|
nif:isString
|
-
This study is a component of a larger mixed-methods research program aimed at increasing understanding of how public reporting may improve quality of care in public and private hospitals in Australia. Previous components of the research program included examining the perspectives of multiple stakeholders including consumer advocates, providers, purchasers [28], public hospital medical directors[29, 30], general practitioners (GPs) [31] and patients [32]. This component used a quantitative approach to understand the effect of national healthcare reforms on various ED time-based process outcomes. This study involved an uncontrolled, interrupted time series (ITS) analysis of Victorian ED presentations data. ITS is a powerful quasi-experimental research design for evaluating the effect of an intervention when random allocation is not feasible. ITS contains a series of observations related to the outcome of interest at multiple time points before and after the introduction of an intervention. The trends before and after the intervention are compared to determine the effect of the intervention from its secular (underlying) trend [33–35]. It is particularly useful in the analysis of ‘natural experiments’ in real world settings, for example the introduction of a national policy or incentive.
A data request was submitted to the Victorian Department of Health and Human services (DHHS) (data requests activities transferred to The Victorian Agency for Health Information in 2017) for data access to the Victorian Emergency Minimum Dataset (VEMD) [36]. De-identified patient-level data from VEMD spanning years 2006 to 2016 were provided by DHHS. The VEMD records all presentations to EDs in Victorian public hospitals that have a designated 24-hour ED. The VEMD includes de-identified demographic, administrative and clinical data. Data are collected by individual hospitals using standard definitions and protocols, then transferred to the Data Collections Unit (DCU) which manages VEMD operations. Thirty-nine hospitals currently provide data to the VEMD, of those, one hospital provided data only from 2011 onwards (n = 78,139) and therefore was excluded from the analyses.
We selected all ED presentations between 2006 and 2016. Cases were excluded if the: a) type of visit was a planned return visit, pre-arranged admission or patient in transit (n = 314,857); b) patient was dead on arrival (n = 20,274); c) patient did not wait to be attended by a healthcare professional (n = 909,012); c) hospital was others than public acute hospitals (e.g. specialised hospital—women or children hospitals) (n = 1,505,268); or d) waiting time to treatment was greater than eight hours (n = 2,957). With regard to (c), we excluded hospital that had only small numbers of their type. With regard to (d), a wait time more than eight hour was considered a potential data error or may have represented a patient who did not require emergency care.
Australian national healthcare reforms relevant to this study began in August 2011. The VEMD provided by DHHS did not include the full date of patient presentations, only year of presentations. For these analyses, we defined the pre-reform period as 2006 through to 2010, and the post-reform period as 2011 through to 2016.
Outcomes of interest were waiting time to treatment, treatment within recommended time, and departing ED within four hours of arrival. Waiting time to treatment was defined as the time between a patient’s arrival at the ED and the commencement of their clinical care and measured in minutes. Treatment within recommended time referred to the recommended maximum waiting times for commencement of clinical care based on patient’s urgency need for care. There are five urgency categories defined by the Australasian Triage Scale [16]: 1) resuscitation (immediate treatment–defined as within two minutes in this study [as per the MyHospitals website]); 2) emergency (within 10 minutes); 3) urgent (within 30 minutes); 4) semi-urgent (within 60 minutes); and 5) non-urgent (within 120 minutes). Treatment within recommended time was coded as yes or no and was derived from waiting time to treatment and urgency category variables. Departing ED within four hours of arrival was defined as the time between a patient’s arrival at the ED and their physical departure from the ED. Departing ED within four hours of arrival was coded as yes or no and was derived from length of stay in ED variable.
Gender, age, triage and diagnosis were all based on VEMD variables as defined at the time of the ED presentation. Gender was categorised into three groups: male; female; and intersex. Age was categorised into 11 groups: 0–4; 5–14; 15–24; 25–34; 35–44; 45–54; 55–64; 65–74; 75–84; 85–94; and ≥95. Triage was based on the five Australasian Triage Scale [16] classification categories described above. Diagnosis was based on the best information available after the patient’s ED presentation using the International Classification of Diseases, 10th Revision, Australian Modification (ICD 10-AM) diagnosis codes [37]. Diagnosis was grouped into 23 categories (S1 Text).
Descriptive statistics of key variables were used to characterise the differences before and after the introduction of national healthcare reforms stratified by hospital peer groups. Hospitals were categorised into four groups as defined by the AIHW [38]: 1) major hospitals (principal referral); 2) large metropolitan and regional hospitals (public acute group A); 3) medium metropolitan and regional hospitals (public acute group B); and 4) small hospitals all areas (public acute group C). Segmented linear regression analyses, adjusted for demographic and clinical factors, were conducted to assess the significance of change in level and slope of the regression lines of waiting time to treatment, before and after the introduction of national healthcare reforms. Interaction effects between triage category and year of presentation were conducted. Similarly, segmented logistic regression analyses, adjusted for demographic and clinical factors, were conducted to estimate treatment within recommended time, and departing ED within four hours of arrival. All models were stratified by hospital peer groups to allow valid comparisons across similar hospitals. Residual analyses were conducted to examine the presence of serial autocorrelation. A p-value of less than 0.05 was considered statistically significant in all analyses. Data analyses were conducted using STATA version 14 (StataCorp, College Station, TX, USA). Ethical approval for this study was obtained from the Melbourne School of Population and Global Health Human Ethics Advisory Group, The University of Melbourne.
|
![[This page as RDF]](http://data.gesis.org/somesci/static/rdf-icon.gif){kind=icon}