3.2. Substance Use among the Subjects and Comorbidities
Among those who used substances, 33.3% (
n = 14) used alcohol (based on the active DSM-IV diagnosis and supported by urine tests/breath analyser), and the remainder used other psychoactive substances. Alcohol dependence was present in 23.8% (
n = 10) of cases and alcohol abuse was 7.1% (
n = 3). Among the 10 subjects with alcohol dependence, one subject was dependent on multiple substances and was therefore grouped under substance use with more than two substances (
Table 2). The majority of the subjects had a diagnosis of opioids (
n = 20; 47.6%) or benzodiazepines dependence (31%,
n = 13), among which eight subjects were dependent on both opioid and benzodiazepine. These subjects were included under the opioid and benzodiazepine group. Amphetamine and solvent use contributed to 2.4% (
n = 1) each and cases with more than two substances were grouped as “substance use with more than two substances” and they contributed to 7.2% (
n = 3,
Table 2). For the purpose of the descriptive analysis, subjects were grouped into six groups: (1) Alcohol alone (dependence, abuse and use) (2) Opioid alone (dependence) (3) Benzodiazepine alone (dependence) (4) Opioid and Benzodiazepine together (5) Substance dependence involving more than two substances which includes polysubstance dependence with alcohol and without alcohol (6) others (includes amphetamine and solvent abuse).
Approximately 31% (
n = 31) of the sample had Axis I psychiatric comorbidities. Seven subjects (16.6%) with alcohol use (abuse or dependence) had psychiatric comorbidities compared to three subjects with opioid dependence. Other illnesses such as depression (
n = 2), anxiety (
n = 1), schizophrenia (
n = 1), psychosis (
n = 1), gambling (
n = 2) and behavioral problems (
n = 1) were observed in trivial proportions. Approximately 16.7% had an Axis II diagnosis of personality disorder and 57.1% (
n = 24) had an Axis III diagnosis: hepatitis C (
n = 6; 14.3%), hypertension (
n = 5; 11.9%), asthma (
n = 4; 9.5%) and cardiovascular conditions (
n = 3; 7.1 %) were the most frequent illnesses. Majority (
n = 36; 85.7%) had psychosocial problems (Axis IV). Global Assessment of Functioning score (GAF, Axis V) was recorded for 31% (
n = 13) of the subjects and it showed a mean score of 62.3 (SD ± 6). A detailed description of various Axis IV conditions can be found in
Table 3.
3.3. Substance Use Profile and Manner of Death
Among the substances reported (self-reported by the subject, not supported by urine tests or diagnostic criteria) opioids, benzodiazepines and alcohol were the most abused drugs (
Table 4). The age of onset for the substance use ranged from 11 to 59 years, with a mean age of 30 (± 11), 32.7 (± 12), and 22.8 (± 12) for opioid, benzodiazepines and alcohol respectively. Kruskal-Wallis test was conducted to examine the distribution of age of onset of substance across different categories of substances. No statistically significant differences (χ
2 (2) = 5.9,
p = 0.05) were seen in the age of onset between different substance types. Heroin was the most abused drug among the opioid group (54.4%,
n = 12) followed by buprenorphine (
n = 5). The majority of the opioid users were daily users (88.9%,
n = 16) with an average use on 28.2 days (mode: 30, median: 30). Among the benzodiazepine users, midazolam (dormicum) was the most abused benzodiazepine (68.4%,
n = 13) with an average monthly use of 22.8 days (
n = 12, mode: 30, median: 30). Subjects who used alcohol reported an average monthly use of 19.6 days (mode 30 days, median 20).
History of polysubstance use was reported by 65.9% (
n = 27; 65.9%). Heroin (
n = 17; 63%), cannabis (
n = 9; 33.3%), buprenorphine (
n = 11; 40.7%) and midazolam (
n = 8; 29.6 %) were the commonly used drugs reported by polysubstance users (
Table 4).
Of the total number of unnatural deaths during the given period, suicide formed a large proportion (
n = 27; 64.3%) followed by accidental death (
n = 15; 35.7%). Among the accident cases, seven (16.7%) were due to accidental overdose. The substances used by the suicide group included alcohol (33.3%,
n = 9), opioid (25.9%,
n = 7) and opioid-benzodiazepine combinations (14.8,
n = 4). Opioid was the most abused drug in accidental deaths (
n = 5; 33.3%) followed by alcohol and opiate-benzodiazepine combination (
n = 4; 26.7% each,
Table 5). Intentional fall from high-rise buildings was the common mode of suicide in 85% (
n = 23) of the cases.
The mean age for accidental deaths and completed suicides were 44.4 and 45.1 years, with a median of 47 and 44 years respectively. The mean and median years to death after the onset of substance use was 19.4 and 17 years for accidental deaths and 16.3 and 10.5 years for suicides respectively.
Notably, when measured from their last visits, the mean years at death for accidents and suicides were 1.07 and 2.2 years, respectively. In both cases the median years were 0, which indicates that the death had occurred within 12 months from their last visit to NAMS clinic. More than half of the cases (n = 24, 57.1 %) had died within the first year from their last visit to the clinic. The remaining deaths had occurred between wider ranges (2 to 17 years) after the last visit to NAMS clinics. For those deaths that had occurred within the first year from the last appointment, 12 deaths occurred within 30 days of their last visit. Of this, six deaths had occurred within 7 days of their last visit. Suicide was the common cause of death (n = 9) among those who died within the 30 days of their last appointment.
Among those who sought admission to the hospital, 15 had undergone inpatient detoxification programme for their substance abuse related difficulties and 23 were admitted for other psychiatric conditions. The mean number of admissions to inpatient detoxification programme and other general psychiatric wards were 0.8 (SD ± 1.4) and 2.2 (SD ± 3.5) respectively, which showed that the inpatient admissions were predominantly for general psychiatric conditions than for substance abuse related conditions.
Pearson’s correlation was employed to determine the relationships between the collected variables. There was a moderate negative correlations between the age of onset of the main substance and years to death (rp(35) = −0.48,
p = 0.004), total number of clinic sessions defaulted and time to death after the last visit (rp(39) = −0.35,
p = 0.029), and number of admissions to general psychiatric ward and time to death after the last visit (rp(41) = −0.34,
p = 0.032). After controlling for the sociodemographic factors, a strong negative correlation was observed between years to death after the onset of substance use and age of onset of substance use. Gender, marital status, employment and age of onset were significant predictors of years to death after the onset of substance use (
Supplementary Tables S1–S3). Females tend to have a significantly higher chance to die early when compared to males (B = −20.4, CI: −36.8–3.9,
p = 0.02). Marriage (B = 11.9, CI: 2.9−20.9,
p = 0.01) and employment (B = 8.5, CI = 0.76–16.3,
p = 0.03) were protective against early death when compared to those who were single and unemployed. Ethnicity and employment status were stronger predictors of unnatural death following the last visit to the clinic than the number of defaulted clinic sessions and admissions to general psychiatric wards. Compared to Malays, Chinese (B = −959.63, CI: −1743.2–176.1,
p = 0.02) and Indians (B = −1209.8, CI: −2181.5–238.1,
p = 0.02) tend to succumb to unnatural death early after their last visit to the clinic. Employment was protective against early death (B = 511.3, CI: 5.5–1017.2,
p = 0.04).