Faculty Bibliography

Objective: To develop a dataset based on open data sources reflective of community-level social determinants of health (SDH). Materials and Methods: We created FACETS (Factors Affecting Communities and Enabling Targeted Services), an architecture that incorporates open data related to SDH into a single dataset mapped at the census-tract level for New York City. Results: FACETS (https://github.com/mcantor2/FACETS) can be easily used to map individual addresses to their census-tract-level SDH. This dataset facilitates analysis across different determinants that are often not easily accessible. Discussion: Wider access to open data from government agencies at the local, state, and national level would facilitate the aggregation and analysis of community-level determinants. Timeliness of updates to federal non-census data sources may limit their usefulness. Conclusion: FACETS is an important first step in standardizing and compiling SDH-related data in an open architecture that can give context to a patient's condition and enable better decision-making when developing a plan of care.

As population health becomes more of a focus of health care, providers are realizing that data outside of traditional clinical findings can provide a broader perspective on potential drivers of a patient's health status and can identify approaches to improving the effectiveness of care. However, many challenges remain before data related to the social determinants of health, such as environmental conditions and education levels, are as readily accessible and actionable as medical data are. Key challenges are a lack of consensus on standards for capturing or representing social determinants of health in electronic health records and insufficient evidence that once information on them has been collected, social determinants can be effectively addressed through referrals or other action tools. To address these challenges and effectively use social determinants in health care settings, we recommend creating national standards for representing data related to social determinants of health in electronic health records, incentivizing the collection of the data through financial or quality measures, and expanding the body of research that measures the impact of acting on the information collected.

BACKGROUND: Diabetic medications are approved by the Food and Drug Administration (FDA) based on efficacy in controlled clinical settings. Real world effectiveness may often vary from the results of clinical trials. We attempted to quantify the "efficacy versus effectiveness" gap at our institution. METHODS: We performed a retrospective analysis of all patients prescribed oral diabetic medications at New York University Langone Medical Center, a large urban tertiary care center, using data extracted from our electronic health record from October 2009 to September 2014. We limited the dataset to include newly diagnosed diabetic patients who were started on monotherapy or a combination pill. We excluded patients who were prescribed insulin or multiple diabetic agents. We used two-tailed Welch's t-tests to compare the initial Hemoglobin A1c (HbA1c), final HbA1c, and total HbA1c reduction to clinical trial data used for FDA approval, with statistical significance defined as p<0.05. RESULTS: Among the 566 of the 2452 patients who met our inclusion criteria, the mean age was 62.9 +/- 12.2 years, the majority of whom were male (51.1%), white (61.1%), non-Hispanic (80.4%), and prescribed metformin monotherapy (70.3%). The mean treatment time for each drug class ranged from 32-39 weeks. The mean initial HbA1c for all groups was 8.95%, the mean final HbA1c was 7.44%, and the mean HbA1c reduction was 0.51%. For almost all drug classes, the initial HbA1c was lower than the initial HbA1c from efficacy trials. For most drug classes, effectiveness was about half of the expected value from efficacy trials. Most patients achieved their goal HbA1c levels [Table]. CONCLUSIONS: In our study population, real world effectiveness did not reflect measures of efficacy from controlled clinical settings. The results highlight the need for clinicians to realize that a medication may not perform as well as it does in a controlled setting and they may need to increase the intensity of treatment more quickly than expected. Also, replicating the results of clinical trials in an actual clinical setting is challenging, both from the perspective of data availability and clinical practice patterns. (Table Presented)

Billions of dollars spent, millions of subject-hours of clinical trial experience and an abundance of archived study-level data, yet why are historic data underutilized? We propose that historic data can be aggregated to provide safety, background incidence rate and context to improve the evaluation of new medicinal products. Here, we describe the development and application of the eControls database, which is derived from the control arms of studies of licensed products, and discuss the challenges and potential solutions to the proper application of historic data to help interpret product safety.

Previous research on standardization of eligibility criteria and its feasibility has traditionally been conducted on clinical trial protocols from ClinicalTrials.gov (CT). The portability and use of such standardization for full-text industry-standard protocols has not been studied in-depth. Towards this end, in this study we first compare the representation characteristics and textual complexity of a set of Pfizer's internal full-text protocols to their corresponding entries in CT. Next, we identify clusters of similar criteria sentences from both full-text and CT protocols and outline methods for standardized representation of eligibility criteria. We also study the distribution of eligibility criteria in full-text and CT protocols with respect to pre-defined semantic classes used for eligibility criteria classification. We find that in comparison to full-text protocols, CT protocols are not only more condensed but also convey less information. We also find no correlation between the variations in word-counts of the ClinicalTrials.gov and full-text protocols. While we identify 65 and 103 clusters of inclusion and exclusion criteria from full text protocols, our methods found only 36 and 63 corresponding clusters from CT protocols. For both the full-text and CT protocols we are able to identify 'templates' for standardized representations with full-text standardization being more challenging of the two. In our exploration of the semantic class distributions we find that the majority of the inclusion criteria from both full-text and CT protocols belong to the semantic class "Diagnostic and Lab Results" while "Disease, Sign or Symptom" forms the majority for exclusion criteria. Overall, we show that developing a template set of eligibility criteria for clinical trials, specifically in their full-text form, is feasible and could lead to more efficient clinical trial protocol design.

BACKGROUND: Ethical concerns about randomly assigning patients to suboptimal or placebo arms and the paucity of willing participants for randomization into control and experimental groups have renewed focus on the use of historical controls in clinical trials. Although databases of historical controls have been advocated, no published reports have described the technical and informatics issues involved in their creation. OBJECTIVE: To create a historical controls database by leveraging internal clinical trial data at Pfizer, focusing on patients who received only placebo in randomized controlled trials. METHODS: We transformed disparate clinical data sources by indexing, developing, and integrating clinical data within internal databases and archives. We focused primarily on trials mapped into a consistent standard and trials in the pain therapeutic area as a pilot. RESULTS: Of the more than 20,000 internal Pfizer clinical trials, 2404 completed placebo controlled studies with a parallel design were identified. Due to challenges with informed consent and data standards used in older clinical trials, studies completed before 2000 were excluded, yielding 1134 studies from which placebo subjects and associated clinical data were extracted. CONCLUSIONS: It is technically feasible to pool portions of placebo populations through a stratification and segmentation approach for a historical placebo group database. A sufficiently large placebo controls database would enable previous distribution calculations on representative populations to supplement, not eliminate, the placebo arm of future clinical trials. Creation of an industry-wide placebo controls database, utilizing a universal standard, beyond the borders of Pfizer would add significant efficiencies to the clinical trial and drug development process.

Translational informatics (TI) is extremely important for the pharmaceutical industry, especially as the bar for regulatory approval of new medications is set higher and higher. This paper will explore three specific areas in the drug development lifecycle, from tools developed by precompetitive consortia to standardized clinical data collection to the effective delivery of medications using clinical decision support, in which TI has a major role to play. Advancing TI will require investment in new tools and algorithms, as well as ensuring that translational issues are addressed early in the design process of informatics projects, and also given higher weight in funding or publication decisions. Ultimately, the source of translational tools and differences between academia and industry are secondary, as long as they move towards the shared goal of improving health.

To achieve the goal of personalized medicine, advances in healthcare information technology (HIT) must be accompanied by similar advances in data analysis and translation into clinical care. For healthcare providers and their patients, advances in genomics and other related fields must be captured by sophisticated decision support systems that integrate information from multiple sources of data and facilitate their use in the clinic. In addition to addressing privacy concerns, regulators will also need to work toward establishing evidentiary standards for the use of data from electronic clinical sources in areas such as comparative effectiveness and safety, while addressing the complexity behind the genetic basis of the exceptions to their general findings. Overall, the increased use of HIT will allow for a more integrated, efficient system for collecting and analyzing clinical data, providing a fruitful environment for hypothesis testing and generation, as well as leading to more personalized and effective healthcare.