The Hazards of Hazard Ratios

Department of Epidemiology, Harvard School of Public Health, and the Harvard-MIT Division of Health Sciences and Technology, Boston, MA 02115, USA.
Epidemiology (Cambridge, Mass.) (Impact Factor: 6.2). 01/2010; 21(1):13-5. DOI: 10.1097/EDE.0b013e3181c1ea43
Source: PubMed
4 Reads
  • Source
    • "This study also reported only the weighted average of the period-specific scores, despite changes in the HRs over time. To overcome this limitation , sensitivity analyses calculating a series of average HRs for several durations are suggested (Hernán, 2010). "
    [Show abstract] [Hide abstract]
    ABSTRACT: This study aimed to examine the influences of social, attitudinal, and intrapersonal factors at hree levels on tobacco use among female adolescents in South Korea using longitudinal national data. The study analyzed data from the Korean Youth Panel Study, with a study population consisting of middle-school second-graders (N = 1,594). Using time dependent Cox regression, our analyses yielded the following main findings: the TTI model was verified to provide a theoretical framework for smoking among Korean female adolescents. All of the social factors at the three levels, including parental supervision, attachment to friends, and having smoking friends, were found to influence tobacco use among Korean female adolescents. Stigma on the distal level and attitude toward smoking on the proximal level were significant attitudinal factors. Among intrapersonal factors, self-control on the distal level and stress on the proximal level were found to be significant. The study findings suggest that including parental education and promoting attachment to non-smoking friends, as well as enhancing sound relationships with them, would provide an effective strategy for the prevention and cessation of smoking. Prevention and cession should include strategies that alleviate stigma and stress, and improve negative attitude toward smoking and the level of self-control.
    Children and Youth Services Review 01/2015; 50. DOI:10.1016/j.childyouth.2015.01.009 · 1.27 Impact Factor
  • Source
    • "Hazards as a metric of disease occurrence along with semi-parametric methods have provided a means to understanding many diseases. However, hazards have limitations [15] that become compounded in competing-risk problems: CSHs and SUBHs are decoupled; CSHs lack specificity as they are strongly influenced by the competing event; SUBHs are specific but they are intrinsically tethered because their cumulative incidences must add up to one; both CSHs and SUBHs combine frequency and timing of events, making it difficult to identify exposures that modify only the timing of an event but not the frequency. "
    [Show abstract] [Hide abstract]
    ABSTRACT: A competing risk is an event (for example, death in the ICU) that hinders the occurrence of an event of interest (for example, nosocomial infection in the ICU) and it is a common issue in many critical care studies. Not accounting for a competing event may affect how results related to a primary event of interest are interpreted. In the previous issue of Critical Care, Wolkewitz and colleagues extended traditional models for competing risks to include random effects as a means to quantify heterogeneity among ICUs. Reported results from their analyses based on cause-specific hazards and on sub-hazards of the cumulative incidence function were indicative of lack of proportionality of these hazards over time. Here, we argue that proportionality of hazards can be problematic in competing-risk problems and analyses must consider time by covariate interactions as a default. Moreover, since hazards in competing risks make it difficult to disentangle the effects of frequency and timing of the competing events, their interpretation can be murky. Use of mixtures of flexible and succinct parametric time-to-event models for competing risks permits disentanglement of the frequency and timing at the price of requiring stronger data and a higher number of parameters. We used data from a clinical trial on fluid management strategies for patients with acute respiratory distress syndrome to support our recommendations.
    Critical care (London, England) 05/2014; 18(3):146. DOI:10.1186/cc13892 · 4.48 Impact Factor
  • Source
    • "Alternatively, interval-specific HRs could be modeled as a function of time (or logarithm of time) or of time-varying covariates (or smoothed in some other way) [4]. However, a problem with all these approaches is that interval-specific HRs can suffer from an ''evolving selection bias'' [5]. This goes by various other labels including ''survivor bias'', ''survivor cohort bias'', ''frailty'', and ''depletion of susceptibles''. "
    [Show abstract] [Hide abstract]
    ABSTRACT: To describe the use of risk-difference curves for communicating time-dependent absolute treatment effects. Three examples based on individual patient data meta-analyses for adjuvant treatments for early-stage breast cancer are presented. Unit record datasets were re-created from the published Kaplan-Meier curves and numbers at risk or person-years at risk. Risk-difference curves, with corresponding 95% confidence bands, are presented and discussed. Risk-difference curves are useful for communicating the results from trials of adjuvant treatments for early-stage cancer when standard measures of the absolute treatment effect for survival data (ie, difference-in-mean and difference-in-median survival) can be difficult to estimate. They also avoid the problem of "evolving selection bias", which can affect interval-specific hazard ratio (HR)s in trials with long follow-up and where the participants are heterogeneous with respect to prognosis. Clinical epidemiologists should consider reporting risk-difference curves in addition to Kaplan-Meier curves and the HR.
    Journal of clinical epidemiology 04/2014; 67(9). DOI:10.1016/j.jclinepi.2014.03.006 · 3.42 Impact Factor
Show more

Similar Publications


4 Reads
Available from