Journal of Psychosomatic Research 78 (2015) 356–362

Contents lists available at ScienceDirect

Journal of Psychosomatic Research

General hospital-treated self-poisoning in England and Australia: Comparison of presentation rates, clinical characteristics and aftercare based on sentinel unit data Sarah Hiles a, Helen Bergen b, Keith Hawton b, Terry Lewin a, Ian Whyte c,d, Gregory Carter a,⁎ a

Centre for Translational Neuroscience and Mental Health, University of Newcastle, Australia Centre for Suicide Research, University Department of Psychiatry, University of Oxford, Warneford Hospital, Headington, Oxford, UK c Department of Clinical Toxicology and Pharmacology (Hunter Area Toxicology Service — HATS), Calvary Mater Newcastle, Australia d Discipline of Clinical Pharmacology, School of Medicine and Public Health, Faculty of Health, University of Newcastle, Australia b

a r t i c l e

i n f o

Article history: Received 10 September 2014 Received in revised form 30 December 2014 Accepted 14 January 2015 Keywords: Deliberate self-poisoning Suicide attempt Hospital services Epidemiology Self-harm

a b s t r a c t Objective: Hospital-treated deliberate self-poisoning (DSP) is common and the existing national monitoring systems are often deficient. Clinical Practice Guidelines (UK and Australia) recommend universal psychosocial assessment within the general hospital as standard care. We compared presentation rates, patient characteristics, psychosocial assessment and aftercare in UK and Australia. Methods: We used a cross sectional design, for a ten year study of all DSP presentations identified through sentinel units in Oxford, UK (n = 3042) and Newcastle, Australia (n = 3492). Results: Oxford had higher presentation rates for females (standardised rate ratio 2.4: CI 99% 1.9, 3.2) and males (SRR 2.5: CI 99% 1.7, 3.5). Female to male ratio was 1.6:1, 70% presented after-hours, 95% were admitted to a general hospital and co-ingestion of alcohol occurred in a substantial minority (Oxford 24%, Newcastle 32%). Paracetamol, minor tranquilisers and antidepressants were the commonest drug groups ingested, although the overall pattern differed. Psychosocial assessment rates were high (Oxford 80%, Newcastle 93%). Discharge referral for psychiatric inpatient admission (Oxford 8%, Newcastle 28%), discharge to home (Oxford 80%, Newcastle 70%) and absconding (Oxford 11%, Newcastle 2%) differed between the two units. Conclusions: Oxford has higher age-standardised rates of DSP than Newcastle, although many other characteristics of patients are similar. Services can provide a high level of assessment as recommended in clinical guidelines. There is some variation in after-care. Sentinel service monitoring routine care of DSP patients can provide valuable comparisons between countries. © 2015 Elsevier Inc. All rights reserved.

Introduction Non-fatal self-harm (variously defined) is common, although the reported rates vary widely. Community 12 month estimates for suicide attempt are: 300 per 100,000 in developed and 400 per 100,000 in developing countries [1]. 12 month estimates of hospital-treated or community parasuicide (deliberate, nonfatal self-injury or selfpoisoning) are 2.6 to 1100, and lifetime rates 720 to 5930 per 100,000 [2]. General hospital-treated self-harm (SH) is common and costly, and deliberate self-poisoning (DSP) is the most common variant [3,4]. Hospital-treated SH is associated with increased repetition of SH [5], suicide [6], and natural cause mortality [7,8]. However, there are serious limitations to our understanding of rates, clinical characteristics and service provision for SH. Systematic reviews ⁎ Corresponding author at: Locked Bag #7, Hunter Region Mail Centre, NSW 2310, Australia. Tel.: +61 249 211283; fax: +61 249 211870. E-mail address: [email protected] (G. Carter).

http://dx.doi.org/10.1016/j.jpsychores.2015.01.006 0022-3999/© 2015 Elsevier Inc. All rights reserved.

[5] and multicentre studies [9] indicate marked heterogeneity in patient populations, service use and service delivery. Comparison of service use and patient characteristics between countries usually reports on combined SH populations (DSP, self-cutting and other SH) despite the known differences for these forms of SH [10]. The large cross-country comparisons of clinical [9] and community samples [1] identify only broad demographic and limited clinical correlates of SH, are restricted to limited time periods and lack the specific service information to provide context for the reported rates. National data sources for hospitaltreated SH are usually inadequate [4], are considered to produce serious underestimates of rates [11] and cannot provide detailed clinical and service provision data. Clinical Practice Guidelines in the UK [12] and Australasia [13], have recommended the organisation of clinical services for these patients, in order to provide adequate triage, medical and mental health management and after-care; including the availability of integrated physical and mental health care 24 h per day, since most presentations are outside office hours.

S. Hiles et al. / Journal of Psychosomatic Research 78 (2015) 356–362

The Oxford Monitoring System for Attempted Suicide in the UK and the Hunter Area Toxicology Service (HATS) in Newcastle, Australia have been examples of continuously active “sentinel” units, providing clinical services for known regional referral populations, whilst maintaining databases populated with prospectively collected standardised toxicological and psychiatric data representing all hospital-treated SH cases; and reporting on service delivery and clinical outcomes [14,15]. No direct comparisons of rates of presentation, clinical characteristics, or service use in the UK and Australian hospital-treated DSP populations have been reported previously. Understanding the similarities and differences of these clinical populations and the services provided would provide the information for planning for service needs, as well as providing a novel comparison of the delivery of psychiatric care in general hospitals for this important patient group. We aimed to compare rates of hospital-treated DSP, patient characteristics, psychosocial assessment and after care; from two sentinel units representing socio-economically similar but geographically distant English-speaking countries.

357

18]. In this paper the index DSP episode was the first contact recorded in the 10-year time period. The study was approved by the local Ethics Committee at both centres. Measures and assessment procedures In Oxford, the majority of patients received a psychosocial assessment by psychiatric clinicians. Patients not receiving an assessment were identified through scrutiny of emergency department and medical records, from which more limited data were extracted by research clerks. In Newcastle, all DSP patients are admitted by the Department of Clinical Toxicology and have a psychosocial assessment by psychiatric clinicians. Demographic, clinical and hospital management data on each episode were collected by clinicians using standardised forms at both units. Data from these assessments were entered into an electronic database by trained data entry staff blinded to any study hypotheses. Study variables

Materials and methods Setting The study is set in Oxford, UK and Newcastle, Australia. National socio-demographic variables including age, gender ratio and marital status are similar for England and Wales and Australia [16,17]. The cities of Oxford and Newcastle are similar in gender proportion (51% female in both) and are served by major universities (two in Oxford and one in Newcastle), which are important sources of employment in each city. Oxford is geographically smaller (Oxford city 46 km2 and Oxfordshire 1684 km2; Newcastle area 4042 km2 and the remainder of the Hunter region 31,393 km2). Oxford (20.7% of those aged 16–74 years) has a proportionately larger student population than Newcastle (13% of those aged over 15 years), fewer married couple households (Oxford 35%, Newcastle 51%) and a higher proportion aged 15–24 (Oxford 24%, Newcastle 13%). Oxford also has a greater proportion of people born outside the UK than England and Wales' average in 2001 (19% vs. 9%), a pattern reversed in Newcastle, which is lower than the Australian national average (10% vs. 22%). Study participants Participants were all patients (either treated in the emergency department alone or formally admitted to a general hospital ward bed) with hospital-treated DSP at the John Radcliffe Hospital in Oxford, UK and Calvary Mater Newcastle, Australia, between 1997 and 2006. Participants were also required to be a resident in the primary city referral areas of each unit (Oxford City and Newcastle, which included the local government areas of Newcastle, Port Stephens and Lake Macquarie). Both the Oxford unit and the Newcastle unit also treat patients from surrounding areas (the remainder of Oxfordshire and the balance of the Hunter Valley local government areas), although these patients are also potentially serviced by other hospitals. Patients from these surrounding areas were excluded from all analyses in order to provide an epidemiological study of the primary referral areas for each unit. We also excluded other forms of SH (unless concurrent with a DSP event), especially self-injury or self-cutting, because these populations have different characteristics than the much more numerous DSP population; and because the Newcastle unit does not offer a regional service for these patients who may present for care at other hospitals. Study design We used a cross sectional study design with data drawn from the Oxford Monitoring System for Attempted Suicide and the HATS Paracelsus databases, which have both been previously described in detail [15,

Initially all episodes of DSP were identified (and used for the calculation of age-standardised rates); however the other analyses were restricted to the first episode of DSP in the period. We extracted key variables, which were defined in a similar way at Oxford and Newcastle. Participant characteristics were: age, sex, marital status, employment status, previous psychiatric treatment, previous self-harm (hospital-treated or other types), current drug misuse and method of DSP (alone vs. DSP with SH). Service-related characteristics were: admission to a general hospital or clinical decision unit bed, time of presentation, professional completing the psychosocial assessment, discharge destination and after-care services. Toxicological characteristics were: alcohol co-ingestion with DSP and poisoning agent (major drug groups) used in the DSP episode. The exposure to poisoning agents was determined by the regular clinical assessment, which included patient history, empty packaging retrieved by ambulance officers or family and clinical symptoms (toxidromes) exhibited. Statistical analyses Analyses were completed using IBM SPSS Statistics 19. For the comparison of age-standardised rates, we set the criterion for statistical significance a priori at p b 0.01, which we believed was an appropriately conservative level for these four comparisons. For the subsequent comparisons of demographic, service and toxicological variables, we set the criterion for statistical significance a priori at p b 0.001, because of the large sample size and multiple comparisons. We firstly calculated sex-specific age-standardised rates. We calculated age-specific person rates for a typical year for age ranges of 10–19, 20–34 and over 35 in males and females separately. For each age and sex stratum: the numerator was the total number of unique individuals in a calendar year over the 10 year study period divided by 10 to represent the number of individuals presenting in a typical year; the denominator was the 2002 population of Oxford City and Newcastle. We then calculated the sex-specific, age-standardised rates for each site by weighting the age-specific rates by the European standard population sizes [19]. In addition to these rates for individuals per 100,000 of population in a typical year, we calculated the rate of events per 100,000 in the population in the same way. Standardised rate ratios with 99% Confidence Intervals (SRR: CI 99%) were calculated as a ratio of the age-standardised rates of Oxford over Newcastle. Participant, toxicological and service-related characteristics were compared between Oxford and Newcastle using t-test and Chi-square tests. Where there were substantial unknown data for a given variable we made a combined variable of “no/never/none or unknown” and we also ran sensitivity analyses restricted to the population with recorded data (not reported in detail).

358

S. Hiles et al. / Journal of Psychosomatic Research 78 (2015) 356–362

Results

Main findings

[9], comparisons of various UK [20,21] or UK and Dutch hospitals [22] or UK and Hong Kong hospitals [23], but none comparing hospital populations in UK and Australia. From western general hospital registration studies the person rates per 100,000 per year varied: e.g., Canada (male 357, female 534), Scotland (male 200–250, female 250–340), UK (male 150–200, female 250–425) and Norway (male 90, female 149) [2]. A recent study of three UK centres reported age-standardised person rates per 100,000: Oxford (male 310, female 412); Manchester (male 371, female 544); and Derby (male 373, female 510) [20]. The current study restricted to DSP, showed significantly higher patient rates in Oxford (male 236, female 394) than in Newcastle (male 95, female 163), which would have placed Newcastle much lower than other UK centres and closest to the rates for Norway [2]. A closer comparison, the WHO/EURO Multicentre Study on Parasuicide, reported age-standardised event (not person) rates for “overdose of drugs” per 100,000 population, with the highest rates in Helsinki (males 274, females 238) and Oxford (males 265, females 347) [3]. The current DSP study, Oxford (male 355, female 549) and Newcastle (male 113, female 220) would have placed Newcastle at a similar level to centres in Stockholm (male 127, female 221), Odense (male 151, female 197) and Szeged (male 130, female 190). The difference in rates in the current study cannot be attributed to differences in the age structure of the underlying population, since we used age-standardised rates for the comparisons. However, the reasons for the 2.5 fold difference in rates are not clear. Such a discrepancy between geographically and socioeconomically similar countries is not uncommon; e.g., within Nordic countries, rates of acute, hospitaltreated DSP differ, ranging from Finland at 150 per 100,000 to Norway at 255 per 100,000 [24]. The reported age standardised event rate using national hospital admitted datasets (standardised to the Australian population) for hospital admitted SH in Australia was only 118.5 and for major metropolitan centres only 107 per 100,000 [25], which confirms a serious underestimate of rates derived from admitted cases [11]. Service requirements that are planned or evaluated on the basis of these under-enumerated national datasets have an unacceptable margin for error. Sentinel units are a useful alternative, yielding more accurate estimates and more timely reports of rate changes. A network of sentinel units distributed throughout the respective countries would better inform developments in this important public health area. Without sentinel units, which can provide reliable and timely data on a longitudinal basis, we rely on specific studies to inform us about changes over time. A recent study of 32 randomly selected general hospitals in the UK compared the average number of self-harm attendances per 4-week period in 2001/2002 against 2010/2011 and reported an overall 24% increase in episodes and a 15% increase in the median number of episodes per hospital [26].

Rates of hospital-treated DSP Hospital-treated SH is common in western countries, although reported rates vary widely. We are aware of reviews reporting on international clinical populations [1], clinical and community populations [2], national UK sample [4], multicentre clinical European populations

Clinical characteristics of DSP patients There were several similarities between the patient populations in Oxford and Newcastle. The gender ratios were identical; alcohol and cannabis misuse was common; and previous SH was common. However, Oxford patients were younger, more often single and less

During the study period, 12,306 individuals with DSP presented to the Oxford (n = 7939) or Newcastle (n = 4367) hospitals, accounting for 20,337 events. The comparative analyses were restricted to 6534 individuals, Oxford (n = 3042) and Newcastle (n = 3492), which represented the primary city referral areas for each unit, as described in the Materials and methods section. Age-standardised rates and standardised rate ratios Table 1 shows gender specific age-standardised person rates and the standardised rate ratios (SRR). Oxford had significantly higher rates for both females (SRR 2.4) and males (SRR 2.5). There were no significant fluctuations in the number of individuals presenting each year for both Oxford and Newcastle (approximately 10% of the overall presentations occurred in each of the 10 years of study, ranging from 9.1% to 11.3%; Χ2(9) = 10.59, p = 0.31). The pattern for event rates was similar, with Oxford having higher rates for both females (SRR 2.5) and males (SRR 3.2). Participant characteristics and service use As shown in Table 2, there was no difference between the two centres for gender, whilst Oxford had a younger mean and median age. The differences in age structure can also be seen in Table 2, where Oxford has nearly a quarter of their patients under the age of 19 years and more than two thirds under 34 years, whilst Newcastle has only 14% under 19 years and only half under 34 years of age. Oxford had higher ratios of single to married (3.66 vs. 1.76); higher proportions of employed, student/child, retired; and lower proportions of unemployment benefits or other pension benefits. Newcastle had higher proportions with psychiatric treatment in the previous 12 months; whilst Oxford had a greater percentage with previous self-harm (with the same result after sensitivity analyses). Habitual drug misuse patterns (not ingestions in DSP episode) were statistically different but clinically similar, with alcohol and cannabis the most common. Cocaine misuse was more common in Oxford and benzodiazepine more common in Newcastle. As shown in Table 3, around 5% of DSP episodes in both centres were accompanied by other SH in the same episode of treatment. Admission to a general hospital bed occurred with similar frequency in both centres. Although statistically different, the pattern for time of presentation was similar (out of hours, Oxford 75% vs. Newcastle 71%). The proportion of patients receiving a psychosocial assessment was high (Oxford 80% vs. Newcastle 93%) and the ratio of medical officers compared with nurses as assessors varied (Oxford 0.6:1 vs. Newcastle 3.7:1). Most were discharged home in both centres (Oxford 80%, Newcastle 70%). The main differences in discharge destination were referral to a psychiatric hospital for possible admission (Oxford 8%, Newcastle 28%) and absconding from hospital (Oxford 11%, Newcastle 2%). Agents used in poisoning event As shown in Table 4, agents used for poisoning differed, although alcohol co-ingestion was common at both units. In Oxford, poisoning with non-opioid analgesics (paracetamol, or combinations; salicylate, or combinations; dextropropoxyphene; and non-steroidal anti-inflammatories) was more common than in Newcastle. In Newcastle, poisoning with psychotropic drugs (major tranquilisers, minor tranquilisers, antidepressants and mood stabilisers) was more common. Opioid (prescription or illicit) ingestion was not significantly different. Carbon monoxide poisoning via motor vehicle exhaust gas was uncommon in both centres. Inpatient mortality was higher in Newcastle than in Oxford.

Discussion

Table 1 Age standardised rates (ASR) and standardised rate ratios (SRR) for males and females with DSP in a typical year in Oxford and Newcastle Sex

Patients Events

Male Female Male Female

Oxford

Newcastle

Comparison

ASR/100,000

99% CI

ASR/100,000

99% CI

SRR

99% CI

235.7 394.2 355.3 549.0

183.9, 287.6 329.2, 459.2 291.2, 419.3 472.8, 625.2

95.4 162.7 112.7 220.3

75.8, 115.0 137.4, 188.1 91.5, 134.0 190.7, 249.9

2.5 2.4 3.2 2.5

1.7, 3.5 1.9, 3.2 2.3, 4.3 2.0, 3.0

ASR = age standardised rates. SRR = standardised rate ratio (with Newcastle as referent group).

S. Hiles et al. / Journal of Psychosomatic Research 78 (2015) 356–362

359

Table 2 Demographic and clinical characteristics at index DSP episode Oxford n = 3042

Newcastle n = 3492

Total n = 6534

Statistics Chi-square (df), p

n

%

n

%

n

Gender Female Male

1937 1105

63.7 36.3

2176 1316

62.3 37.7

4113 2421

1.3 (1), 0.256

Age 10–19 20–34 35+

725 1314 1003

23.8 43.2 33.0

518 1356 1618

14.8 38.8 46.3

1243 2670 2621

149.2 (2), b0.001

Marital status Single Married/de facto Divorced/separated Widowed Unknown

1913 522 423 62 122

62.9 17.2 13.9 2.0 4.0

1748 991 519 97 137

50.1 28.4 14.9 2.8 3.9

3661 1513 942 159 259

140.9 (4), b0.001

889 736 505 301 128 169 314

29.2 24.2 16.6 9.9 4.2 5.6 10.3

697 292 631 657 56 198 961

20.0 8.4 18.1 18.8 1.6 5.7 27.5

1586 1028 1136 958 184 367 1275

692.6 (6), b0.001

Previous psychiatric treatment Never (or unknown) Past 12 months N12 months

2037 409 596

67.0 13.4 19.6

1338 1572 582

38.3 45.0 16.7

3375 1981 1178

800.5 (2), b0.001

Previous self-harm Yes No (or unknown)

1310 1732

43.1 56.9

1096 2396

31.4 68.6

2406 4128

95.3 (1), b0.001

Current drug misuse Any drug misuse No drug misuse (or unknown)

1046 1996

34.4 65.6

1335 2157

38.2 61.8

2381 4153

10.4 (1), 0.001

794 281 39 111 85 26

75.9 26.9 3.7 10.6 8.1 2.5

971 381 301 139 13 21

72.7 28.5 22.5 10.4 1.0 1.6

1765 662 340 250 98 47

3.1 (1), 0.079 0.8 (1), 0.365 169.7 (1), b0.001 0.02 (1), 0.875 76.0 (1), b0.001 2.5 (1), 0.112

Employment status Employed (FT/PT) Student/child Unemployment benefits Other pension benefits Retired or other Home duties Unknown

Of misusers, drugs used Alcohol Cannabis Benzodiazepines Opioids Cocaine Other drugs Age

Age

a

a

Mean

SD

Range

Mean

SD

Range

T test (df), p

30.7

14.5

12–91

35.1

15.0

12–97

12.2 (6532), b0.001

Median

I-Q

Range

Median

I-Q

Range

Mann–Whitney U, p

26

19

12–91

33

21

12–97

4,262,033, b0.001

Unknown drug use was regarded as “none” for this chi square analysis.

likely to be on unemployment or pension benefits, probably reflecting a difference in the age structures of the underlying populations (younger in Oxford). This difference in age structure probably accounts for much of other differences between the two units, especially previous psychiatric treatment and therefore greater access to psychiatric medications for the self-poisoning event. Newcastle patients more commonly had recent psychiatric treatment, possibly reflecting higher levels of recognised mental illness and treatment engagement as expected in an older population structure. There may be other unmeasured differences in the patterns of recognition of and service provision for mental illness in the two countries. Psychosocial assessment and general hospital service provision Clinical Practice Guidelines in Australasia [13] and the UK [12] recommend 24-hour service provision of medical and psychosocial assessments and follow-up care. Most presentations were after-hours (70%) and almost all (95%) resulted in admission to the general hospital.

Psychosocial assessments were usually done by nursing staff in Oxford (64%) and medical officers in Newcastle (79%), reflecting a difference in service models. Psychosocial assessments were done at Oxford (80%) and Newcastle (93%), with the higher rate of missed assessments at Oxford probably due to a combination of non-referral of patients by emergency department staff, direct transfer to inpatient psychiatric hospital and reduced availability of psychiatric staff after hours or patient absconding. Absconding was more common in Oxford (11%) than in Newcastle (2%) and absconding patients may be at increased mortality risk [27]. These rates are generally lower than a study of 22 UK hospitals which reported a range of self-discharging from 8.4% to 38.6% [28], roughly equivalent to our concept of absconding. The reasons for the differential rate of absconding are not known but there are differences in allocation of Triage category that may be relevant. In Newcastle all DSP patients (regardless of severity of poisoning) are given a Triage category of 2 on the Australasian Triage Scale (ATS),

360

S. Hiles et al. / Journal of Psychosomatic Research 78 (2015) 356–362

which is the second highest category available. Any patient with ATS category 1 or 2 should be taken immediately into an appropriate assessment and treatment area [29]. This means that all DSP cases are brought into the emergency department promptly and the opportunity for absconding is reduced. In addition, in Newcastle, if a DSP patient was to abscond before assessment the usual practice would be to contact the police and ask for the patient to be returned to the ED or transported to the psychiatric hospital. In Oxford the Manchester Triage Scale (MTS) was used during the study period, which utilises 52 flow charts to score a five-point clinical acuity scale similar to the ATS [30]. Since the time of the current study, Oxford now uses an “escalation scale” based on the triage nurse assessment of “how worried am I?” This has shortened the time in the waiting room and has been associated with a reduction in absconding rates. In a comparison of the ATS and the MTS for a series of eight toxicology vignettes, the MTS gave a less acute Triage category for commonly and rarely encountered poisoning vignettes, than the ATS [31]. When a lower acuity category of triage is allocated and hence longer waiting times, there is a greater opportunity for absconding from the waiting room and we would suggest that this might account for some of the differential absconding rate at the two units. Other general hospitals show similar patient characteristics but report lower levels of admission to a hospital bed and provision of psychosocial assessment. An audit of 31 UK hospitals showed: overdose alone (79%), after-hours presentations (80%), admission (46%), psychosocial assessment (56%) and psychiatric hospital on discharge (10%) [4]. A recent follow-up to that study in 32 hospitals found marked variation: psychosocial assessment (22% to 88%); admission (22% to 85%); and psychiatric hospital on discharge (0% to 21%) [26]. In Australia, a study of all public and private hospital admissions for intentional self-injury showed: self-poisoning (82%), SH by a sharp object (13%), with no data for psychosocial assessment [25].

Table 3 Service characteristics at index DSP episode Oxford n = 3042

Newcastle n = 3492

n

n

Total n=

Statistics

6534

Method DSP only DSP + SH

%

%

Chi-square (df), p

2857 93.9 3320 95.1 6177 185 6.1 172 4.9 357

4.2 (1), 0.043

Admission to general or CDU bed Admitted 2862 94.1 3341 95.7 6203 Not admitted 180 5.9 150 4.3 330

8.9 (1), 0.003

Time of admission Office hours After hours

741 25.2 1007 28.8 1748 2197 74.8 2484 71.2 4681

10.6 (1), 0.001

Assessment Any assessment Not assessed

2445 80.4 3262 93.4 5707 597 19.6 230 6.6 827

250.0 (1), b0.001

If assessed, assessed by: Psychiatrist/SHO/GP trainee Nurse

879 36.0 2553 78.9 3432 1565 64.0

Those not assessed: Psychiatric inpatient Absconded Gaol or police custody Unknown

94 40.9 34 14.8 1 0.4 101 43.9

1074.2 (1), b0.001

149 361 8 309

157.8 (3), b0.001

2441 80.2 2426 69.5 4867 236 7.8 980 28.1 1216 345 11.3 69 2.0 414 18 0.6 4 0.1 22 2 0.1 13 0.4 15

628.2 (4), b0.001

After care services if discharged home with GP care Psychiatric outpatient 1208 49.4 1301 39.9 2509 Welfare services 97 4.0 152 4.7 249 Other agencies 67 2.7 0 0 67

51.4 (1), b0.001 1.6 (1), 0.205 90.5 (1), b0.001

Discharge destination Home Psychiatric inpatient Absconded Prison or police custody Deaths

55 9.2 327 54.8 7 1.2 208 34.8

682 21.1 2247

Agents ingested Common non-opioid analgesics were more common in Oxford, reflecting the younger population with access to over the counter analgesics. Major tranquilisers, minor tranquilisers, mood stabilisers and antidepressants were more common in Newcastle, reflecting the higher frequency of recent psychiatric treatment and availability of prescribed medications. The category “all other prescription drugs” was also more common in Newcastle reflecting an older population with increased access.

CDU, clinical decisions unit. SHO, senior house officer. Office hours (8 am–6 pm Mon–Fri). GP, general practitioner.

Table 4 Toxicological characteristics at index DSP episode Oxford

Alcohol at time of DSP Yes No (or unknown)

Total

Statistics

n

%

n

Newcastle %

n

Chi-square (df), p

725 2317

23.8 76.2

1124 2368

32.2 67.8

1849 4685

55.9 (1), b0.001

149 503 690 44 196 273 1293 76 308 14 459 90

4.9 16.5 22.7 1.4 6.4 13.0 52.9 3.8 15.5 0.5 15.1 3.0

399 1152 943 232 246 69 981 34 131 77 699 37

11.4 33.0 27.0 6.6 7.0 2.0 28.1 1.0 3.8 2.2 20.0 1.1

548 1655 1633 276 442 342 2274 110 439 91 1158 127

90.2 (1), b0.001 232.7 (1) b0.001 16.2 (1), b0.001 108.6 (1), b0.001 0.9 (1), 0.334 276.8 (1), b0.001 374.5 (1), b0.001 52.3 (1), b0.001 237.0 (1), b0.001 36.0 (1), b0.001 27.1 (1), b0.001 30.8 (1), b0.001

a

Agent used Major tranquilisers Minor tranquilisers Antidepressants Mood stabilisers Opioids Salicylate/combinations Paracetamol/combinations Dextropropoxyphene NSAIDs Gas All other prescriptions Drug not known

Dextropropoxyphene: Coproxamol®/Digesic®. NSAIDs non-steroidal anti-inflammatory drugs. a Note: Chi square was calculated using valid cases for agents ingested, missing values in Oxford due to non-recording prior to 2000: salicylate/combinations n = 938 (14.4%), paracetamol/combinations n = 598 (9.2%), Coproxamol®/Digesic® n = 1054 (16.1%) and NSAIDs n = 1054 (16.1%).

S. Hiles et al. / Journal of Psychosomatic Research 78 (2015) 356–362

The top five ingestion groups in Oxford: paracetamol, antidepressants, minor tranquilisers, “other” prescription drugs and NSAIDs, and in Newcastle: minor tranquilisers, paracetamol, antidepressants, “other” prescription drugs and major tranquilisers. The difference in mortality between the centres most likely represents the differences in proportion of paracetamol poisoning which has a very low mortality with current management; otherwise the toxicological management requirements were comparable and similar to other reports from the UK [20] and Australia [25]. Discharge destination and follow-up care Discharge to home was more common in Oxford than in Newcastle, whilst more were referred for psychiatric inpatient treatment in Newcastle than in Oxford. Oxford and Newcastle both have access to community-based aftercare and so we would not expect this to explain the marked differences in referral for psychiatric hospitalisation. For those discharged to home, nearly all were referred to the GP and many referred to psychiatric outpatient treatment, Oxford 49% and Newcastle 40%. The Newcastle population had a greater proportion with recent psychiatric treatment, which probably influenced discharge to psychiatric inpatient care. There may be other important differences in patient characteristics or service characteristics (e.g., nurse vs. doctor assessments) which may have contributed to the different rates of referral to psychiatric hospital on discharge. We plan to explore possible reasons for this differential in discharge to a psychiatric hospital in a later paper. We have previously reported for Newcastle that discharge to psychiatric hospital was associated with older age, homelessness, unemployment, previous selfharm, previous psychiatric inpatient treatment, suicidal ideation or plan, mood or psychotic disorders and lower clinician experience; with lower risk for being married/de facto, and having an after-hours presentation [32]. Strengths and limitations of the study We used a cross-sectional design covering a ten-year period to investigate rates and characteristics of hospital treated DSP. The two units are responsible for all DSP cases in their primary referral populations. Both “sentinel” units are well established for clinical service delivery and observational research and both units routinely collect prospective data on all study characteristics. Data entry was undertaken by trained research staff blind to any study hypotheses. The calculation of age standardised rates allows for direct comparison without the influence of different age structures in the underlying populations. Gender rates were calculated separately to reflect the expected differences in rates. There are a number of study limitations and caution should be used when interpreting these results. Almost all of the measures used relied on various clinicians making their usual judgements assisted by standardised clinical record sheets for data recording. We were restricted to clinical data that was common to both units, although the core datasets were similar. There was a different practice for default coding for clinical variables, with Oxford using “unknown” and HATS using “no”, and we therefore combined “unknown” and “no” responses together for analyses and examined results with sensitivity analyses. There was a differential rate of missing data on psychosocial assessments at the two units, which may introduce bias. The results from this study may not be generalisable to other DSP populations in the UK or Australia, although results from Oxford are broadly comparable with other UK centres [20]. Conclusions Services can be organised to provide hospital admission and psychosocial assessment. Previous drug misuse, DSP and psychiatric treatment are common, so access to clinical records and appropriate clinicians

361

would be useful. Hospitals need emergency, toxicology, drug and alcohol and psychiatric services; organised on a 24 hour basis. Sentinel services providing clinical services and systematic monitoring of routine care can provide accurate estimates of the rates of service requirements, clinical care provision and after care needs; whilst providing a standard of care consistent with current Clinical Practice Guidelines. These sentinel units can provide timely reports of changes in rates or clinical characteristics for the purposes of planning and evaluation of national service provision and for comparisons between countries.

Conflict of interest We have no conflicts of interest to declare.

Acknowledgments Ms. Hiles' position as Research Manager of the Suicide Prevention Programme at the Calvary Mater Newcastle is funded by the Burdekin Suicide Prevention programme and administered by the Hunter New England Mental Health Services.

References [1] Borges G, Nock MK, Haro Abad JM, Hwang I, Sampson NA, Alonso J, et al. Twelvemonth prevalence of and risk factors for suicide attempts in the World Health Organization World Mental Health Surveys. J Clin Psychiatry 2010;71:1617–28. [2] Welch SS. A review of the literature on the epidemiology of parasuicide in the general population. Psychiatr Serv 2001;52:368–75. [3] Michel K, Ballinari P, Bille-Brahe U, Bjerke T, Crepet P, De LD, et al. Methods used for parasuicide: results of the WHO/EURO multicentre study on parasuicide. Soc Psychiatry Psychiatr Epidemiol 2000;35:156–63. [4] Gunnell D, Bennewith O, Peters TJ, House A, Hawton K. The epidemiology and management of self-harm amongst adults in England. J Public Health (Oxf) 2005; 27:67–73. [5] Owens D, Horrocks J, House A. Fatal and non-fatal repetition of self-harm. Systematic review. Br J Psychiatry 2002;181:193–9. [6] Reith DM, Whyte IM, Carter GL, McPherson M, Carter NT. Risk factors for suicide and other deaths following hospital treated self-poisoning in Australia. ANZ J 2004;38: 520–5. [7] Bergen H, Hawton K, Waters K, Ness J, Cooper J, Steeg S, et al. Premature death after self-harm: a multicentre cohort study. Lancet 2012;380:1568–74. [8] Carter G, Reith DM, Whyte IM, McPherson M. Non-suicidal deaths following hospital-treated self-poisoning. Aust N Z J Psychiatry 2005;39:101–7. [9] Schmidtke A, Bille-Brahe U, DeLeo D, Kerkhof A. Attempted suicide in Europe: rates, trends and sociodemographic characteristics of suicide attempters during the period 1989–1992. Results of the WHO/EURO multicentre study on parasuicide. Acta Psychiatr Scand 1996;93. [10] Hawton K, Harriss L, Simkin S, Bale E, Bond A. Self-cutting: patient characteristics compared with self-poisoners. Suicide Life Threat Behav 2004;34:199–208. [11] Steenkamp M, Harrison J. Suicide and hospitalised self-harm in Australia. Injury Research and Statistics Series (AIHW cat no.INJ 30), 1-123. Adelaide: AIHW; 2001[1-10-1444]. [12] National Collaborating Centre for Mental Health. Self-harm: the short-term physical and psychological management and secondary prevention of self-harm in primary and secondary care. National clinical practice guideline (DRAFT) number 16, 1-199. The British Psychological Society; 2004. [13] Australian and New Zealand College of Psychiatrists Clinical Practice Guidelines Team for Deliberate Self-harm. Australian and New Zealand clinical practice guidelines for the management of adult deliberate self-harm. Aust N Z J Psychiatry 2004; 38:868–84. [14] Hawton K, Fagg J, Simkin S, Mills J. The epidemiology of attempted suicide in the Oxford area, England (1989–1992). Crisis 1994;15:123–35. [15] Whyte IM, Dawson AH, Buckley NA, Carter GL, Levey CM. A model for the management of self-poisoning. Med J Aust 1997;167:142–6. [16] Australian Bureau of Statistics. Census QuickStats: Newcastle (NSW) (statistical district). Internet. 9–3–2006. 17–9–2012; 2001. [17] Office for National Statistics. Census 2001: local authority profiles and population pyramids. Internet. 13–2–2003. 17–9–2012; 2001. [18] Hawton K, Harriss L, Hall S, Simkin S, Bale E, Bond A. Deliberate self-harm in Oxford, 1990–2000: a time of change in patient characteristics. Psychol Med 2003;33: 987–95. [19] Ahmad OB, Boschi-Pinto C, Lopez AD, Murray CJL, Lozano R, Inoue M. Age standardization of rates: a new WHO standard. GPE discussion paper series. World Health Organization; 2001. [20] Bergen H, Hawton K, Waters K, Cooper J, Kapur N. Epidemiology and trends in nonfatal self-harm in three centres in England: 2000–2007. Br J Psychiatry 2010;197: 493–8.

362

S. Hiles et al. / Journal of Psychosomatic Research 78 (2015) 356–362

[21] Platt S, Hawton K, Kreitman N, Fagg J, Foster J. Recent clinical and epidemiological trends in parasuicide in Edinburgh and Oxford: a tale of two cities. Psychol Med 1988;18:405–18. [22] Grootenhuis M, Hawton K, van Rooijen L, Fagg J. Attempted suicide in Oxford and Utrecht. Br J Psychiatry 1994;165:73–8. [23] Yip PSF, Hawton K, Liu K, Liu K, Ng PWL, Kam P, et al. A study of deliberate self-harm and its repetition among patients presenting to an emergency department. Crisis 2011;32:217–24. [24] Andrew E, Irestedt B, Hurri T, Jacbosen P, Gudjonsdottir G. Mortality and morbidity of poisonings in the Nordic countries in 2002. Clin Toxicol 2008;46:310–3. [25] McKenna K, Harrison JE. Hospital separations due to injury and poisoning. Australia 2008–09. Injury research and statistics series no. 65[Cat. no. INJCAT 141], 1-141. Canberra: AIHW; 2012. [26] Cooper J, Steeg S, Bennewith O, Lowe M, Gunnell D, House A, et al. Are hospital services for self-harm getting better? An observational study examining management, service provision and temporal trends in England. BMJ Open 2013;3.

[27] Hickey L, Hawton K, Fagg J, Weitzel H. Deliberate self-harm patients who leave the accident and emergency department without a psychiatric assessment: a neglected population at risk of suicide. J Psychosom Res 2001;50:87–93. [28] Bennewith O, Peters TJ, Hawton K, House A, Gunnell D. Factors associated with the non-assessment of self-harm patients attending an accident and emergency department: results of a national study. J Affect Disord 2005;89:91–7. [29] Australasian College of Emergency Medicine. Guidelines on the implementation of the Australasian Triage Scale in emergency departmentIn: Australasian College of Emergency Medicine, editor. ; 2013. p. G24. [30] Manchester Triage Group. Emergency triage. 3rd ed. Oxford, UK: BMJ Books; 2013. [31] Jayaweera D, Mitter S, Grouse A, Strachan L, Murphy M, Douglass D, et al. A comparison of emergency triage scales in triaging poisoned patients. Australas Emerg Nurs J 2014;17:184–9. [32] Carter GL, Safranko I, Lewin TJ, Whyte IM, Bryant JL. Psychiatric hospitalization after deliberate self-poisoning. Suicide Life Threat Behav 2006;36:213–22.