WEST AFRICAN JOURNAL OF

ORIGINAL ARTICLE

Improving Emergency Attendance and Mortality – The Case for Unit Separation

Améliorer la fréquentation des urgences et de la mortalité - le cas pour l’unité de séparation

E. Asumanu*, R. Richardson †

ABSTRACT RÉSUMÉ BACKGROUND: Emergency attendance and mortality which CONTEXTE: L’Assistance de secours et mortalité qui sont les are reliable indicators of quality of care, have been of concern indicateurs fiables de la qualite de soin, ont été une to many health institutions. Different models are being proposed préoccupation de beaucoup des institutions de la santé. to improve emergency outcomes in different parts of the world. Different modeles sont proposés d’améliorer des résultats dans différentes régions du monde. Un modèle pour séparer une A model to separate a single emergency Unit into multiple unité simple de secours dans les unités multiples de secours a emergency units has been tried in . été essayé au Ghana OBJECTIVE: The objective of the study was to find the effect of OBJECTIF: L’objectif de l’étude est de trouver l’effet de la the Unit Separation Model (USM) on the quality of emergency Modele séparation d’Unité (MSU) sur la qualité de la livraison care delivery in a developing country. de soin de secours dans un pays en voie de developpement METHODS: The study compared the outcomes (attendance and MÉTHODE: L’etude a comparé les resultants(assistance et mortality) in a Single Emergency Model (SEM) with a USM, mortalité ) dans une Modele d’Urgence Unite(MUU) et MSU, over a two-year period. Two groups of patients were studied - sur une période de 2 ans. Deux groupes de patients ont été the SEM phase comprising 809 patients, and the USM phase étudiés dans deux phases - la phase de MUU comportant 809 comprising 3,505 patients. Data on patients’ attendance and patients, et la phase de MSU comportant 3505 patients. Des mortality in the two groups were analysed. données sur l’ assistance des patients et le motality dans les deux groupes ont été analysés. RESULTS: Attendance increased four fold in the USM period RÉSULTATS: L’ assistance a augmenté quatre replient pendant compared to the SEM period. This also reflected in increases la période d’étude. Ceci s’est également reflété dans les in individual unit attendance. There was almost a three fold augmentations d’ assistance d’unité individuel. Il y avait increase in medical and surgical emergencies, over ten fold presque une augmentation de trois fois des urgences médicales increase in gynaecological emergencies and over twenty fold et chirurgicales, plus d’augmentation de dix fois des urgences increase in paediatric emergencies. There was a statistically gynécologiques et plus d’augmentation de vingt fois des significant reduction in emergency mortality in the USM period urgences pédiatriques. Il y avait une réduction statistiquement compared to the SEM period. Reductions in mortality were from significative de la mortalité de secours au cours de la période 27.8% to 7.9% for , 46.3% to 23.2% for Medicine, d’étude dans l’hôpital. Les réductions de la mortalité étaient 17.5% to 0.8% for , and 50.0% to 8.4% for de 27,8% à 7,9%, pour la chirurgie, 46,3% à 23,2% pour la Paediatrics (p-value <0.001 for each unit); and overall from médecine, 17,5% à 0,8% pour la gynécologie, et 50,0% à 8,4% pour Paediatrics(p-value<0.001 pour chaque unité); et 36.7% to 10.9%. globalement de 36,7% à 10,9%. CONCLUSION: It is concluded that Unit Separation Model with CONCLUSION: On le conclut que la modèle de séparation Specialty based emergency care has a positive impact of d’unité avec soin basé par spécialité de secours, a un impact increasing emergency admissions and reducing mortality, positif des admissions croissantes de secours et réduction de without compromising quality. This also means that avoidable mortalité, sans qualité compromettante. Ceci signifie également mortalities occur in the Single Emergency Model system. que les mortalités évitables se produisent dans le Système WAJM 2009; 28(1): 318–322. Simple de Modèle de Secours. WAJM 2009; 28(1): 316–320.

Keywords: Attendance, Emergency, Mortality, Unit Mots clés: Assistance De Secours, Mortalité, Séparation Separation. D’Unité.

*Post Graduate Unit, 37 Military Hospital, , Ghana. e-mail: easumanu @yahoo.com. †Department of Community Health, University of Ghana Medical School, Accra, Ghana. email: [email protected]

Abbreviations: SEM, Single Emergency Model; USM, Unit Separation Model.

West African Journal of Medicine Vol. 28, No. 1 January 2009 E. Asumanu and R. Richardson Effect of Separating Emergency Unit

INTRODUCTION March to September 2006. In the first Paediatrics Emergency, Gynaecology Emergency units of many health phase of March to September 2005, Emergency and a Trauma and Surgical facilities in Ghana have been experiencing patients were seen in a single emergency Emergency Units. The Trauma and increasing admissions in recent times. unit under the SEM. This was followed Surgical Unit created a DST to attend to Rapid population growth, persistence of by a period of preparation of the Unit all surgical cases (including accidents preventable childhood diseases, Separation Model where consultations and trauma). Each unit determined its duty malnutrition, trauma, the H.I.V pandemic were made and concerns addressed by team composition and its operation and re-emergence of tuberculosis are the model teams. Units were given time system. factors contributing to the increased for the establishment of their separate emergency admissions. The changing emergency units. The process of Patient Recruitment and Selection pattern of disease towards non implementation took two months from In the SEM, all patients who reported communicable ones has added its burden January to February 2006. This was with emergencies were seen at the central to emergency care. followed by the second phase during point for triaging and further management. Mortality among patients presenting which period patients were seen in Under the USM, patients were triaged in emergency is high and may be a multiple specialty specific units under the based on their referral documents at the consequence of the quality of USM. The months of study (March– reception points. Patients without institutional care.1,2 Emergency units September) were selected for each year referrals were triaged at a 24-hour worldwide as part of quality assurance to account for seasonal variations in Outpatient polyclinic while the few who have adopted various models and emergency attendance and mortalities, are rushed to any of the emergencies are measures to improve mortality and the periods when complete data were triaged at the receiving units. All patients outcomes.2,3 In Ghana, the increasing obtained. seen at the emergency units of the hospital burden of emergency care may have were included in the study. Records of contributed to the increasing mortality in Pre-Intervention or Single Emergency attendance and mortality were retrieved emergency units. Lack of adequate pre Model from the hospital database system. hospital care has contributed to high An emergency department of the 400-bed Patients in SEM were assigned to Specific trauma mortality in our emergency units.4 capacity hospital served as a single unit Units by their diagnosis. Patients in the Single, dual or multifunctional emergency all purpose emergency for surgery, USM group were assigned by the unit of units are not uncommon, and the medicine, and gynaecology. All final admission. Paediatric patients were increasing admissions to these units have patients reporting with any emergency to those 12 years or below. created a challenge for many institutions. this department were first seen by the duty Exclusion criteria included the This has resulted in expansion of structure doctors of the department who would following: i) non civilian patients admitted and/or resources as a common option to then triage and provide first line care to directly to special wards and Intensive meet the current challenge. The all such emergencies. The patients were Care Unit; ii) patients whose diagnoses effectiveness of such systems of subsequently reviewed by the specialist were not stated or could not be assigned; emergency care adopted in secondary team as appropriate. Nurses and allied iii) patients with multiple diagnoses; iv) and tertiary facilities is yet to be staff supported the duty doctors all of patients admitted to the Neonatal determined. whom worked on a shift basis. Intensive Care Unit. An alternative approach employs the industrial concept of operations Ethical Consideration: After a quality Definitions and Outcome Measures: The management in managing the emergency assurance assessment, a proposal was outcome measures for the study were problem where Unit Separation Model made to run this emergency department attendance/admissions and mortality in (U.S.M), based on operations by specialty units. The hospital’s emergency services for both SEM and management principle, replaces Single approval was obtained for the USM. Emergency Model (S.E.M) in the implementation of the Unit Separation management of emergency services. Model and subsequent collection of data Emergency Mortality was defined as The objective of the study was to related to the study. deaths that occurred at the emergency find the effect of Unit Separation Model units and /or deaths that occurred within (USM) on admissions and mortality at Unit Separation Model 24 hours of admission into the hospital the 37 Military Hospital in Accra, Ghana. Unit Separation Model (USM) through the emergency units. requires that emergency services are PATIENTS AND METHODS provided by specialty teams. Such Unit Fatality Rate was determined as the specialty teams may follow specialist/ number of deaths in a unit over total Study Design sub-specialty teams (ST) where available emergency admissions to the unit. This This was a comparative study of the or dedicated specialty teams (DST) that reflected the mortality of the specific unit. effect of Unit Separation on emergency are created to provide such service. This outcome. The study period comprised resulted in the creation of four separate RESULTS two phases: March to September 2005 and emergency units – Medical Emergency, A total of 4,314 patients with valid

West African Journal of Medicine Vol. 28, No. 1 January 2009 319 E. Asumanu and R. Richardson Effect of Separating Emergency Unit

700 The pattern of monthly mortality also shown in Table 1. All the units had a 600 500 rates for the hospital over the period of statistically significant reduction in 400 study for all emergencies is shown in emergency fatality rates in the USM over 300 Figure 2. Total deaths for SEM phase were the study period. Reductions in unit 200

ATTENDANCE 297(36.7%) while USM phase recorded fatality rates were from 27.8% to 7.9% for 100 372(10.9%) deaths. There was a decline Surgery, 46.3% to 23.2% for Medicine, 0 MARCH MAY JULY SEPTEMBR in mortality of 70% over the study period. 17.5% to 0.8% for Gynaecology, and 50.0% CALENDAR MONTH The highest mortalities in both phases to 8.4% for Paediatrics (p-value <0.001 for were in June. each unit); and overall from 36.7% to Figure 1: Monthly Emergency Attendance Table 1 shows the emergency 10.9%. The percentage decline in unit SEM USM attendance and deaths for the various fatality rates was 71.6, 49.9, 95.4 and 83.2% units in both the SEM and USM phases. respectively for Surgery, Medicine, All units recorded an increase in Gynaecology, and Paediatrics. The records seen during the study period attendance over the period. Medicine had highest unit fatality rates in both phases were analysed. Eight hundred and nine the highest attendance and Paediatrics were in Paediatrics and Medicine while patients were seen in SEM period while the lowest in the SEM phase. In the USM Gynaecology had the lowest unit fatality 3505 patients were seen in the USM phase, Paediatrics recorded the highest rates. period. The average monthly emergency and Gynaecology the lowest attendance. admissions were 116 cases in 2005 and Surgery and Medicine recorded almost a DISCUSSION 500 cases in 2006. The highest attendance 3-fold increase in attendance while There has been growing worldwide in SEM was in March and April and the Gynaecology and Paediatrics recorded debate about the rise in hospital lowest was in June and July. The highest more than a 10-fold increase in emergencies and their outcomes.2-7 attendance in USM was in April and June attendance. Emergency room overcrowding is an ever and the lowest was in March and August Unit fatality rates in both phases are increasing problem even in advanced (Figure 1). countries.8 This has resulted in many proposed models for emergency care 70 worldwide.2-6 Regions and institutions are 60 modeling emergency units based on their 50 resources and capacity to achieve the best outcomes. The success of the 40 industrial concept of Operations 30 Management has been gainfully applied to emergency unit care. Generally, 20 Operations Management involves the 10 transformation of processes with focus

MO RTALITY0 RATE on efficiency and effectiveness. Changes MAR APR MAY JUN JUL AUG SEPT in operations have been shown to have a positive impact on the outcomes. Unit CALENDAR MONTH Separation Model is one such Operations management system that could be applied Figure 2: Monthly Mortality Rates for the Two Periods to emergency care. This study has shown Single Emergency Model Unit Separation Model an increase in attendance and reduction in fatality rates using the Unit Separation Model. Table 1: Comparative Unit Emergency Attendance and Fatality Rates There was a four fold increase in overall hospital emergency attendance after Unit separation. The increase in SEM 2005 USM 2006 attendance was occasioned by the Unit Attendance Deaths UFR* Attendance Deaths UFR* Z- p- cumulative effect of multiple emergency %% score value units. During the pre intervention phase, Surgery 338 94 27.8 922 73 7.9 7.7 <0.001 Medicine and Surgery contributed almost Medicine 380 176 46.3 941 219 23.2 8.0 <0.001 90% of emergency attendance. Gynaecology 57 10 17.5 641 5 0.8 3.3 <0.001 Gynaecology and Paediatric emergency Paediatrics 34 17 50.0 1001 84 8.4 4.8 <0.001 attendances were not significant. In the USM phase, Medicine, Surgery and Total 809 297 36.7 3505 381 10.9 8.9 <.001 Paediatrics contributed almost equal *UFR – Unit Fatality Rate proportions to overall emergency

320 West African Journal of Medicine Vol. 28, No. 1 January 2009 E. Asumanu and R. Richardson Effect of Separating Emergency Unit attendance. Most of these new remarkable improvement. Emergency separation did provide adequate care for emergency units of above 20-bed capacity trauma mortality in well organized and well those who needed such services. More had a high bed turnover. The trend of resourced institutions has been reported patients could be attended to over the increased attendance that occurred after at 5.8 %.7 same time period when multiple units Unit Separation suggests that a lot of Medical admission increased almost operate simultaneously. Thirdly, patients who required emergency services three fold and fatality declined by almost remarkable increases in admissions can in the past were unable to receive care for 50%. This is an indication that many more be achieved without increases in lack of emergency space. The study also patients who need emergency care are mortality. shows that the human resource capacity either being denied assess or are victims of the departments was underutilized. of avoidable deaths. Most of these CONCLUSION Although the change created an initial medical emergencies in patients who The results of this study show that Unit increased workload on staff, improvement sought emergency care were due to Separation Model can increase in unit staff management brought hypertension and diabetes related emergency admissions and reduce efficiency to manage the increased load. complications, and this may be partly due mortality. Separation on the basis of sub- This supports previous assertion that to an increase in non communicable specialisation in facilities with specialist hospital resources can be more efficiently diseases and probably late reporting. services is likely to reduce mortality utilized.1 Gynaecology recorded over a 10- further. In the absence of adequate Mortality or fatality is a clear end fold increase in attendance and fatality resources, the industry concept of point used in quality assurance decline of 95%over the study period. The operations management which includes assessment and may reflect the quality of creation of a dedicated team with improvement in hospital staff care in any health institution.1 Apart from specialized nurses helped attend to the management, can improve the efficiency patient delays, mortality in the emergency increasing emergencies. The unit was able of emergency units and thus lower department is accounted for by other to reduce mortality significantly despite mortality. Further research into the clinical factors including systems and operations a tremendous increase in patient and epidemiological profile of emergency at the emergency departments. Overall the population. Mortality of about 1% was mortality in the various units will be mortality rates were reduced by 70%. recorded in the post intervention phase. useful. Using the USM, the two groups showed Paediatric attendance increased over 20 similar patterns in mortality trends – the fold with a fatality decline of 83%. The ACKNOWLEDGEMENT peaks of June and August were preceded increase in attendance was accounted for This work was supported from the and followed by reductions in mortalities. by high turnover and greater willingness Internally Generated Fund (IGF) of the 37 The apparent high mortality rates of parents to access a user-friendly Military Hospital, Ghana. We wish to recorded in the pre-intervention phase specialist paediatric care. This resulted in thank the relevant staff of the hospital for can be attributed to the small space and an observed reduction in attendance of their various contributions, and Professor the small sample size, while the low paediatric patients to the polyclinic. These R. B. Biritwum for reading through the mortality in the post-intervention phase findings appear similar to a Japan manuscript. is attributable to the positive impact of International Cooperation Agency (JICA) the USM. The low mortality achieved is project in Egypt which achieved a REFERENCES also an indication of the potential of the reduction in emergency mortality from 1. Biritwum RB, Nyame PK. Korle Bu hospital to provide care to many under 19% to 13% despite tremendous increase Teaching Hospital: Bed Occupancy and improved conditions. Additionally, Unit in admissions.9 As with our model Manpower Statistics. Ghana Medical Separation allows for effective audit and patients had access to specialist teams. Journal 1993/94; 27: 499–509. 2. Goodacre S, Turner J, Nichol J. quality assessment making training and Paediatric mortality reduction from 50.0% Prediction of mortality among protocol implementation easier. to 8.4% is significant compared to Emergency Medical Admissions. In the USM phase, the pattern of emergency mortality of 12.2% in Southern Emergency Medical Journal 2006; 23: increased unit admissions and reduced India.10 372–375. mortality occurred in all departments of These results were achieved with 3. McGlynn EA, Asch SM, Adams J, the hospital. Surgery recorded approxi- marginal increase in resources and some Keesey J, Hicks J, DeCristofao A, Kerr mately a three fold increase in attendance lessons can be learnt from this study. First, EA. The quality of health care delivery and a percentage decline of 70% in unit avoidable mortalities do occur in our to adults in the United States. N Engl J fatality. The high attendance for surgery emergency units and there is the need for Med 2003; 348: 2635–2645. 4. Mock CN, Jorkovich GJ, Nii-Amon – in the pre-intervention period was largely institutions to recognize these Kotei D, Arreola-Risa C, Maiver RV. contributed to by trauma patients who preventable deaths. Some of these Trauma Mortality Patterns on three were admitted in large numbers. mortalities can be reduced by re- nations at different economic levels- Emergency Surgical mortality was examining the operations of emergency Implications for Global Trauma System reduced from 27.8% to 7.9% over the units. A study of patterns of mortality will Development. J Trauma 1998; 44: 804– study period. The inclusion of trauma provide useful guides to manage 12. mortality in this outcome indicates a institutional mortalities. Secondly, Unit 5. Capwell S. The continuing rise in

West African Journal of Medicine Vol. 28, No. 1 January 2009 321 E. Asumanu and R. Richardson Effect of Separating Emergency Unit

Emergency Admissions. B M J 1996; Surg. 2006; 244: 371–380. archives13-1-25-html (Accessed 20th 312: 991–2. 8. Steven LB, Brent AA. Emergency December 2007). 6. Manish NS. The Formation of the Department Crowding: Old Problem, 10. Santhanam I, Pai M, Katsuri K, Emergency Medical Services System New Solutions. Emergency Medicine Radhamani MP. Mortality after AJPH 2006; 96: 414–423. Clinics of North America 2006; 24: 821– Admission in the Paediatric Emergency 7. Russell LG, Gregory JJ, Lisa KM, Hugh 837. Department : a prospective study from MF, Ronald VM. Patterns of Errors 9. JICA-Evaluation-The Pediatric Care a referral children’s Hospital in Southern Contributing to Trauma Mortality: Project in Egypt. [Web Page]. Available India. Paed. Crit. Care Med. 2002; 3: Lessons Learned from 2594 Deaths. Ann at jica.go.jp/evaluation/project/term/me/ 358–63.

322 West African Journal of Medicine Vol. 28, No. 1 January 2009