Introduction: Motorcyclists in New Zealand suffer a disproportionate number of road traffic crashes. In 1995, 2% of all registered vehicles in NZ were motorcycles, but 9% of reported injuries and 13% of all fatalities were in motorcycle riders. Generally there is a lack of good evidence about many postulated risk factors, with previous studies reporting inconsistent findings. Aim: To identify the role of age, gender, experience, training and alcohol consumption as risk factors for moderate to fatal injury resulting from a motorcycle crash. These factors were chosen because they are the subject of current policy debate in New Zealand. Methods: A population-based case-control study was conducted in the Auckland region over a three year period from Feb 1993. The study base was defined as motorcycles being ridden on non-residential public roads of the Auckland region over a three year period between the 15th of February 1993 and the 14th of February 1996 between 6am and midnight. A case was defined as a motorcycle crash occurring within the study base in which either or both the motorcycle driver or pillion passenger met the injury criteria: were admitted to a public hospital within Auckland; were treated in the Public Hospital Emergency Department in the region with an Injury Severity Score of 5 or greater; or died as a result of a motorcycle crash. The aim was to identify all cases using comprehensive case finding procedures included surveillance of emergency departments, the Coroner's office and injury crashes reported to the Police. Controls, or the sample of the study base, were identified from roadside surveys which were conducted at random times from 150 random sites chosen from non-residential roads. Each road was sampled in proportion to its total length. This produced a random sample of motorcycle riding from the study base. If possible, the motorcyclist was stopped at the survey site. If this was not possible, they were identified by following up their registration plate number. Interviewer-administered questionnaires were used, either face-to-face or by telephone, to obtain exposure data. Objective alcohol data were also obtained from blood and breathalyser tests. Results: Information was obtained on 477 cases and 1518 controls, with interviews completed for 94.5% of case drivers and 81.2% of control drivers. Motorcyclists aged 25 years and over had the lowest risk of moderate to fatal injury from a motorcycle crash when compared to drivers aged from 15 to 19 years (adjusted RR=0.45, 95%Cl=0.33-0.62); the risk for 20 to 24 year olds was intermediate (adjusted RR=0.72, 95%Cl=0.52-1.00). There was insufficient evidence from this study to assess whether there was a changing pattern of risk for motorcycle riders aged over 50 years. The population attributable risk for 15 to 16 year olds in this study was only 0.72%. There was no evidence of an association between gender and risk, indicating that most motorcycle injuries occur in males because most motorcycle riding is done by males. Motorcyclists with more than five years of regular on-road motorcycle riding experience were associated with some increased risk compared to those with less than two years (adjusted RR=1.57, 95%Cl=0.96-2.58), particularly among the 20 to 24 year age group. Motorcyclists who had ridden their motorcycle 10,000 kilometres or more had a lower risk compared to those who had ridden less than 1000 kilometres (adjusted RR=0.52, 95%Cl=0.35-0.79). There was no clear effect observed for other measures of experience including off-road motorcycle riding, other motor vehicle driving and familiarity with the road. There was some evidence that motorcyclists who had attended motorcycle training were at lower risk (adjusted RR=0.74, 95%Cl=0.49-1.14), but the precision of this estimate was poor. There was less clear evidence of an association with recent non-motorcycle training and risk. Motorcyclists who reported having consumed alcohol within 12 hours were at higher risk (adjusted RR=1.53, 95%Cl=1.05-2.23) with a population attributable risk of 11%. Discussion: The methods used in this study to identify and sample the study base by obtaining a random sample of motorcycle riding are likely to have provided a valid estimate of the risks associated with the exposures measured. However some limitations of the study design include the general reliance on self-reported measures of exposures, the lack of data collected on other motor vehicles on the road and the poor quality of data available on alcohol and other environmental factors. This study has identified the following groups of motorcyclists to be at increased risk of moderate to fatal injury: riders aged less than 20 years, riders with more than five years of riding, riders who have not undertaken a motorcycle training course, riders who have ridden less than 1000 kilometres on their current motorcycle and those who have recently consumed any alcohol. Given these results, combined with the information obtained from the systematic literature review, the provisions of the Graduated Drivers Licensing System should be reviewed. This should include consideration of basing restrictions for riding primarily on the age rather than the experience of the motorcyclist and the completion of specific motorcycle training courses only, rather than a training course for any motor vehicle. Further research should consider some of the shortcomings of this study that have been outlined above and also consider methods of increasing the proportion of female riders and older riders in research. The limited evidence of benefit from motorcycle training courses, while encouraging, indicates that further investigation of their benefit in controlled trials is required.

Coronary heart disease is the leading cause of death and a major cause of morbidity in New Zealand. Although the mortality rates are now declining, they remain high by international standards and there is considerable scope for the prevention of coronary heart disease. There is a paucity of New Zealand data on the aetiology of this disease. The Auckland Heart Study is a case-control study of coronary heart disease which was designed to determine whether a range of variables including; smoking, passive smoking, hypertension, exercise, alcohol, serum lipids, and dietary and psychosocial factors, are related to coronary heart disease in New Zealand. Other factors examined included respiratory infection, sex hormones and serum vitamins. The study also documented the prevalence of the major risk factors in the Auckland adult population so as to examine trends in risk factors since the 1982 Auckland Risk Factor Study. The study was conducted between 1 March 1986 and 3 May 1988. There were two case groups: non fatal myocardial infarction cases and coronary death cases; and two control groups: myocardial infarction controls and coronary death controls. Participants were aged 25-64 years.

This thesis traces the public health debate about venereal disease in New Zealand from 1920, when the first venereal disease clinics were established, to 1980 before the first AIDS/HIV cases emerged. Studies of venereal disease in New Zealand have concentrated on issues of morality and on the political and social debates; this thesis focuses on treatment procedures and Health Department campaigns. The thesis explores the role of doctors in relation to venereal disease. While advancements in drug therapy benefited patients, medical authority was undermined by demanding and defaulting patients, inadequate medical education, and a low status of the profession. The medical profession developed epidemiological studies and defined 'at risk' groups in post-war decades. Despite claims to be 'scientific', the assessments were informed by stereotypes which had changed little over time. The thesis evaluates the scope of preventative health campaigns. Defined as a public health issue by the 1920s, venereal disease was seen as an individual responsibility by the 1960s. During this time the use of legislation declined, and education and contact tracing increased. The control of infection was limited owing to financial and administrative problems, defaulters and opposition from doctors. Those deemed most at risk were not reached by government educational campaigns, leaving much to the work of welfare groups and individual doctors. The health campaigns targeted groups like Maori and servicemen. The historiography has tended to overlook Maori, and, when military campaigns are discussed, to focus on females. This thesis attempts to redress the balance. Maori had, at least until the 1950s, different treatment experiences from non-Maori patients, although this did not necessarily imply discrimination. The military did attempt to control servicemen, though each Service had different experiences. This thesis stresses the complexity of the gender issue. There was a change from blaming females for infection in the early twentieth century to increasingly pointing to male responsibility. Despite these changes, even with the concept of individual risk pattern by the 1960s, and the understanding that men could be asymptomatic carriers, women were persistently seen as the 'reservoir'. A gender bias persisted. / Note: Thesis now published. (2007) Kampf, Antje. Mapping Out the Venereal Wilderness: Public Health and STD in New Zealand, 1920-1980. Berlin: Lit-Verlag. http://www.lit-verlag.de/isbn/3-8258-9765-9. Whole Document not available at the request of the author.

The Reaccreditation Programme of the Royal New Zealand College of General Practitioners (RNZCGP) became compulsory for all members in 1994. Ongoing reaccreditation of medical specialists is a new, but rapidly growing trend on which little research has been done. The aim of this thesis is to investigate the effects of reaccreditation on general practice in New Zealand in the first three years of the programme. The history and current status of reaccreditation programmes throughout the world is summarised. The requirements of the RNZCGP Reaccreditation Programme are described and contrasted with the requirements of other programmes. A basic assumption of the Reaccreditation Programme that general practitioners can accurately identify their own areas of deficiency, is tested and found to be incorrect. A very low correlation between general practitioners' self assessments of knowledge on a given topic, and their results on a written objective test was found. The attitudes of general practitioners to the Reaccreditation programme, both shortly before it began and three years post-implementation are tested using a validated semantic differential questionnaire. Initial acceptance was reasonably high, but this decreased with time. Qualitative and quantitative studies of the ways in which general practitioners have altered their educational activities because of the programme showed improvements in the quality of the activities selected, but little increase in overall time spent. Analysis of the effects of the compulsory audit aspect of the Reaccreditation Programme showed that this activity has resulted in significant improvements to patient care. The effects of the programme on general practice are analysed in the light of current change theories and finally recommendations are made regarding the possible future development of the RNZCGP Reaccreditation Programme.

Introduction This thesis describes two studies: The Auckland Heart Study Validation Project which validated the physical activity questionnaire from an earlier case-control study (known as the Auckland Heart Study); The Auckland Blood Pressure Control Study which was a factorial design randomised controlled trial of physical activity and salt restriction as a means of further lowering blood pressure in a community sample of patients treated with anti-hypertensive medication. This study used the physical activity questionnaire validated in the Auckland Heart Study Validation Project. Coronary heart disease is the leading cause of death and a major cause of morbidity in New Zealand. The Auckland Heart Study was a case-control study examining coronary heart disease and its risk factors and was conducted in Auckland from 1986 to 1988 (Jackson, 1989). This study incorporated a three month recall questionnaire on physical activity. The results showed that a high level of physical activity was associated with a low rate of coronary heart disease. For example the odds ratio for those with high levels of moderate leisure time physical activity was 0.78 for men and 0.39 for women. To assess the validity of the physical activity questionnaire, the Auckland Heart Study Validation Project was conducted in 1988. Both physical inactivity and hypertension are risk factors for coronary heart disease and physical activity is known to lower blood pressure. The Auckland Blood Pressure Control study was conducted during 1989-90 in order to assess the effectiveness of physical activity as a means of lowering blood pressure in a community setting. The physical activity questionnaire used in the Auckland Blood Pressure Control study was the same one validated in the Auckland Heart Study Validation Project. This thesis describes both the Auckland Heart Study Validation Project and the Auckland Blood Pressure Control study. The Auckland Heart Study Validation Project The 186 Participants for the Auckland Heart Study Validation Project were randomly selected from the control group of the Auckland Heart Study. Of those who could be contacted, 152 completed a seven day physical activity and food intake diary. The seven day diary was the gold standard for the three month physical activity recall questionnaire used in the Auckland Heart Study. The response rate for completing the seven day diary was 82%. The original control group had been randomly chosen from the community and hence the sub-sample of 152 participants represented a reasonable cross-section of the community. The correlations for the three month recall questionnaire compared with the seven day dairy, were 0.61, 0.49 and 0.86 for moderate, vigorous and total activity respectively. These findings were consistent with other validation studies in the literature. One of the strengths of the Auckland Heart Study Validation project was that it was undertaken in the community population for which it was intended. It was concluded that the three month physical activity recall measured physical activity in general and over the three recall period. Auckland Blood Pressure Control study Low levels of physical activity have been shown in observational studies to be associated with a high incidence of both coronary heart disease and hypertension. A concern with observational studies is that the findings may be due to confounding factors which are not able to be controlled, either in the design or the analysis. The best method of controlling for confounding is through the use of randomised controlled trials. The literature on physical activity as a means of lowering blood pressure contains many randomised trials and almost all have methodological weaknesses. Moreover, most of those studies have been conducted in laboratory settings; very few trials of physical activity and blood pressure have been conducted in community settings. The literature on salt restriction as a means of lowering blood pressure contains numerous well designed randomized controlled trials showing that salt restriction can lower blood pressure. While significant results have been achieved from salt restriction most of these studies have been involved intensive input from dietitians. None of the community based studies have demonstrated significant blood pressure reductions. The aim of the Auckland Blood Pressure Control study was to assess the effectiveness of physical activity and/or salt restriction as therapies to lower blood pressure in treated hypertensive patients in a community setting. The research design was a factorial design randomised controlled trial of physical activity and salt restriction as therapies for lowering blood pressure. Participants were recruited for the study from general practitioners and a variety of public advertisements. The study was conducted over six months and 181 of the baseline 208 participants completed the study. The two interventions were brisk walking for 40 minutes, three times a week and salt restriction advice. The main outcome measures were blinded blood pressures measured at three and six months. The average age of the participants was 55 years and there were approximately equal numbers of men and women. At the three month interview there was a statistically reduction in systolic blood pressure for salt restriction and physical activity as separate therapies, but not for the combination. There was no significant reduction in diastolic blood pressure at the three or six month assessment. Although the Auckland Heart Study three month recall questionnaire was valid for the case-control study there was concern that it was not sensitive enough for the randomised controlled trial. It was concluded that both physical activity and salt restriction lowered systolic blood pressure, at least in the short term, in persons with hypertension treated with medication in a community setting.

The thesis examines general medical practice with particular attention to patient outcomes. Sociological descriptions of healing are presented and compared with official views of medical institutions. A theoretical position is taken capable of interaction with the disciplines of both sociology and medicine. Definitions of health are reviewed and used to specify desirable outcomes to patient-practitioner encounters. The context of modern medical care is then discussed; the influence of the intellectual, social and economic environment on the development of medical care since 1800 are evaluated, and modern reformist pressure on medical practice is examined. From this discussion a list of qualities of interest in the sociological investigation of medical practice is developed. A research project, undertaken in New Zealand, is described which gathered information on a random sample of 9477 general practice patient-practitioner encounters. A survey methodology was used with data supplied by practitioners. In addition, a patient survey was carried out in a sub-sample of 763 visits. Data on practitioner, patient, problem presented, process of the encounter and outcome was recorded. Outcome was measured as empowerment at the visit and improvement after two weeks. The results of the project are presented and analyzed. It was found that most patients are satisfied with their consultation and that a majority have improved health state at two weeks. Much of the variation in improvement is explained by the severity of the problem and the age of the patient. Social variables make a small contribution to the prediction of good outcomes. In particular, empowerment at the consultation is related to improved subsequent health state. The influence of occupation, gender and ethnicity on problems presented to the practitioner and on outcome are discussed. A proportion of patients had delayed obtaining medical care for financial reasons. This was associated with more severe problems and poorer outcome. A psycho-social approach to problems was rare and patient counseling did not appear to contribute to a sense of empowerment. Practitioners reported highest rapport with simpler consultations and rapport was negatively related to empowerment. The significance of these findings is discussed. In a final section it is suggested that medical treatment and counseling may be antithetical skills and that both practitioner and patient may have vested interests in a materialistic view of the content of their interaction. The forms of social control implicit in the encounter are elaborated.

The main purpose of this study was to examine empirically the trends and reasons for rising acute adult medical admissions at two major public hospitals in Auckland from 1997 to 2004. According to recent national and international literature published on the topic, there has been unsustainable growth in the adult medical admissions both in NZ and most of the other developed countries. Overall, the causes of this increase have not been explored sufficiently in the literature reviewed. The NZ research has largely focused on the macro-analysis of hospital throughput data from health policy points of view. Methodology: A mixed methodology research design was applied to address the problem. Phase 1 quantitatively analysed adult medical hospital admission data (N = 277,416) obtained from the two hospitals (Middlemore and Auckland Public Hospitals), and phase 2 qualitatively explored the responses and views of the health professional expert panel (n = 16) in relation to the findings of phase 1 of the study. Findings: Overall, the crude number of admissions and age-standardised admission rates at both hospitals increased more rapidly than actual population increases. Approximately 1/3 of the patients accounted for 2/3 of the total admissions. Five major diagnostic categories accounted for 70%-80% of total acute admissions, with circulatory and respiratory system disorders being the leading causes of medical admissions. There was a strong relationship between age and increased admissions. MMH hospital overall, and its ethnic groups separately, had significantly higher admission rates than APH. Comparison of ethnic groups highlighted significant variations in the admission rates at the two hospitals despite adjusting for age, morbidity and deprivation. Conclusions: Overall the increase and variation in admission rates between the hospitals and ethnic groups was dependent on factors such as the characteristics of the population and patients, hospital admission and administration processes, availability of hospital beds, medical management at the hospital, and availability of primary and community care services. By making changes to those factors in the control of hospitals and District Health Boards, hospitals can potentially influence the trajectory of rising medical admissions. These factors include systems for managing patients with chronic illness, and pathways from community services to hospital. Finally, a number of future research areas, such as a large-scale study to explore the health service utilisation of the 55+ age groups, have been proposed.

A study of cleft lip and cleft palate was carried out in order to determine whether or not any differences in incidence between Europeans and Polynesians were accompanied by differences in recurrence risks, and to test the genetic hypotheses currently favoured as explanations of familial agggregation of these disorders. An incidence study vas undertaken on all live births in the Auckland urban area for the years 1960 to 1976.Family information was obtained from these probands and from other affected persons or their close relatives, by interview at the cleft palate clinic at Middlemore Hospital. The ascertainment probability for cleft 1ip and cleft palate probands was about 95% and was not correlated with any of the demographic characteristics measured on the probands. After correction for ascertainment, the incidence of cleft palate in Maoris was estimated to be 1.867/1000 1lve births. For Europeans the estimate was 0.643/1000. The corresponding figures for cleft lip with or without cleft palate were 0.397/1000 and 1.195/1000. The sex ratio for cleft palate was 0.485 with heterogeneity between the races. For cleft lip the sax ratio was 0.649 overall. There were no secular or seasonal trends in the incidence of facial clefts and no significant effects of maternal age, or paternal age. The mean birth rank for probands with cleft lip with or without cleft palate was higher than expected. For probands with cleft palate, mean birth rank was not significantly elevated. The pattern of additional malformations in these probands was similar to those reported in similar studies from other centres. The recurrence risk for cleft palate was 1.8% overall. Although it was s1ightly higher in polynesian families than in European families, the difference was nowhere near statistical significance. For cleft lip the recurrence risk was 2.6% overall, with the risk being slightly higher in Polynesian families, but again not significantly higher than in European families. Using current analytical techniques, no discrimination was possible between a generalized single autosomal locus model and a multifactoriar threshold model. A consideration of the parameter estimates for both models suggests that the multifactorial threshold model is the more appropriate one to use for the calculation of recurrence risks in complicated family situations. It is concluded that further family studies of this nature would no longer be warranted unless hypotheses can realistically be tested on the samples available. However, incidence studies in special populations will remain important for hypothesis testing. Following on the work using animal models, a study of face shape within and among races in New Zealand may provide clues to the aetiology of facial clefts, particularly isoleted cleft palate. It will be important to follow changes in incidence over time and discover what effects intermarriage and cultural changes might have on the incidence of facial clefts.

Whole document restricted, see Access Instructions file below for details of how to access the print copy. / Background: Giardia is a leading cause of human gastrointestinal illnesses globally and is the most commonly notified waterborne disease in New Zealand. The national incidence rate of 46.6 per 100,000 is thought to be one of the highest among developed countries, peaking in the 1-4 and the 25-44 age groups. Risk factors for infection among vulnerable groups have not been explored systematically in New Zealand, although environmental factors and person-to-person transmission have been suspected. The true burden of Giardia infecion in the community has been difficult to estimate due to suspected under-notification of the disease. An important component of disease surveillance and the validation of disease incidence rates is an estimation of the degree of undercount. AIMS: To describe the epidemiological patterns of Giardia infection in the Auckland region and in New Zealand, compare them with local and international patterns of infection, and explore environmental links. To estimate the level of completeness of giardiasis notification in the Auckland adult population by using a simple capture-recapture method. To identify risk factors for giardiasis among Auckland children under 5 years of age. Methods: Analysis of Auckland data: Anonymised giardiasis notification data from Auckland Regional Public Health Services (ARPHS) for the period of July 1996 to June 2000 were analysed by person, place and time. Infection rates and relative risks were calculated and compared with national and international information. Analysis of New Zealand data: A study of national surveillance data utilised anonymised information for 7818 notified cases throughout New Zealand between July 1996 and June 2000. A weighted average of drinking water grades was estimated using the Community Drinking Register. Pearson's coefficient was used to measure the correlation between average drinking water grades and notified cases. Daily climate data were plotted against daily case notifications and modelled, using Poisson's regression, to predict any influence of climate on infection. Data were presented by age, gender, ethnicity and area using statistical and spatial methods. Estimation of under-notification: The capture-recapture technique is now being used in many countries to evaluate the completeness of disease ascertainment. Comparison of disease ascertainment involves two or more datasets. Two independent datasets of giardiasis cases aged 15 years or over were generated. Of them one was generated from the demographic information of cases recruited during the Auckland Giardiasis Study in 1998-99 and the other from giardiasis cases notified to the ARPHS for the same period of time. The area of residence of cases was geo-coded, mapped and overlaid by water distribution zones. Cases were matched and under-notification was estimated using a two-sample capture-recapture method. Case-control study: A case-control methodology was used to analyse the exposure history of 69 cases and 98 controls under 5 years of age in Auckland. Significant risks of infection were estimated and their attributable risks. Results: Giardiasis in Auckland: Auckland had a significantly higher rate of Giardia notification (58/100,000) than New Zealand (46/100,000) as a whole. Notification rates, which peaked during February-May, were significantly higher in Pakeha/Europeans and Asian/others, compared with Maori/Pacificans. Adjusted notification rates were higher for residents of North Shore and Auckland cities than for other areas of Auckland. The crude regional and national notification rates were almost 6 times the rate of laboratory identification of positive isolates in the UK and 4 times US reported rates. Giardiasis in New Zealand: At the national level, most cases occurred in the 1-4 year age group followed by the 25-44 year age group, and most cases were Pakeha/European. Ethnicity was not known for 18% of cases, affecting demographic calculations. Infection rates were high for a number of Health Districts (West Coast, Wanganui, Waikato and Tauranga) compared to the national average. Over 50% of the population received Aa-graded drinking water. No correlation between infection and the weighted average grades for water treatment plants (r = -0.12) or the reticulation systems (r = -0.11)) was found. A significant correlation with the mean daily maximum (r = 0.05) and minimum (r = 0.06) temperature was observed. Poisson's regression modelled minimum-temperature (chi2 = 5.40, p<0.05) and relative humidity (chi2 = 5.37, p<0.05) as predictors of a significant number of Giardia infections on a given day. Under-notification of giardiasis: The estimation of under-notification during the 12 month period, compared 413 cases who were notified to the ARPHS with 199 cases who had participated in a case-control study over the same period of time. North Auckland had slightly higher notification rates and also study participation rates. The giardiasis notification rate was higher in un-reticulated water zones (72.4/100,000 population). The capture-recapture calculation indicated that only 49% of cases were notified. Risk of giardiasis in children: In the case-control study, 95% of cases and 86% of controls used water from the Auckland Metropolitan Mains (AMM) supply for domestic purposes, 44 cases and 42 controls swam, and 59 cases and 54 controls wore nappies. Children wearing nappies were at significantly increased risk of the disease (OR=3.0, 95%CI 1.01-8.9), as were those from households which had more than one child wearing a nappy (OR=6.5, 1.8-23.4). The AMM water supply was associated with a reduced giardiasis risk compared to other drinking water sources. Significantly increased risks were associated with drinking water consumed away from home (OR=4.7, 2.2-10.1), swimming at least once a week (OR=2.4, 1.1-5.3) and travelling in side New Zealand (OR=2.5, 1.03-6.0). Conclusions: The higher rates of giardiasis observed in Auckland and New Zealand, in comparison with other developed countries, may be related to environmental or social factors. Missing ethnicity information precludes clear interpretation of variations in notification rate by ethnic group and suggests a need for improvement in data collection. There are opportunities to investigate the influence of risk factors on seasonal changes in notification rates both locally and nationally. Reported improvements in some areas could be due to local health measures or to random variation. Time-trend analysis suggests a seasonal pattern. The weighted average drinking water grading is a novel and readily available measure, and may not be truly representative of local supplies. Modelling of climate data showed an association with giardiasis but was inconclusive. Giardia notifications in Auckland are under-notified by half. This has obscured the true burden of Giardia infection. This has implications for estimates of the burden of disease in New Zealand. Notification rates vary inversely with socio-economic conditions and the presence of a reticulated water system. The case-control study identified vulnerable groups and modifiable risk factors for Giardia infection. Nappy-wearing was an independent risk factor for infection. Further study is advocated to ensure better protection of public health, especially for children. These studies identified vulnerable groups and major data-gaps. The risk of infection in children needs further attention for effective infection control measures to be developed for this disease. Recommendations for improvements in disease surveillance and data quality are discussed. GIS (Geographical Information System) is a useful tool for disease monitoring. Capture-recapture techniques are useful tools in evaluating the completeness of disease surveillance.

Background Road traffic injury is projected to rank as the third largest contributor to the global burden of disease by 2020. Disability is a significant component of the burden of disease ranking. Most published data on traffic crash outcomes, however, focus primarily on deaths and hospitalisations. Reliable estimates of post-crash disability and information on factors that modify the disabling process are essential to prioritise and allocate appropriate resources for road traffic injury prevention and interventions that reduce the risk of secondary disability. Aims To quantify the risk of disability associated with serious injury crashes in car drivers in a defined population; to explore the extent to which this risk is modified by chronic alcohol abuse; and to critically review methodological approaches that can redress the inadequate epidemiological attention to injury-related disability. Methods Systematic reviews were conducted to examine the available epidemiological evidence quantifying the association of car crashes with disability and the effect of alcohol on the risk of post-injury disability. Studies published or presented between January 1980 and April 2003 were reviewed. No language restriction was imposed. A population-based prospective cohort study conducted in the Auckland region of New Zealand recruited drivers exposed to serious injury crashes (identified through a surveillance system monitoring hospital admissions of injured car occupants). A representative sample of car drivers in the region was identified through roadside surveys (controls). The participants were interviewed at recruitment (to obtain pre-crash information from crash drivers and baseline data from controls) and re-interviewed at five and eighteen-months follow-up. Structured interviews on all three occasions included the Short Form-36, a global health change indicator, and the Alcohol Use Disorders Identification Test. Information on a range of potential confounders was sought at baseline through the interview, alcohol measurements and clinical records. Results Studies identified in the systematic reviews revealed that published estimates of the risk of post-crash disability ranged from 2% to 57%. The evidence regarding the effect of alcohol on post-injury disability is inconclusive largely because none examined this association directly. Most studies identified in the reviews were limited by several methodological problems including the absence of appropriate comparison groups, inadequate or no adjustment for confounding, significant potential for selection bias due to the study setting, high levels of loss to follow-up, and missing data. In the prospective cohort study, 215 crash drivers (75% follow-up) and 254 controls (69%) completed the 18-month interview. Overall, 40% of the drivers who were hospitalised, 20% of the crash drivers not hospitalised, and 7% of the controls reported deteriorated health at 18 months relative to their baseline health. This represents a tenfold excess risk of disability among hospitalised drivers and a three-fold excess risk among non-hospitalised crash drivers, relative to drivers in the general population. Among crash drivers reporting an overall decline in health, clinically important reductions in general and mental health were apparent over the follow-up period despite improving physical health and function. This trend was more evident among non-hospitalised than hospitalised crash drivers. Compared with drivers who were neither involved in a crash nor defined as hazardous drinkers, crash drivers who were hazardous drinkers had a seven-fold excess risk of a clinically significant (≥ 10%) decline in the SF-36 general health score (OR 6.85; 95% CI: 1.84-25.43). Crash drivers who were not hazardous drinkers had a three-fold risk (OR: 3.00; 95% CI: 1.14-7.89). The results indicated an important interaction between crash involvement and chronic alcohol abuse in potentiating the risk of disability. Conclusion Serious traffic crashes are associated with significant longer-term disability in a substantial proportion of survivors with an apparent worsening of mental health over time. Definitions of disability and estimates of the burden of disability following traffic injury remain highly variable in the published literature and it is timely for the international research community to develop a more systematic and consistent approach to this major and increasing component of the global burden of disease. By addressing the main methodological limitations of previous studies, this study revealed that chronic alcohol abuse potentiates the risk of post-crash disability. The findings suggest that measures for preventing road traffic crashes as well as efforts to identify problem drinkers among crash survivors should be intensified. The thesis highlights the need for robust indicators of non-fatal injury to monitor the impact of road safety programs and large-scale epidemiological studies to investigate the spectrum and determinants of post-injury disability.