Wendy M. White, M.D., a specialist in Maternal and Fetal Medicine at Mayo Clinic’s campus in Minnesota and the director of perinatology services at Olmsted Medical Center, and Yvonne S. Butler Tobah, an obstetrician at Mayo Clinic in Minnesota, review an article they authored and published in the December 2018 issue of Mayo Clinic Proceedings. The paper presents an algorithm they developed that proved more effective in diagnosing hypertension in pregnant women, a condition that has longer term ramifications than previously thought.
Hi, I'm Doctor Yerevan Butler TOBA, an obstetric physician at Mayo Clinic in Rochester, Minnesota, and instructor at the Mayo Clinic College of Medicine. And I'm Wendy White, an assistant professor at the Mayo Clinic College of Medicine, supplemental faculty and maternal fetal medicine and the director of paleontology services at Homestead Medical Center. Now, today we're discussing our manuscript. Electronic algorithm is superior to hospital Discharge codes for Diagnosis of Hypertensive Disorders of Pregnancy in Historical Chords, which will be published in the December 2018 issue of Mayo Clinic Proceedings. This is a work of of a collaborative team led by Dr Vesna Gojkovic. So you might be wondering, why is this important and why this tool is relevant in today's clinical practice? Pre eclampsia and other related hypertensive disorders of pregnancy have long been recognized as a significant cause of morbidity and mortality for mothers and their babies, both globally and in the US. Increasingly, they are also recognized as an important risk factor for future adverse cardiovascular outcomes, such as early onset of hypertension, myocardial infarction and stroke. In fact, a history of hypertensive disorders of pregnancy is now incorporated into some risk stratification schemes for women Ongoing work. Describing the strength of the association between a history of hypertensive disorders of pregnancy and future cardiovascular outcomes will be important to allow us to understand and manage this risk. Now. Unfortunately, most epidemiologic studies in this area assigned exposure status based upon international classification of diseases or I C D codes that are assigned at discharge from the hospital. Other sources may include recall or registry data. All of these methods lack sensitivity and specificity. The gold standard for establishing exposure outcome diagnosis for a given epidemiology study would be to have a physician review the clinical data and arrive at a diagnosis of interest. But Dr White, you know, we both know that this isn't really feasible for large studies involving thousands of women. Uh, so there's a major need for an accurate, standardized electronic tool for assessing large epidemiologic data when evaluating long term outcomes of hypertensive disorders of pregnancy. Our study had two objectives. The first was to develop that electronic research algorithm to create the algorithm, a multidisciplinary group involving obstetricians, experts in nephrology and hypertension, epidemiologists and statisticians designed a computer algorithm which closely replicates physician judgment for the purposes of research. The algorithm can analyze clinical data that can be extracted from either paper records or the electronic medical record by trained abstract URZ examples of relevant clinical data. Our blood pressures, laboratory data, medications and presence or absence of symptoms, as well as the time course of this data and the context such as did it occur at a prenatal visit. An emergency department visit or on labor and delivery Analysis of all of this data then leads to the assignment of an exposure status in one of the following categories. Normal times of pregnancy, chronic hypertension, gestational hypertension, preeclampsia, eclampsia or one of these disorders superimposed upon chronic hypertension. These categories parallel those used clinically for the diagnosis and management of hypertensive disorders of pregnancy as laid out by the American College of Obstetricians and Gynecologists Task Force on Hypertension in pregnancy. Our algorithm was programmed by our Study Statistician in collaboration with our co investigators from the University of Belgrade in Serbia and is currently available online now. Next, we applied the algorithm to a large historical cohort derived from the Rochester Epidemiology Project, commonly known as rap. Now this is a unique and rich resource for epidemiology studies. The rep was created in 1966 and links all healthcare information from all medical providers for the entire population of Homestead County. Here in Rochester, Minnesota, This rich resource now includes approximately 6. 3 million person years of patients. Contact. Approximately 7544 women in the rep who delivered at more than 20 gestational weeks between 1976 in 1982 were included in our cohort for screening and analysis, and this court was then followed forward to present day to assess for cardiovascular diseases. For our second objective, we created a sub cohort within the larger historical cohort to test for the accuracy of our algorithm utilizing two obstetricians blinded to the algorithm based diagnosis, We independently review the sub to her medical records and then classify them as either in normal terms of pregnancy, gestational hypertension or preeclampsia, utilizing your professional opinion in the American College of Obstetrics and Gynecology Task Force criteria. Now, consensus between the two independent providers was amazing. 100% for our algorithm, similarly had 100% sensitivity and 100% specificity for pre eclampsia, and 100% sensitivity in 94% specificity for normal tons of pregnancies. Similarly, our electronic algorithm fared pretty well with a more contemporary cohort who delivered from 2012 to 2015. We encourage other researchers studying long term outcomes that are associated with hypertensive disorders of pregnancy to utilize our algorithm, which will be made publicly available on the Mayo Clinic Proceedings website. After validating our algorithm and their own data set, it can be applied to cohorts that number in the thousands. And as long as the appropriate clinical variables have been abstracted, it can be quite useful. Algorithm allows a standardized approach to assessment of hypertensive disorders of pregnancy. For especially large epidemiologic data, It's more sensitive than the commonly used by CD nine and building codes. It's relevant for historical as well as contemporary co hurts, and it also retains the large scale applicability of computer based methods. The clinical relevance of this approach is not only important for accuracy, but allows a more nuanced approach that involves accuracy of studying the long term clinical implications of hypertensive disorders of pregnancy. We hope this algorithm will strengthen the quality of research delineating the role that hypertensive disorders of pregnancy have on future risk for cardiovascular disease in women and highlight that a woman's reproductive history does not cease to be important when she is no longer in her childbearing years. In fact, this is a factor that primary care providers and specialists should make sure they include in the past medical history of all women for their lifetime. Thanks for joining in. We hope you found this presentation from the content of Mayo Clinic proceedings valuable. Our journal's mission is to promote the best interests of patients by advancing the knowledge and professionalism of the physician community. If you are interested in more information about us, our homepage is www dot mayo clinic proceedings dot org. There you will find access information for our social media content, such as additional videos on our YouTube channel or journal updates on Facebook. You can also follow us on Twitter. More information about health care at Mayo Clinic is available at www dot mayoclinic dot org. This video content is copyrighted by Mayo Foundation for Medical Education and Research
