The relativity of relative risks: disadvantage or opportunity?
Explore the source record for details and available documents.
Humans↗
Biomedical subjects
J Neeleman
Publications and source records attributed to J Neeleman.
Quantifying psychiatric comorbidity--lessions from chronic disease epidemiology.
BACKGROUND: Comorbidity research in psychiatric epidemiology mostly uses measures of association like odds or risk ratios to express how strongly disorders are linked. In contrast, chronic disease epidemiologists increasingly use measures of clustering, like multimorbidity (cluster) coefficients, to study comorbidity. This article compares measures of association and clustering. METHODS: Narrative review, algebraical examples, a secondary analysis of an existing dataset and a pooled analysis of published data. RESULTS: Odds and risk ratios, but the former more than the latter, confound clustering with coincidental comorbidity. Multimorbidity coefficients provide a pure estimate of clustering which is the proportion of the association between disorders that is of etiological interest. Odds and risk ratios can express comorbidity between no more than two disorders, whilst clustering coefficients, although computationally laboursome, can capture multimorbidity of any number of disorders. Cluster coefficients depend less on the prevalence of illness in study groups than measures of association. CONCLUSION: Odds and risk ratios are well suited for comorbidity research which focuses on which sets of disorders or syndromes tend to occur in combination and the implications of this for, for instance, nosological classification, a traditional interest of psychiatric epidemiology. However, the cluster coefficient is to be preferred if the interest is more aetiological, addressing for example why certain individuals are prone to multiple health problems.
Alcoholism↗
Propensity to psychiatric and somatic ill-health: evidence from a birth cohort.
BACKGROUND: Somatic and psychiatric morbidity may cluster because of reciprocal effects between them but also as a result of common underlying factors. METHODS: The data come from the 1946 MRC birth cohort (N = 5362). Clustering of 20 chronic medical conditions at the participants' 43rd year was analysed using multi-morbidity coefficients. Associations of somatic and psychiatric morbidity at 43 years, with temperament and behaviour, assessed from adolescence to early adulthood, and background variables like paternal social class, were examined using structural equation modelling. The focus was on whether links applied to both morbidity types or one only, and were direct or were indirect, mediated by prior morbidity. RESULTS: Reported chronic medical conditions clustered strongly. Somatic multi-morbidity and psychiatric ill-health at 43 years were associated with temperamental and behavioural features assessed between the subjects' 13th and 26th birthday, like neuroticism and aggression, as well as with external variables like parental death before the participants' 16th birthday. However, only neuroticism holds direct links with somatic as well as with psychiatric ill-health, 28% of the former and 52% of the latter association being independent of the simultaneous presence of the other morbidity type and of participants' health status 7 years earlier. CONCLUSIONS: Personality traits like neuroticism not only raise the risk of psychiatric disorder but also, irrespective of whether manifest psychiatric disorders have developed, of a broad spectrum of chronic somatic diseases. This suggests that clinicians and researchers should focus not only on the psychiatric disorders associated with such personality traits but also on their medical consequences.
Adult↗
Beyond risk theory: suicidal behavior in its social and epidemiological context.
The effect of exposure to risk factors for suicidal behavior varies from place to place and from period to period. This may be due to contextual influences, which arise if individuals' suicide risk depends not only on their personal exposure to risk or protective factors, but also on how these are distributed in their social, cultural, economic, or even physical environments. There has been relatively little explicit attention in suicide research for such contextual influences even though they are an important component of the cross-level bias, which can arise when aggregate level associations are assumed to also apply in individuals and vice versa. Contextual effects are conceptually related to the issues of social selection vs. causation, population density, and network effects. Because of a lack of prospective multilevel research, it is unclear exactly which mechanisms underlie the phenomenon that the distribution of risk factors in an individual's context may affect their suicide risk above and beyond their personal exposure. A number of mechanisms, like deviancy amplification, formalization of restraints, and buffering effects of social support are proposed. Contextual effects may result in a concentration of suicide risk in persons when the risk factors they are exposed to become rare--whether spontaneously or through focused prevention. This has important but mostly overlooked implications for population-based prevention strategies.
Causality↗
[Suicidality].
Prevention of suicidal behaviour remains difficult, despite increasing knowledge of its determinants. Health service efforts hardly affect suicide rates. Recent shifts in the epidemiology of suicidal behaviour are rising rates among the young and increasing use of violent methods; these can be linked to emerging aetiological insights. Early stages of the suicidal process often manifest themselves at an early age, as emotional and behavioural instability. Epidemiological evidence suggests that the suicidal process increasingly concentrates itself in a vulnerable minority in whom it progresses rapidly to a phase of relative unresponsiveness to environmental influences including preventive efforts. Thus, prevention should focus not only on persons in the later stages of the suicidal process like psychiatric patients, but especially on individuals who, as youngsters, show signs of entering its very first stages. Since high suicide risk implies high risk of other adverse health outcomes as well, this should, in time, yield health gains in more domains than suicide reduction only.
Age Factors↗
A continuum of premature death. Meta-analysis of competing mortality in the psychosocially vulnerable.
BACKGROUND: Suicide may be an extreme expression of liability to death of any type. If true, suicide risk factors should also increase other mortality, and, given exposure, excess risk should be higher for suicide than for other mortality. METHODS: Of 304 publications identified in Index Medicus (1966-June 1988) by the string (suicide) and (mortality or death) and (accidental or natural), 24 reported total and cause-specific mortality associated with exposure to 16 established suicide risk factors; reference scanning yielded 122 more. These 146 publications reported on 163 cohorts (total subjects = 1179 126) mortality. Meta-analysis gave random effects standardized mortality ratios (SMR) for natural, accidental and suicidal death, stratified over the 16 risk groups. RESULTS: Overall, SMR were 8.6 (95% CI : 7.1-10.4) for suicide, 3.4 (95% CI : 2.9-4.0) for accidental and 2.1 (95% CI : 1.9-2.3) for natural death. Compatible with the first hypothesis, in most groups, mortality of any type was raised. Supporting the second hypothesis, excesses increased from lowest for natural death to highest for suicide. This trend was most pronounced following deliberate self-harm, intermediate in substance abusers, and weakest, but present, in bereaved and low social class cohorts and reversed in smokers and epileptic people. CONCLUSIONS: Many suicide risks apply to any type of premature death, whilst also retaining some specificity for suicide. Primary prevention, targeting such generic risk factors, will not only reduce rates of suicide but also of other types of death. Conversely, when prevention focuses on specific outcomes-such as suicide-only, other types of mortality may increase.
Humans↗
Ethnic density and deliberate self harm; a small area study in south east London.
STUDY OBJECTIVE: Relative risks are frequently used to convey how strongly outcomes like mental illness and suicidal behaviour are associated with personal characteristics like ethnic background. This study examined whether RRs for deliberate self harm (DSH) in ethnic groups vary between small areas according to their ethnic mix. DESIGN: Small area study of DSH rates in ethnic groups, by local ethnic density, using negative binomial regression. SETTING: 73 south London electoral wards, 1994-1997. SUBJECTS: 1643 people attending casualty after DSH. MAIN RESULTS: African-Caribbean and Asian DSH rates, relative to the white population, varied between wards. A linear model indicated a decline by factors (relative rate ratios) 0.76 (95% confidence intervals (CI) 0.64 to 0.90) and 0.59 (95% CI 0.36 to 0.97) respectively per SD increase in the local size of these minority populations. However, for both groups, an inverted U shaped curve provided a better fit for the link between the relative DSH rate in these groups and their local population density. CONCLUSIONS: The DSH rate of minority groups relative to the white group is low (suggesting protection) in some areas, and high (suggesting risk) elsewhere. This has implications for management of suicidal behaviour in ethnic groups but also for interpretation, and policy implementation, of research on risk for suicidal behaviour, and, probably, many other outcomes. Relative risks or rates are not stable indicators of association in psychiatric epidemiology.
Adult↗
Predictors of time to remission from depression in primary care patients: do some people benefit more from positive life change than others?
The authors examined the role of personality, social support, and coping strategies and their interaction with positive life change (PLC) in influencing time to remission from depression in 86 depressed primary care patients (69% women, mean age = 36.6 years), who were followed for 3.5 years. Diagnostic information was obtained by the Present State Examination (J. K. Wing, J. Cooper, & N. Sartorius, 1974), and life change by the Life Events and Difficulties Schedule (G. W. Brown & T. O. Harris, 1978). The association of PLC and other predictors with the time to remission was studied by means of Cox regression. PLC reduced time to remission in women, but not in men. Other predictors that expedited remission were low severity of premorbid difficulties, high self-esteem, and a tension-reducing coping style. Neuroticism modified the effect of PLC in that the remission forward bringing effect of PLC was stronger for highly neurotic people. The results suggests that women and psychologically vulnerable persons benefit most from PLC.
Adult↗
Ethnic minority suicide: a small area geographical study in south London.
BACKGROUND: The relationship between ethnicity and suicide risk is ill-understood. It is unclear whether, and if so, how, the ethnic mix of local areas affects risk in local individuals. METHODS: Coroners' records of 329 suicides were used to obtain ethnic (White, Afro-Caribbean, Asian) suicide rates in South London (population 902,008) for 1991-3. Geographical variation and associations of ethnic suicide rates with small area (mean population 8274) ethnic densities (proportion of residents of given ethnic groups) and deprivation, were examined with random effects Poisson regression. RESULTS: Adjusted for deprivation, age and gender, suicide rates in wards with larger minority groups were higher among Whites (relative rate (RR) per standard deviation (S.D.) increase in minority density 1.18; 95% CI 1.02-1.37) but lower among minority groups (RR 0.75 (0.59-0.96)) (LR-test for interaction chi2 = 9.2 (df = 1); P = 0.003). Similar patterns were also apparent for Afro-Caribbeans and Asians separately. With White suicide rates as baseline, ethnic minority status is a risk factor for suicide in wards with small, but a protective factor in neighbourhoods with large minority populations. The RR of minority versus White suicide declines with a factor (relative RR) 0.67 (0.51-0.87) per S.D. increase in local minority density. CONCLUSIONS: Minority suicide rates are higher in areas where minority groups are smaller. This effect is ethnic-specific and not due to confounding by gender, age, deprivation or unbalanced migration. Dependent on address, a suicide risk factor for a White individual may protect an ethnic minority individual and vice versa. This has implications for research and prevention.
Adult↗
Suicide, religion, and socioeconomic conditions. An ecological study in 26 countries, 1990.
STUDY OBJECTIVE: Relative risks are frequently assumed to be stable across populations but this may not apply in psychiatric epidemiology where sociocultural context may modify them. Such ecological effect modification will give curved associations between aggregated risk factor and outcome. This was examined in connection with the ecological association between suicide rates and an aggregate index of religiosity. DESIGN: Ecological study of associations between suicide rates and an index of religiosity, adjusted for socioeconomic variation. The effect of stratification of the study sample according to levels of religiosity, was examined. SETTING: 26 European and American countries. SUBJECTS: Interview data from 37,688 people aggregated by country. OUTCOME MEASURES: Age and sex specific (1986-1990) suicide rates. MAIN RESULT: Adjusted for socioeconomic variation, negative associations of male suicide rates with religiosity were apparent in the 13 least religious countries only (test for interaction F (1, 25) = 5.6; p = 0.026). Associations between religiosity and female suicide rates did not vary across countries. CONCLUSION: The bent ecological association was apparent only after adjustment for socioeconomic variation suggesting that, rather than confounding, ecological modification of individual level links between religion and male (but not female) suicide risk is the responsible mechanism. This concurs with micro-level findings suggesting that suicide acceptance depends not only on personal but also on contextual levels of religious belief, and that men are more sensitive to this phenomenon than women. In psychiatric epidemiology, relative risks vary with the exposure's prevalence. This has important implications for research and prevention.
Adolescent↗
Predictors of suicide, accidental death, and premature natural death in a general-population birth cohort.
BACKGROUND: Whether putative suicide risk factors, such as conduct and emotional disorders, are specific to suicide or are general associations of a continuum between subintentional and intentional self-destruction is not clear. We undertook an investigation of this issue in a UK population-based birth cohort. METHODS: Using competing-risks analysis, we examined links between prospectively collected childhood and adolescent temperamental and behavioural variables and the risk of natural, accidental, and suicidal death, occurring between the ages of 16 and 50 years. Of the 5362 members of the cohort, full data were available for 3591. A panel of psychiatrists scored deaths on likely suicidal intention. These scores were used in a weighted logistic regression to examine independence of risk factors for (sub)intentional self-destruction. FINDINGS: There were 167 deaths among the risk-set between the 16th and 50th birthdays. 120 were due to natural causes; of the other 47, the panel of psychiatrists judged that 36 were accidental deaths and 11 were suicides (certainty rating > or = 51%). Adolescent emotional instability and conduct problems had different associations with the various causes of premature death, being lowest for natural death (odds ratio 1.0 [95% CI 0.8-1.2] and 1.2 [1.0-1.5]), intermediate for accidental death (1.1 [0.8-1.5] and 1.3 [1.0-1.7]), and highest for suicidal death (2.0 [1.2-3.6] and 1.8 [1.3-2.5]). Emotional instability (boys 1.3 [1.0-1.7], girls 1.4 [1.0-1.9]) increased risk for subintentional or intentional self-destruction, as did low anxiety in adolescence (1.7 [1.3-2.5]) and nocturnal enuresis at age 4 (1.4 [1.2-1.7]) in boys, and conduct problems in girls (1.4 [1.0-2.0]). INTERPRETATION: Suicide shares important risk factors with other causes of premature death. These findings suggest an aetiological continuum of self-destruction from subintentional to intentional. Research on premature mortality associated with psychological disorder should include all causes of death.
Accidents↗
Regional suicide rates in the Netherlands: does religion still play a role?
BACKGROUND: This study examined the nature of ecological associations between 'religiousness' and suicide rates (1985-1994) in the 11 provinces in the Netherlands. METHODS: Indices of religiousness, obtained from a nationwide survey, were used as aggregate predictors of provincial suicide rates in weighted linear regressions, and as individual-level predictors of suicide acceptance in logistic regressions. Sociodemographic confounding was controlled for. RESULTS: Orthodox beliefs and religious affiliation were the best predictors of lower suicide acceptance in individuals and of lower suicide rates in provinces. The ecological association was most pronounced in the least religious parts of the country giving rise to a curvilinear ecological regression line. CONCLUSIONS: Curvilinear ecological regression lines arise when mean levels of exposure affect individual risk above and beyond personal exposure i.e. when there is ecological effect modification. This study demonstrates that such contextual effects, responsible for cross-level bias, apply to the association between suicide and religiousness. Variation, from context to context, of the effects of exposure to psychosocial risk or protective factors for outcomes such as suicide, has important implications for research and prevention.
Adolescent↗
Suicide acceptability in African- and white Americans: the role of religion.
Rates of suicidal behavior are lower among African- than white Americans. We analyzed the association of suicide acceptability with religious, sociodemographic, and emotional variables in representative samples of African- and white Americans (1990). Adjusted for ethnic response bias, the former were less accepting of suicide than the latter (odds ratio .60; 95% confidence interval .41, .88). Orthodox religious beliefs and personal devotion predicted rejection of suicide best; this effect was equally strong in both groups. The comparatively low level of suicide acceptability among African-Americans was mostly attributable to their relatively high levels of orthodox religious beliefs and devotion, as opposed to practice and affiliation, although sociodemographic and emotional differences contributed as well. These results are interpreted using the cognitive dissonance model. Given rapid secularization among the young in the United States, these findings may help explain the rising suicide rates among white and, especially, African-American young people.
Black or African American↗
Changes in classification of suicide in England and Wales: time trends and associations with coroners' professional backgrounds.
BACKGROUND: The legal definition of suicide in England and Wales (E & W) gives rise to a high proportion of open verdicts and an underestimated suicide rate. We examined whether the ratio between open and suicide verdicts in E & W has changed between 1974 and 1991 and whether it varies according to coroners' qualifications. METHODS: Temporal changes of the ratio of open and suicide verdicts were examined using logistic regression adjusting for confounders such as changing age and gender distributions of suicide victims and the methods they use. RESULTS: Adjusted for age at death and suicide method, the ratio between open and suicide verdicts had, over successive 3-year time periods, increased with a factor 1.21 (95% CI 1.20-1.23) for male and 1.15 (95% CI 1.14-1.17) for female deaths. Medical coroners were 1.25 (1.08-1.44) times more likely than non-medical coroners to return open rather than suicide verdicts. CONCLUSIONS: As a likely result of factors in the death registration system, the ratio between open and suicide verdicts has increased substantially in E & W since the early 70s. In 1990 it was higher than in any other comparable country. This has important implications for comparisons of time trends in suicide between E & W and other countries.
Adolescent↗
Tolerance of suicide, religion and suicide rates: an ecological and individual study in 19 Western countries.
BACKGROUND: Negative associations between religion and suicide, in individuals and countries, may be mediated by the degree to which suicide is tolerated. METHODS: Linear regression was used to examine ecological associations between suicide tolerance, religion and suicide rates in 19 Western countries in 1989/90. Logistic regression was used to study associations between suicide tolerance and strength of religious belief in 28085 individuals in these countries. The concept of effect modifying function was used to examine whether the strength of the association between suicide tolerance and religious belief in individuals depended on the extent of religious belief in their country. RESULTS: Higher female suicide rates were associated with lower aggregate levels of religious belief and, less strongly, religious attendance. These associations were mostly attributable to the association between higher tolerance of suicide and higher suicide rates. In the 28085 subjects suicide tolerance and the strength of religious belief were negatively associated even after adjustment for other religious and sociodemographic variables and general tolerance levels (odds ratios: men 0.74 (95% CI 0.58-0.94), women 0.72 (95% CI 0.60-0.86)). This negative individual-level association was more pronounced in more highly religious countries but this modifying effect of the religious context was apparent for men only. CONCLUSIONS: Ecological associations between religious variables and suicide rates are stronger for women than men, stronger for measures of belief than observance and mediated by tolerance of suicide. In individuals, stronger religious beliefs are associated with lower tolerance of suicide. Personal religious beliefs and, for men, exposure to a religious environment, may protect against suicide by reducing its acceptability.
Adult↗
Drugs taken in fatal and non-fatal self-poisoning: a study in south London.
This study compared the number and type of substances taken in deliberate self-poisoning with fatal (n = 127) and non-fatal (n = 521) outcome. The aims were (i) to describe substances typically involved in self-poisoning in England and Wales, (ii) to examine the role of drug "cocktails' and (iii) to examine whether toxic substances are over-represented in cases with fatal outcome. Over-the-counter (OTC) analgesics, minor tranquillizers and antidepressants accounted for about 70% of substances taken, irrespective of outcome. Compared with survivors, cases who died had taken a higher mean number of substances. Among self-poisonings with a single substance, antidepressants and paracetamol-opiate combinations were over-represented in fatal-outcome cases. This report emphasizes the role of OTC analgesics and antidepressants in overdose-related mortality in England and Wales.
Adult↗
Fatal methadone and heroin overdoses: time trends in England and Wales.
STUDY OBJECTIVE: Although the total number of self poisonings in England and Wales has dropped by 32%, the number involving methadone and/or heroin rose by 900% in 1974-92. Because of concern about the role of methadone in this increase, the part played by methadone and heroin in poisoning deaths in England and Wales in 1974-92 was investigated. DESIGN: A proportional mortality design was used to study whether the ratio between deaths involving methadone or heroin and other substances had increased. The time trend was examined with logistic regression. SETTING: England and Wales, 1974-92. SUBJECTS: Accidental, undetermined, and suicidal poisoning deaths (n = 43,231). MAIN RESULTS: The proportions of poisoning deaths involving methadone (alone or in combination with heroin) rose by 80% (95% CI 69%, 92%) per 3 year period. The proportion of poisoning deaths involving heroin without methadone rose by 76% (95% CI 60%, 93%) per 3 year period. Similar results were obtained when poisoning deaths were examined in relation to gender and legal category (suicide and undetermined versus accidental deaths). CONCLUSIONS: The impact of opiate addiction on rates of death by poisoning is rising quickly. This may reflect the growth of the addict population and is an important public health problem. There is no evidence that methadone's involvement in these deaths has risen disproportionately in relation to that of heroin up to 1992.
Accidents↗