Victorian Emergency Minimum Dataset: factors that impact upon the data quality

Changes in paediatric Ambulatory Care Sensitive Conditions in Victoria, 2018-20: the COVID-19 effect?

Objectives This study sought to describe the patterns in emergency department (ED) presentations and hospital admissions in children with Ambulatory Care Sensitive Conditions (ACSCs) before and during the coronavirus disease 2019 (COVID-19) pandemic restrictions in Victoria, Australia, to assess if changes in out-of-hospital care resulted in an increase in delayed/severe presentations. Methods This study involved secondary analysis of Victorian Emergency Minimum Dataset and Victorian Admitted Episode Dataset data. Patients (<18 years) attending EDs with a diagnosis of an ACSC were included. We compared the number and severity of ASCS presentations pre-COVID-19 (1January 2018-27 March 2020) and during COVID-19 (28 March-31 October 2020). A linear regression prediction model was built to compare the observed versus predicted presentation number in the 2020 period. Results In total, there were 108 104 paediatric ACSC ED presentations in Victoria during the study period. Females accounted for 51 462/108 104 (47.6%) of all presentations, with a median age of 3 years. A significant decrease in ED presentations was seen in 2020 (41 319 in 2018; 44 978 in 2019; and 21 807 until October 2020), predominantly due to reductions in conditions that are typically mediated by viruses in childhood (i.e. asthma, convulsions/epilepsy and ear, nose and throat conditions). The proportion of high-urgency presentations and those requiring admission was stable in 2020. Conclusions An overall reduction in the number of ED presentations and admissions with paediatric ACSCs was seen in Victoria in 2020, indicating that rates of delayed or more severe presentations did not occur as a result of changes in out-of-hospital care.

Predicting Patient Length of Stay in Australian Emergency Departments Using Data Mining

Length of Stay (LOS) is an important performance metric in Australian Emergency Departments (EDs). Recent evidence suggests that an LOS in excess of 4 h may be associated with increased mortality, but despite this, the average LOS continues to remain greater than 4 h in many EDs. Previous studies have found that Data Mining (DM) can be used to help hospitals to manage this metric and there is continued research into identifying factors that cause delays in ED LOS. Despite this, there is still a lack of specific research into how DM could use these factors to manage ED LOS. This study adds to the emerging literature and offers evidence that it is possible to predict delays in ED LOS to offer Clinical Decision Support (CDS) by using DM. Sixteen potentially relevant factors that impact ED LOS were identified through a literature survey and subsequently used as predictors to create six Data Mining Models (DMMs). An extract based on the Victorian Emergency Minimum Dataset (VEMD) was used to obtain relevant patient details and the DMMs were implemented using the Weka Software. The DMMs implemented in this study were successful in identifying the factors that were most likely to cause ED LOS > 4 h and also identify their correlation. These DMMs can be used by hospitals, not only to identify risk factors in their EDs that could lead to ED LOS > 4 h, but also to monitor these factors over time.

Improving injury surveillance data quality: a study based on hospitals contributing to the Victorian Emergency Minimum Dataset

OBJECTIVE

In this paper, we describe the design and baseline data of a study aimed at improving injury surveillance data quality of hospitals contributing to the Victorian Emergency Minimum Dataset (VEMD).

METHODS

The sequential study phases include a baseline analysis of data quality, direct engagement and communication with each of the emergency department (ED) hospital sites, collection of survey and interview data and ongoing monitoring.

RESULTS

In 2019/20, there were 371,683 injury-related ED presentations recorded in the VEMD. Percentage unspecified, the indicator of (poor) data quality, was lowest for 'body region' (2.7%) and 'injury type' (7.4%), and highest for 'activity when injured' (29.4%). In the latter, contributing hospitals ranged from 3.0-99.9% unspecified. The 'description of event' variable had a mean word count of 10; 16/38 hospitals had a narrative word count of <5.

CONCLUSIONS

Baseline hospital injury surveillance data vary vastly in data quality, leaving much room for improvement and justifying intervention as described. Implications for public health: Hospital engagement and feedback described in this study is expected to have a marked effect on data quality from 2021 onwards. This will ensure that Victorian injury surveillance data can fulfil their purpose to accurately inform injury prevention policy and practice.

Patient-specific record linkage between emergency department and hospital admission data for a cohort of people who inject drugs: methodological considerations for frequent presenters

Background

People who inject drugs (PWID) have been identified as frequent users of emergency department (ED) and hospital inpatient services. The specific challenges of record linkage in cohorts with numerous administrative health records occurring in close proximity are not well understood. Here, we present a method for patient-specific record linkage of ED and hospital admission data for a cohort of PWID.

Methods

Data from 688 PWID were linked to two state-wide administrative health databases identifying all ED visits and hospital admissions for the cohort between January 2008 and June 2013. We linked patient-specific ED and hospital admissions data, using administrative date-time timestamps and pre-specified linkage criteria, to identify hospital admissions stemming from ED presentations for a given individual. The ability of standalone databases to identify linked ED visits or hospital admissions was examined.

Results

There were 3459 ED visits and 1877 hospital admissions identified during the study period. Thirty-four percent of ED visits were linked to hospital admissions. Most links had hospital admission timestamps in-between or identical to their ED visit timestamps (n = 1035, 87%). Allowing 24-h between ED visits and hospital admissions captured more linked records, but increased manual inspection requirements. In linked records (n = 1190), the ED ‘departure status’ variable correctly reflected subsequent hospital admission in only 68% of cases. The hospital ‘admission type’ variable was non-specific in identifying if a preceding ED visit had occurred.

Conclusions

Linking ED visits with subsequent hospital admissions in PWID requires access to date and time variables for accurate temporal sorting, especially for same-day presentations. Selecting time-windows to capture linked records requires discretion. Researchers risk under-ascertainment of hospital admissions if using ED data alone.

Comparison of the Victorian Emergency Minimum Dataset to medical records for emergency presentations for acute cardiovascular conditions and unspecified chest pain

OBJECTIVE

The Victorian Emergency Minimum Dataset (VEMD) collects administrative and clinical data for all presentations to Victorian public ED. The present study aimed to examine the level of agreement between the VEMD data and the medical record for a sample of patients coded as having acute cardiovascular conditions (acute coronary syndrome, stroke and transient ischaemic attack [TIA]) and unspecified chest pain in the VEMD.

METHODS

Six months of data provided to the VEMD from a large metropolitan hospital was obtained, and a random sample of 10% of cases (n = 310) were selected for review. Data for eight VEMD items were compared for concordance to data recorded in the ED medical record.

RESULTS

Complete concordance between the VEMD and medical records for all eight items was observed only for 101 (33%) presentations. Overall, the least concordant variables were those with a high number of coding options: usual type of accommodation (76%), referral pattern (84%) and primary diagnosis (85%). The concordance of the VEMD primary diagnosis varied when examined as individual codes (range 75%-100%) and when combined (acute coronary syndrome = 94%, stroke or TIA = 85% and chest pain unspecified = 75%). The level of agreement for some items improved when VEMD codings were combined.

CONCLUSION

When compared to the medical record, our data suggest there is likely variation in the accuracy of some VEMD items, and suggests a larger prospective validation of the VEMD is warranted. For researchers using existing VEMD data, combining of some codes may be necessary.

Drug overdose in the ED: a record linkage study examining emergency department ICD-10 coding practices in a cohort of people who inject drugs

Background

Drug overdose is a leading cause of mortality and morbidity amongst people who inject drugs (PWID). Drug overdose surveillance typically relies on the International Classification of Diseases (ICD-10) coding system, however its real world utilisation and the implications for surveillance have not been well characterised. This study examines the patterns of ICD-10 coding pertaining to drug overdoses within emergency departments for a cohort of known PWID.

Methods

Cohort data from 688 PWID was linked to statewide emergency department administrative data between January 2008 and June 2013. ICD-10 diagnostic codes pertaining to poisonings by drugs, medicaments and biological substances (T-codes T36-T50) as well as mental and behavioural disorders due to psychoactive substance use (F-codes F10-F19) were examined.

Results

There were 449 unique ED presentations with T or F code mentions contributed by 168 individuals. Nearly half of the T and F codes used were non-specific and did not identify either a drug class (n = 160, 36%) or clinical reaction (n = 46, 10%) and 8% represented withdrawal states. T and F codes could therefore be used to reasonably infer an illicit drug overdose in only 42% (n = 188) of cases. Majority of presentations with T or F overdose codes recorded only one diagnostic code per encounter (83%) and representing multiple-drug overdose (F19.- = 18%) or unidentified substances (T50.9 = 17%) using a single, broad diagnostic code was common.

Conclusions

Reliance on diagnoses alone when examining ED data will likely significantly underestimate incidence of specific drug overdose due to frequent use of non-specific ICD-10 codes and the use of single diagnostic codes to represent polysubstance overdose. Measures to improve coding specificity should be considered and further work is needed to determine the best way to use ED data in overdose surveillance.

Emergency demand, repeat and frequent presentations by older patients in metropolitan Melbourne: A retrospective cohort study using routinely collected hospital data

OBJECTIVE

To describe patterns for potentially avoidable general practice (PAGP)-type and non-PAGP-type ED presentations by older patients during 2008 and 2012.

METHODS

Retrospective analysis of ED presentations by patients ≥70 years for 2008 and 2012. Metropolitan Melbourne public hospital data were obtained from the Victorian Emergency Minimum Dataset. Outcomes were characteristics of PAGP-type and non-PAGP-type presentations as defined by the Australian Institute of Health and Welfare; numbers and rates per 1000 population ≥70 years of repeat (×2-3/year) and frequent (≥ ×4/year) PAGP-type and non-PAGP-type presentations.

RESULTS

The older metropolitan Melbourne population increased by 10.3% between 2008 and 2012, whereas the number of ED presentations increased by 12.7%. The volume of PAGP-type presentations decreased by 2.6%, with declining rates per 1000 population ≥70 years of repeat (7.2-6.2) and frequent (0.7-0.4) presentation. In contrast, the volume of non-PAGP-type presentations grew by 15.4%, with increasing repeat (57.6-60.7) and frequent (13.1-14.2) presentation rates per 1000 population ≥70 years. The majority (39%) of non-PAGP-type presentations by frequent ED attenders were due to cardiovascular or respiratory problems.

CONCLUSION

The rate of repeat and frequent PAGP-type presentations by older people decreased in 2012 compared with 2008, suggesting that initiatives implemented to reduce avoidable presentations may have had an effect. However, an increase in the rate of frequent non-PAGP-type presentations, predominately for acute exacerbation of cardiovascular and respiratory conditions, has important implications for planning future healthcare delivery; hence, the importance of initiatives such as the Health Care Home.

Assessing the completeness of coded and narrative data from the Victorian Emergency Minimum Dataset using injuries sustained during fitness activities as a case study

BACKGROUND

Injury surveillance systems support the ongoing systematic collection, analysis and interpretation of health information vital to the prevention, planning and evaluation of injury prevention strategies. One key measure of the success of such systems is their reliability. Data completeness is a major component of system reliability, and is an indicator of a system's data quality. The Victorian Emergency Minimum Dataset (VEMD) is a state-wide record of injury presentations to emergency departments in Victoria, Australia. For each case, it provides information on the injury cause, place of occurrence, activity at time of injury, body region affected and nature of injury, as well as a free-text narrative of the injury event. The aim of this study was to assess the completeness of data in the VEMD using injuries sustained in fitness facilities as a case study.

METHODS

Analysis of VEMD coded parent injury variables (nature of injury, injured body region, cause of injury, place where injury occurred, activity at time of injury) and detailed narratives were reviewed for completeness over the ten-year period July 2003 to June 2012, inclusive. Narratives were text analysed manually to determine which items of injury information they contained and compared to the parent injury variables.

RESULTS

There were 2936 identified cases related to injuries sustained during fitness activities. Two percent of cases had all coded injury variables unspecified. Overall, 95.8 % of narratives had at least one piece of injury information missing. The nature of injury and body region variables were coded in 92.6 and 96.6 % of cases, yet were only mentioned in 27.1 and 75.4 % of narratives, respectively. The cause variable was allocated a specified code in 47.7 % of cases and was mentioned in 45.9 % of narratives. The cause was missing in both in 42.8 % of cases. In approximately half of all cases, the activity and place were specified in both the coded injury variable and narrative; they were missing in both in 7.4 and 13.6 % of cases, respectively.

CONCLUSIONS

The reliability of the VEMD as an injury surveillance system, varied depending on the injury variable being examined.

Comparison of data sets for surveillance of work-related injury in Victoria, Australia

OBJECTIVE

To investigate differences and similarities between three sources of work-related injury information: workers compensation claims, emergency department (ED) presentation data and hospital admissions data.

METHODS

This population-based, retrospective descriptive analysis of non-fatal, work-related injuries of workforce participants in Victoria, Australia, has compared data from workers compensation claims and ED presentation and hospital admission data sets for the period 2004-2011. Work-related injury case frequency and rate were compared across study years according to gender, age, geographical location and injury type. Injury rates were expressed as cases per million hours worked.

RESULTS

Rates of hospital admissions for treatment of work-related injury increased over the study period, compared with decreasing rates of injury in compensation claims and ED data. The highest rate of injuries to younger workers was captured in ED data. There was greater capture of musculoskeletal injuries by workers' compensation data, and of open wound and burn injury by the ED data. Broad similarities were noted for temporal trends according to gender, for the distribution of cases across older age groups and for rates of fracture injuries.

CONCLUSIONS

These study findings inform use of workers' compensation, ED presentation and hospital admission data sets as sources of information for surveillance of work-related injuries in countries where these types of data are routinely collected. Choice of data source for investigation of work-related injury should take into consideration the population and injury types of interest.

Implementing injury surveillance systems alongside injury prevention programs: evaluation of an online surveillance system in a community setting

Background

Previous research aimed at improving injury surveillance standards has focused mainly on issues of data quality rather than upon the implementation of surveillance systems. There are numerous settings where injury surveillance is not mandatory and having a better understanding of the barriers to conducting injury surveillance would lead to improved implementation strategies. One such setting is community sport, where a lack of available epidemiological data has impaired efforts to reduce injury. This study aimed to i) evaluate use of an injury surveillance system following delivery of an implementation strategy; and ii) investigate factors influencing the implementation of the system in community sports clubs.

Methods

A total of 78 clubs were targeted for implementation of an online injury surveillance system (approximately 4000 athletes) in five community Australian football leagues concurrently enrolled in a pragmatic trial of an injury prevention program called FootyFirst. System implementation was evaluated quantitatively, using the RE-AIM framework, and qualitatively, via semi-structured interviews with targeted-users.

Results

Across the 78 clubs, there was 69% reach, 44% adoption, 23% implementation and 9% maintenance. Reach and adoption were highest in those leagues receiving concurrent support for the delivery of FootyFirst. Targeted-users identified several barriers and facilitators to implementation including personal (e.g. belief in the importance of injury surveillance), socio-contextual (e.g. understaffing and athlete underreporting) and systems factors (e.g. the time taken to upload injury data into the online system).

Conclusions

The injury surveillance system was implemented and maintained by a small proportion of clubs. Outcomes were best in those leagues receiving concurrent support for the delivery of FootyFirst, suggesting that engagement with personnel at all levels can enhance uptake of surveillance systems. Interview findings suggest that increased uptake could also be achieved by educating club personnel on the importance of recording injuries, developing clearer injury surveillance guidelines, increasing club staffing and better remunerating those who conduct surveillance, as well as offering flexible surveillance systems in a range of accessible formats. By increasing the usage of surveillance systems, data will better represent the target population and increase our understanding of the injury problem, and how to prevent it, in specific settings.

Electronic supplementary material

The online version of this article (doi:10.1186/s40621-014-0019-y) contains supplementary material, which is available to authorized users.

Emergency demand and repeat attendances by older patients

BACKGROUND

Population ageing is projected to impact on health services utilisation including Emergency Departments (ED), with older patients reportedly having a high rate of return visits. We describe and compare patterns in ED utilisation between older and younger adults, and quantify the proportion and rate of return visits.

METHODS

Population-based retrospective analysis of metropolitan Melbourne public hospital ED data, 1999/2000 to 2008/2009. Numbers of patients, presentations, re-presentations and rates per 1000 population were calculated, with comparison of older (aged ≥ 70 years) and younger (15-69 years) attendances.

RESULTS

Population growth in each age group was similar over the study period, yet ED presentations rose by 72% for older adults compared with a 59% increase for younger adults. Rates per 1000 population rose with increasing age. Of the population aged ≥ 70 years, 39% presented to ED compared with 17% of the population aged 15-69 years in 2008/2009. Twenty-seven per cent of the increase in older adult presentations was driven by a cohort who attended ≥ 4 times in 2008/2009. The number of older patients presenting ≥ 4 times doubled over the decade, contributing to 23% of all older presentations in 2008/2009. ED length of stay rose with increasing age; 69% of older adults remained in ED for ≥ 4 h compared with 39% of younger adults in 2008/2009. The number of older adult ED hospital admissions doubled over the decade.

CONCLUSIONS

Older patients are disproportionately represented among ED attendances. They also have an increasing propensity to re-present to ED, indicating a need to identify the clinical, social and health system-related risk factors for re-attendance by specific patients.

Emergency Department Presentations of Victorian Aboriginal and Torres Strait Islander People

This paper explains how routinely collected data can be used to examine the emergency department attendances of Victorian Aboriginal and Torres Strait Islander people. The data reported in the Victorian Emergency Minimum Dataset (VEMD) for the 2006/2007 financial year were analysed. The presentations of Aboriginal and Torres Strait Islander and non-Aboriginal people were compared in terms of age, gender, hospital location (metropolitan and rural) and presenting condition. Aboriginal and Torres Strait Islander people were found to attend the emergency department 1.8 times more often than non-Aboriginal people. While the emergency department presentation rates of metropolitan Aboriginal and Torres Strait Islander and non-Aboriginal people were similar, rural Aboriginal and Torres Strait Islander people presented to the emergency department 2.3 times more often than non-Aboriginal people. The injuries or poisonings, respiratory conditions and mental disorders presentation rates of the Aboriginal and Torres Strait Islander and non-Aboriginal population were compared. No previous studies have assessed the accuracy of the Indigenous status and diagnosis fields in the VEMD; therefore the quality of this data is unknown.

Hospital admissions following presentations to emergency departments for a fracture in older people

The aim of this paper is to estimate the proportion of older people who are hospitalised following a presentation to an emergency department for hip, pelvic and wrist fractures. The findings indicate that hospitalisation data do not accurately reflect the incidence of low-trauma fractures, particularly wrist and pelvic fractures, in older people.

Challenges of recruiting farm injury study participants through hospital emergency departments

BACKGROUND

Hospital emergency departments are common recruitment sites for injury studies. Yet recruitment method details, capture rates and response fractions are not consistently reported. As privacy legislation increasingly impinges on research activity, these parameters become even more important. The authors describe their experience with recruitment via emergency departments and outline subsequent adjustments to the recruitment approach.

METHODS

The FIRM study was an Australian case-control study of serious farm work-related injury. Cases were identified prospectively by hospital staff on presentation to emergency departments. Consistent with the Victorian Health Records Act, potential cases were initially approached by hospital staff, and full recruitment was subsequently undertaken by study staff. Manual hospital record audits were conducted at five recruitment sites to determine the proportion of eligible cases approached.

RESULTS

Among 660 medical records audited, 19 eligible cases were confirmed, 9 of whom were approached by hospital staff (47%, 95% CI 25 to 70%). In response, an additional process was established to capture missed cases, who were sent a letter from the hospital providing the opportunity to opt out of telephone contact by study staff. Early indications were that 34% (41/122) of missed cases actively declined to be contacted. Among those who were contacted and eligible, 84% (21/25) agree to study participation.

CONCLUSIONS

Recruitment of injury research participants via hospital emergency departments is challenging, particularly where authorities require an intermediary to make the initial contact. Removal of some constraints imposed by privacy legislation would considerably simplify recruitment and enhance scientific rigour in conducting epidemiological research.