MINI suicidal module was administrated at entry and discharge from the Emergency Room. The MINI suicidal module (Fig. 2) is a validated submodule of the MINI, consisting of a short structured diagnostic interview with nine items.23 In the CSRC, a shorter version was used, with six items that were scored yes or no (Fig. 2). Whilst items 1–5 record whether an event has occurred during the last month, item 6 records the lifetime occurrence of the event (Fig. 2). Each item score is weighted according to its estimated contribution to risk level. The aggregated score range of the short version is 0–33 points.24

Fig. 2.

MINI suicidal submodule (MINI Neuropsychiatric Interview, short version).

(0.17MB).

The follow-up information was collected: the number of days until the CRSC proposed follow-up consultation with a psychiatrist at the outpatient clinics; the number of completed suicides; patients attending the follow-up visit and the reasons for not attending were also collected. Specific follow-up outcomes considered and obtained in the present study were (1) the number of Emergency Room evaluations due to any reason; (2) the number of Emergency Room evaluations due to a suicide attempt/suicidal ideation; (3) the number of Psychiatric Hospitalizations. These follow-up measures were recorded at 6 months and one year, when available.

Descriptive statistical analysis for demographic and clinical characteristics of the sample was performed. Normality of distribution for continuous variables was evaluated with the Kolmogorov–Smirnov test, visually and with the skewness and kurtosis values.

McNemar's test, the chi-square test for repeated measures in the same sample, was conducted in order to assess the differences in the proportion of patients visiting the Emergency room for any reason or for a suicide attempt or being hospitalized in the first semester versus the second semester after the activation of the CSRC.

Statistical analyses were performed using the Statistical Package for Social Sciences (Statistical Package for Social Science-SPSS, 23.0 version for Windows Inc., Chicago, IL, USA).

A sample of 365 patients was enrolled. Two-hundred eighteen (59.7%) were female. The mean age was 44.9 (±15.7), ranging from 18 to 92 years.

Fifty patients (14.7%) had no previous psychiatric diagnosis whilst 290 patients (85.3%) had at least one psychiatric diagnosis.

Among the 290 patients with a previous diagnosis, 64 patients (22%) had more than one diagnosis. The different psychiatric diagnoses are reported in Table 1.

From the whole sample, in 270 patients (74%) the prevention program was activated in the context of a suicide attempt. To a lesser extent, the CSRC was activated in 95 patients (26%) because of suicidal ideation, without a suicide attempt.

Among the 270 suicide attempters, 19 (7%) mentioned previous planning of the suicide attempt.

In the sample (data available for 226 patients), 15 (6.6%) patients were hospitalized at discharge from the Emergency Room.

Amongst the different types of suicide attempts, the most common were intoxication by any kind of noxious agent, occurring in 84.4% of the patients, followed by cutting 10% and other less used methods such as jumping from a high place 1.1%, hanging 0.8%, drowning 0.4%. In 3.3% of cases, the method was not specified.

Out of 351 patients (14 patients with missing information), 50 patients (13.7%) presented drug intoxication during the suicide attempt. In particular, alcohol intoxication was the most frequent (76%), followed by cocaine (12%), morphics (8%), cannabis and amphetamines (2% each).

The scores of the MINI suicidal module were evaluated and recorded in different phases, namely at the Emergency Room, hospital discharge and after the first external consultation at the psychiatric outpatient clinic. Also, scores based on clinical criteria at the first Emergency Room visit were collected. The results are detailed in Table 2.

Among the 365 patients with activated CSRC, 66.6% (n=243) attended the first follow-up visit. The 33.4% (n=122) of the total did not attend the follow-up visit (reasons for not attending are described in Table 3). The mean waiting time between the activation of the CSRC at the discharge from the Emergency Room and the follow-up visit at the Mental Health Service was 5.9 days (±8.4).

The proportion of patients and the number of visits at follow-up in the two different time points (6 and 12 months) after the activation of the CSRC are listed in Table 4.

The comparison between those patients visiting the Emergency room for any reason or for a suicide attempt or being hospitalized in the first semester versus the second semester has been conducted on 206 patients out of 365, whose data on these variables was available.

A significant reduction has been observed in the proportion of patients visiting the Emergency room, for any reason and being hospitalized in the first semester in comparison with the second semester after the CSRC activation (30.1% versus 19.9%, p=0.006; 14.1% versus 5.8%, p=0.002). No significant difference was observed in the proportion of patients visiting the Emergency room, for a suicidal attempt/ideation in the first semester in comparison with the second semester, although the rate was numerically smaller (9.7% versus 6.8%, p=0.307).

Similarly, to previous studies,25–28 the prevalence of female patients attempting suicide was higher than male patients.

Most patients in our study had a previous psychiatric diagnosis. This result has also been previously reported in studies supporting the frequent association between psychiatric past history and suicidal behavior.29–33

The most frequent diagnoses in our sample were major depressive disorder not otherwise specified or adjustment disorder, as well as substance use disorder and personality disorder. Similarly, in previous literature anxiety, depressive, and alcohol use disorders were the most prevalent disorders among suicide attempters in general.33

Moreover, symptoms such as low mood (measured as a continuous variable rather than a discrete mental disorder), hopelessness, and impulsivity, which are common in those specific diagnoses, are also considered to be particularly important predictors of suicidal thoughts and attempts.3 In the same way, previous studies remark the importance of cluster B personality disorder and borderline symptoms as risk factors for suicide attempts, in accordance with our results.34,35

In addition, more than one out of ten patients attempted suicide with drug overdose, and in almost two out of ten patients, a dual diagnosis was present. This is consistent with previous evidence showing that suicidal behavior is a significant problem for people with co-occurring substance use disorders. The combination of different precipitating risk factors such as recent heavy substance use and intoxication as well as personality traits and mental illnesses has been seen to intensify the risk for suicidal behavior in dual diagnosis patients.36

A noteworthy finding was that, the prevalence of severe mental illness diagnosis such as schizophrenia and bipolar disorder was quite low in our sample and definitely lower than expected according to previous studies, where severe mental disorders are considered an important risk factor for suicide behavior.29,37 A possible explanation for these differences could be that patients suffering from bipolar disorder or schizophrenia generally commit more severe and lethal suicide attempts so they do not generally attend the psychiatric emergency room because they require medical assistance.38 It is important to highlight that, the present study relies on a registry of data collected at the psychiatric emergency room, that does not include all the data on suicide of our territory. Moreover, the Hospital Clinic has two very prominent units, one for Bipolar Disorder and another one for Schizophrenia, that provide not only tertiary care but also community care for patients in the catchment area.39,40 Hence, it is likely that any suicidal ideation or event may have been captured and addressed by those units, out of the scope of the Emergency room.

In our study, the most common type of suicide attempt was by drug overdose, similarly to previous literature. Drug intoxications, are by far the most frequently used method, mainly among females.25 Interestingly, research on suicide methods could lead to the development of gender-specific intervention strategies.41

Regarding suicidal risk scores, more than a half of the patients were initially evaluated according to the MINI suicidal module from moderate to high risk, at the Emergency Room, which was similar to the clinical criteria evaluation. At discharge the risk was lower, with only one out of six patients with a moderate to high risk evaluation. This could be due to the fact that the CSRC should be activated when the patient comes to the Emergency Room and, after a clinical observation period, the suicidal risk is re-evaluated and a clinical decision should be made. If the patient continues to be suicidal and keeps scoring high in the MINI suicidal module, psychiatric hospitalization was eligible. If the suicidal risk is low or absent, the patient is discharged by the Emergency Room and a follow-up visit is planned at the community mental health service, within the CSRC. This is consistent with the low rates of hospitalizations in the present sample. An important aspect is that one out of four patients still presents high or moderate risk at the evaluation with the community mental health team, which underlines the main important aspect of this program which is that the patient feeling suicidal has fast access to the psychiatric follow-up.

As a consequence, suicidal assessment by means of specific assessing tools allows the quantifying of the severity of patients’ symptoms, and could be then implemented, as a complementary tool, in routine use in emergency departments in order to tailor case management.19

Nonetheless, a high percentage of patients did not attend the follow-up visit, despite the fact that the time period from the discharge from the emergency room to the follow-up visit in the outpatient clinic was less than one week.

A significant reduction of consultation in the Emergency Room and of the hospitalization rate was found in the second semester after the activation of the CSRC in comparison with the first one. However, the number of Emergency Room visits due to a suicidal attempt was not significantly reduced. It can be hypothesized that if the patient is involved in a specific follow-up program, he or she will less possibly need to be visited at the Emergency Room because he/she is already involved in a longitudinal care system, with important implications in terms of mental health costs.12 As underlined by Inagaki and colleagues in their recent meta-analysis on interventions to prevent repeated suicide attempts, active contact and follow-up type interventions could reinforce the therapeutic connection between patients and care providers, reducing the risk of a repeated suicide attempt within 6 months.12 Indeed, one of the major issues in the prevention of suicide reattempts is the lack of continuity of care following a suicide attempt.11 The time window for preventing a suicide reattempt should be considered. In a Danish case-control study performed between 1981 and 1997, the risk of reattempt of suicide was higher during the first week following psychiatric hospital discharge.42 Interestingly, the risk declined in the following weeks and months.42 As a consequence, policy makers should prioritize the promotion of continuity of mental health care for those patients at high risk of suicide during the period immediately following hospital discharge.11 Similarly to our results, a British observational study that provided a 7-day follow-up intervention after psychiatric hospital discharge was associated with a significant decline in suicide reattempts during the 3-month period following hospital discharge.43

One of the important limitations of the CSRC is the quite high dropout rate, which has also been outlined in a recent article explaining the development of the protocol in Catalonia.2 Furthermore, there might be an underutilization of the protocol by the mental health providers. As a consequence, further analyses on the process of implementation of the CSRC should be provided. Another important limitation to consider is the lack of comparison group since this study was done as an observational study. Finally, the CSRC is mostly focused on non-fatal suicidal behavior, and therefore its impact on suicide rates is still unknown. An epidemiological analysis of the comparison between the outcomes of the CSRC intervention and the regional rates of death by suicide should be further explored.

Despite these limitations, the CSRC protocol is an example of the attempt to improve surveillance and monitoring of suicidal patients, that is one of the most important challenges according to the World Health Organization.