The Accident Compensation Corporation (ACC) is a Crown organisation providing no-fault personal cover for injuries to residents and visitors to New Zealand according to the Injury Prevention, Rehabilitation, and Compensation Act 2001.1 In the 2007/08 year, ACC spent $1.9 billion on support for 215,676 entitlement claimants.2 Entitlement claims are for injuries that may result in a week or more off paid employment, and include payments such as weekly compensation (salary), rehabilitation, equipment and modifications required to housing.Fifty-eight percent of 2007/08 expenditure supported one-third of claimants with problems resulting from ‘old' injuries sustained at least one year earlier.2 ACC's expenditure represents only one aspect of the costs of disabling injury; yet provides sufficient evidence that longer-term disability following injury is a significant issue in New Zealand with a population of only 4.3 million people. Furthermore, the 2006 New Zealand Disability Survey identified injury was the reported cause of disability for 166,300 (29%) disabled adults.3To identify strategies to improve post-injury outcomes for New Zealanders we need to know the barriers to, and facilitators of, good outcomes. The Prospective Outcomes of Injury Study (POIS) was designed to quantitatively determine factors leading to disability following injury in New Zealand, and, to qualitatively explore with injured Māori and non-Māori New Zealanders their ‘lived experiences' and perceptions of injury outcome.4Few prospective cohort studies of outcomes following injury have been completed in New Zealand; consequently we did not know if the preferred study design (obtaining the cohort from sampling ACC claimants) was feasible.5 We thus applied and obtained funding for a feasibility study.Key research questions were: Will injured claimants participate in an injury outcomes study when a letter of invitation is sent by the ACC insurer; Will we be able to recruit a range of people with different types and causes of injury; Will injured people be willing to respond to detailed questions about anatomical region, type and cause of injury, alcohol and drug use, disability, health and household composition, home ownership and finances; and Will they be satisfied with the interview and the process? This paper describes the development and piloting of POIS, insights obtained in relation to the research questions and other issues important to the conduct of the proposed prospective cohort study.Methods Before the pilot study, developmental work included reviewing literature related to injury and disability outcomes; and consultation with Māori, lay and scientific advisors and stakeholders within ACC, Ministries of Health and Social Development, and Statistics New Zealand. In addition, a focus group was held with people with injury-related disabilities, and a preliminary study questionnaire was pre-tested with recently-injured people. Developmental work supported the need for investigation into the experiences and outcomes of people after injury. To facilitate a high Māori participation rate in the proposed POIS project, recommendations included having a qualitative component and ensuring regions with a high proportion of Māori residents were selected.6 Telephone interviews of less than one hour were preferred to postal questionnaires—but Māori and focus group participants also recommended making available face-to-face interviews and postal questionnaires as optional alternatives. It was recommended that ethnicity be collected according to the Census questions and that interviews be available in Te Reo Māori.7 The focus group advised collecting information about: pre-injury disability; psychological well-being; sexuality, and positive outcomes following injury. The importance of not sharing participants' individual data with agencies was emphasised, as was the need for the analysis to be undertaken independently of the ACC. A $10 voucher for participants in appreciation of their taking part was recommended to increase participation. Following ethical approval the POIS pilot study commenced in 2006. ACC randomly selected eligible claimants from four regions of New Zealand (Manukau City, Gisborne, Otago and Southland). These regions were chosen because they included: people who lived rurally, regions where Māori comprised a significant proportion of the population and a range of socioeconomic groups. People were eligible for the Pilot Study if they were aged 18 to 65 years, were New Zealand residents, had been placed on the entitlement claimant's register in November 2006 for an acute injury event, lived in one of the four regions and had not already been invited to take part in another survey via the ACC. People were ineligible if the cause of their injury was self-harm, if they were on ACC's Sensitive Claims register (which includes people who have been sexually assaulted), or if they were otherwise unable to participate in a telephone interview administered in English. All selected entitlement claimants were first written to by the ACC. This postal contact included a covering letter from ACC introducing the study and notifying people about the opportunity to ‘opt out' (by phoning a toll-free number at ACC), together with a Participant Information Sheet, provided by the researchers, describing the study. Contact details of those not opting out were then provided to the researchers by ACC. If people were willing to take part, their oral consent was obtained and contact details given to one of five interviewers. People were then sent a response sheet to help with answering the interview questions, and also a written consent form and a reply-paid envelope to return the consent form to the researchers. Interviewers arranged a telephone interview at a mutually acceptable time. Participants' answers to interview questions were entered onto paper questionnaires. On completion of the interview participants were sent a thank you letter and $10 voucher. Interviews included a range of self-reported set-response measures selected to reflect the World Health Organization model of functioning, disability and health.8 These included the short-form WHODAS II - a measure developed by WHO specifically to measure disability and health.9 The WHODAS was recommended for inclusion in studies of injury outcome by an international working group, along with a brief measure of general health status, the EQ-5D, which was also included.10 Other questions included health and ACC service satisfaction, the general self-efficacy scale,11 the Positive and Negative Affect Scale (PANAS),12 chronic health and disability,7 exercise,13 sleep, height and weight, alcohol and drug use,14 general health, unpaid activities,7 work characteristics before and after injury,15 general happiness, social and life satisfaction based on General Social Survey questions,16 expectations of recovery17 and return to work, independence and demographics. Open-ended questions allowing free responses asked participants about: the injury event, treatment provided, and factors influencing perceived satisfaction with health and ACC services. Participants were asked to report their pre-injury and current status for certain measures (WHODAS, EQ-5D, global health and happiness). Additionally, interviewers were asked for their perceptions of the questions to help improve the questionnaire for the intended main POIS study. Following written consent, ACC provided injury information from electronic records for participants and anonymous summary information for non-participants. Injury information for participants included: accident description, injury diagnoses, days on earnings-related compensation, days off work and claim costs. Results ACC sent 285 letters of invitation to people selected randomly from ACC's entitlement claimant register (Figure 1). Six people did not want to be contacted about the study, three were ineligible and 108 could not be contacted by telephone. Of 168 people who were both eligible and contactable, 111 (66%) participated. In addition to providing oral consent, all participants were asked to return, in reply-paid envelopes, a written consent form; 91 (82%) did. Figure 1. Flow diagram of recruitment to the POIS Pilot Study Participants tended to be older than non-participants and a smaller proportion of men participated (Table 1). Responses from two participants were excluded from the study because they incorrectly assumed the researchers were interested in an older, more serious injury, rather than the injury that led to their selection. Four participants were not asked many interview questions because of difficulties understanding the questions in English language, leaving 105 participants for most interview questions. Participants reported a range of anatomical regions injured and injury types including multiple trauma following vehicle crashes, burns, blood poisoning, sprains and strains. Missing responses were highest for self-reported personal and household income—14% and 30% respectively. Questions about difficulty living on household income, likelihood of financial hardship and reduced standard of living resulted in only 2% missing responses. A question about household composition was answered by all; home ownership and highest educational qualification questions had missing responses for 1% and 2% respectively. Table 1. Characteristics of non-participants and participants Characteristics Non-participants (n=168) Participants* (n=110) P value n % n % Age at time of interview 18-34 years 35-49 years 50-64 years Mean (SD) 83 58 27 35.9 (12.9) 49 35 16 37 43 30 41.4 (12.6) 34 39 27 .0006 Sex Male 132 79 67 61 .001 ACC-reported ethnicity New Zealand European/Pākehā Māori Pasifika Peoples Asian Other 93 38 16 7 14 55 22 10 4 8 71 12 10 5 12 65 11 9 5 11 .16 Injury at work Yes No 66 102 39 61 40 70 36 64 .63 Employment Employed Non-earner Self-employed 146 5 17 87 3 10 92 6 12 84 5 11 .56 *The total number of participants equals 110, rather than 111 in Table 1, because descriptive data were inadvertently not provided by ACC for one participant. The PANAS affect scale and WHODAS disability questions had four percent missing responses. The full range of response options were used for all WHODAS activities apart from ‘Learning a new task' and ‘Concentrating' where the greatest level of difficulty reported was Severe (rather than Extreme/Cannot Do). Questions asking about employment situation, exercise and sleep were answered by all participants. The average time for the interview was 57 minutes, ranging between 20 minutes (with a participant who was not asked the full set of questions because of English-as-second-language) and 183 minutes (Table 2). The overall impression of the interview and questions was positive for two-thirds and negative for 6%. When asked about the initial contact coming from ACC, 4% reported they felt negatively about this. The procedure for taking part was unclear for one person; receipt of the $10 voucher was perceived negatively by one person who suggested it go to charity instead. Interviewers provided feedback about: the need to collect quantifiable information about injury treatment received by participants; difficulty answering specific questions about the exact medications received; ease of administration of the self-efficacy questions and list of chronic health conditions; that the WHODAS resulted in a comprehensive range of responses, and difficulty administering the home help and PANAS questions to people for whom English is not their first language. Interviewers also reported the length of interview was most problematic for those with few injury-related problems at the time of interview—but these people also reported understanding the need to include all question areas. Table 2. Satisfaction with interview and process of recruitment Characteristics Participants (n=109) n % Length of interview (minutes) 20-60 61-90 >91 72 33 3 67 31 3 Impression of interview length and questions Positive Mixed Negative 68 26 6 68 26 6 Satisfaction with initial approach via ACC Positive Mixed Negative 89 4 4 92 4 4 Procedure for taking part? Clear Unclear 89 1

The Accident Compensation Corporation (ACC) is a Crown organisation providing no-fault personal cover for injuries to residents and visitors to New Zealand according to the Injury Prevention, Rehabilitation, and Compensation Act 2001.1 In the 2007/08 year, ACC spent $1.9 billion on support for 215,676 entitlement claimants.2 Entitlement claims are for injuries that may result in a week or more off paid employment, and include payments such as weekly compensation (salary), rehabilitation, equipment and modifications required to housing.Fifty-eight percent of 2007/08 expenditure supported one-third of claimants with problems resulting from ‘old' injuries sustained at least one year earlier.2 ACC's expenditure represents only one aspect of the costs of disabling injury; yet provides sufficient evidence that longer-term disability following injury is a significant issue in New Zealand with a population of only 4.3 million people. Furthermore, the 2006 New Zealand Disability Survey identified injury was the reported cause of disability for 166,300 (29%) disabled adults.3To identify strategies to improve post-injury outcomes for New Zealanders we need to know the barriers to, and facilitators of, good outcomes. The Prospective Outcomes of Injury Study (POIS) was designed to quantitatively determine factors leading to disability following injury in New Zealand, and, to qualitatively explore with injured Māori and non-Māori New Zealanders their ‘lived experiences' and perceptions of injury outcome.4Few prospective cohort studies of outcomes following injury have been completed in New Zealand; consequently we did not know if the preferred study design (obtaining the cohort from sampling ACC claimants) was feasible.5 We thus applied and obtained funding for a feasibility study.Key research questions were: Will injured claimants participate in an injury outcomes study when a letter of invitation is sent by the ACC insurer; Will we be able to recruit a range of people with different types and causes of injury; Will injured people be willing to respond to detailed questions about anatomical region, type and cause of injury, alcohol and drug use, disability, health and household composition, home ownership and finances; and Will they be satisfied with the interview and the process? This paper describes the development and piloting of POIS, insights obtained in relation to the research questions and other issues important to the conduct of the proposed prospective cohort study.Methods Before the pilot study, developmental work included reviewing literature related to injury and disability outcomes; and consultation with Māori, lay and scientific advisors and stakeholders within ACC, Ministries of Health and Social Development, and Statistics New Zealand. In addition, a focus group was held with people with injury-related disabilities, and a preliminary study questionnaire was pre-tested with recently-injured people. Developmental work supported the need for investigation into the experiences and outcomes of people after injury. To facilitate a high Māori participation rate in the proposed POIS project, recommendations included having a qualitative component and ensuring regions with a high proportion of Māori residents were selected.6 Telephone interviews of less than one hour were preferred to postal questionnaires—but Māori and focus group participants also recommended making available face-to-face interviews and postal questionnaires as optional alternatives. It was recommended that ethnicity be collected according to the Census questions and that interviews be available in Te Reo Māori.7 The focus group advised collecting information about: pre-injury disability; psychological well-being; sexuality, and positive outcomes following injury. The importance of not sharing participants' individual data with agencies was emphasised, as was the need for the analysis to be undertaken independently of the ACC. A $10 voucher for participants in appreciation of their taking part was recommended to increase participation. Following ethical approval the POIS pilot study commenced in 2006. ACC randomly selected eligible claimants from four regions of New Zealand (Manukau City, Gisborne, Otago and Southland). These regions were chosen because they included: people who lived rurally, regions where Māori comprised a significant proportion of the population and a range of socioeconomic groups. People were eligible for the Pilot Study if they were aged 18 to 65 years, were New Zealand residents, had been placed on the entitlement claimant's register in November 2006 for an acute injury event, lived in one of the four regions and had not already been invited to take part in another survey via the ACC. People were ineligible if the cause of their injury was self-harm, if they were on ACC's Sensitive Claims register (which includes people who have been sexually assaulted), or if they were otherwise unable to participate in a telephone interview administered in English. All selected entitlement claimants were first written to by the ACC. This postal contact included a covering letter from ACC introducing the study and notifying people about the opportunity to ‘opt out' (by phoning a toll-free number at ACC), together with a Participant Information Sheet, provided by the researchers, describing the study. Contact details of those not opting out were then provided to the researchers by ACC. If people were willing to take part, their oral consent was obtained and contact details given to one of five interviewers. People were then sent a response sheet to help with answering the interview questions, and also a written consent form and a reply-paid envelope to return the consent form to the researchers. Interviewers arranged a telephone interview at a mutually acceptable time. Participants' answers to interview questions were entered onto paper questionnaires. On completion of the interview participants were sent a thank you letter and $10 voucher. Interviews included a range of self-reported set-response measures selected to reflect the World Health Organization model of functioning, disability and health.8 These included the short-form WHODAS II - a measure developed by WHO specifically to measure disability and health.9 The WHODAS was recommended for inclusion in studies of injury outcome by an international working group, along with a brief measure of general health status, the EQ-5D, which was also included.10 Other questions included health and ACC service satisfaction, the general self-efficacy scale,11 the Positive and Negative Affect Scale (PANAS),12 chronic health and disability,7 exercise,13 sleep, height and weight, alcohol and drug use,14 general health, unpaid activities,7 work characteristics before and after injury,15 general happiness, social and life satisfaction based on General Social Survey questions,16 expectations of recovery17 and return to work, independence and demographics. Open-ended questions allowing free responses asked participants about: the injury event, treatment provided, and factors influencing perceived satisfaction with health and ACC services. Participants were asked to report their pre-injury and current status for certain measures (WHODAS, EQ-5D, global health and happiness). Additionally, interviewers were asked for their perceptions of the questions to help improve the questionnaire for the intended main POIS study. Following written consent, ACC provided injury information from electronic records for participants and anonymous summary information for non-participants. Injury information for participants included: accident description, injury diagnoses, days on earnings-related compensation, days off work and claim costs. Results ACC sent 285 letters of invitation to people selected randomly from ACC's entitlement claimant register (Figure 1). Six people did not want to be contacted about the study, three were ineligible and 108 could not be contacted by telephone. Of 168 people who were both eligible and contactable, 111 (66%) participated. In addition to providing oral consent, all participants were asked to return, in reply-paid envelopes, a written consent form; 91 (82%) did. Figure 1. Flow diagram of recruitment to the POIS Pilot Study Participants tended to be older than non-participants and a smaller proportion of men participated (Table 1). Responses from two participants were excluded from the study because they incorrectly assumed the researchers were interested in an older, more serious injury, rather than the injury that led to their selection. Four participants were not asked many interview questions because of difficulties understanding the questions in English language, leaving 105 participants for most interview questions. Participants reported a range of anatomical regions injured and injury types including multiple trauma following vehicle crashes, burns, blood poisoning, sprains and strains. Missing responses were highest for self-reported personal and household income—14% and 30% respectively. Questions about difficulty living on household income, likelihood of financial hardship and reduced standard of living resulted in only 2% missing responses. A question about household composition was answered by all; home ownership and highest educational qualification questions had missing responses for 1% and 2% respectively. Table 1. Characteristics of non-participants and participants Characteristics Non-participants (n=168) Participants* (n=110) P value n % n % Age at time of interview 18-34 years 35-49 years 50-64 years Mean (SD) 83 58 27 35.9 (12.9) 49 35 16 37 43 30 41.4 (12.6) 34 39 27 .0006 Sex Male 132 79 67 61 .001 ACC-reported ethnicity New Zealand European/Pākehā Māori Pasifika Peoples Asian Other 93 38 16 7 14 55 22 10 4 8 71 12 10 5 12 65 11 9 5 11 .16 Injury at work Yes No 66 102 39 61 40 70 36 64 .63 Employment Employed Non-earner Self-employed 146 5 17 87 3 10 92 6 12 84 5 11 .56 *The total number of participants equals 110, rather than 111 in Table 1, because descriptive data were inadvertently not provided by ACC for one participant. The PANAS affect scale and WHODAS disability questions had four percent missing responses. The full range of response options were used for all WHODAS activities apart from ‘Learning a new task' and ‘Concentrating' where the greatest level of difficulty reported was Severe (rather than Extreme/Cannot Do). Questions asking about employment situation, exercise and sleep were answered by all participants. The average time for the interview was 57 minutes, ranging between 20 minutes (with a participant who was not asked the full set of questions because of English-as-second-language) and 183 minutes (Table 2). The overall impression of the interview and questions was positive for two-thirds and negative for 6%. When asked about the initial contact coming from ACC, 4% reported they felt negatively about this. The procedure for taking part was unclear for one person; receipt of the $10 voucher was perceived negatively by one person who suggested it go to charity instead. Interviewers provided feedback about: the need to collect quantifiable information about injury treatment received by participants; difficulty answering specific questions about the exact medications received; ease of administration of the self-efficacy questions and list of chronic health conditions; that the WHODAS resulted in a comprehensive range of responses, and difficulty administering the home help and PANAS questions to people for whom English is not their first language. Interviewers also reported the length of interview was most problematic for those with few injury-related problems at the time of interview—but these people also reported understanding the need to include all question areas. Table 2. Satisfaction with interview and process of recruitment Characteristics Participants (n=109) n % Length of interview (minutes) 20-60 61-90 >91 72 33 3 67 31 3 Impression of interview length and questions Positive Mixed Negative 68 26 6 68 26 6 Satisfaction with initial approach via ACC Positive Mixed Negative 89 4 4 92 4 4 Procedure for taking part? Clear Unclear 89 1