Department of International Health, Bloomberg School of Public Health
The Johns Hopkins University
Epidemiology, epidemiologic methods, biostatistics, observational study, cohort study, competing risks, survival analysis, longitudinal analyses, HIV/AIDS
There is often a gap between the research questions and the analytical methods being used in epidemiologic research. My training in biostatistics and epidemiology allows me to bridge this gap. Therefore a primary research interest is in helping to foster the transition of statistical methodologies into the field of epidemiology especially as applied to the setting of HIV research. For example, competing risk methods have been in the statistical literature for decades, but it has only been recently that these methods are being widely applied in epidemiologic research. However, just because there is a sophisticated analytical tool that is available (e.g. causal inference methods) does not mean that this is necessarily the right tool for the research question. Thus a guiding principle for me is to use the simplest but most appropriate tool to answer the research question in a suitable manner. Nevertheless, my methodological interests are in the areas of competing risks, causal inference including analytical methods for controlling for confounding (e.g. propensity scores, g-methods, instrumental variables), selection bias, and cohort study design.
My interests has have been in applying and developing new novel methods to public health and clinical research. I have a long standing interest in time-to-event analyses including application and development of competing risk methodology (Lau, JAIDS 2007; Lau, Statistics in Medicine 2008; Lau, American Journal of Epidemiology 2009; Brune, Journal of the National Cancer Institute 2010; Lau, Statistics in Medicine 2011; Ishwaran, Biostatistics 2014; Cole, American Journal of Epidemiology 2015; Lesko, Epidemiology, In Press). Additionally, I have expertise in biomarker especially analyses of longitudinal biomarkers (Gange, AIDS Research and Human Retroviruses, 2001; Lau, AIDS, 2003; Lau, JAIDS, 2005; Lau, Archives Internal Medicine, 2006; Lucas, Journal of Infectious Diseases, 2008; Cole, American Journal of Epidemiology, 2010; Mugavero, Clinical Infectious Diseases, 2011; Kowalski, JAIDS 2012; Kalayjian, Journal of Acquired Immunodeficiency Syndromes, 2012). I am also experienced in causal inference methods including the so-called g-methods which includes marginal structural models, parametric g-formula (g-computation), and g-estimation (Lau, Epidemiology 2009; Mugavero, Clinical Infectious Diseases 2011; Kalayjian, AIDS 2012; Buchanan, AIDS Res Hum Retroviruses 2014).
The majority of the analytical methods presented above has been in the field of HIV research. I currently am involved in several consortium projects including the CFAR Network of Integrated Clinical Systems (CNICS) which is comprised of multiple existing clinical cohort studies. My current role in CNICS is as a co-investigator as well as a co-chair of the Epidemiology and Biostatistics core and voting member of the Research and Coordinating Committee. Furthermore, I am involved with several HIV collaborative cohort studies including the North American-AIDS Cohort Collaboration on Research and Design which is a part of NIAIDs International Epidemiological Databases to Evaluate AIDS (IeDEA) project. I am the PI of the data analysis and management center for Investigating HIV/AIDS Lung Disease (INHALD) consortium and Alcohol Research Consortium in HIV: Epidemiological Research Arm (ARCH-ERA). These projects have spanned the breadth of HIV research including analysis of longitudinal biomarker measurements over time as they relate to HIV disease progression or development of co-morbidities (e.g., Lau, Archives of Internal Medicine 2006; Kalayjian, AIDS 2012; Kowalski, JAIDS 2012), cancer (e.g., Silverberg, Annals of Internal Medicine 2015; Silverberg, CID 2012) and other HIV associated non-AIDS co-morbidities (e.g., Lesko JAIDS 2016; Sterling, JID 2011), HIV care continuum (e.g., Lesko, AIDS 2016, Rebeiro, AJE, 2015), and alcohol use (Chander, JAIDS 2006; Bilal, AIDS Patient Care and STDs 2016; Monroe, JAIDS 2016).
Mentoring, study design development & consulting, analytical consultation
1. Competing Risk Methodology in Clinical and Public Health Research Competing risk settings commonly occur in clinical and epidemiologic research. This situation is in which there are two or more events and occurrence of one event precludes the occurrence of the others. The canonical example is cause-specific mortality, whereby a death due to primary cause of interest is precluded by death due to other causes. Competing risk methodology has been increasing in the scientific literature. It is obvious that use and understanding of competing risk methods will be important for individuals conducting research among aging populations. My contributions in this area includes application of these methods to clinical and epidemiological research, development of new statistical methodology, and a didactic and methodology paper to promulgate competing risk methods in epidemiology. My work in this area (applied and methodological work) has resulted in increased recognition of competing risk as an issue in research.
2. Epidemiologic Methods Epidemiology has been termed a “basic science of Public Health” as the methods (approach to scientific question, study design, control of confounders and other threats to validity) are critical to understanding the strengths and weaknesses of addressing important scientific questions at a population level. Nevertheless, as technology and statistical methodology makes advances, methods for epidemiology must also evolve. My contributions in this area include the competing risk methods above as well as adapting new assays into ongoing cohort studies, and investigations into comparison of clinical and the more traditional interval cohort study designs.
3. Investigating Longitudinal Biomarkers Biomarkers have become the most important staging tool for HIV infected individuals. Yet investigation of biomarkers in disease is not necessarily straightforward with many complex issues in the use and description of biomarkers such as dealing with repeated longitudinal measurements in cohort studies, measurement error, and changing technology in assays. My work on biomarkers include describing the longitudinal patterns and use as predictors for disease progression.
4. Investigating Co-morbidities in HIV Infected Populations HIV infection has become a long term chronic disease as treated HIV infected individuals now have life expectancy approaching HIV uninfected individuals. Yet HIV infected individuals have increased co-morbidities that threaten the gains since introduction of effective antiretroviral treatment. This is specifically of concern with an aging HIV population. Specifically non-AIDS related cancers, renal kidney disease, and others are resulting in multi-morbidity in this aging HIV population. My work in this area has included examining the risk of anal cancer among HIV infected individuals, examining the risk factors and incidence for chronic renal disease, and investigating multi-morbidity.
5. Investigating the effects of alcohol use among HIV infected individuals Hazardous alcohol consumption is prevalent among HIV infected individuals. Yet the implications of alcohol consumption on HIV outcomes and co-morbidities is not fully realized. My work in this area has included identification of heavy alcohol use as a risk factor for non-adherence to HIV antiretrovirals, as impacting timing of antiretroviral initiation and survival, CD4 response to treatment and trials of intervention studies to reduce drinking behaviors.