September 23, 2015|Categories: Analytics and Research
The NIH (National Institutes of Health) recently released initial results from its SPRINT (Systolic Blood Pressure Intervention Trial) study, suggesting more aggressive treatment for hypertension may reduce risk for cardiovascular events by 30 percent and the risk of death by 25 percent.
The SPRINT group recruited over 9,000 non-diabetic subjects aged 50 or older with systolic blood pressure (SBP) over 130. Individuals were then randomly assigned to either a standard treatment goal of lowering SBP below 140 or a more intensive treatment goal of lowering SBP below 120. The findings were sufficiently compelling that results were released more than a year in advance of the study’s planned conclusion.
While the SPRINT researchers have only released very high level findings, the general sentiment appears to be supportive of releasing this information early as the New York Times Op-Ed by Eric Topol, MD, so rightly notes. However, the lack of detail has generated sharp questions on several issues, including whether a reduction of cardiovascular risk by 30 percent actually translates into substantial numbers of individuals affected, potential side effects of increasing the number of medications, and general caution regarding findings released before peer reviewed publication.
Even before the final SPRINT study results are released, however, some physicians will want to understand how their overall patient panels may be affected. Others will want to contact individual patients who may benefit from tighter blood pressure management.
To shed light on this issue, we queried data from athenaNet1, athenahealth’s cloud-based network of 67,000 providers and 69 million patients. athenaResearch identified 7.7 million patients 50 or older, who visited a health care provider using athenaClinicals, our EHR service, between July 2013 and June 2015. Among these individuals, approximately 9 percent of patients with medical profiles similar to the SPRINT trial criteria had blood pressure levels that might benefit from more intensive management. Slightly less than half of these patients had SBP readings between 120 and 139 (see figure below).
The vast majority of patients with hypertension were excluded because they met exclusion criteria, including stroke, diabetes, congestive heart failure, proteinuria, and chronic kidney disease.
ImplicationsAlthough matching trial criteria to EHR data is challenging, the ability to identify potentially at risk patients through efficient data technologies is an important capability in medicine. Physicians with ready access to this information can then evaluate a patient’s specific medical history, in combination with SPRINT’s study results, to determine the best course of treatment. Although SPRINT’s complete findings have not yet been released, we believe that some of our physicians will want to identify and consult with patients who fall within study criteria. In the coming weeks, we will be modifying our electronic health record, athenaClinicals, to allow physicians to activate point of care clinical decision support functionality that will alert them when they are caring for a patient who might benefit from more aggressive hypertension management, as per the SPRINT Trial results. We will also modify our Quality Management EngineSM to allow physicians to generate a real-time report on the number and percent of patients who fall outside of defined clinical parameters. The physician could then elect to activate an outreach campaign to contact all appropriate patients via automated phone call, email, or text to suggest the patients schedule an office visit for a consultation. This new functionality, as so much of our work, will be designed to allow physicians to provide effective care to at-risk patients with minimal disruption to their workflow. We invite readers (including medical practitioners) to provide suggestions on how our data could be utilized more effectively in light of the SPRINT results, either in comments through our athenahealth Cloudview blog, or directly to email@example.com, or @JoshGray_HIT and @IyueSung on Twitter. Iyue Sung, Sr. Data Engineer Manager, and Stewart Richardson, Sr. Data Engineer Associate, both of athenaResearch, developed analytics for this study.
1. Methodology: Individuals who may satisfy SPRINT study criteria (We identified patients with similar diagnosis and lab values, based on Problem List diagnoses and lab values stored in athenaClinicals, our EHR. We did not included the criteria for Framingham Risk Score for 10-year CVD risk ≥ 15%) are at least 50 years old and have SPB > 130. Furthermore, subjects must have one of the following health risks: cardiovascular disease, chronic kidney disease with eGFR between 20-59 mL/min/1.73m2), Framingham Risk Score for 10-year cardiovascular disease ≥ 15%, or at least 75 years old. Individuals are excluded from the study if they have the following history: stroke, diabetes, congestive heart failure, proteinuria, chronic kidney disease with eGFR < 20 mL/min/1.73m2, and having factors that may make it difficult for a subject to adhere to treatments (e.g. substance abuse, history of poor medication compliance, etc.). The study protocol also specifies precise methods and periods of measurement (e.g. exclude subjects with one minute standing SBP < 110).
The data used to develop our estimate are based on diagnosis history and laboratory values stored in athenaNet®, the database for athenahealth’s EHR. Therefore, caveats are in order with respect to our estimate. Data for patients in athenaNet® may have missing values; for example, that a patient may have diabetes although diabetes has not yet been documented in the patient’s record. Furthermore, due to their complexity, we do did attempt to include the criteria for Framingham Risk Score ≥ 15% and treatment adherence. The purpose of our exercise was to give our physician clients a sense of how many of their patients might benefit from tighter blood pressure management and to lay the groundwork to develop tools that will allow providers to reach out to individual patients. ↩