Death and serious injury involving transport are important metrics for monitoring and evaluating road safety initiatives. New Zealand has a higher rate of road fatalities than comparable countries and, after a downward trend in fatalities and serious injuries since the mid-1980s, the number of people killed or seriously injured on New Zealand roads has been increasing since 2013.
Full article available to subscribers
Death and serious injury involving transport are important metrics for monitoring and evaluating road safety initiatives. New Zealand has a higher rate of road fatalities than comparable countries1 and, after a downward trend in fatalities and serious injuries since the mid-1980s, the number of people killed or seriously injured on New Zealand roads has been increasing since 2013.2 To reverse this trend, investment in road safety strategies have been launched with the aim to reduce annual deaths and serious injury caused by road transport.3 The focus of these strategies are motorised modes of personal transport such as cars and motorbikes, and also public transport, cycling and walking. Policies are influenced by the successes observed in countries such as the Netherlands and Sweden, which have changed approaches from being an individual’s responsibility to take a systems-approach to addressing road safety issues. The “safe system” approach is focussed on three principles: people make mistakes, roads and vehicles need to be designed to minimise the impact of crashes and road safety is a shared responsibility.4
Accurate and comprehensive data is important to monitoring performance against targets, evaluating road safety interventions and setting future priorities. However, complete transport injury information is not centralised in any single source. While serious car crashes in New Zealand are typically documented in the Crash Analysis System,5 moderate and minor road incidents, particularly those involving a non-motorised mode of transport such as cycling and walking, are not always captured by police.6
Covering a 4-year period using data from hospital admission and the New Zealand Trauma Registry (NZTR), the present study investigated temporal trends in injury-causing crashes in New Zealand and explored whether specific population groups have an increased incidence of injury. These trends provide insight into changes in injury patterns from road trauma.
The study period included the COVID-19 pandemic and resultant changes to daily life following the introduction of the COVID-19 Alert System. This study explored the impact of the lockdown on injury rates during and after the restrictions.
This observational study of hospitalised patients in New Zealand used existing data from national collections and no additional information was obtained from individuals. The study covers the period between 1 July 2017 to 30 June 2021.
Hospitalisations were identified within the Hospital Events National Minimum Dataset (NMDS),7 a national collection of all hospital admission data. Using ICD-10-AM8 external cause codes, which are linked to individual hospital events, admissions coded as being caused by road traffic injury were identified. Road traffic injury was defined as “traffic” codes in the range V010 to V899, as well as a small number of X and Y codes. Non-traffic codes (e.g., mountain biking) were excluded.
Events were excluded if the patient was discharged home from the emergency department, or the patient was admitted to hospital over 14 days after the crash, or the admission was arranged and the clinical specialty was Maternity Services, or no overnight stay occurred, or the clinical specialty was Mental Health Services or arranged or waitlist.
Mode of transport was identified using the ICD-10-AM coding in the NMDS and categorised into car (driver or passenger, and passenger cars only), motorcycle (driver or passenger), pedal cycle and pedestrian. The last category was “Other”, and included any mechanised vehicle not classified as a car such as a truck, van, light or bus, and incidents involving horses or other animals on a road.
Major trauma incidents were linked to the hospitalisation dataset using National Health Index (NHI) in the NZTR,9 a population-based registry that collects data on all major trauma patients who have a threat to life and are admitted to acute hospitals. Matching cases between the NMDS and the NZTR enabled stratification analyses by major and non-major trauma.
Statistics New Zealand population projections for relevant periods and for all population sub-groups and rural/urban classifications were used. The 2013 Census estimated population for Māori was used as the reference population.
Prioritised ethnicity data associated with the NHI was used.
Rural/urban classification is based on University of Otago’s Geographic Classification of Health10 using the patient’s domicile code of residence and Census data. This does not reflect where the injury occurred.
The study period included the COVID-19 pandemic and the introduction of the lockdowns that were first put in place from 26 March 2020 to 27 April 2020, with strong restrictions on transport. Legally, individuals had to stay home other than for essential personal movement. These restrictions resulted in a dramatic decrease in traffic volume to around 15% of usual rates during the lockdown period.11 Auckland city was subject to further lockdowns at other periods during the pandemic.
The Health and Disability Ethics Committee approved this study (2022 EXP 12993) and the Data Governance Groups of the NZTR and Manatū Hauora – Ministry of Health approved the use of their respective data.
Age standardisation was applied to all ethnicity rates, with 95% confidence intervals calculated using the Dobson method.12 Incidence rates were calculated using the 2022 Census data and presented as an event rate per 100,000 population, as robust exposure data were unavailable for all modes of transport. Temporal trends were not analysed for pedestrians and cyclists due to insufficient numbers.
Due to the disruption to usual transport activity by the COVID-19 restrictions in 2020, an interrupted time series analysis (ITSA) was used to determine changes in the number of incidents over the study period.13 The ITSA models accounted for seasonality and were compared to the counterfactual to investigate if the disruption altered the post-COVID trends. Two models were used for the modes of transport with sufficient incidents to analyse: one for motorcycle, and one for car occupant. Further details are described in the Appendices. Analysis was performed using R statistical software14 and statistical significance was defined as p <0.05.
Over the 4-year period there were 20,607 incidents of transport-related road trauma that resulted in a hospital admission in New Zealand. Of these, 14.5% (2,992) involved injuries that were classified as major trauma.
Car occupants accounted for 62% of incidents, followed by motorcyclists (23%), pedestrians (9%) and pedal cyclists (4%) (Table 1). Males made up 63% of the cohort and the median age was 38 years (IQR: 25, 59). Those of European/other ethnicities made up 62% of incidents, followed by Māori (25%), Asian peoples (7.4%) and Pacific peoples (6.2%).
View Table 1–3, Figure 1–3.
A sharp decline in deaths and serious injury in early 2020 was observed and coincided with the period of the COVID-19 Alert Level 4 lockdown, which was in place from March 2020 to April 2020 (Figure 1).
The incidence rate of all transport-related hospitalisation was lowest in 2019/2020, at 96.6 per 100,000 people (95% CI: 93.6–99.4), although motorcycle incidents were highest in 2020/2021 than in any other period prior (Table 2).
The rate of transport-related hospitalisation varied between age groups within different modes of transport. Those aged 80 years and older had the highest rate of incidents as car occupants, followed by those aged 20–24 years. Pedal cyclist incidents were highest for those aged 55–59 years. An increase in the incidence rate of motorcycling injuries was observed for males aged 10–19 and 60–69 years (Appendix Table 1).
Asian peoples had a substantially lower rate of transport-related hospitalisation than other ethnicity groups, with around half the incidence rate of NZ European/other for both injury severity categories (Table 3). Māori had an age-standardised incidence rate almost 3.5 times higher than the rate for Asian peoples.
Incidence rates for road trauma hospitalisation were higher in people living in rural areas relative to people living in urban areas. People living in the remotest rural areas (rural 3) had over twice the rate of hospitalisation (198.3, 95% CI: 179.7–218.3) than those living in the densest urban areas (92.0, 95% CI: 90.3–93.7; Table 3).
The seasonal model best captures the car injury hospitalisations patterns for both the pre-COVID and post-COVID periods (refer red line predictions in Figure 2). Again, the sharp drop-off of car injury hospitalisations at the start of the COVID-19 period appeared to be only a minor interruption as the expected pre-COVID seasonal pattern returned.
During the pre-COVID period, there was no change over time in the number of car occupant hospitalisation injuries (IRR= 1.000, 95% CI: 0.999–1.000, p= 0.326). The car model detected a fall in the rate of car injury hospitalisations of between 3% to 68% into the initial week of COVID-19 (IRR= 0.565, 95% CI: 0.320–0.997, p= 0.049). However, the average car rate returned to the counterfactual by first week into the post-COVID period, with no difference between the post-COVID and counterfactual detected (IRR= 1.032, 95% CI: 0.918–1.161, p= 0.595) (Appendix Table 2).
The seasonal model captured the seasonal pattern in motorcycle hospitalisation counts for both the pre-COVID and post-COVID periods, as seen by the predictions in Figure 3. The sharp drop-off of motorcycle-related hospitalisations at the start of the COVID-19 period appeared to be only a minor interruption as the expected pre-COVID seasonal pattern returned, albeit initially lower than the counterfactual model, but returned by the end of the period.
During the pre-COVID period there was an increase in the number of weekly motorcycle injury hospitalisations (IRR= 1.001, 95% CI: 1.000–1.002, p= 0.009). As expected, the COVID-19 period was associated with a large decrease in the count of motorcycle injury hospitalisations into the first week of between 33% to 83% (IRR= 0.334, 95% CI: 0.167–0.667, p= 0.002), which was an initial deviation from the counterfactual. However, post-COVID, motorcycle hospitalisations were similar to the predicted rate using data from a pre-COVID period. (Appendix Table 3).
This study has highlighted the difference in rates across modes, and across ethnicity groups, and has described the impact of COVID-19. The results are unique to the New Zealand context, although many results are consistent with the findings in international studies of this nature.15–17
The rate of motorcycle injury hospitalisations increased over the study period and were over 10% higher in 2020/2021 than in 2017/2018, and largely occurred in men. This is despite investment in initiatives to reduce motorcycle injuries, such as rider competency training.18 The incidence rate is nearly double for motorcycle injury for those who live rurally compared to urban domicile, but to the best of our knowledge there is scant research to understand why the difference occurs. The results suggest two quite distinct demographic groups who are being injured on motorcycles: males aged between 10–19 and 60–69 years, where significant increases in the rate of injury were observed, and males aged between 20–59 that had high rates of injury. Motorcycle exposure data shows a reduction of 18 million kilometres over the 4 years of this study.19 This suggests that while motorcycling injuries are increasing, that increase is not necessarily associated with increased exposure.
The converse pattern is observed for pedal cyclists. The number of cycling injuries has remained stable, but the number of cyclists has increased, at least in the urban areas on key cycling routes.20–22 One possible explanation is that the infrastructure spending by local councils and the transport agencies to build dedicated cycle lanes separated from cars and other vehicles may be successful in reducing cycling injuries. Another possibility is that cycling injuries are not being captured well in the data, which is discussed further below.
There is distinct variation in the patterns of transport-related injury by ethnicity. We found that Asian peoples have a much lower rate of transport-related road trauma. There is a dearth of research to provide insight into what influences among Asian peoples lead to this low rate of injury. While Asian peoples may be more urban than the total population, this does not explain the difference, as both major and non-major trauma rates in Asian peoples are approximately half of the rates in the urban 1 category.
A study by Randal et al. showed a faster rate of increase of transport-related deaths and serious injury for Māori than non-Māori between 2014 and 2017.23 Our study shows this upward trend has continued unabated over the subsequent 4 years after their study. There is a substantial long-term negative impact on Māori health at a similar level due to the effects of tobacco and obesity.24 The relationship between age, rurality and Māori ethnicity indicate there are specific geographical locations where serious injury is particularly high.
The differences of hospitalisation rate across ethnicities, sex and age groups provide some impetus to the aims of Road to Zero. Given the evidence that there is variation between some groups within the population on the magnitude of the 40% target, this implies that the aims are achievable in principle, and that a focus on reducing inequities is a worthy endeavour.
The use of ICD-10-AM external cause codes alone appears to be insufficient for identifying all relevant hospitalisations that occurred during the study period. The coding system distinguishes between “traffic” and “non-traffic” events, and analysis suggests hospitalised pedal cyclists and pedestrians not involved in a collision with another vehicle are often coded as “non-traffic” events, even if the incident occurred on a public road. All but one of the 910 pedal cycle incidents captured in this study have an external cause code indicating a collision with another vehicle. The NMDS collects information on place of occurrence (which can be identified as a road/street or highway) of injury and external cause codes for falls, which taken together could be used to identify transport incidents that have been coded as “non-traffic” in future studies. Previous work investigating pedestrian injuries in Victoria, Australia using this method identified approximately 65 times more pedestrian falls than were present in the Victorian Police report Road Crash Data.25
Better exposure data, and agreement on how to interpret that data, is needed for all modes of transport to understand whether changes over time can be attributed to changes in usage. Differences between modes with respect to the number of kilometres, travel time, number of trips and vehicle registrations contribute to analysis of exposure rates. Notwithstanding this limitation, trend analysis, such as analysed in this paper, is important particularly in the context of reducing deaths and serious injuries for all road users.
This study has highlighted critical differences in road trauma rates and explored trends over time. The increases in motorcycling injuries and no changes in car, pedestrian and cycling injuries, as well as demographic variation, highlight the need to focus on vulnerable road users. Effective and targeted initiatives for vulnerable road users will support the aims to reduce deaths and serious injury on New Zealand roads. More clearly defined exposure data is needed to provide an accurate denominator. Linked data across population-based datasets is an important asset that enhances our understanding of road traffic injuries and allows evidence-based countermeasures to be developed.
View Appendices.
This observational study aimed to investigate temporal trends in transport-related injuries in New Zealand by mode of transport and explore whether specific population groups and localities have a relatively higher incidence of injury. These trends provide insight into changes in injury patterns from road trauma.
A retrospective study of hospitalised road trauma in New Zealand was conducted between 1 July 2017 to 30 June 2021. Data were obtained from the National Minimum Dataset of hospital admissions and the New Zealand Trauma Registry (NZTR). Road trauma was identified using ICD-10 coding, and major trauma using Abbreviated Injury Scale (AIS) coding. Analysis included road trauma by mode, ethnicity, rurality and population rates. Statistical analysis included Interrupted Time Series (ITS) analysis to account for the impact of COVID-19 on road trauma.
Over the 4-year period there were 20,607 incidents of transport-related injury that resulted in admission to a New Zealand hospital. Of these, 14.5% (2,992) involved injuries that were classified as major trauma. Car occupants accounted for 62% of hospitalisations, followed by motorcyclists (23%), pedestrians (9%) and pedal cyclists (4%). Temporal trends showed no reduction in injuries from cars, pedal cyclists and pedestrian injuries, but an increase in motorcycling injuries. Māori had an age-standardised incidence rate almost 3.5 times higher than the rate for Asian peoples.
The increases in motorcycling injuries and no changes in pedestrian and cycling injuries, as well as demographic variation, highlight the need to focus on vulnerable road users. Effective and targeted initiatives on vulnerable road users will support objectives to reduce deaths and serious injury on New Zealand roads. Enhanced exposure data is needed for vulnerable road users to account for mobility changes over time. Linked data across population-based datasets is an important asset that enhances our understanding of road traffic injuries and allows evidence-based countermeasures to be developed.
Siobhan Isles: Independent Researcher.
Michael Kean: Data Analyst, Te Tāhū Hauora Health Quality & Safety Commission, New Zealand.
Joanna F Dipnall: Senior Research Fellow & Lecturer, Pre-hospital, Emergency and Trauma Research, School of Public Health and Preventive Medicine, Monash University, Melbourne Victoria; Honorary Associate Professor, School of Medicine, Deakin University, Geelong, Victoria, Australia.
Ben Beck: Head of Sustainable Mobility and Safety Research, School of Public Health and Preventive Medicine, Monash University, Australia.
Siobhan Isles: Independent Researcher. Ph: 021 578 654.
Nil.
1) World Health Organization. Global Status Report on Road Safety 2018 [Internet]. Geneva (CH): World Health Organization; 2018 [cited 2023 Jun 9]. Available from; https://apps.who.int/iris/handle/10665/276462.
2) Ministry of Transport. Safety — Road deaths [Internet]. NZ: Ministry of Transport; 2023 [cited 2023 Apr 21]. Available from: https://www.transport.govt.nz/statistics-and-insights/safety-road-deaths/.
3) Waka Kotahi NZ Transport Agency. Road to Zero [Internet]. NZ: Waka Kotahi NZ Transport Agency; 2023 [cited 2023 Apr 19]. Available from: https://www.nzta.govt.nz/safety/what-waka-kotahi-is-doing/nz-road-safety-strategy/.
4) International Transport Forum. Zero Road Deaths and Serious Injuries: Leading a Paradigm Shift to a Safe System. OECD; 2016. doi:10.1787/9789282108055-en.
5) Waka Kotahi NZ Transport Agency. Crash Analysis System (CAS) [Internet]. NZ: Waka Kotahi NZ Transport Agency; 2023 [cited 2023 Jun 9]. Available from: https://www.nzta.govt.nz/safety/partners/crash-analysis-system/.
6) Koorey G, Sim P, Diorni G. How do we measure harm in transport? [Internet]. NZ: Transportation Group New Zealand; 2023 [cited 2023 Jun 9]. Available from: https://viastrada.nz/sites/default/files/2023-04/Transportation2023-ThinkPiece-KooreyEtAl-revised.pdf.
7) Te Whatu Ora – Health New Zealand. National Minimum Dataset (hospital events) [Internet]. Wellington (NZ): Ministry of Health; 2023 [cited 2013 Apr 21]. Available from: https://www.health.govt.nz/nz-health-statistics/national-collections-and-surveys/collections/national-minimum-dataset-hospital-events.
8) Te Whatu Ora – Health New Zealand. Introduction to Classification – ICD-10-AM/ACHI/ACS [Internet]. Wellington (NZ): Ministry of Health; 2023 [cited 2013 Apr 21]. Available from: https://www.health.govt.nz/nz-health-statistics/classification-and-terminology/icd-10-am-achi-acs.
9) National Trauma Network. New Zealand Trauma Registry [Internet]. Wellington (NZ): National Trauma Network; 2024 [cited 2023 Apr 21]. Available from: https://www.majortrauma.nz/nz-mtr/.
10) Whithead J, Davie G, de Graaf B, et al. Defining rural in Aotearoa New Zealand: a novel geographic classification for health purposes. N Z Med J. 2022;135(1559):24-40.
11) Automobile Association. Road report: The impact Covid-19 had on motoring in New Zealand [Internet]. NZ: AA; 2020 [cited 2023 Apr 21]. Available from: https://www.aa.co.nz/membership/aa-directions/driver/road-report-the-impact-covid-19-had-on-motoring-in-new-zealand/.
12) Dobson AJ, Kuulasmaa K, Eberle E, Scherer J. Confidence intervals for weighted sums of Poisson parameters. Stat Med. 1991;10(3):457-462. doi: 10.1002/sim.4780100317.
13) Bernal JL, Cummins S, Gasparrini A. Interrupted time series regression for the evaluation of public health interventions: a tutorial. Int J Epidemiol. 2017;46(1):358-355. doi: 10.1093/ije/dyw098.
14) R Core Team. R: A language and environment for statistical computing. R Foundation for Statistical Computing [Internet]. Vienna (AT): R Foundation; 2024 [cited 2023 Jun 9]. Available from: https://www.R-project.org/.
15) Leijdesdorff HA, Gillissen S, Schipper IB, Krijnen P. Injury Pattern and Injury Severity of In-Hospital Deceased Road Traffic Accident Victims in The Netherlands: Dutch Road Traffic Accidents Fatalities. World J Surg. 2020;44(5):1470-1477. doi: 10.1007/s00268-019-05348-6.
16) Fagerlind H, Harvey L, Humburg P, et al. Identifying individual-based injury patterns in multi-trauma road users by using an association rule mining method. Accid Anal Prev. 2022;164:106479. doi: 10.1016/j.aap.2021.106479.
17) Koch DA, Hagebusch P, Lefering R, et al. Changes in injury patterns, injury severity and hospital mortality in motorized vehicle accidents: a retrospective, cross-sectional, multicenter study with 19,225 cases derived from the TraumaRegister DGU®. Eur J Trauma Emerg Surg. 2023;49(4):1917-1925. doi: 10.1007/s00068-023-02257-1.
18) Ride Forever [Internet]. NZ: Ride Forever; 2024 [cited 2014 Apr 19]. Available from: https://www.rideforever.co.nz/.
19) Te Manatū Waka Ministry of Transport. RD017 Vehicle kilometers travelled by motorcycle [Internet]. NZ: Ministry of Transport; 2021/2017 [cited 2023 Apr 19]. Available from: https://www.transport.govt.nz/statistics-and-insights/road-transport/.
20) Cycling in Christchurch. Christchurch Cycling Counter [Internet]. Christchurch (NZ): Cycling in Christchurch; 2022 [cited 2023 Apr 19]. Available from: http://cyclingchristchurch.co.nz/topic/cycle-numbers/.
21) Bike There. Monthly biking stats [Internet]. Wellington (NZ): Wellington City Council; 2023 [cited 2023 Apr 19]. Available from: https://www.bikethere.org.nz/monthly-biking-stats/.
22) Auckland Transport. Monthly cycle monitoring, 2023 monthly cycle monitoring report [Internet]. Auckland (NZ): Auckland Transport; 2023 [cited 2023 Apr 21]. Available from: https://at.govt.nz/cycling-walking/research-monitoring/monthly-cycle-monitoring/.
23) Waka Kotahi NZ Transport Agency. Māori road safety outcomes report [Internet]. NZ: Waka Kotahi NZ Transport Agency; 2021 [cited 2023 Apr 21]. Available from: https://www.nzta.govt.nz/resources/maori-road-safety-outcomes-report/.
24) Randal E, Shaw C, McLeod M, et al. The Impact of Transport on Population Health and Health Equity for Māori in Aotearoa New Zealand: A Prospective Burden of Disease Study. Int J Environ Res Public Health. 2022;19(4):2032. doi: 10.3390/ijerph19042032.
25) Oxley J, O’Hern S, Burtt D, Rossiter B. Falling while walking: A hidden contributor to pedestrian injury. Accid Anal Prev. 2018;114:77-82. doi: 10.1016/j.aap.2017.01.010.
26) Zeileis A, Kleiber C, Jackman S. Regression Models for Count Data in R. J Stat Softw. 2008;27(8). doi: 10.18637/jss.v027.i08.
27) Bhaskaran K, Gasparrini A, Hajat S, et al. Time series regression studies in environmental epidemiology. Int J Epidemiol. 2013;42(4):1187-1195. doi: 10.1093/ije/dyt092.
28) Stock JH, Watson MW. Introduction to Econometrics. 3rd ed. US: Addison-Wesley; 2011.
29) Nelson M. Tests and Confidence Intervals on Directly Standardized Rates for Several Methods [Internet]. R package version 1.0.0; 2022. [cited 2023 Jun 9]. Available from: https://CRAN.R-project.org/package=dsrTest.
30) Hothorn T, Bretz F, Westfall P. Simultaneous Inference in General Parametric Models. Biom J. 2008;50(3):346-363. doi: 10.1002/bimj.200810425.
Sign in to view your account and access
the latest publications by the NZMJ.
Don't have an account?
Let's get started with creating an account.
Already have an account?
Become a member to enjoy unlimited digital access and support the ongoing publication of the New Zealand Medical Journal.
The New Zealand Medical Journal is fully available to individual subscribers and does not incur a subscription fee. This applies to both New Zealand and international subscribers. Institutions are encouraged to subscribe. The value of institutional subscriptions is essential to the NZMJ, as supporting a reputable medical journal demonstrates an institution’s commitment to academic excellence and professional development. By continuing to pay for a subscription, institutions signal their support for valuable medical research and contribute to the journal's continued success.
Please email us at nzmj@pmagroup.co.nz