Browsing by Subject "Bias (Epidemiology)"

Now showing 1 - 6 of 6
  • Thumbnail Image
    ItemOpen Access
    A brief history of research synthesis.
    (Eval Health Prof, 2002-03) Chalmers, Iain; Hedges, Larry V; Cooper, Harris
    Science is supposed to be cumulative, but scientists only rarely cumulate evidence scientifically. This means that users of research evidence have to cope with a plethora of reports of individual studies with no systematic attempt made to present new results in the context of similar studies. Although the need to synthesize research evidence has been recognized for well over two centuries, explicit methods for this form of research were not developed until the 20th century. The development of methods to reduce statistical imprecision using quantitative synthesis (meta-analysis) preceded the development of methods to reduce biases, the latter only beginning to receive proper attention during the last quarter of the 20th century. In this article, the authors identify some of the trends and highlights in this history, to which researchers in the physical, natural, and social sciences have all contributed, and speculate briefly about the "future history" of research synthesis.
  • Thumbnail Image
    ItemOpen Access
    High and far: biases in the location of protected areas.
    (PLoS One, 2009-12-14) Joppa, Lucas N; Pfaff, Alexander
    BACKGROUND: About an eighth of the earth's land surface is in protected areas (hereafter "PAs"), most created during the 20(th) century. Natural landscapes are critical for species persistence and PAs can play a major role in conservation and in climate policy. Such contributions may be harder than expected to implement if new PAs are constrained to the same kinds of locations that PAs currently occupy. METHODOLOGY/PRINCIPAL FINDINGS: Quantitatively extending the perception that PAs occupy "rock and ice", we show that across 147 nations PA networks are biased towards places that are unlikely to face land conversion pressures even in the absence of protection. We test each country's PA network for bias in elevation, slope, distances to roads and cities, and suitability for agriculture. Further, within each country's set of PAs, we also ask if the level of protection is biased in these ways. We find that the significant majority of national PA networks are biased to higher elevations, steeper slopes and greater distances to roads and cities. Also, within a country, PAs with higher protection status are more biased than are the PAs with lower protection statuses. CONCLUSIONS/SIGNIFICANCE: In sum, PAs are biased towards where they can least prevent land conversion (even if they offer perfect protection). These globally comprehensive results extend findings from nation-level analyses. They imply that siting rules such as the Convention on Biological Diversity's 2010 Target [to protect 10% of all ecoregions] might raise PA impacts if applied at the country level. In light of the potential for global carbon-based payments for avoided deforestation or REDD, these results suggest that attention to threat could improve outcomes from the creation and management of PAs.
  • Thumbnail Image
    ItemOpen Access
    How common are common mental disorders? Evidence that lifetime prevalence rates are doubled by prospective versus retrospective ascertainment.
    (Psychol Med, 2010-06) Moffitt, TE; Caspi, A; Taylor, A; Kokaua, J; Milne, BJ; Polanczyk, G; Poulton, R
    BACKGROUND: Most information about the lifetime prevalence of mental disorders comes from retrospective surveys, but how much these surveys have undercounted due to recall failure is unknown. We compared results from a prospective study with those from retrospective studies. METHOD: The representative 1972-1973 Dunedin New Zealand birth cohort (n=1037) was followed to age 32 years with 96% retention, and compared to the national New Zealand Mental Health Survey (NZMHS) and two US National Comorbidity Surveys (NCS and NCS-R). Measures were research diagnoses of anxiety, depression, alcohol dependence and cannabis dependence from ages 18 to 32 years. RESULTS: The prevalence of lifetime disorder to age 32 was approximately doubled in prospective as compared to retrospective data for all four disorder types. Moreover, across disorders, prospective measurement yielded a mean past-year-to-lifetime ratio of 38% whereas retrospective measurement yielded higher mean past-year-to-lifetime ratios of 57% (NZMHS, NCS-R) and 65% (NCS). CONCLUSIONS: Prospective longitudinal studies complement retrospective surveys by providing unique information about lifetime prevalence. The experience of at least one episode of DSM-defined disorder during a lifetime may be far more common in the population than previously thought. Research should ask what this means for etiological theory, construct validity of the DSM approach, public perception of stigma, estimates of the burden of disease and public health policy.
  • Thumbnail Image
    ItemOpen Access
    Liberal bias and the five-factor model.
    (Behav Brain Sci, 2015) Charney, Evan
    Duarte et al. draw attention to the "embedding of liberal values and methods" in social psychological research. They note how these biases are often invisible to the researchers themselves. The authors themselves fall prey to these "invisible biases" by utilizing the five-factor model of personality and the trait of openness to experience as one possible explanation for the under-representation of political conservatives in social psychology. I show that the manner in which the trait of openness to experience is conceptualized and measured is a particularly blatant example of the very liberal bias the authors decry.
  • Thumbnail Image
    ItemOpen Access
    Model validity and risk of bias in randomised placebo-controlled trials of individualised homeopathic treatment.
    (Complement Ther Med, 2016-04) Mathie, Robert T; Van Wassenhoven, Michel; Jacobs, Jennifer; Oberbaum, Menachem; Frye, Joyce; Manchanda, Raj K; Roniger, Helmut; Dantas, Flávio; Legg, Lynn A; Clausen, Jürgen; Moss, Sian; Davidson, Jonathan RT; Lloyd, Suzanne M; Ford, Ian; Fisher, Peter
    BACKGROUND: To date, our programme of systematic reviews has assessed randomised controlled trials (RCTs) of individualised homeopathy separately for risk of bias (RoB) and for model validity of homeopathic treatment (MVHT). OBJECTIVES: The purpose of the present paper was to bring together our published RoB and MVHT findings and, using an approach based on GRADE methods, to merge the quality appraisals of these same RCTs, examining the impact on meta-analysis results. DESIGN: Systematic review with meta-analysis. METHODS: As previously, 31 papers (reporting a total of 32 RCTs) were eligible for systematic review and were the subject of study. MAIN OUTCOME MEASURES: For each trial, the separate ratings for RoB and MVHT were merged to obtain a single overall quality designation ('high', 'moderate, "low", 'very low'), based on the GRADE principle of 'downgrading'. RESULTS: Merging the assessment of MVHT and RoB identified three trials of 'high quality', eight of 'moderate quality', 18 of 'low quality' and three of 'very low quality'. There was no association between a trial's MVHT and its RoB or its direction of treatment effect (P>0.05). The three 'high quality' trials were those already labelled 'reliable evidence' based on RoB, and so no change was found in meta-analysis based on best-quality evidence: a small, statistically significant, effect favouring homeopathy. CONCLUSION: Accommodating MVHT in overall quality designation of RCTs has not modified our pre-existing conclusion that the medicines prescribed in individualised homeopathy may have small, specific, treatment effects.
  • Thumbnail Image
    ItemOpen Access
    Power and sample size calculations for the Wilcoxon-Mann-Whitney test in the presence of death-censored observations.
    (Stat Med, 2015-02-10) Matsouaka, Roland A; Betensky, Rebecca A
    We consider a clinical trial of a potentially lethal disease in which patients are randomly assigned to two treatment groups and are followed for a fixed period of time; a continuous endpoint is measured at the end of follow-up. For some patients; however, death (or severe disease progression) may preclude measurement of the endpoint. A statistical analysis that includes only patients with endpoint measurements may be biased. An alternative analysis includes all randomized patients, with rank scores assigned to the patients who are available for the endpoint measurement on the basis of the magnitude of their responses and with 'worst-rank' scores assigned to those patients whose death precluded the measurement of the continuous endpoint. The worst-rank scores are worse than all observed rank scores. The treatment effect is then evaluated using the Wilcoxon-Mann-Whitney test. In this paper, we derive closed-form formulae for the power and sample size of the Wilcoxon-Mann-Whitney test when missing measurements of the continuous endpoints because of death are replaced by worst-rank scores. We distinguish two approaches for assigning the worst-rank scores. In the tied worst-rank approach, all deaths are weighted equally, and the worst-rank scores are set to a single value that is worse than all measured responses. In the untied worst-rank approach, the worst-rank scores further rank patients according to their time of death, so that an earlier death is considered worse than a later death, which in turn is worse than all measured responses. In addition, we propose four methods for the implementation of the sample size formulae for a trial with expected early death. We conduct Monte Carlo simulation studies to evaluate the accuracy of our power and sample size formulae and to compare the four sample size estimation methods.