DOI: 10.1111/ajag.12142

Research Lower rates of appropriate initial diagnosis in older emergency department patients associated with hospital length of stay Tee Juan Ong Aged Care, Royal Melbourne Hospital, Parkville, Victoria, Australia

Yohanes Ariathianto, Rabindra Sinnappu and Wen Kwang Lim Aged Care, Northern Hospital, Epping, Victoria, Australia

Aim: Emergency department (ED) doctors are under time pressure to expedite decision-making. This task would seem more difficult in older patients who present atypically, have multiple comorbidities and require more diagnostic tests. This study aimed to investigate the rate of appropriate initial diagnosis of older ED patients admitted under medical units, and whether time was a factor. Methods: Retrospective review of all patients admitted under medical units from ED over a one-month period was conducted. Results: Four hundred ninety-three records were reviewed. Mean time to ED review was 87 minutes, and to medical registrar review, 409 minutes. Overall rate of appropriate initial diagnosis made by ED was 85.8%, with significantly lower rate detected in older patients. Overall rate for medical registrar was 94.5%. Conclusions: Admitted older ED patients received lower rates of appropriate initial diagnosis. Time may be a contributing factor to this lower rate. Length of stay was prolonged if initial diagnosis was inappropriate. Key words: accuracy, diagnosis, emergency department, time.

Introduction The patterns of use of emergency department (ED) among older patients (older 65 years old) differ from younger patients [1]. They are more likely to arrive by ambulance, have higher rates of hospital admission, have longer ED stays, have greater diagnostic tests and procedures and has higher risk of adverse health outcomes after an index ED visit [1]. Common challenging conditions that older patients present with to ED include delirium, dementia, falls, coronary disease, abdominal pain, polypharmacy/adverse drug effects, infections and social cases [2]. Hospitals have focused on patient safety by addressing adverse events amenable to system-wide solutions such as infection control and medication errors. However, diagnostic

Correspondence to: Associate Professor Wen Kwang Lim, Aged Care, Northern Hospital. Email: [email protected] Australasian Journal on Ageing, Vol 34 No 2 June 2015, 121–126 © 2014 ACOTA

bs_bs_banner

errors, although common and often serious, have not received comparable attention [3,4]. Across a wide variety of clinical conditions, diagnostic error rates average about 10%[3]. Safer patient treatment should also be based on better diagnosis – diagnoses with fewer delays, mistakes and process errors [3]. The passage of time often enables doctors to make a more accurate diagnosis. This is due to the evolution of symptoms, further information gathering or results of investigations becoming available. ED crowding is common [5], associated with poorer patient-oriented outcomes [6–8], and in turn delays entry for new patients, leading to prolonged waiting times, ambulance ramping and hospital bypass. ED doctors are under time pressure to expedite decision-making, and this task would seem more difficult in the older patient who presents atypically, have multiple comorbidities and require more diagnostic tests and procedures [1]. Accordingly, the aim of this study was to investigate the rate of appropriate initial diagnoses made in ED, whether older patients receive lower rates of appropriate initial diagnoses and whether time was a factor.

Methods Approval from the Northern Hospital Ethics Committee was received. The Northern Hospital is a 330-bed hospital in Victoria, Australia. Its ED saw 63 314 patients for the year 2011 [9]. On presentation to ED, patients are seen by the triage nurse and assigned a triage category [10] (1 – most acute, 5 – least acute). ED doctors review patients in order of triage category. Should the patient require admission under a medical unit, ED doctors refer patients to one medical registrar who then admits for all medical units. This includes general medical units and specialty units (cardiology, oncology, nephrology and gastroenterology). Patient care is then handed over to the respective units as soon as appropriate, usually the following day. The admitting medical registrar works only in ED and does not provide further clinical input once patient care is handed over. This study was a retrospective case review performed on all ED patients who were admitted under all medical units in August 2010. This period was selected because it was a winter month, reflecting a time when the hospital is stretched in its resources. Also, most junior medical staff employment contracts begin in February, allowing for six months of development and adjustment. We recorded patient demographics, calculated Charlson’s Comorbidity Index (age not included in index) [11], length of stay, time of review and diagnosis 121

O n g

T J ,

A r i a t h i a n t o

made by ED doctor and medical registrar and compared that to the diagnosis at discharge from acute hospital. Time of review was the time the patient was first seen, not the time diagnosis was made. Time recorded was dependent on doctor or nursing notes. If neither were available, then a surrogate time marker was taken as time of review (scan or treatment that required doctor authorisation). In this study, the following four criteria were used to define an appropriate initial diagnosis: (i) Diagnosis was included in the differential diagnoses; (ii) Diagnosis of a syndrome. For example, diagnosis of sepsis or delirium even if a cause was not identified; (iii) Appropriate working diagnosis was made. For example, a patient with chest pain was admitted with a diagnosis of possible non-ST segment elevation myocardial infarction. It eventuated to be non-cardiac chest pain after a return of serial serum troponin. This initial diagnosis was felt to be appropriate under those initial circumstances; and (iv) Appropriate initial management was made to suggest a particular diagnosis. For example, breathlessness in a patient with a history of both chronic obstructive pulmonary disease and congestive cardiac failure were treated with both nebulised medications and intravenous diuretics. Diagnosis of the presenting condition at discharge was decided by physician chart review. The primary investigator reviewed all records (TJO). If an inappropriate diagnosis was detected, the secondary investigator (YA) reviewed the record. If there was a disagreement, both investigators met to reach a consensus. Diagnosis was considered inappropriate if wrong diagnosis or no diagnosis was offered, or if there was failure to recognise a critical symptom, sign or investigation. Statistical program SPSS (version 11.5.1) was used to analyse the data. Pearson’s χ2 test was used for categorical variables, Student’s t-test to compare means and logistic regression analysis to look for variables that determined binary outcomes.

Results In August 2010, 549 records were identified as admissions under medical units. Fifty-six were excluded as they were admitted directly to the ward (without review by ED) or were statistical separations for the same acute admission. Four patients admitted initially under the medical unit had surgical cause of their presenting complaint found. Three records were not completely scanned, therefore omitted from the study. Four hundred ninety-three records were reviewed and analysed. Descriptive statistics of the ED patients who were admitted under medical units are shown in Table 1. Two-thirds of the patients were older than 65 years. Two-thirds of the patients spoke English as their first language, but 87% of people who did not speak English as their first language were older than 65 years. This reflected the diverse cultural background of patients at Northern Hospital. Eighty-three per cent of the 122

Y ,

S i n n a p p u

R

e t

a l .

Figure 1: Rate of appropriate initial diagnosis made by ED and medical registrar. ED, emergency department. 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0%

94.5% 85.8%

ED

M e d Re gi s t r ar

patients lived at home, the remaining in residential care. Eighty per cent of the admitted patients arrived by ambulance. On average, patients had a Charlson’s Comorbidity Index score of 2.9, took 6.5 medications and stayed 7.5 days in the hospital. On discharge, most patients returned home, 15% were discharged to rehabilitation services and 7% died in the hospital. Mean time from triage to ED review was 1 hour and 27 minutes. Mean time from triage to medical registrar review was 6 hours and 49 minutes (Table 1). Before the primary and secondary investigators met for a consensus, the level of agreement between reviewers was 87% of ED diagnoses and 82% of medical registrar diagnoses. The average rate of appropriate initial diagnosis for ED was 85.8% and 94.5% for medical registrar (Figure 1). ED made lower rates of appropriate initial diagnosis in older patients (≥65 years old) when compared to younger patients (65 years old), and a subset of older than 80 years trended to even lower rates. This may be due to older patients having atypical presentations and multiple comorbidities [1,2]. It could also be due to a greater reliance on collateral history and investigation results; time may be a surrogate for these factors. There was no significant difference in the medical registrar’s rate of appropriate initial diagnosis in older patients. Hospital length of stay was prolonged if ED or medical registrar made inappropriate initial diagnosis. This may have been due to incorrect treatment commenced for the condition. The longer length of stay may also reflect a diagnostically difficult case, the patient requiring other specialty unit consultation or requiring more complex investigations. Making an appropriate initial diagnosis from the outset in ED seems likely to provide downstream benefits of shorter lengths of stay, cost savings to the hospital and likely clinical benefits to the patient. Length of stay reflected days spent in the acute ward only, rather than the entire stay within the hospital system. This was because there may be a myriad of reasons patients require rehabilitative services, which may not relate to the initial diagnosis being appropriate. The Australian government is gradually implementing the National Emergency Access Target, whereby 90% of all

Table 3: Association between length of stay and appropriate initial diagnosis Length of stay

Length of stay

ED appropriate

n

Mean

SD

Std. Error Mean

P-value

Yes No

413 80

7.21 9.29

5.49 6.31

0.27 0.71

0.01

Med appropriate

n

Mean

SD

Std. Error Mean

P-value

Yes No

460 33

7.26 11.58

5.37 8.02

0.25 1.40

0.01

ED, emergency department; SD, standard deviation.

124

Australasian Journal on Ageing, Vol 34 No 2 June 2015, 121–126 © 2014 ACOTA

A p p r o p r i a t e

i n i t i a l

Table 4: Logistic regression for mortality Multivariate analysis Age Charlson's score Triage category 1 Triage category 2

P-value

OR

0.00 0.03 0.00 0.02

1.08 1.17 13.43 5.21

OR, odds ratio.

Table 5: Summary of common inappropriate initial diagnosis System Sepsis (17 cases)

Neurology (17 cases)

Pain (13 cases)

Cardiac (11 cases)

Surgical (4 cases)

Remarks (number of cases) • Failure to consider infection as underlying cause of vital signs abnormality (tachyarrhythmia, hypotension, hyperglycaemia, vomiting, dehydration) (10) • H1N1 influenza infection diagnosed as asthma (2) • Others (5) including drug-induced fever, liver abscess • Failure to recognise delirium (7) • Presentation attributed to delirium, failure to recognise stroke (4) • Ataxia due to medication (2) • Others (4) including Parkinson's disease, Guillain–Barre, Ramsay Hunt. • Crush fractures and musculoskeletal pain diagnosed as pyelonephritis, cellulitis, urinary tract infection, viral upper respiratory tract infection (8) • Others (5) including polymyalgia, metastasis, gout, • Failure to recognise NSTEMI and tachyarrhythmia as significant component of dyspnoea (6) • Failure to recognise significant ECG changes (1) • Others (4) including pericardial effusion, pericarditis, metastatic disease • Initial admission under medical unit. Subsequent surgical cause for presenting complaint found. • Incarcerated hernia thought to be fracture (1) • Cholecystitis presenting as breathlessness in morbidly obese (1) • Bowel perforation presenting as breathlessness (1) • Pseudobowel obstruction presenting as dehydration (1)

ECG, electrocardiogram; NSTEMI, non-ST segment elevation myocardial infarction.

d i a g n o s i s

Australasian Journal on Ageing, Vol 34 No 2 June 2015, 121–126 © 2014 ACOTA

o l d e r

p e o p l e

medical registrar review was difficult to ascertain because doctors frequently did not denote the time on their notes. If they did, there was no standard practice to indicate if this denoted the time of their review or time they wrote notes. Nursing notes were frequently helpful. It was estimated that a surrogate time marker (investigation or treatment requiring doctor’s authorisation) was required in approximately 30–50% of cases to determine time of ED or medical registrar review. We acknowledge that the discharge diagnosis potentially could be biased towards the medical registrar assessment. Even though there was no further clinical input once the patient was handed over, the admitting medical registrar may influence decisions made by the receiving unit at handover. Initial management commenced by the admitting medical registrar may be numerous, and some may require completion of treatment course. Receiving units may be obliged to include the response to these treatments in deciding the diagnosis. This study showed association between time with rate of appropriate initial diagnosis and rate of appropriate initial diagnosis with length of stay. However, it was not designed to show a cause and effect. Mortality was an infrequent outcome, and this study was not powered to measure an association between rate of appropriate initial diagnosis and mortality.

Conclusions This study showed that ED doctors made lower rates of appropriate initial diagnosis in older patients. Time may be a factor in the rate of appropriate initial diagnosis, as it allows for further history taking, evolution of symptoms and signs, and return of investigation results. When ED doctors or medical registrar make an inappropriate diagnosis at admission, it was associated with longer patient length of stay.

patients presenting to a public hospital ED will either physically leave the ED for admission to hospital, be referred to another hospital for treatment or be discharged within four hours [13]. Average time from triage to ED doctor review was 1 hour and 27 minutes, leaving on average of 2 hours and 33 minutes to obtain collateral history and investigation results to return, before making a decision to admit. Strong consideration must be given for ED priority access to pathology and radiology investigations. Scanned/electronic medical records may expedite decision-making and improve diagnostic ability by having immediate access to complete patient clinical history. There were a number of limitations of this study. As a retrospective study, we were not able to adjust for numerous factors that may affect review times or rate of appropriate initial diagnosis (such as the level of ED crowding, level of staff training/experience, general or specialist units). Triage time was always accurate because it was entered on presentation before being printed on ED notes, but time of ED or

i n

Key Points • Older ED patients admitted to medical units received significantly lower rates of appropriate initial diagnosis. • Time may be a factor in the rate of appropriate initial diagnosis in older ED patients. • When initial diagnosis was inappropriate, length of stay was prolonged.

References 1

2

Aminzadeh F, Dalziel WB. Older adults in the emergency department: A systematic review of patterns of use, adverse outcomes, and effectiveness of interventions. Annals of Emergency Medicine 2002; 39: 238– 247. Samaras N, Chevalley T, Samaras D, Gold G. Older patients in the emergency department: A review. Annals of Emergency Medicine 2010; 56: 261–269. 125

O n g

3

4 5 6

7 8

126

T J ,

A r i a t h i a n t o

Schiff GD, Kim S, Abrams R et al. Diagnosing Diagnosis Errors: Lessons from a Multi-institutional Collaborative Project and Methodology. 2005; [Cited 1 August 2012.] Available from URL: http://www.ahrq.gov/ professionals/quality-patient-safety/patient-safety-resources/resources/ advances-in-patient-safety/index.html Wachter RM. Why diagnostic errors don't get any respect – and what can be done about them. Health Affairs (Project Hope) 2010; 29: 1605–1610. Bernstein SL, Asplin BR. Emergency department crowding: Old problem, new solutions. Emergency Medicine Clinics of North America 2006; 24: 821–837. Bernstein SL, Aronsky D, Duseja R et al. The effect of emergency department crowding on clinically oriented outcomes. Academic Emergency Medicine: Official Journal of the Society for Academic Emergency Medicine 2009; 16: 1–10. Richardson DB. Increase in patient mortality at 10 days associated with emergency department overcrowding. The Medical Journal of Australia 2006; 184: 213–216. Sprivulis PC, Da Silva JA, Jacobs IG, Frazer AR, Jelinek GA. The association between hospital overcrowding and mortality among patients admitted via Western Australian emergency departments. The Medical Journal of Australia 2006; 184: 208–212.

9

10

11 12 13

Y ,

S i n n a p p u

R

e t

a l .

Australian Institute of Health and Welfare. Australian Hospital Statistics 2010–11: Emergency Department Care and Elective Surgery Waiting Times. 2011 [Cited 1 August 2012.] Available from URL: http:// www.aihw.gov.au/publication-detail/?id=10737420662&tab=2 Australasian College Emergency Medicine. Policy on the Australian Triage Scale. [Cited 1 August 2012.] Available from URL: http://www.acem .org.au/media/policies_and_guidelines/P06_Aust_Triage_Scale___Nov _2000.pdf Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: Development and validation. Journal of Chronic Diseases 1987; 40: 373–383. Berner ES, Graber ML. Overconfidence as a cause of diagnostic error in medicine. The American Journal of Medicine 2008; 121 (Suppl. 5): S2–S23. Department of Health and Ageing A. National Partnership Agreement on Improving Public Hospitals. 2011 [Cited 1 August 2012.] Available from URL: http://www.yourhealth.gov.au/internet/yourhealth/publishing.nsf/ Content/npa-improvingpublichospitals-agreement/$File/National%20 Partnership%20Agreement%20on%20Improving%20Public%20 Hospital%20Services.pdf

Australasian Journal on Ageing, Vol 34 No 2 June 2015, 121–126 © 2014 ACOTA