Injury is the leading cause of disability worldwide.1,2 Post-injury burden can be further exacerbated by subsequent injuries (ie, injuries that occur after, but not necessarily because of, an earlier ‘sentinel’ injury).3–5 Subsequent injuries may also be more detrimental, both financially and physically than a sentinel injury.2,5,6 Therefore, investigating and developing injury prevention interventions and initiatives aimed at preventing subsequent injuries provides a specific avenue for reducing the overall injury burden. Despite this, there is limited knowledge of the pathways and predictors of subsequent injuries among general ‘all injury’ populations. An extensive search failed to uncover specific literature relating to subsequent injuries for Māori or other indigenous populations. However, one US study found differences in repeat trauma admissions according to ethnicity.11

Recent studies have highlighted that Māori, the indigenous population of New Zealand, experience a disproportionate burden of injury compared to non-Māori,1,3,7,8 and the health-related loss due to injury for Māori is at least twice that for non-Māori.7 Our previous research has found that Māori experience poorer outcomes, at both three and 12 months post-injury compared with non-Māori.9,10 Additionally, Māori who have been hospitalised for injury have a 70% increased risk of disability 24 months post-injury compared with non-Māori.1

New Zealand researchers and healthcare workers have particular responsibilities and requirements to address health inequities for Māori.12 These are outlined in a number of key documents, including legislation such as the Public Health and Disability Act 2000.13,14 Our injury outcomes research is underpinned by such responsibilities.12,15 The current study aims to increase our understanding of subsequent injuries, specifically for Māori, in order to identify key characteristics and potential for future interventions to improve Māori post-injury pathways.3 Injured Māori were recruited following an Accident Compensation Corporation (ACC; New Zealand’s no fault injury insurer) entitlement claim; a type of claim for injuries likely to require income compensation for more than a week off work or other additional rehabilitation such as home-help.16 The specific objectives of this paper are to describe, for the first time for injured Māori, the: 1) proportions experiencing at least one subsequent injury (SI) that involved an ACC claim, in the 12 and 24 months following their sentinel injury event, 2) frequency of SI claims over 12 and 24 months, 3) number of SI per person over 24 months, 4) time periods people are at higher risk of SI, and 5) nature of SI.

Methods

This paper uses data from the Subsequent Injury Study (SInS),3,8,17 which combined data from three sources: 1) Interview data from the Prospective Outcomes of Injury Study (POIS),12,15,18,19 a study of 2,856 injured New Zealanders (including 566 Māori), 2) administrative data from ACC for SI claims in the 24 months following each participant’s sentinel injury, and 3) hospital discharge data from the National Minimum Dataset (NMDS) for those who were hospitalised. Participants aged between 18 and 64 years were recruited following a sentinel injury event involving an ACC entitlement claim between 2007 and 2009. People who experienced injuries as a result of self-harm or sexual assault were not recruited, however SI claims of this nature were included in our analyses. Ethical approval was obtained from the New Zealand Health and Disability Multi-Region Ethics Committee (MEC/07/07/093).

The analyses presented here use data collected from the 566 Māori POIS participants, 20% of the cohort.18 During the first interview, all participants were asked to report their ethnicity using the New Zealand Census question, which allows participants to self-identify with more than one ethnic group.20 Those who identified Māori as one of their ethnic groups were included in the Māori cohort.

Information about a variety of pre-injury, injury-related and post-injury factors was collected during interviews held, on average, three, 12 and 24 months post-injury.18,21–23 Age, sex and occupation were collected from participants at the first interview using questions from the New Zealand Census.20 Participants were asked if they were in paid employment at the time of their sentinel injury, and if so they were asked about the nature of their main job, classified as professional, technical, trade/manual or unclassifiable.24 Participants were asked about their adequacy of household income25 (classified as ‘adequate’ if participants reported having ‘enough’ or ‘more than enough’ household income to meet every day needs and ‘inadequate’ if they reported ‘not enough’ or ‘just enough’).26 Additionally, participants were asked whether the sentinel injury was unintentional or intentional (ie, assault), was work-related, and if they perceived their sentinel injury to be a threat to life or long-term disability at time of injury.17,21,27 Information was collected about pre-injury depressive-type episodes (using two screening questions from the Diagnostic and Statistical Manual of Mental Disorders-III),23,28 disability (using the World Health Organization Disability Assessment Schedule II; WHODAS),29 health-related quality of life (using the EQ-5D),29,30 and alcohol use (using the brief Alcohol Use Disorders Identification Test; AUDIT-C) where participants were grouped as ‘no or low’ (males AUDIT-C score 0–4; females 0–3), ‘moderate’ (males 5–7; females 4–6) or ‘high’ (males 8–12; females 7–12).21,29–31

Information about the nature, body region and severity of sentinel injuries and SIs were derived from ACC injury diagnosis codes.21 In these analyses, 12 injury type variables, derived from the most common sentinel injury nature and body region groupings, have been used for both sentinel injuries and SIs. Participants could have had more than one injury type resulting from each injury event (both sentinel and subsequent) and could also have more than one SI event. New Injury Severity Scores (NISS) derived to measure the severity of an injury event were categorised as 1–3 (least severe), 4–6 and >6 (most severe).32,33 Hospitalisation for sentinel injuries and SIs was determined using probabilistic data linkage to the NMDS of hospital discharges,34 with participants classified as being hospitalised if they had been admitted to hospital or treated at an emergency department for ≥3 hours within seven days of the injury event.21

Sentinel injuries were all entitlement claims, however, SI events could be of any ACC claim type; categorised as ‘entitlement claims’, ‘medical fees only claims’ (whereby participants received treatment from a health professional but no additional rehabilitation support),4 ‘other claims’ (eg, those involving additional assessments), and ‘unclassified’ (those claims without a specified type, eg, where district health board bulk funding was associated).17,35 Statistical analyses are descriptive, as per the aims of this paper, and were carried out using Stata® version 14.2.36

Results

Of the 566 Māori participants, 238 (42%) had at least one SI in the 12 months after their sentinel injury event, and by 24 months, 349 (62%) had at least one SI. The mean age of those who did not have an SI in the 24-month period was 40.0 years (standard deviation (SD) 11.9) compared to 38.2 years (SD 12.7) for those who had at least one SI.

Table 1 presents pre-sentinel injury sociodemographic and health characteristics of Māori participants according to whether or not they had at least one ACC SI claim in the 24 months after their sentinel injury. In the 24-month period, those with lowest proportions of SI were: females compared to males (58% versus 63%); those in professional occupations compared to other occupations (57% versus 61–76%); and those living with family compared to those living alone or with non-family (60% versus 68% and 72%). However, there was insufficient evidence to conclude that any of these observed differences were statistically significant. Those experiencing cognitive problems prior to their sentinel injury were more likely to have a SI compared to those reporting no cognitive problems (82% versus 60%). Those reporting moderate or high alcohol use pre-sentinel injury were also more likely to have a SI compared to those reporting no or low alcohol use (69% and 62% compared to 56%, respectively).

Table 1: Pre-sentinel injury sociodemographic and health characteristics of Māori participants according to ACC-reported subsequent injury (SI) status in the 24 months after sentinel injury.

‍

*P-values obtained from chi-squared tests; tests exclude those with missing data.
Notes: Some percentages sum to more than 100, due to rounding. Results presented are descriptive only and have therefore not been adjusted for any potential confounding factors.

Those who had a sentinel injury event involving a lower extremity fracture were less likely to have a SI (44% versus 65%); as were those who had a sentinel injury event involving a lower extremity open wound (45% versus 63%) (Table 2). Participants hospitalised for their sentinel injury were also less likely to have a SI (55% versus 64%). Those whose sentinel injuries were categorised as ‘other’ were less likely to have a SI (49% versus 65%), however this category includes a mixture of different and less common injury types.

Table 2: Sentinel injury-related characteristics of Māori participants according to ACC-reported subsequent injury (SI) status.

‍

*P-values obtained from chi-squared tests; tests exclude those with missing data.
**Twelve separate variables. Participants could have more than one type of injury in a given injury event.
Notes: Some percentages sum to more than 100, due to rounding. Results presented are descriptive only and have therefore not been adjusted for any potential confounding factors.

Of the 566 participants, 349 (62%) experienced at least one SI in the 24-month period after their sentinel injury; 27% (n=152) experienced one SI event, 18% (n=103) experienced two SI, and 8% (n=47) experienced three or more SI. The total number of SI events was 755 with a range of 0 to 14 SI events per person. Four percent of the SI events resulted in hospitalisation. Overall, 12% of the SI were entitlement claims, 77% were medical fees only claims, and 11% were other or unclassified claims.

The distribution of SI increased between each of the first three three-month periods (ie, from 0–3 months to 3–6 months and from 3–6 months to 6–9 months after the sentinel injury; p<0.001). After which, the frequency of SI remained relatively stable, except for 12–15 months after the sentinel injury where the distribution of SI was very similar to the 6–9-month period and for 21–24 months where the distribution was more similar to the 3–6-month period (Figure 1).

Figure 1: Distribution of ACC-reported subsequent injury (SI) events over the 24-month period following the sentinel injury event for Māori participants.

Note: P-value for the increasing trend from 0–12 months is <0.001.

Table 3 presents information about the types of SI sustained by participants in the 24 months post-sentinel injury. The 755 SI events resulted in 962 injuries, since multiple injuries were possible from one event. Spine dislocations/sprains/strains were the most common SI type (23%), followed by lower extremity dislocations/sprains/strains (18%). Participants could also have multiple SI events of the same injury type during the 24-month period. For example, 155 SI events involved lower extremity dislocations/sprains/strains, yet these were experienced by only 114 people showing that some participants had multiple lower extremity dislocations/sprains/strains during this period (Table 3). Spine dislocations/sprains/strains had the highest number of claims per person (1.55) during the 24 months followed by both upper and lower extremity dislocations/sprains/strains (each with 1.36 claims per person) (Table 3).

Table 3: Frequency of ACC-reported subsequent injury (SI) in the 24 months following the sentinel injury event by the type of injury for Māori participants.

‍

*Each SI event could involve more than one type of injury.
**Column total sums to 842 because each SI event could involve more than one injury diagnosis of the same injury type in a single injury event (eg, an ankle dislocation and a knee sprain would be classified as two lower extremity dislocations/sprains/strains).
***This is the total number of participants who had at least one SI. Participants could have more than one SI event and each event could involve more than one type of injury. Therefore, the total percentages exceed 100%.
†Claims per person for a particular injury type is the total number of claims with that type of injury divided by the number of claimants with that type of injury.

Discussion

A considerable proportion of injured Māori had at least one SI event in the 12- and 24-month periods after their sentinel injury event (42% and 62%, respectively). We found that those who reported moderate or high alcohol use, or who were experiencing cognitive difficulties prior to their sentinel injury, were more likely to have an SI in the 24 months following a sentinel injury. High use of alcohol is a strong risk factor of injuries both worldwide37 and in New Zealand,38 so finding it to be associated with SI is not unsurprising. Cognitive impairment or difficulties can affect physical performance or function. A systematic review and meta-analysis showed that those who had experienced a sports-related concussion were at increased odds of sustaining a later musculoskeletal injury compared to those who had not sustained a concussion.39 The authors of that review pose that physical effects of concussion are often resolved more readily than other more persistent effects such as cognitive and behavioural changes, therefore those concussed require additional attention to reduce the risk of subsequent musculoskeletal injuries.

Those with a lower extremity fracture or lower extremity open wound sentinel injury were less likely to have an SI in the following 24 months, as were those who were hospitalised for their sentinel injury. These findings align with our previous non-Māori-specific work investigating self-reported subsequent injuries.4 It seems plausible that those with a lower extremity injury, or hospitalised, might be less mobile than others after such a sentinel injury and might be less able to participate in their usual activities, therefore reducing their likelihood of SIs.

The timing of SI events for Māori participants in the 24 months after a sentinel injury is similar to our observation for the entire SInS cohort17 (ie, a lower frequency of SI was observed in the first three months after a sentinel injury, this then increased for 3–6 months and then again for 6–9 months after a sentinel injury). These variations may be due to a number of reasons, for example in the earlier periods after a sentinel injury, people may not be fully participating in their everyday activities and are therefore at less risk of SI. Alternatively, people might be more careful or aware of SI after their ‘entitlement claim’ sentinel injury, which warranted additional rehabilitation and support beyond medical fees only, or may have received injury prevention information or advice at the time of their sentinel injury. One Canadian study40 investigating subsequent injuries found that the majority of such injuries among a cohort of performing artists occurred between 2–12 months after someone’s ‘return to full participation’ (measured either by a performance after their injury or the last medical treatment for their initial injury). Another study investigating the timing of subsequent injuries after being hospitalised for injury found that the highest rate ratio of subsequent injuries was 6–12 months after the sentinel injury. However, this study only examined the first subsequent injury experienced by participants, and only injuries that resulted in further hospitalisation or death.41

A strength of our study is that it addresses a topic about which very little is known, especially for Māori, a significant population group that experiences significant injury burden and injury outcome inequities.7,8,42 Our study contributes to the limited body of knowledge in this important area by describing subsequent injuries over a 24-month period for a cohort of injured Māori who have previously accessed ACC for an entitlement claim. Additional strengths include that our study is of injured Māori who have experienced a range of different types of injuries, who were both hospitalised and not hospitalised for injury and both work-related and non-work-related injuries. Limitations of our study include the descriptive nature of the analyses presented, however, this is a very complex area and so a detailed understanding of the proportions, frequencies, numbers, time periods and nature of SI is required before examining predictive factors and potential confounders, which was beyond the scope of this paper. Other limitations include that the ‘sentinel’ injury for which participants were recruited is unlikely to be their ‘first ever’ injury. Despite this, the focus our study is to examine SIs after an ACC entitlement claim injury, not to capture participants’ first ever ACC entitlement claim. While our cohort is relatively small, it appears to be the largest research cohort of injured Māori who have sustained a range of injury types, causes and severities, and importantly we have been able to use ACC information about SI claims and not rely on participants’ self-report information for these.

In this study we have analysed SIs from a cohort of Māori who had already accessed ACC for their sentinel injury. It is a limitation that we have not been able to investigate SIs for those who did not access ACC for injury. Like other health services in New Zealand, Māori experience barriers to accessing ACC.43 Importantly though, our findings show that there are differences in pre-sentinel and sentinel injury characteristics for Māori who experienced a SI. This highlights the importance of understanding SIs for Māori, particularly as those who do access ACC present a potential point for future interventions.

We have described the proportions, frequencies, nature and time periods of subsequent injuries for Māori. Despite their descriptive nature, our findings point to both the complexity of SI among an ‘all injury’ Māori cohort, and to the need for a greater research, ACC and health service focus on SI if the burden of injury for Māori is truly to be addressed. That 62% of Māori who had already experienced a profound sentinel injury went on to experience at least one SI reported to ACC within a period of only 24 months suggests both that the burden is considerable and that preventive opportunities are being missed. Additional analyses are currently underway by our team to investigate factors predicting SI, while accounting for potential confounders, in order to assist in the development of SI preventive initiatives at multiple points in the complex post-injury pathway specifically for injured Māori.

‍