Prospective benchmarking of an observational analysis in the SWEDEHEART registry against the REDUCE-AMI randomized trial.

Prospective benchmarking of an observational analysis against a randomized trial increases confidence in the benchmarking process as it relies exclusively on aligning the protocol of the trial and the observational analysis, while the trials findings are unavailable. The Randomized Evaluation of Decreased Usage of Betablockers After Myocardial Infarction (REDUCE-AMI, ClinicalTrials.gov ID: NCT03278509) trial started recruitment in September 2017 and results are expected in 2024. REDUCE-AMI aimed to estimate the effect of long-term use of beta blockers on the risk of death and myocardial following a myocardial infarction with preserved left ventricular systolic ejection fraction. We specified the protocol of a target trial as similar as possible to that of REDUCE-AMI, then emulated the target trial using observational data from Swedish healthcare registries. Had everyone followed the treatment strategy as specified in the target trial protocol, the observational analysis estimated a reduction in the 5-year risk of death or myocardial infarction of 0.8 percentage points for beta blockers compared with no beta blockers; effects ranging from an absolute reduction of 4.5 percentage points to an increase of 2.8 percentage points in the risk of death or myocardial infarction were compatible with our data under conventional statistical criteria. Once results of REDUCE-AMI are published, we will compare the results of our observational analysis against those from the trial. If this prospective benchmarking is successful, it supports the credibility of additional analyses using these observational data, which can rapidly deliver answers to questions that could not be answered by the initial trial. If benchmarking proves unsuccessful, we will conduct a "postmortem" analysis to identify the reasons for the discrepancy. Prospective benchmarking shifts the investigator focus away from an endeavour to use observational data to obtain similar results as a completed randomized trial, to a systematic attempt to align the design and analysis of the trial and the observational analysis.

Benchmarking Observational Analyses Before Using Them to Address Questions Trials Do Not Answer: An Application to Coronary Thrombus Aspiration.

To increase confidence in the use of observational analyses when addressing effectiveness questions beyond those addressed by randomized trials, one can first benchmark the observational analyses against existing trial results. We used Swedish registry data to emulate a target trial similar to the Thrombus Aspiration in ST-Elevation Myocardial Infarction in Scandinavia (TASTE) randomized trial, which found no difference in the risk of death or myocardial infarction by 1 year with or without thrombus aspiration among individuals with ST-elevation myocardial infarction. We benchmarked the emulation against the trial at 1 year and then extended the emulation's follow-up to 3 years and estimated effects in subpopulations underrepresented in the trial. As in the TASTE trial, the observational analysis found no differences in risk of outcomes by 1 year between groups (risk difference = 0.7 (confidence interval, -0.7, 2.0) and -0.2 (confidence interval, -1.3, 1.0) for death and myocardial infarction, respectively), so benchmarking was considered successful. We additionally showed no difference in risk of death or myocardial infarction by 3 years, or within subpopulations by 1 year. Benchmarking against an index trial before using observational analyses to answer questions beyond those the trial could address allowed us to explore whether the observational data can be trusted to deliver valid estimates of treatment effects.

Comparing Effect Estimates in Randomized Trials and Observational Studies From the Same Population: An Application to Percutaneous Coronary Intervention.

Background To understand when results from observational studies and randomized trials are comparable, we performed an observational emulation of a target trial designed to ask similar questions as the VALIDATE (Bivalirudin Versus Heparin in ST-Segment and Non-ST-Segment Elevation Myocardial Infarction in Patients on Modern Antiplatelet Therapy) randomized trial. The VALIDATE trial compared the effect of bivalirudin and heparin during percutaneous coronary intervention on the risk of death, myocardial infarction, and bleeding across Sweden. Methods and Results We specified the protocol of a target trial similar to the VALIDATE trial, then emulated the target trial in the period before the VALIDATE trial took place using data from the SWEDEHEART (Swedish Web System for Enhancement and Development of Evidence-Based Care in Heart Disease Evaluated According to Recommended Therapies) registry-the same registry in which the trial was undertaken. The target trial emulation and the VALIDATE trial both estimated little or no effect of bivalirudin versus heparin on the risk of death or myocardial infarction by 180 days (target trial emulation risk ratio for death, 1.21 [95% CI, 0.88 - 1.54]; VALIDATE trial hazard ratio for death, 1.05 [95% CI, 0.78 - 1.41]). The observational data, however, could not capture less severe cases of bleeding, resulting in an inability to define a bleeding outcome like the trial, and could not accurately estimate the comparative risk of death by 14 days, which may be the result of intractable confounding early in follow-up or the inability to precisely emulate the trial's eligibility criteria. Conclusions Using real-world data to emulate a target trial can deliver accurate effect estimates. Yet, even with rich observational data, it is not always possible to estimate the short-term effect of interventions or the effect on outcomes for which data are not routinely collected.

Burden and prevalence of prognostic factors for severe COVID-19 in Sweden.

The World Health Organization and European Centre for Disease Prevention and Control suggest that individuals over the age of 70 years or with underlying cardiovascular disease, cancer, chronic obstructive pulmonary disease, asthma, or diabetes are at increased risk of severe COVID-19. However, the prevalence of these prognostic factors is unknown in many countries. We aimed to describe the burden and prevalence of prognostic factors of severe COVID-19 at national and county level in Sweden. We calculated the burden and prevalence of prognostic factors for severe COVID-19 based on records from the Swedish national health care and population registers for 3 years before 1st January 2016. 9,624,428 individuals were included in the study population. 22.1% had at least one prognostic factor for severe COVID-19 (2,131,319 individuals), and 1.6% had at least three factors (154,746 individuals). The prevalence of underlying medical conditions ranged from 0.8% with chronic obstructive pulmonary disease (78,516 individuals) to 7.4% with cardiovascular disease (708,090 individuals), and the county specific prevalence of at least one prognostic factor ranged from 19.2% in Stockholm (416,988 individuals) to 25.9% in Kalmar (60,005 individuals). We show that one in five individuals in Sweden is at increased risk of severe COVID-19. When compared with the critical care capacity at a local and national level, these results can aid authorities in optimally planning healthcare resources during the current pandemic. Findings can also be applied to underlying assumptions of disease burden in modelling efforts to support COVID-19 planning.

Serum IL8 is not associated with cardiovascular events but with all-cause mortality.

BACKGROUND: The aim of this study is to investigate if IL8 levels were associated with incident cardiovascular (CV) events (CVE) and mortality (all-cause, CV, and cancer) in a cohort of 60 years old men and women from Stockholm (60YO). METHODS: The 60YO comprises 4232 participants; baseline period: 1997-1999. The cohort is matched annually to population registries to record deaths and incident CVE. Serum IL8 was measured in 4011 participants and categorized in quartiles. Cox proportional hazard models were used to estimate the CVE and mortality risk, expressed as hazard ratios (HR) with 95% confidence intervals (CI). Potential confounding was addressed by adjusting for traditional CV risk factors (CVE estimates) and by sex, life style habits, metabolic factors (mortality estimates). Laplace regression was used to calculate the difference in time until a certain percentage of the cohort died according to IL8 levels. RESULTS: During 16.5 years follow up, 522 incident CVE were recorded and 647 study participants died. IL8 was not associated with CVE risk (IL8 Q4 vs Q1, HR of 0.95; 95% CI 0.75-1.22). Compared to Q1, IL8 Q4 was associated with all-cause mortality (adjusted HR 1.28; 95% CI 1.02-1.63). No association was observed with CV and cancer related mortality in the fully adjusted model. Participants with IL8 above the median died of any cause ≈1.3 years before the 15% of the population had died. CONCLUSION: Elevated IL8 levels were not associated with CVE risk and CV mortality, but were associated with an increased risk of all-cause mortality regardless of the underlying cause.

Postmenopausal hormone therapy and risk of stroke: A pooled analysis of data from population-based cohort studies.

BACKGROUND: Recent research indicates a favourable influence of postmenopausal hormone therapy (HT) if initiated early, but not late, on subclinical atherosclerosis. However, the clinical relevance of timing of HT initiation for hard end points such as stroke remains to be determined. Further, no previous research has considered the timing of initiation of HT in relation to haemorrhagic stroke risk. The importance of the route of administration, type, active ingredient, and duration of HT for stroke risk is also unclear. We aimed to assess the association between HT and risk of stroke, considering the timing of initiation, route of administration, type, active ingredient, and duration of HT. METHODS AND FINDINGS: Data on HT use reported by the participants in 5 population-based Swedish cohort studies, with baseline investigations performed during the period 1987-2002, were combined in this observational study. In total, 88,914 postmenopausal women who reported data on HT use and had no previous cardiovascular disease diagnosis were included. Incident events of stroke (ischaemic, haemorrhagic, or unspecified) and haemorrhagic stroke were identified from national population registers. Laplace regression was employed to assess crude and multivariable-adjusted associations between HT and stroke risk by estimating percentile differences (PDs) with 95% confidence intervals (CIs). The fifth and first PDs were calculated for stroke and haemorrhagic stroke, respectively. Crude models were adjusted for age at baseline only. The final adjusted models included age at baseline, level of education, smoking status, body mass index, level of physical activity, and age at menopause onset. Additional variables evaluated for potential confounding were type of menopause, parity, use of oral contraceptives, alcohol consumption, hypertension, dyslipidaemia, diabetes, family history of cardiovascular disease, and cohort. During a median follow-up of 14.3 years, 6,371 first-time stroke events were recorded; of these, 1,080 were haemorrhagic. Following multivariable adjustment, early initiation (<5 years since menopause onset) of HT was associated with a longer stroke-free period than never use (fifth PD, 1.00 years; 95% CI 0.42 to 1.57), but there was no significant extension to the time period free of haemorrhagic stroke (first PD, 1.52 years; 95% CI -0.32 to 3.37). When considering timing as a continuous variable, the stroke-free and the haemorrhagic stroke-free periods were maximal if HT was initiated approximately 0-5 years from the onset of menopause. If single conjugated equine oestrogen HT was used, late initiation of HT was associated with a shorter stroke-free (fifth PD, -4.41 years; 95% CI -7.14 to -1.68) and haemorrhagic stroke-free (first PD, -9.51 years; 95% CI -12.77 to -6.24) period than never use. Combined HT when initiated late was significantly associated with a shorter haemorrhagic stroke-free period (first PD, -1.97 years; 95% CI -3.81 to -0.13), but not with a shorter stroke-free period (fifth PD, -1.21 years; 95% CI -3.11 to 0.68) than never use. Given the observational nature of this study, the possibility of uncontrolled confounding cannot be excluded. Further, immortal time bias, also related to the observational design, cannot be ruled out. CONCLUSIONS: When initiated early in relation to menopause onset, HT was not associated with increased risk of incident stroke, regardless of the route of administration, type of HT, active ingredient, and duration. Generally, these findings held also for haemorrhagic stroke. Our results suggest that the initiation of HT 0-5 years after menopause onset, as compared to never use, is associated with a decreased risk of stroke and haemorrhagic stroke. Late initiation was associated with elevated risks of stroke and haemorrhagic stroke when conjugated equine oestrogen was used as single therapy. Late initiation of combined HT was associated with haemorrhagic stroke risk.

Estimating incidence and prevalence from population registers: example from myocardial infarction.

AIM: To illustrate how the fundamental epidemiological measures, incidence rate and prevalence proportion, can be estimated based on Swedish population registers using acute myocardial infarction (MI) as an example, together with a discussion about the analytical decisions. METHODS: All individuals in Sweden aged 60-89 (born 1904-1954) during the study period 1994-2014 were identified through the Total Population Register. Cases of MI were defined and identified from information on hospital admissions and causes of death. Incidence rates of all, first, and recurrent MI were calculated together with prevalence proportions. RESULTS: The incidence rate of all, first, and recurrent MI declined over the study period. While the incidence rates of first MI are lower for women than men, the incidence rates of recurrent MI are considerably higher but similar for men and women. The prevalence calculated with duration of disease set at 28 days also declined. This was despite improved survival from MI and increased life expectancy over the same period meaning that the decline in incidence was large enough to compensate for increased survival. CONCLUSIONS: Calculating incidence and prevalence of diseases using population registers requires detailed and well-reasoned definitions. The definitions will affect both the study population and the number of disease events and it is essential that the cases and the study population are defined in a coherent way. Different measures of disease occurrence contribute with different aspects of the disease panorama and a joint interpretation contributes to a thorough understanding of the disease development in a population.

What's in a Word? Concept mapping: a graphical tool to reinforce learning of epidemiological concepts.

Epidemiology is founded on central concepts and principles, essential for conducting, reporting and critically assessing epidemiological studies. Definitions of the many concepts used in the field can be found in textbooks and via the Dictionary of Epidemiology. However, central epidemiological concepts are labelled and used in multiple ways, leading to potential misunderstanding when communicating in different fora. The aim here is to describe collaborative concept mapping, and illustrate how it can be used in teaching and learning epidemiology. Concept mapping is a cognitive technique that is widely used in the education of medical and allied health professions as a tool for critical thinking, and to assimilate new knowledge, but it is still under-utilised in epidemiology. A specific concept map is defined by the aim and question in focus; it is thus framed by a context. The concept map is constructed using a set of concepts (nodes) that are linked with arrows or lines (links). Words and phrases (connective terms) are used to explain relationships between the concepts linked. Different domains can be interconnected by linking concepts in different areas (cross-links). The underlying structure of knowledge is often complex, and consequently concept maps can be constructed using different topological features. Here we provide an illustrative example of concept mapping, based on a set of 'basic' concepts introduced in a doctoral course in epidemiology. In summary, concept mapping is a compelling, active learning tool, which can promote shared deeper knowledge of concepts and their complex interconnections, thereby facilitating a better understanding of epidemiological research.

Duffy antigen receptor genetic variant and the association with Interleukin 8 levels.

UNLABELLED: The aim of this study is to identify loci associated with circulating levels of Interleukin 8 (IL8). We investigated the associations of 121,445 single nucleotide polymorphisms (SNPs) from the Illumina 200K CardioMetabochip with IL8 levels in 1077 controls from the Stockholm Heart Epidemiology Program (SHEEP) study, using linear regression under an additive model of inheritance. Five SNPs (rs12075A/G, rs13179413C/T, rs6907989T/A, rs9352745A/C, rs1779553T/C) reached the pre-defined threshold of genome-wide significance (p<1.0×10(-5)) and were tested for in silico replication in three independent populations, derived from the PIVUS, MDC-CC and SCARF studies. IL8 was measured in serum (SHEEP, PIVUS) and plasma (MDC-CC, SCARF). The strongest association was found with the SNP rs12075 A/G, Asp42Gly (p=1.6×10(-6)), mapping to the Duffy antigen receptor for chemokines (DARC) gene on chromosome 1. The minor allele G was associated with 15.6% and 10.4% reduction in serum IL8 per copy of the allele in SHEEP and PIVUS studies respectively. No association was observed between rs12075 and plasma IL8. CONCLUSION: rs12075 was associated with serum levels but not with plasma levels of IL8. It is likely that serum IL8 represents the combination of levels of circulating plasma IL8 and additional chemokine liberated from the erythrocyte DARC reservoir due to clotting. These findings highlight the importance of understanding IL8 as a biomarker in cardiometabolic diseases.

Does menopausal hormone therapy reduce myocardial infarction risk if initiated early after menopause? A population-based case-control study.

OBJECTIVE: This study aims to assess whether the timing of menopausal hormone therapy initiation in relation to onset of menopause and hormone therapy duration is associated with myocardial infarction risk. METHODS: This study was based on the Stockholm Heart Epidemiology Program, a population-based case-control study including 347 postmenopausal women who had experienced a nonfatal myocardial infarction and 499 female control individuals matched for age and residential area. Odds ratios (with 95% CIs) for myocardial infarction were calculated using logistic regression. RESULTS: Early initiation of hormone therapy (within 10 y of onset of menopause or before age 60 y), compared with never use, was associated with an odds ratio of 0.87 (95% CI, 0.58-1.30) after adjustments for lifestyle factors, body mass index, and socioeconomic status. For late initiation of hormone therapy, the corresponding odds ratio was 0.97 (95% CI, 0.53-1.76). For hormone therapy duration of 5 years or more, compared with never use, the adjusted odds ratio was 0.64 (95% CI, 0.35-1.18). For hormone therapy duration of less than 5 years, the odds ratio was 0.97 (95% CI, 0.63-1.48). CONCLUSIONS: Neither the timing of hormone therapy initiation nor the duration of therapy is significantly associated with myocardial infarction risk.

Association of interleukin 8 with myocardial infarction: results from the Stockholm Heart Epidemiology Program.

BACKGROUND: Interleukin 8 (IL8) has been contradictorily associated with the risk of myocardial infarction (MI). AIM: To investigate the association of IL8 serum levels with the risk of MI and the association of the IL8 (IL8) and IL8 receptors (CXCR1 and CXCR2) genetic variants with IL8 levels and MI risk in a large case control study, the Stockholm Heart Epidemiology Program. METHODS AND RESULTS: IL8 levels (pg/mL) were divided into quartiles and the MI risk was calculated by logistic regression and expressed as odds ratio (OR) and 95% CI. Two IL8 SNPs (rs4073A/T, rs2227306C/T) and three SNPs tagging CXCR1 and CXCR2 (rs4674258C/T, rs1008563C/T, rs6723449T/C) were analyzed for association with IL8 levels and with MI risk. Multivariate adjusted ORs for MI risk by IL8 levels in the highest quartiles indicated reduced point estimates in both women (OR 0.37; 95% CI 0.2-0.8) and men when compared to the lowest quartile. In female cases, IL8 levels decreased progressively in the six months after MI (p=0.03). IL8, CXCR1 and CXCR2 genetic variants were not associated with IL8 levels. In men, the T allele at the IL8 SNP rs4073 was associated with a slight increase in the MI risk under an additive and a recessive model of inheritance. CONCLUSIONS: IL8 serum levels were associated with a reduced occurrence of MI among women, whereas IL8 was associated with a slightly increased risk among men, possibly through different mechanisms. These data suggest that the biological effects of IL8 on MI risk may vary over time and warrant further cohort studies with repetitive IL8 measurements.

Patterns of background factors related to early RA patients' conceptions of the cause of their disease.

The aim of the present study was to identify patterns of background factors related to the early RA patients' conceptions of the cause of the disease. Conceptions from a qualitative study formed the basis for the stratification of 785 patients from the Swedish EIRA study answering a question about their own thoughts about the cause to RA. Logistic regression analyses were used to explore the associations between patients' conceptions and relevant background factors: sex, age, civil status, educational level, anti-cyclic citrullinated peptide antibody (anti-CCP) and smoking habits. The results were presented as odds ratios (OR) with 95% confidence intervals (CI). A conception of family-related strain was strongly associated with being young (OR 0.50; 95% CI 0.33-0.78 for age 58-70 vs. 17-46), female (OR 0.38; 95% CI 0.25-0.60 for male vs. female) and having a high level of education (OR 2.15; 95% CI 1.54-3.01 for university degree vs. no degree). A conception of being exposed to climate changes was associated with being male (OR 1.99; 95% CI 1.24-3.22 for male vs. female), having a low level of education (OR 0.33; 95% CI 0.18-0.58 for university degree vs. no degree) and positive Anti-CCP (OR 1.72; 95% CI 1.03-2.87 for positive vs. negative Anti-CCP). Linking patients' conceptions of the cause of their RA to background factors potentially could create new opportunities for understanding the complexity of the aetiology in RA. Furthermore, this information is important and relevant in the care of patients with early RA.

Patients' conceptions of the cause of their rheumatoid arthritis: a qualitative study.

BACKGROUND: Patients' perspective of the causes and consequences of rheumatoid arthritis (RA) can conflict with that of healthcare professionals and lead to misunderstanding, difficulties in management and a poorer outcome. OBJECTIVES: The aim of this study was to describe the variation in how patients conceive the cause of their RA. METHODS: An open written question from the Epidemiological Investigation of Rheumatoid Arthritis (EIRA) study, aimed at patients recently diagnosed with RA, was answered by 38 strategically selected patients during 2003 and analysed using the phenomenographic approach. RESULTS: Two descriptive categories and six concepts emerged: the category 'consequences beyond personal control' comprised not having a clue, being exposed to climatic change, being genetically exposed and unexpected effects of events; the category 'overloaded circumstances' involved work and family-related strain. Consequences beyond personal control implied that the patients could not prevent the disease and expressed their lack of understanding as to why they contracted it. Overloaded circumstances were described as strained situations that were both work and family related and could be influenced by the patient. CONCLUSIONS: The patient's perspective of the cause of their RA includes aspects that complement the current pathogenetic models and should therefore be considered in the management of the disease. When dealing with rheumatic diseases, it is necessary to be aware of the patient's perspectives in order to new management strategies. In addition to epidemiological studies, further studies of patients' own experience are needed in order to achieve a more tailored care model.

The influence of prognostic factors on neck pain intensity, disability, anxiety and depression over a 2-year period in subjects with acute whiplash injury.

The influence of potential prognostic factors (occupant- and crash-related factors, initial neck pain intensity and headache, whiplash injury severity, helplessness, locus of control, socioeconomic status) on neck pain intensity (VAS), disability (DRI), anxiety and depression (HADS) was estimated in a cohort of 3704 subjects with whiplash injury following a motor vehicle crash. Questionnaires were administered (baseline, 1-, 6-, 12-, 24-month follow-ups). VAS was trichotomized; "low" (0-30), "moderate" (31-54), "severe" (55-100). A cumulative logit model with a proportional odds assumption was applied. Results regarding depression differed somewhat from the other outcomes. Overall, initial neck pain intensity was an important prognostic factor, but acted also as an evident effect modifier. Females had slightly increased odds for all outcomes but depression, for which no gender differences were shown. Injury severity was associated with all outcomes, but was most pronounced regarding disability among those who perceived numbness/pain in arms/hands and also had severe initial neck pain (proportional odds ratio [OR] 6.5; 95% confidence interval [CI] 2.5-17.0). Initial headache influenced all outcomes. Income was not related to any of the outcomes, whereas a lower level of education was associated with all outcomes but depression. Locus of control was not a factor of importance. In contrast, helplessness was related to all outcomes, but was most pronounced regarding neck pain intensity and depression for subjects with severe initial neck pain (OR 4.8; 95% CI 2.9-7.8; OR 6.6; 95% CI 2.6-17.0). Associations seem to be established early, and then to be relatively constant over time.

Low back pain after traffic collisions: a population-based cohort study.

STUDY DESIGN: A population-based, incidence cohort study was conducted. OBJECTIVE: To measure the incidence and prognosis for collision-related low back pain before and after a change in the insurance compensation system. SUMMARY OF BACKGROUND DATA: Low back pain is a common and costly occupational injury. It also occurs after traffic collisions, but less is known about its frequency and recovery in this setting. METHODS: An incidence cohort of 4473 low back pain injury claims was formed between July 1, 1994 and December 31, 1995 in Saskatchewan. On January 1, 1995 the public insurance system changed from a tort system to a no-fault system, eliminating compensation for pain and suffering. The incidence of claims and the time to claim closure were calculated before and after this change. Prognostic models were built using baseline and follow-up data. RESULTS: The 6-month incidence of claims decreased from 256 to 176 per 100,000 after the insurance change. The median time to claim closure dropped from 505 days for tort claims to 210 days and 216 days for claims made during the first and second 6 months of the no-fault period. Improvements in bodily pain and physical functioning and the absence of depressive symptoms were associated with faster claim closure. High pain intensity, female gender, full-time employment, concentration problems, and lawyer involvement early in the claim process delayed claim closure. CONCLUSIONS: Low back pain is a common traffic injury with a prolonged recovery. Its incidence and prognosis are affected by multiple factors, including the type of compensation system. Our study suggests that biopsychosocial factors are important in determining prognosis.

Occupant- and crash-related factors associated with the risk of whiplash injury.

PURPOSE: Given that a motor vehicle crash (MVC) had occurred, to evaluate whether occupant- and crash-related factors, such as age, gender, seating position and type of MVC are associated with the risk of whiplash injury. METHODS: A study of occupants in cars covered by motor insurance at one of the largest insurance companies in Sweden, was undertaken during a one-year period. The study population comprised all occupants in cars exposed to an MVC in which at least one occupant was injured (n = 7120). Adjusted estimates of the relative risk of whiplash injury, associated with the different factors, were calculated by means of binomial regression analysis. RESULTS: Considering different MVCs, rear-end collisions were associated with the highest relative risk of whiplash injury when compared with side impacts (1.82; 95% CI 1.68-1.96), while drivers showed the strongest association with respect to seating position when compared with passengers in the rear seat (1.78; 95% CI 1.60-1.97). Females had a somewhat higher relative risk of whiplash injury than males (1.20; 95% CI 1.16-1.25). Regarding age, the relative risk was moderately increased across the different age groups when compared with the oldest age group. No interaction was observed on the additive scale. CONCLUSIONS: Given that an MVC had occurred, subjects exposed to a rear-end collision and drivers had a substantial increased risk of whiplash injury, while age and gender were of minor importance.