**1. Introduction**

The immediate period following hospital discharge after acute traumatic injury holds substantial risk for ongoing health, recovery and welfare concerns [1]. Ongoing challenges can include physical limitations, continuing pain and the experience of secondary conditions. These factors can impair participation in previous activities, complicate or hamper the return to work [2] and contribute to financial stress, relationship strain and social exclusion. These factors predispose individuals to deterioration in mental health [3,4]; further impeding recovery from injury. Direct and measurable costs to an injured worker incorporate the loss or change in employment earnings, in addition to legal and medical costs. Indirect and less easily measured costs include pain, impaired function, reduced quality of life, potential for chronic injury and stress on interpersonal relationships [5,6]. Harmful substance use has previously been reported as high in injured populations, and likely compounds this multitude of issues [7,8]. Context-specific costs to the healthcare service for readmissions after acute injury are less well described, yet important to measure.

The incidence and cost of traumatic spinal injuries (TSI) sustained in workplaces in New South Wales (NSW) was recently estimated for the first time. Over a 3-year period, 824 persons sustained TSI in work-related incidents, occupying 13,302 acute care bed days and costing a total of \$19,500,000 (95%CI \$16 M–\$23 M) [9]. The total cost of work-related injury and disease in Australia was estimated at \$61.8 billion in 2012–2013, of which NSW bore 28% (\$17.3 billion) of the total cost from 31% of cases nationwide [10]. Injuries sustained in this cohort comprised column fractures, spinal cord injury or both; 21% of persons also sustained concomitant head, chest or abdominal traumatic injury. High numbers of these injuries occurred in the construction industry, particularly due to falling from height. This study did not report on any incidence, nature and cost of readmissions to hospital following the acute care period.

Gabbe and Nunn recently reported 40% of patients after traumatic spinal cord injury to experience readmissions within the first 2 years in Victoria [11]. Similar proportions were described by Ruseckaite et al. [12] studying a cohort of patients injured in compensable transport incidents. Over one third of these patients experienced acute care facility readmission within 28 days of injury. Investigating unplanned readmissions in certain injury populations should be routinely undertaken to benchmark discharge planning efforts and therefore assist trauma care coordination.

Unplanned hospital readmissions occurrences within 28 days post the acute-care episode are progressively being used in various jurisdictions internationally as a measure of the quality of hospital care or treatment. In some systems, rates of unplanned readmissions are being used as an indicator of hospital performance that is then linked with funding reimbursements. Previous analysis in NSW has indicated that at least one quarter of unplanned readmissions are linked to "deficiencies in care" [13]. In 2013–2014, the NSW average for unplanned readmissions across all conditions and acute care facilities was 6.8 per cent; higher than the NSW 2021 target of 5.5 per cent [13].

The burden of TSI occurring in workplace settings is of particular interest for national injury prevention bodies, such as Safe Work Australia, who have prioritised a national target of a 30% reduction in serious work-related injury compensation claims by 2022 [14]. Quantification of the extended burden of TSI within the acute care setting complements the evidence informing injury prevention efforts in occupational health and safety, describing more specifically the extent of enduring disability associated with work-related injury.

Therefore, the aim of this population-based study was to measure the post-acute care burden in a cohort of hospitalised traumatic spinal column and cord injuries that occurred while "working for an income" in NSW, Australia. Specifically, following those patients previously identified as sustaining a work-related TSI [9], we describe the incidence of 28- and 90-day readmissions, subsequent to their discharge from acute hospitalisation, investigating readmission etiology and healthcare system costs.

#### **2. Methods**

The epidemiology and occupational context of persons who sustained a TSI while "working for an income" in NSW has been previously and fully reported [9]. The Centre for Health Record Linkage (CHeReL) linked NSW Admitted Patient Data Collection (APDC), Registry of Births, Deaths and Marriage and Activity Based Funding (ABF) costing records for all people aged ≥16 years who were hospitalised between 1 June 2013 and 30 June 2016 with a TSI recorded in their index admission due to a work-related incident defined as ICD10-AM [15] code U73.0 or funding by workers compensation in the index admissions. Spinal injuries included all spinal cord and/or column injuries, defined using specific ICD-10-AM codes (Appendix A). NSW is the most populous Australian state with approximately 7.5 million inhabitants residing across 800,000 km2 in suburban, rural and very remote areas [16].

For the current study, readmissions were identified from APDC records linked with these index admissions. We defined hospital readmissions as being within 28 and 90 days after discharge with a primary diagnosis code related to the index admission. This benchmark was chosen in accordance with definitions from the NSW Bureau of Health Information [17], and the Independent Price Health Authority [18], who determine that readmissions to hospital within 28 days of discharge should be reviewed and considered as "potentially avoidable". These were identified by manual review of the primary diagnosis codes of all admissions within 28 and 90 days, respectively. Readmissions within the first 28 days were consequently nested within those identified within 90 days.

Estimation of the costs of readmission was based on the formula used to calculate the National Weighted Activity Units (NWAUs) assigned to each ABF activity [19]. Weighted Activity Units (WAU) are the weights assigned per hospital separation; we used the 2013/2014 National Pricing Model technical specifications published by the Independent Hospital Pricing Authority [19] (for compatibility with AR-DRG version 6.0x). The WAUs in this study were adjusted for private patient service and private patient accommodation adjustments; a slight difference to NWAUs to avoid distortion by differential funding of public and compensable patients. All patients were assumed to be funded by Medicare for the WAU estimation. The "per separation" cost was defined as the WAU multiplied by the National Efficient Price 2013/2014 (\$4993); with higher WAUs thus being more resource intensive.

Descriptive statistics were used to report the prevalence of various factors. Values were reported as mean and standard deviation (SD) for normally distributed continuous variables, proportions for categorical variables and median and interquartile range (IQR) for non-normally distributed continuous variables. All statistical analyses were performed using Stata version 15.0 (Stata Corporation, College Station, TX, USA). Standardised reporting of demographic and other variables, as recommended by De Vivo et al. [20], was followed where possible.

This study was approved by the Cancer Institute NSW, Population & Health Services Research Ethics Committee: AU RED Reference: HREC/16/CIPHS/19, Cancer Institute NSW reference number: 2016/07/647.

#### **3. Results**

Between 1 June 2013 and 30 June 2016, 824 individuals sustained a traumatic spinal injury in NSW while working for an income, of which 740 (89.8%) had sufficient follow-up data to analyse readmissions <90 days post-acute hospital discharge (up to 1 April 2016) and were included in the analyses. Of these 740 individuals, 61 (8.2%) experienced a total of 119 (16.1%) readmissions in the first 28 days after their acute care discharge. By 90 days after discharge from acute care, 102 (13.8%) individuals had experienced 250 readmissions.

Characteristics of the primary injury cohort compared with individuals requiring readmissions within 28 and 90 days are described in Table 1. The cohort were predominantly male (86.2%); those aged 45–59 years were the largest age group (34.7%), this was also true for the readmitted proportions (37.7% and 36.3% for 28- and 90-day readmissions, respectively). Almost half of original work-related injuries had occurred as a result of a fall in their workplace (49.2%); over half of these (55.1%) were falls from building structures, scaffolding or ladders.


