HIV/AIDS research advances through Pamela Shaw's work

Shaw's project to reduce the impact of data errors just received an NIH MERIT Award

On World AIDS Day, we would like to recognize the work of Dr. Pamela Shaw, PhD, MSc, a senior biostatistician. Shaw's outstanding research in HIV and AIDS has been recognized with a rare honor: a National Institutes of Health (NIH) MERIT Award.

Shaw, along with Dr. Bryan Shepherd of Vanderbilt University, has been awarded a five-year, $4.2 million NIH grant for statistical methods and design in HIV/AIDS research. This award will extend the scope of funding for an additional five years through MERIT (Methods of Extended Research). The award is based on researchers' "extraordinary competence and productivity," as well as their confidence in future research findings.

Xiao is delighted with the award. It ensures long-term, stable funding for the project, as renewal requires only a short report directly evaluated by the NIH, rather than a full, externally reviewed proposal. “This saves us at least six months, allowing us to continue our research!” she says.

You have expertise in aging, immunology, clinical trials, nutritional and physical activity epidemiology, and more. Why does your MERIT-award project focus on HIV/AIDS?

People with HIV can live a long time with anti-HIV therapy. Extended use of these medications sometimes creates risks, though — for example, for liver fibrosis. We also need to consider the medications’ impact on frailty, with people living longer.

But HIV/AIDS research is not easy. It occurs among hard-to-reach populations. Clinical trials are costly, and study designs are not always appropriate or feasible.

Many HIV/AIDS studies use existing data, often from readily available electronic health records (EHRs), which contain extensive observational data on people living with HIV. However, EHRs are not perfect research tools.

Electronic health record (EHR) data are not specifically used to support research but are collected by physicians as part of patient care. These data may not capture lifestyle factors or other information of research interest, such as physical activity. Furthermore, due to the sheer size of EHR data resources, research often relies on automated computer algorithms to retrieve data. This can miss important information from records and mislead researchers with irrelevant information. Without statistical adjustment, these data errors can bias research findings.

We advance EHR data research by developing study designs and analytical methods to identify and adjust for the potential impact of error-prone data. We strive to increase the precision and accuracy of research findings. More precisely answering research questions can guide clinical and decision-making, thereby improving the care of HIV-infected individuals by accurately predicting risk and treatment response.

We advance EHR data research by developing study designs and analytical methods to identify and adjust for the potential impact of error-prone data. We strive to increase the precision and accuracy of research findings. More precisely answering research questions can guide clinical and decision-making, thereby improving the care of HIV-infected individuals by accurately predicting risk and treatment response.

How does your work contribute to improved clinical research?

The gold standard for using electronic health records (EHRs) in research is to manually extract data through chart review. This is time-consuming and costly.

We are developing resources, such as free software, to help researchers identify which patient records are most relevant to the health outcomes and populations they are studying. We will also provide statistical methods for fine-tuning analyses that rely on EHR data—by incorporating gold-standard information from chart reviews of smaller numbers of records, or from studies that regularly collect information from patient samples—to achieve more accurate results.

Our goal is to provide researchers with a followable protocol for utilizing observational data, validated against gold-standard data, to enhance HIV/AIDS research and ensure optimal clinical decision-making. Our methods are also applicable to other studies. We recently used these methods to investigate how excessive gestational weight gain affects the risk of asthma and obesity in early childhood.

Since graduate school, I've been working with error-prone data and developing statistical methods to improve the accuracy of data-based studies. This HIV/AIDS research is the result of a long-standing collaboration with Bryan Shepherd of Vanderbilt University, a leading figure in HIV/AIDS biostatistics. We are using the International Epidemiology Database on AIDS (IeDEA), a National Institutes of Health (NIH)-supported database with decades of data covering over 2.2 million individuals in 44 countries.

Since graduate school, I've been working with error-prone data and developing statistical methods to improve the accuracy of data-based research. This HIV/AIDS study is the result of a long-standing collaboration with Brian Sheppard, a leading figure in HIV/AIDS biostatistics at Vanderbilt University. We used the International Epidemiology Database on AIDS (IeDEA), a National Institutes of Health (NIH)-supported database with decades of data covering over 2.2 million people in 44 countries.

Since graduate school, I've been working with error-prone data and developing statistical methods to improve the accuracy of data-based research. This HIV/AIDS study is the result of a long-standing collaboration with Brian Sheppard, a leading figure in HIV/AIDS biostatistics at Vanderbilt University. We used the International Epidemiology Database on AIDS (IeDEA), a National Institutes of Health (NIH)-supported database with decades of data covering over 2.2 million people in 44 countries.

ACT allows us to study how to best utilize EHR and other types of research data, the availability of which varies across participants. For example, only some participants may have undergone an MRI or detailed physical activity data. We can develop methods to adjust for varying amounts of research data so that the findings are representative of all ACT participants and Kaiser Permanente members.

When I joined KPWHRI, ACT's new funding had just begun.

One of ACT's main goals was to make research more representative of historically marginalized racial and ethnic groups, and I was able to help design methods to achieve this goal. ACT is already showing results: previously, 80% to 90% of new participants were white. Now, about 50% of new participants are from historically marginalized groups, so we're moving towards a more diverse group.

What do you enjoy doing when you're not at work?

Knitting sweaters and hats and things for the little ones I know. I'm an avid cycler, so that's another draw for the Pacific Northwest. The East Coast is freezing in winter and hot as heck in summer, and I don't mind the rain, so I'm happy to be back to Seattle for cycling and hiking.

I also love birding. I have about 10 birdfeeders. It's like a farm — every day I go out and fill all the birdfeeders and watch the birds that come visit throughout the day. Birding is a great way to stay connected to all the amazing nature around us in Seattle.