Article Text

Download PDFPDF

Prospective Outcomes of Injury Study: recruitment, and participant characteristics, health and disability status
  1. Sarah Derrett1,
  2. Gabrielle Davie1,
  3. Shanthi Ameratunga2,
  4. Emma Wyeth3,
  5. Sarah Colhoun1,
  6. Suzanne Wilson1,
  7. Ari Samaranayaka1,
  8. Rebbecca Lilley1,
  9. Brendan Hokowhitu4,
  10. Paul Hansen5,
  11. John Langley1
  1. 1Department of Preventive and Social Medicine, Injury Prevention Research Unit, Dunedin School of Medicine, University of Otago, Dunedin, New Zealand
  2. 2School of Population Health, Faculty of Medical and Health Sciences, University of Auckland, Auckland, New Zealand
  3. 3Department of Preventive and Social Medicine, Te Roopū Rakahau Hauora Māori a Kāi Tahu (Ngāi Tahu Māori Health Research Unit), Dunedin School of Medicine, University of Otago, Dunedin, New Zealand
  4. 4Te Tumu-School of Māori, Pacific and Indigenous Studies, University of Otago, Dunedin, New Zealand
  5. 5Department of Economics, University of Otago, Dunedin, New Zealand
  1. Correspondence to Dr Sarah Derrett, Department of Preventive and Social Medicine, Injury Prevention Research Unit, Dunedin School of Medicine, PO Box 56, University of Otago, Dunedin, New Zealand; sarah.derrett{at}


The Prospective Outcomes of Injury Study aims to identify predictors of disability following injury. Participants were selected from the entitlement claims register of New Zealand's no-fault compensation insurer, the Accident Compensation Corporation, and followed up by interview for 2 years. This report describes changes to intended Prospective Outcomes of Injury Study methods and key characteristics of the cohort, with an emphasis on general health and disability before injury and soon afterwards. There were 2856 injured participants in the first interview, which occurred 3.2 months (median) after injury. The recruitment period was extended to enable inclusion of sufficient Māori participants. At the first interview, most participants were experiencing worse health status and increased disability compared to before injury, despite less than one-third reporting admission to hospital because of their injury. Analysis of outcome predictors related to post-injury function, disability and return-to-work soon after injury and 1 year later is now under way.

  • Injuries
  • outcome assessment
  • cohort studies
  • prospective studies
  • methods
  • outcome evaluation
  • disability
  • functional outcome
  • longitudinal
  • methodology
  • outcome of injury
  • mechanism
  • risk/determinants
  • injury diagnosis
  • health services
  • international
  • MVTC
  • disability
  • systematic review
  • public health
  • race
  • qualitative
  • aboriginal
  • disability

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


The Prospective Outcomes of Injury Study (POIS) aims to identify predictors of disability following injury.1 Injured participants were randomly selected from the entitlement claims register of New Zealand's no-fault compensation insurer, the Accident Compensation Corporation (ACC), and followed up for 2 years. People on this register have injuries that may entitle them to, for example, compensation while away from paid employment, rehabilitation, treatment costs and home help. ACC becomes aware of these injuries following hospitalisation, emergency department presentation or consultation with general practitioners, physiotherapists and other ACC-authorised health professionals.

This brief report aims to describe (1) changes to the previously reported intended methods of POIS and (2) key characteristics of the cohort, with an emphasis on general health and disability before injury and soon (3 months) afterwards.


On behalf of the POIS research team, each month, ACC invited (by post) people on their entitlement claims register to participate. If people did not decline further contact, ACC provided the POIS team with their names, brief injury description and contact details. If consent was given, participants were interviewed by telephone or face-to-face. Changes in recruitment and data collection circumstances necessitated some changes to the intended methods, as summarised below.

To be eligible, participants had to reside within one of four regions of New Zealand.1 Prior to commencing the study, another large urban region—Auckland City—was added at the suggestion of ACC. First interviews (Interview 1) were to take place 1 month after injury; however, they occurred, on average, 3 months (median=3.2 months) after injury. Although it had been intended that no more than 1 month would elapse between the injury event and ACC invitation, this threshold was extended to 4 months. This was primarily to enable people with a range of injury severities to be included within the study, as people admitted to hospital for lengthy periods are not necessarily placed on ACC's register immediately. Each month between December 2007 and June 2009, contact details for potential participants were provided by ACC. Twelve interviewers (on average) attempted to make telephone contact with, and interview, 50 people each per month. Those who could not be contacted after five attempts were sent postal questionnaires, adding to delays.

Recruitment was expected to take 10 months, but took 19 months. Recruitment was slow during the two summer holiday periods, but the main delay arose from the need to recruit a sufficient number of Māori (New Zealand's indigenous population).2 POIS had been explicitly designed to address injury outcomes for New Zealanders, including Māori specifically. At least 460 Māori were required for independent analysis. The potential sample of Māori was reduced because some had already participated in another ACC survey (ACC has a policy of inviting claimants to only one survey per year to manage possible participant fatigue). In addition, an ACC database error in recording ethnicity led to fewer potential Māori participants being identified. An unintended consequence of the extended recruitment period was that more participants were recruited in total (n=2856, including n=566 Māori) than originally intended (n=2500).

Interviews included questions on a variety of pre-injury, injury and post-injury characteristics. Self-reported injury region and type were categorised according to a modified version of the Barell matrix.3 These included measures of general health status (EQ-5D) and disability (WHO Disability Assessment Schedule II 12-item (WHODAS)), which are the focus of the analysis in this brief report. The EQ-5D measures health on five dimensions (Mobility, Self-Care, Usual Activities, Pain/Discomfort and Anxiety/Depression),4 and an additional question about Cognition was asked. For analysis, responses were grouped as ‘No’ or ‘Any Problems’. The EQ-5D also contains a 0–100 visual analogue scale (VAS) of overall health state where 0 corresponds to the worst health state imaginable, and 100, to the best. The WHODAS contains 12 questions, each with five level-of-difficulty responses, which were grouped as ‘No’ or ‘Any Difficulty’ for analysis.5 A WHODAS disability score was derived using the summed approach (where 0=no disability and 48=most disability).6 When one WHODAS item was missing, the individual's average was imputed; when more items were missing, the disability score was not derived.5 Pre-injury and post-injury EQ-5D problems and WHODAS difficulty on each dimension were compared using χ2 tests, and EQ-5D VAS scores pre-injury and post-injury were compared via paired t tests using Stata version 11.7


ACC invited 7875 people to participate in POIS, of whom 2994 were unable to be contacted by us after at least five attempts and did not return a postal questionnaire (figure 1). Of the 4881 able to be contacted, 2025 either declined to participate or were ineligible because they completed Interview 1 more than 6 months after injury, failed to answer at least 70% of the questions or did not speak sufficient English or Māori. Thus, there were 2856 in the POIS cohort, corresponding to a response proportion of 36% and participation proportion of 59%.8 Most (89%) participated by telephone, 11% by postal questionnaire and 0.5% by face-to-face interview.

Figure 1

Flow chart of participation in POIS Interview 1.

Table 1 summarises participants' characteristics. The mean age at Interview 1 was 41.4 years (SD=13.0 years). A range of ethnicities is represented, and 660 (23%) were born outside New Zealand (table 1). At least one specified pre-injury comorbidity was reported by 1326 (46%), and 577 (20%) were affected by a previous injury at the time of the injury event that led to their participation in POIS. Of the 83% who reported their personal income, the median in the year before injury was $42 500 (in New Zealand dollars; the mean=$50 465).

Table 1

Baseline socio-demographic and pre-injury characteristics of participants (n=2856) in the POIS

Table 2 presents self-reported anatomical region and type of injury. Thirty percent reported being admitted to hospital as a consequence of their injury.

Table 2

Injury regions, type of injury and hospital admissions reported as a consequence of injury

Table 3 presents pre-injury and post-injury EQ-5D and WHODAS as reported at Interview 1. After injury, and compared to pre-injury, participants experienced overall (1) poorer health status across all five EQ-5D dimensions and Cognition; (2) worse EQ-5D VAS, dropping from a mean of 86 pre-injury to 72 post-injury (95% CI for mean difference=13.50 to 14.95; n=2836); (3) increased disability across all WHODAS items and (4) higher WHODAS disability scores, with the proportion with scores >2 increasing from 15% pre-injury to 70% post-injury.

Table 3

Participants reporting any problems with general health status (EQ-5D) and any difficulty with disability (WHODAS) dimensions before and soon after injury


To understand the burden of injury, information about injury-related disability outcomes is required. Studies examining characteristics related to disability outcomes following injury have been, and are being, undertaken in a number of countries.9 10 However, compared to other studies, a strength of POIS is that recruitment was not confined to those who had been either treated at an Emergency Department or admitted to a hospital.1 Another strength is that participants reported a wide range of injury types. Also, there was sufficient heterogeneity in the baseline characteristics to be able to explore these as predictors of outcomes of interest in future analyses. Further, recruitment of a large subsample of Māori permits analyses of outcomes specifically for Māori. Study limitations include variable times from injury to first interview; explanations of changes to our intended methods provide insights in this regard. Future analyses will include time to Interview 1 within multivariate models where appropriate. Differences were apparent in the quality of self-reported injury body region and type, most likely due to differences in participant knowledge of anatomy and injury diagnoses. Future analyses will also consider relationships between participants' reports of injury type and hospital discharge and ACC diagnostic fields.

The majority of participants were men, reflecting the greater proportion of eligible men on ACC's entitlement register (68%, n=5408). Most participants were in paid employment at the time of injury. This was expected because of New Zealand's rates of labour-force participation for people aged 18–64 years.11

POIS is one of only a few injury outcome studies to describe pre-injury comorbidity.12 13 Despite almost half the POIS cohort reporting one or more chronic conditions, 95% reported that their overall pre-injury health was ‘good’, ‘very good’ or ‘excellent’.

Less than one-third of participants reported hospital admission as a consequence of their injury, yet deterioration in health status and disability after injury was apparent across all dimensions measured. Research undertaken 15 years ago also found considerable burden for injured people who had not been admitted to hospital.14 We found that more EQ-5D problems are reported after injury compared to the general New Zealand population.15 Poor outcomes are particularly notable for problems with EQ-5D Usual Activities and Pain/Discomfort, with over half the cohort experiencing problems. Although reductions in health status observed in this analysis cannot be attributed entirely to the injury event (and could reflect recall bias), similar patterns have been reported in other studies with varying times to follow-up.13 16–18 To our knowledge, no other studies of injury outcome have been published reporting disability measured by the WHODAS. More than half of the cohort reported having difficulty soon after injury with standing, household responsibilities, emotions, walking and their day-to-day work, indicating a considerable post-injury disability burden. Compared to an Australian general population, our participants had greater disability post-injury.6 Analysis of outcome predictors related to post-injury function, disability and return-to-work soon after injury and 1 year later is now under way.

What is already known on this subject

  • Injury outcome studies tend to focus on a narrow range of injury types, predictors and outcomes; collect limited pre-injury characteristics and include only injuries requiring treatment by secondary or tertiary health services

  • Cohort studies seldom report changes to published protocols or key health and disability characteristics

What this study adds

  • Changes to intended cohort study methods may be necessary due to recruitment and data collection circumstances

  • Many, and varied, types of injuries (including ones not resulting in hospital admission in the acute phase) represent significant threat of disability


We are most grateful to the study participants for sharing their information with us and to the study interviewers for their role in collecting this information. We thank Dr Kypros Kypri for helpful comments on an earlier draft of this paper and the ACC for sending out the letters of invitation to potential participants on behalf of the research team.



  • Funding This study is funded by the Health Research Council of New Zealand (2007–2013) and was co-funded by the Accident Compensation Corporation, New Zealand (2007–2010). The views and conclusions expressed herein are the authors' and may not reflect those of the funders. Scholarships: Dr Wyeth was supported by a Health Research Council of New Zealand Eru Pomare Research Fellowship and Dr Lilley by an ACC Early Research Career Post-Doctoral Fellowship.

  • Competing interests None.

  • Ethics approval This study was undertaken following approval from the New Zealand Health and Disability Multi-Region Ethics Committee and was approved by the ACC Research Ethics Committee.

  • Provenance and peer review Not commissioned; externally peer reviewed.