Emergency Medicine Australasia (2017)
doi: 10.1111/1742-6723.12755
ORIGINAL RESEARCH
Establishing a dedicated toxicology unit reduces length of stay of poisoned patients and saves hospital bed days Katherine Z ISOARDI,1,2 Melanie C ARMITAGE,1,2 Keith HARRIS1 and Colin B PAGE1,2,3 1
Clinical Toxicology Unit, Princess Alexandra Hospital, Brisbane, Queensland, Australia, 2School of Medicine, The University of Queensland, Brisbane, Queensland, Australia, and 3Clinical Toxicology Research Group, The University of Newcastle, Newcastle, New South Wales, Australia
Abstract Objective: This study evaluates the effect on the average length of stay (LOS), relative stay index (RSI), bed days and costs saved following the establishment of a dedicated clinical toxicology unit in an Australian tertiary referral hospital. Methods: This retrospective descriptive study uses Health Roundtable and other state and federal data to compare the average LOS, RSI, estimated bed days and costs saved by patients admitted with a diagnosisrelated group (DRG) of X62 (Poisoning/Toxic Effects of Drugs and Other Substances), over the 4 year period 2012–2015. This period corresponds to before and after the introduction of the clinical toxicology unit in February 2014 at the Princess Alexandra Hospital, a tertiary referral teaching hospital in Brisbane, Queensland, Australia. Results: There was a reduction in the average LOS and RSI from 2.1 days and 122% in 2012 to 0.9 days and 52% in 2015, respectively. This reduction correlates with a reduction in 1350 bed days and a saving of $2.25 million over the 2 year period 2014–2015 since the clinical toxicology unit was established. Conclusion: The reduction in average LOS is similar to results
previously published by two Australian toxicology units over 15 years ago. Despite changes in healthcare delivery since this time, these results continue to support the efficiency and associated cost saving of a dedicated toxicology unit in managing poisoned patients. Key words: cost savings, inpatients, length of stay.
Introduction Intentional self-poisoning is a frequent problem worldwide, be it deliberate self-harm or recreational drug use. In Australia the estimated rate of hospitalisation secondary to self-poisoning is 122.5 per 100 000 population per annum1 and this is thought to account for 1–5% of all hospital admissions.2 Importantly, this figure does not include presentations solely managed within the ED that are not admitted. In Australia, most patients presenting to the ED with intentional self-poisoning who need ongoing care beyond the ED and the ED short-stay unit (SSU) are admitted under general or subspeciality medical units. However, a small number of poisoned patients are managed by dedicated toxicology services. In 1986 the Hunter Area Toxicology Service (HATS) was the
Correspondence: Dr Colin B Page, Clinical Toxicology Unit, Princess Alexandra Hospital, Ipswich Road, Woolloongabba, QLD 4102, Australia. Email:
[email protected] Katherine Z Isoardi, BMed, FACEM, GradDipClinTox, Emergency Physician, Clinical Toxicology Fellow; Melanie C Armitage, MBChB, FACEM, Emergency Physician; Keith Harris, MBChB, Clinical Toxicology Registrar; Colin B Page, MBChB, FACEM, MMedSci (Clin Epid), Clinical Toxicologist, Emergency Physician.
Key findings • The development of a clinical toxicology unit has reduced the average length of stay of toxicology presentations by 57%. • Over the two years since its introduction there was a saving of 1350 bed days at an estimated cost of $2.25 million. first toxicology unit established in Australia and since this time toxicology units have also been established in other metropolitan hospitals in state capital cities. The HATS service published its model of care in 1997 and demonstrated a reduction in average length of stay (LOS) of 0.53–1.22 days (24–47% decrease) in comparison to all Australian hospitals.3 Overall this saved 518 bed days annually, with a cost saving of $468 000. The reduction was seen more in complicated patients who had a longer average LOS. The only other Australian toxicology unit to publish its efficiency data was the Austin and Repatriation Medical Centre toxicology service in 2001.4 This unit was able to reduce its average LOS for noncomplicated patients from 1.97 to 1.4 days (29% decrease) and for complicated patients from 5.59 to 1.92 days (66% decrease). There was no cost saving calculated. Units overseas have also shown similar reductions in the average LOS.5,6 A dedicated clinical toxicology unit (CTU) was formally established with admission rights to the Princess Alexandra Hospital (PAH) being
Accepted 10 January 2017 © 2017 Australasian College for Emergency Medicine and Australasian Society for Emergency Medicine
2
KZ ISOARDI ET AL.
granted in February 2014 after a transitional period during 2013. The aim of the unit is to coordinate and improve the quality of care of the toxicological patient by providing 24/7 consultation and inpatient management for patients who have been poisoned or envenomed. Prior to its development, multiple units within the hospital managed these patients. In 2012 this group of patients was identified by the PAH executive as a patient group that had a comparatively longer average LOS and relative stay index (RSI) when compared to other peer hospitals within Australia.7 Since these two seminal Australian publications,3,4 there has been considerable change within the Australian public hospital healthcare system, with increasing utilisation and budgetary pressures driving an emphasis on efficiency and decreasing inpatient average LOS. Over the past 10 years, the average LOS of all patients at the PAH has fallen steadily from 4.1 days in 2005 to 3.2 days (22% reduction) in 2015.8 Following the establishment of a dedicated CTU, the aim of this study was to determine if the reduction in the average LOS of poisoned patients demonstrated in previous studies is still evident in the current healthcare system. In addition, we aimed to calculate a cost saving on any reduction in average LOS achieved.
Methods Study setting and population The PAH is one of four tertiary referral teaching hospitals in Queensland and provides healthcare for all major adult specialities except obstetrics and gynaecology. It is located in Brisbane within the Metro South Hospital and Health Service. Prior to 2013, approximately 70% of toxicological presentations were managed within the ED with or without a brief admission through the ED SSU. Those patients whose management required a longer stay were admitted to the wards – usually under a general medical team but occasionally under a subspeciality medical unit. This comprised
approximately 25% of toxicological presentations. The remaining 5% of toxicological presentations required an ICU admission from the ED, following which their care was usually transferred to a general medical unit for ongoing management. In early 2013, one of the authors (CBP) took over the management of the ED SSU toxicological patients. Over the course of 2013 he managed an increasing proportion of the toxicological patients who previously would have been admitted medically, within the SSU environment. ICU admissions, once discharged to the wards remained under medical units. After the formal introduction of the CTU in February 2014, patients requiring admission both to the SSU within the ED or an inpatient ward, including the ICU, remained the responsibility of the CTU. Patients not requiring admission and managed entirely within the ED were managed in conjunction with the ED medical staff.
Study design and data analysis This is a retrospective descriptive study of Health Roundtable (HRT) data with additional information from federal and state departments. The HRT is a non-profit membership organisation of health services across Australia and New Zealand. There are 90 health service organisations encompassing 160 facilities including all major teaching hospitals across Australia and New Zealand who are currently members and provide data for comparative analysis. Its aims are threefold: ‘to provide opportunities for health executives to learn how to achieve best practice in their organisations; to collect, analyse and publish information comparing organisations and identify ways to improve operational practices; and to promote interstate and international collaboration and networking amongst health organisation executives’. HRT data was utilised to compare the median LOS, average LOS, RSI and estimated bed days saved, by patients admitted with a diagnosis-related group (DRG) of X62 (Poisoning/Toxic Effects of
Drugs and Other Substances) over a 4 year period from 2012 to 2015,7 for PAH compared to all hospitals contributing to the HRT. Inpatient briefings for this DRG were released for the financial years of 2012–2014 as well as for the calendar years of 2014 and 2015.7 These reports allowed analysis, using descriptive statistics, of data from before, during and after the introduction of the CTU. The first complete year of the CTU solely managing the patients is reflected first in the December 2014 calendar year briefing. The RSI is used in these reports for comparison among institutions and is calculated from the average LOS adjusted for age, case complexity, admission source and discharge destination and is therefore a more accurate measure for comparison of LOS between hospitals than the average LOS. Ingested drug classes for the same time period were also extracted from this data. Additional data on average LOS at PAH over the same time period was accessed via a Queensland Health website8 and the RSI at PAH was accessed from HRT. Bed day costs for X62 were extracted from the Independent Hospital Pricing Authority (IHPA; https://www.ihpa. gov.au) Australian Public Hospital Cost Report 2013–2014 published in February 2016.9 The IHPA is an independent government agency established as part of the National Health Reform Agreement in 2011. Its primary function is to implement activity-based funding for public hospital services. All analysis and graphics were performed using Graphpad Prism (version 6.0h; GraphPad Software, San Diego, CA, USA).
Ethics approval HRT, Queensland Health and IHPA data are freely available online and provided in a de-identified format. In addition, permission was granted from HRT to use their data in this study.
Results The number of patients (Fig. 1) admitted to the PAH with the DRG
© 2017 Australasian College for Emergency Medicine and Australasian Society for Emergency Medicine
3
TOXICOLOGY UNITS AND REDUCED LENGTH OF STAY
1000
Admissions
900
800
700
600
500 2 -1
n Ju
2 -1
ec
D
3 -1
n Ju
3 -1
ec
D
4 -1
n Ju
4 -1
D
ec
5 -1
n Ju
5 -1
ec
D
Year to date
Figure 1. Inpatient episodes of diagnosis-related group X62. The arrow corresponds to the introduction of the clinical toxicology unit.
(b) 150
2.0
125 100
1.5
%
ALOS (Days)
(a) 2.5
75
decrease from 1.9 to 0.9 days (53% decrease). Over the same time period, the ‘all HRT hospital average’ for average LOS for the DRG X62 (Fig. 2a) remained unchanged at 1.8 days and the RSI (Fig. 2b) decreased from 101 to 88%. The reduction in RSI for the DRG X62 at the PAH when compared with the RSI for the DRG X62 at all HRT hospitals over the 2 year period 2014–2015 that the CTU has been in operation, equates to 1350 bed days saved (720 and 630 bed days in 2014 and 2015, respectively). Based on a bed day cost of $1667,9 this amounted to a potential saving of $2.25 million over the 2014–2015 period. This was the highest of the top 10 actual savings at PAH in comparison to all HRT averages. Over the same time period, the proportion of long-stay (complicated) patients, defined by HRT as a LOS beyond the outlier cut point (2.2 days for lower acuity X62B and 11.1 days for the higher acuity X62A), reduced from 15 to 2.4% following the introduction of the CTU (Fig. 4). Similarly the associated bed days occupied by long-stay patients fell from 56 to 18% (Fig. 4).
1.0
50 0.5
0.0
12 -12 -13 -13 -14 -14 -15 -15 c c c nn n n ec Ju De Ju De Ju De Ju D
25 0 12 -12 -13 -13 -14 -14 -15 -15 c c c n n n nec Ju De Ju De Ju De Ju D
Figure 2. (a) Diagnosis-related group (DRG) X62 average length of stay. (b) DRG X62 relative stay index. The arrow corresponds to the introduction of the clinical toxicology unit (CTU). ( ), X62 all Health Roundtable average; ( ), X62 CTU.
X62 has steadily increased over the 2012–2015 period from 602 episodes in 2012 to 985 episodes in 2015 (16% annual increase over 4 years). Ingestions of benzodiazepines, paracetamol, unspecified antipsychotics and unspecified antidepressants accounted for the majority of X62 admissions. The relative proportions of these ingestions did not change significantly over the corresponding time period. Over the 2012–2015 time period, the average LOS (Fig. 2a) for DRG
X62 episodes decreased from 2.1 to 0.9 days (57% decrease), while the RSI (Fig. 2b) decreased from 122 to 52%. For comparative purposes, over the same time period the average LOS for all patients at PAH decreased from 3.8 to 3.2 days (16% decrease) and the RSI decreased from 100 to 91%. The median LOS (Fig. 3) for the DRG X62 episodes was relatively constant over the period; however, in association with the fall in average LOS, the 75th percentile also showed a similar
Discussion The development of a CTU has seen a 57% reduction in the average LOS of toxicology patients since its introduction and the saving of 1350 bed days at a cost of $2.25 million over the 2 year period 2014–2015 based on the reduction in RSI. During the same period, the average LOS of all patients at the PAH fell by 16% (approximately one-third of the fall in average LOS of toxicology patients), reflecting the significant impact of the introduction of the CTU on efficiency savings compared to the average hospital improvements. The number of toxicology admissions increased by 16% per year over the time period and this is much higher than other hospitals within HRT whose toxicology admissions increased by approximately 5% per year. The definitive reasons for this are not fully known. It is possible an
© 2017 Australasian College for Emergency Medicine and Australasian Society for Emergency Medicine
4
KZ ISOARDI ET AL.
2.0
LOS (Days)
1.5
1.0
0.5
0.0 2 -1
2 -1
n Ju
ec
D
3 -1
3 -1
n Ju
ec
D
4 -1
4 -1
n Ju
ec
D Year to date
5 -1
5 -1
n Ju
ec
D
Figure 3. Diagnosis-related group X62 median (25–75%) length of stay. The arrow corresponds to the introduction of the clinical toxicology unit. 70 60 50
%
40 30 20 10 0 2 -1
2 -1
n Ju
D
ec
3 -1
3 -1
n Ju
ec
D
4 -1
n Ju
4 -1
D
ec
5 -1
5 -1
n Ju
ec
D
Year to date
Figure 4. Diagnosis-related group X62 long-stay patients. The arrow corresponds to the commencement of the clinical toxicology unit. ( ), % Bed days; ( ), % longstay patients.
‘if you build it they will come’ phenomenon might account for the increase in toxicological admissions to the PAH, with the ambulance service preferentially taking toxicology patients in the region to our hospital, knowing that there is a dedicated toxicological service. The Austin toxicology service documented a similar increase (15%) with the introduction of their service.4 An improvement in diagnostic coding (CTU staff performing coding rather than ED and ward medical staff)
may also be contributing. Of note, the increase in admissions makes the efficiency gains more significant. It could be postulated that the reduction in average LOS seen in this study following the establishment of the CTU might reflect an increase in the number of admissions of relatively shorter average LOS (i.e. mild poisonings), which previously were managed without an admission. However, if this was the case, then the median LOS should correspondingly fall. The reduction in the
average LOS corresponds better with the fall in percentage of long-stay patients and their bed occupancy, which is supported by the median LOS remaining relatively unchanged over the time period. Data showing a decrease in 75th percentile with an unchanged median LOS also supports this premise. The overall reduction in the average LOS seen in this study is of similar magnitude to the findings of the two previous Australian studies.3,4 Likewise, this reduction was seen more in the complicated patient subgroup that had correspondingly longer average LOS. On the whole, the majority of toxicology patients have a short LOS, with the median LOS of 12 h in our study. This is consistent with clinical experience that the majority of toxicological presentations require only supportive care for a relatively brief duration. This median LOS is slightly less than the median LOS of 16 h published by the HATS in 1997.3 There are a number of possible reasons for this improvement in average LOS. The toxicology unit provides a 24/7 phone consultative service with 7 days a week consultant/senior registrar ward rounds, providing evidencebased specialist toxicological management. Almost all non-intensive care patients are managed within the ED SSU environment, where repeated/frequent senior toxicological review throughout the course of the admission streamlines care. This allowed earlier definitive decisions on required duration of treatment or observation, and facilitates final disposition or discharge endpoints. Earlier mental health assessment is achieved by close communication and cooperation with the ED mental health service. This allows patients to be evaluated in parallel with toxicological management rather than in sequence, with the mental health team not waiting for a patient to complete medical management or to be ‘medically cleared’ prior to review. The major limitations of the study are related to the methodology. The data was retrospective, and had subsequently been handled and interpreted by an external agency. Also, inaccurate coding could result in relevant patient episodes being missed,
© 2017 Australasian College for Emergency Medicine and Australasian Society for Emergency Medicine
5
TOXICOLOGY UNITS AND REDUCED LENGTH OF STAY
which may impact on the quality of LOS data and subsequent interpretations. The potential bias relating to these issues is not quantifiable. Furthermore the X62 DRG is not a perfect surrogate for toxicology admissions and includes diagnoses, such as alcohol intoxication, which are not exclusively managed by the CTU. However, in 2014 the CTU managed 90% of X62 episodes and in 2015 this proportion increased to 95%,7 suggesting the impact of patients managed by other units on the average LOS is small. In summary, a dedicated CTU at the PAH has seen the average LOS of the poisoned patient fall from 2.1 to 0.9 days equating with 1350 bed days and $2.25 million saved in the 2 year period since its establishment. These results add to the evidence supporting the efficiency of a dedicated toxicology unit in managing poisoned patients. Hospitals who manage poisoned patients with relatively long average LOS should consider the formation of a toxicology unit to improve governance and patient flow.
Acknowledgements The HRT kindly provided permission for the use of their data in this
study. CBP is supported by an Emergency Medicine Foundation Research Fellowship.
5.
Competing interests
6.
CBP is a section editor for Emergency Medicine Australasia.
References 1.
2. 3.
4.
Australian Institute of Health and Welfare; Harrison JE, Henley G. Suicide and Hospitalised Self-Harm in Australia: Trends and Analysis, 2014. Injury Research and Statistics Series No. 93. Cat. No. INJCAT 169. Canberra, Australia. [Cited May 2016.] Available from URL: http://www.aihw.gov.au/WorkArea/ DownloadAsset.aspx?id=60129549727 Pond SM. Prescription for poisoning. Med. J. Aust. 1995; 162: 174–5. Whyte I, Dawson A, Buckley N et al. A model for the management of self poisoning. Med. J. Aust. 1997; 167: 142–6. Lee V, Kerr JF, Braitberg G et al. Impact of a toxicology service on a metropolitan teaching hospital. Emerg. Med. Australas. 2001; 13: 37–42.
7.
8.
9.
Chung A, Tsui SH, Tong HK. The impact of an emergency department toxicology team in the management of acute intoxication. Hong Kong J. Emerg. Med. 2007; 14: 134–43. Curry SC, Brooks DE, Skolnik AB et al. Effect of a medical toxicology admitting service on length of stay, cost and mortality among inpatients discharged with poisoning-related diagnoses. J. Med. Toxicol. 2015; 11: 65–72. The Health Roundtable. Inpatient Briefing for X62: Poisoning/Toxic Effects of Drugs. [Cited May 2016.] Available from URL: http:// healthroundtable.org Queensland Hospital Admitted Patient Data Collection, Department of Health. Statistical Reporting and Coordination, Health Statistics Unit, Department of Health, 13 January 2016. [Cited June 2016.] Available from URL: http://qheps.health.qld. gov.au/hsu/Infobank/Activity_morb/ hosp_summ_data.xls Independent Hospital Pricing Authority. Australian Public Hospital Cost Report 2013–2014, February 2016. [Cited July 2016.] Available from URL: https://www. ihpa.gov.au/publications/australianpublic-hospitals-cost-report-20132014-round-18
© 2017 Australasian College for Emergency Medicine and Australasian Society for Emergency Medicine
