Epi today

Articles, reports and books that are selected to highlight emerging epidemiological scholarship in each of the following domains: study design, data collection, measurement, data analysis and visualization and modeling/simulation.
Aug 21
2014
Publication type: Article
Author(s): J Bor, E Moscoe, P Mutevedzi, ML Newell, T Barnighausen
Publication: Epidemiology
Year published: 2014
Why this Article: Regression discontinuity designs are a quasi-experimental design that allow for the estimation of causal effects when exposure or treatment is determined by a threshold of a continuous variable. This article provides a great overview of regression discontinuity designs and how they may be useful for epidemiologists.
Aug 8
2014
Publication type: Article
Author(s): HW Lin, YH Chen
Publication: American Journal of Epidemiology
Year published: 2014
Why this Article: Missing or incomplete data on confounders is a common problem in epidemiology. This article discusses a two-stage calibration method using propensity scores to address missing data on confounders in observational studies. The main drawback is that this approach requires a small subset with complete data. However, in situations where gathering complete data on a subset is feasible this a method worth considering.
Aug 7
2014
Publication type: Article
Author(s): S Neupane, D Doku
Publication: Journal of Epidemiology and Global Health
Year published: 2014
Why this Article: A great application of a multi-level analysis to assess the important individual, community and district level factors that contribute to country-specific neonatal mortality
Jul 7
2014
Publication type: Article
Author(s): N Agahi, B Shaw, S Fors
Publication: Journal of Epidemiology and Community Health
Year published: 2014
Why this Article: This paper is an unusual panel study that is able to link childhood experiences to adult health. It shows that social and economic disadvantages in childhood are associated with an earlier onset and a faster progression of functional health problems from midlife into old age. Mediational analyses suggest that it is differences in foundational factors like education, rather than differences in behaviors, like smoking, that primarily explain some of these associations.
Jul 3
2014
Publication type: Article
Author(s): C Eriksson, A Hilding, A Pyko, G Bluhm, G Pershagen, CG Ostenson
Publication: Environmental Health Perspectives
Year published: 2014
Why this Article: We highlight this paper for discussion. A prospective cohort study finds an association between noise exposure and wait circumference. It is not so clear that there is an association with BMI or type 2 diabetes. This raises a challenge we have noted here before that the epidemiological effort to isolate causes leads to some inferential cul de sacs. Why would noise exposure be associated with waist circumference? The paper suggests that stress processes may be at play. Perhaps. But then why not a link with BMI? We suggest that this represents more the isolation of one very particular component cause but that this isolation is only of consequence if we could understand the full causal architecture underlying the causal relations of interest.
Jun 30
2014
Publication type: Article
Author(s): J Cylus, M Glymour, M Avendano
Publication: American Journal of Epidemiology
Year published: 2014
Why this Article: A good example of the use of additive interactions in epidemiology. The paper shows a negative additive interaction robust over multiple model specifications. The paper also tackles an interesting question showing how some of the consistent findings in epidemiology, i.e., the potential harmful consequences of unemployment, can be mitigated through programmatic action.
Jun 25
2014
Publication type: Article
Author(s): M Chang, RA Hahn, SM Teutsch, LC Hutwagner
Publication: Journal of Clinical Epidemiology
Year published: 2001
Why this Article: This paper, unusually, moves beyond isolating individual causes to note that "Analysis that does not examine risk factor combinations may greatly overestimate PARs associated with individual risk factors.” This is almost certainly the case for all aspects of population health in which we might be interested, but is seldom tackled empirically in our work.
Jun 19
2014
Publication type: Article
Author(s): KJ Psoter, M Rosenfeld, AJ De Roos, J Mayer, J Wakefield
Publication: American Journal of Epidemiology
Year published: 2014
Why this Article: This paper summarizes a robustly carried out geospatial analysis. These analyses are becoming ever more common in Epidemiology. Importantly this paper shows that place matters in shaping health even for a disorder that is genetically determined such as Cystic Fibrosis. The manifestation of a particular phenotype is almost inevitably a product of biological and social exposures.
Jun 17
2014
Publication type: Article
Author(s): A Sarkadi, F Sampaio, MP Kelly, I Feldman
Publication: Journal of Clinical Epidemiology
Year published: 2014
Why this Article: The goal of public health is to improve health at a population-level. Yet we still very often assess the impact of population-level interventions at the individual-level. The authors of this paper propose comparing the distribution of the outcome at baseline with the distribution of the outcome at follow-up as a way to assess the population-level impact of public health interventions. Methods for calculating the difference in distributions, with confidence intervals, are provided in the online appendix. In addition, their approach provides an analytic framework for evaluating the impact of interventions on closing health equity gaps at the population level.
Jun 12
2014
Publication type: Article
Author(s): MS Farvid, E Cho, W Chen, AH Eliassen, WC Willett
Publication: BMJ
Year published: 2014
Why this Article: Women who consume an average of 1.5 servings of red meat per day over 15 years have 1.2 times the risk of breast cancer incidence compared with women who consume almost no red meat, and the increase in risk is only apparent after adjusting for many factors, according to the Nurses Health Study. Concerns about comparability and consequentialism?
Jun 5
2014
Publication type: Article
Author(s): C Akdeniz, H Tost, F Streit et al.
Publication: JAMA Psychiatry
Year published: 2014
Why this Article: This article is a great example of translational epidemiology in action. While an association between ethnic minority status and increased risk of schizophrenia has long been established in the epidemiological literature, the mechanisms underlying these links has remained elusive. New research uses neuroimaging technology to support the hypothesis that chronic stress and discrimination due to marginalized social status may play a role in explaining epidemiological patterns.
Jun 4
2014
Publication type: Article
Author(s): KR de Camargo Jr, F Ortega, CM Coeli
Publication: Revista de Saúde Pública
Year published: 2013
Tags: ,
Why this Article: This paper catalyzes several arguments that have challenged epidemiology as a discipline over the past two decades. It captures the ontological, epistemological, axiologic, and pragmatic critiques facing the field and offers some propositions for ways forward. While the conclusions the authors draw may be debatable, it is hard to disagree with the challenges they articulate.
May 27
2014
Publication type: Article
Author(s): JP Robinson-Cimpian
Publication: Educational Researcher
Year published: 2014
Why this Article: The rigor of our conclusions is often based on the accuracy of survey reporting. A small group of adolescents who mis-report on surveys may drive estimates of some health disparities. Robinson-Cimpian proposes a solution in this recently published study.
May 21
2014
Publication type: Article
Author(s): JE Goldstick, J Trostle, JNS Eisenberg
Publication: American Journal of Epidemiology
Year published: 2014
Why this Article: When studying risk factors for infectious diseases very often they are seen as not changing over time and not dependent on contextual factors such as whether a village is located upstream or downstream from a major city. This study assesses how the regional context, both spatially and temporally, influences local risk factors for diarrheal disease. The results offer important insight into the local prevention of diarrheal disease and the conclusions may be applicable to many other infectious disease risk factors.
May 20
2014
Publication type: Article
Author(s): KM Keyes, R Nicholson, J Kinley, S Raposo, MB Stein, EM Goldner, J Sareen
Publication: American Journal of Epidemiology
Year published: 2014
Why this Article: Age-period-cohort effects are often very difficult to tease apart, with a major limiting factor being a lack of data covering enough time periods and cohorts. This study analyzes data over a 14 year period in order to isolate recent cohort effects of depressive and anxiety symptoms in nontreated populations. The results do support a cohort effect in that more recently born birth cohorts do exhibit higher levels of psychological distress.
May 3
2014
Publication type: Article
Author(s): PC Burkner, P Doebler
Publication: Statistics in Medicine
Year published: 2014
Why this Article: Very often once numerous studies have been published on a particular topic a meta-analysis is conducted in order to generate a summary effect estimate from all the published evidence. However, not all studies conducted are published, and if the non-published studies are not published for systematic reasons, e.g. non-significant results, publication bias may arise. Publication bias takes on a unique twist when it comes to meta-analysis of diagnostic studies, and this article takes an in-depth look at the methods to detect publication bias in these settings.
May 2
2014
Publication type: Article
Author(s): M Shere, XY Zhao, G Koren
Publication: PLoS One
Year published: 2014
Why this Article: This paper uses a novel approach for patient recruitment. Using an interrupted time series approach, this study assesses the effectiveness of advertising using a wide range of social media channels, from Facebook to Craigslist, to recruit patients to participate in clinical trials. The results showed a significant increase in recruitment using social media.
May 1
2014
Publication type: Article
Author(s): W Mutale, J Stringer, N Chintu et al.
Publication: PLoS One
Year published: 2014
Why this Article: Improving health systems in low income countries is extremely important. However, given the complexity of health systems, complex interventions are required with methods of evaluation that account for the complexity. This paper uses a balanced scorecard approach to evaluate multiple domains of the health system after implementation of two interventions targeted at different levels of the health system.