Inclusion of Patient Headshots in Electronic Health Records Decreases Order Errors

In a study of Brigham patients, rate of wrong patient order entry decreased by 35 percent

Each year, health care practitioners at Brigham and Women’s Hospital place over a million orders through the electronic health records (EHR) system. Even though studies indicate that practitioners place more than 99.9 percent of orders for the correct patients, researchers at the Brigham analyzed that remaining 0.1 percent to determine and address the root causes of wrong-patient order errors. In an effort to improve patient safety, the Brigham required headshots for participating patients to be displayed in their EHR as part of a quality improvement program in the Emergency Department. Analysis of the millions of orders placed for participating patients over a two-year span showed the rate of wrong patient order entry to be 35 percent lower for patients whose photos were included in their EHR. Results are published in JAMA Network Open.

“There’s one specific solution to mitigating wrong-patient errors that turned out to be really effective: displaying patient’s photos in their electronic chart. As a provider, these are patients that you know personally — you’ve cared for them and you’re going to quickly recognize that face,” said Hojjat Salmasian, MD, MPH, PhD, of the Department of Quality and Safety at the Brigham.

Salmasian had previously collaborated on a project in which pop-up alerts were used to reduce wrong-patient errors. Unlike interruptive pop-up alerts, including patient photos in EHRs enables uninterrupted navigation and utilizes the natural human affinity for facial recognition. Promising results from smaller-scale studies looking at the implementation of patient photos to decrease wrong patient order entry (WPOE) inspired Salmasian and his colleagues to pursue this larger-scale test at the Brigham.

The researchers focused on the Emergency Department, where providers often multitask and, consequently, have a higher rate of errors. In a retrospective cohort of patients admitted between July 2017 and June 2019, photos taken of willing patient participants and corresponding orders placed were analyzed for error. Of 2.5 million total orders placed across 71,851 unique patients, there was a decrease in errors of 35 percent. Salmasian emphasized the sheer volume of orders this 35 percent amasses to when considering the millions of orders placed per year at the Brigham, saying without this photo implementation an estimated 2 in every 1,000 orders may be placed incorrectly in the ED.

This improvement in error risk was slightly more detectable in white patients, a finding that illuminates implicit bias, treatment inequities, and the patient care impact of having a predominantly white patient population. The Brigham and Mass General Brigham have plans to include photos of all participating patients in their electronic health records. Despite the barrier COVID-19 has caused, with masks being required of all patients, hospitals locally and across the country plan to integrate this photo feature as soon as they are able. As requests to include a headshot in one’s electronic health record increase, patients will begin to realize their actions — even as small as uploading a headshot to a healthcare portal — can have a huge impact on their health outcomes.

“It’s important for all of us to realize that there are things that we can do as patients that directly impact the appropriateness and safety of care that we receive,” said Salmasian. “If more patients engage in the care they receive, our health care system improves in both safety and quality.”

Funding for this work was provided by the CRICO Risk Management Foundation and grant R01HS024713 from the AHRQ. Salmasian reported receiving grants from the Agency for Healthcare Research and Quality (AHRQ) during the conduct of the study.

Paper Cited: Salmasian, H et al. “Association of Display of Patient Photographs in the Electronic Health Record with Wrong-Patient Order Entry Errors” JAMA Network Open DOI: 10.1001/jamanetworkopen.2020.19652

Rapid Changes in Biomarker of Inflammation May Be a Key Predictor of COVID-19 Outcomes

Results provide insight into underlying mechanisms of inflammation that may help anticipate clinical trajectories of COVID-19 patients

Predicting the course of a COVID-19 patient’s disease after hospital admission is essential to improving treatment. Brigham and Women’s Hospital researchers analyzed patients’ levels of inflammation, known to be associated with severity of illness, by looking at C-reactive protein (CRP) trends in 100 COVID-19 patients admitted to the hospital. They found that a rapid rise in CRP levels during the first 48-to-72 hours of hospitalization was predictive of subsequent respiratory deterioration and intubation, while steadier CRP levels were observed in patients whose condition remained stable. Findings were published in Cell Reports Medicine.

“We realized that whereas a single CRP lab value from hospital admission wasn’t very practical as a predictor of who might get sicker, tracking the rate of change from Day 1 to Day 2 or 3 was a very powerful and very clinically predictive test,” said corresponding author Edy Yong Kim, MD, PhD, of the Division of Pulmonary and Critical Care Medicine at the Brigham. “Even though all of these patients looked clinically similar upon admission, as early as 24 hours after hospitalization, the immune systems of patients who would go on to the ICU multiple days later were already inflamed, as measured by these biomarkers.”

Inflammation is a broad term that describes the release of chemicals involved in immune responses. CRP tests integrate signals from a number of different proteins involved in inflammation, called cytokines, to provide physicians with a snapshot of a patient’s inflammatory activity within a matter of hours. Other tests, like cytokine assays, can provide more specific information about which proteins may be active in inflammatory pathways, but these tests can take one to two days to process, and COVID-19 patients’ conditions can worsen before the results are received. CRP tests can therefore serve as a practical addition to standard protocols for assessing the anticipated clinical trajectories of COVID-19 patients.

“Because of our findings, we changed our guidelines at the Brigham to mandate CRP tracking every day for the first three days of hospitalization so that we could try to identify vulnerable patients and keep a close eye on their inflammation,” said Kim, who stressed the importance of putting findings into practice as early as possible amid the current uptick of positive cases.

The results, from a study population of 100 Brigham patients, also provide insight into the underlying mechanisms at play in COVID-19 infections. In particular, an increase in a cytokine called IL-6 during the first 24-48 hours was correlated to CRP levels and the progression of the disease. Fifteen patients treated during this acute period with the drug tocilizumab, an IL-6 receptor, had rapid, sustained reductions in their CRP levels. In larger, randomized trials, tocilizumab was not shown to provide benefits to COVID-19 patients, but Kim states that this could be because the drug was not administered early enough to the subset of patients who stand to benefit most. Alternatively, while CRP is associated with IL-6, CRP can reflect other inflammatory pathways besides IL-6, so targeting other inflammatory cytokines or pathways besides IL-6 could be considered.

“Even if you gave immunomodulatory drugs, which reduce rising inflammation, as early as Day 3 — which is pretty early for a clinical trial — that may already be too late,” Kim said. “But here we have some evidence that a rise in inflammation directly drives respiratory failure, which implies that the immunomodulatory drugs might be able to prevent respiratory failure if given very, very early — as early as hospital Day 1 and 2.”

Ultimately, Kim hopes that the findings will help front-line workers better understand the volatility of COVID-19 patients’ conditions. “Doctors’ and nurses’ clinical instincts about COVID-19 are not fully developed because the disease is still so new,” Kim said. “But when we showed these results to frontline doctors and nurses at the Brigham, they felt like it matched what they intuitively saw in the spring. It’s always nice to hear that what you do in the lab reflects what goes on in the real world, too.”

This study was supported by the American Heart Association Award 2014D007100 and NIAMS T32 AR007530-35. Brigham and Women’s Hospital has a portfolio of patents characterizing an individual’s risk of experiencing a future Major Adverse Cardiovascular Event (MACE) by using CRP and IL-6 as crucial biomarkers. This invention also involved methods for evaluating the likelihood that an individual will benefit from treatment with an agent for reducing the risk of future cardiovascular disorders.

Paper cited: Mueller, AA et al. “Inflammatory biomarker trends predict respiratory decline in COVID-19 patients” Cell Reports Medicine DOI:10.1016/j.xcrm.2020.100144

Brigham Researchers Make Strides in Detecting Preeclampsia Risk

Investigators identify possible subtypes of preeclampsia, advancing search for therapeutics

Preeclampsia, a pregnancy complication characterized by high blood pressure, occurs in about 7 percent of pregnancies and is a leading cause of maternal mortality and premature birth. The condition can appear in a variety of forms both early and late in pregnancy, making it difficult to test and therefore hard to prevent or treat. To improve early detection of preeclampsia risks and better classify potential subtypes of the disease, a team led by researchers from Brigham and Women’s Hospital analyzed an array of maternal and placental proteins, identifying several biomarkers that, from a blood sample drawn at 12 weeks gestation, can help predict the risk of preeclampsia and, equally importantly, can assess who may develop more severe forms of the condition. Findings are published in Scientific Reports.

“For many, many years, preeclampsia was assumed to be one disease,” said corresponding author Thomas McElrath, MD, PhD, of the Division of Maternal-Fetal Medicine in the Department of Obstetrics and Gynecology at the Brigham. “This study is starting to show that there are potentially multiple different forms of preeclampsia that come together in one common clinical presentation, and once we understand what the subgroups are, we may be more effective at testing or treating them.”

The researchers were the first to study the relationship between preeclampsia and proteins associated with circulating microparticles (CMPs), which are virus-sized cellular secretions that contain small proteins and RNA, among other cellular products. In humans and ancient life forms alike, the transmission of CMPs facilitates communication between cells by modifying the receiving cells’ functions. The researchers found that at 10-12 weeks of pregnancy, women who would later develop preeclampsia already had different CMP-associated proteins than women who ultimately had healthy pregnancies. They identified four proteins, in particular, that seemed to be indicative of preeclampsia risk.

“If we can eavesdrop on the conversations that cells are having with each other about what they’re trying to modify in the patient, then we might have the ability to predict what the future holds for that patient and what her risks are,” McElrath said.

Furthermore, when the researchers examined how CMP proteins differed across the patients who got preeclampsia, they found that more severe forms, with higher blood pressure and worse lab testing, were associated with proteins indicative of the complement system — a branch of the immune system — dysfunction. They identified another possible subgroup of preeclampsia linked to platelet dysfunction.

“To develop therapies, we need better information about what pathophysiologic pathways to target,” McElrath said. “The idea that we might be able to identify women most at risk early on in pregnancy means that we might be able to develop more efficient testing and therapies.”

This study utilized data from the Brigham’s LIFECODES Cohort, which includes records of approximately 6,000 pregnancies and is one of the nation’s largest pregnancy cohorts and specimen banks. The team analyzed CMP-associated proteins in 23 women diagnosed with severe forms of preeclampsia at or before 34 weeks of gestation, most of whom delivered before 35 weeks, and in 50 women who delivered after at least 37 weeks of gestation without evidence of any hypertensive diseases, including preeclampsia.

By comparing the most severe cases of preeclampsia to the healthiest pregnancies, the researchers were able to complete a preliminary, proof-of-principle investigation, but acknowledged that their small, specific sample makes it harder to generalize the patterns they’ve observed in CMP-associated proteins to the entire population.

Going forward, the researchers hope to validate their results by analyzing thousands of pregnancies, aggregating data from the LIFECODES Cohort and other birth cohorts nationwide. They hope to study proteins in women with later-stage preeclampsia, and in doing so, identify other subgroups of preeclampsia.

“We really don’t have a prognostic test for preeclampsia,” McElrath said. “Ideally, this will help develop a test that could be used at the woman’s first prenatal visit to help us decide if she’s at higher or lower risk of preeclampsia and put her in a plan of care in accordance with that risk.”

This work was supported by donation from the Sidman Family as well as support for specimen processing by NxPrenatal of Louisville KY.

Paper cited: McElrath, TF et al. “Late First Trimester Circulating Microparticle Proteins Predict the Risk of Preeclampsia <35 Weeks and Suggest Phenotypic Differences Among Affected Cases” Scientific Reports DOI:


Combining Population Health Management and Online Program May Help Patients Lose Weight

In a study of Brigham primary care patients, nearly one-third of participants receiving the combined intervention lost 5 percent or more of their weight

Over 70 percent of U.S. adults have overweight or obesity. While primary care physicians (PCPs) are often the most immediate source of counseling for those trying to lose weight, most PCPs do not have the time, training, or support to guide their patients through a comprehensive weight loss program. Additionally, studies of online weight loss programs have found mixed levels of success. Population health management programs, which are delivered by non-clinical staff members, have emerged as supplemental support systems that offer outreach for preventive and chronic condition care needs, outside of primary care visits. Researchers from Brigham and Women’s Hospital paired an online weight loss program with a phone- and email-based population health management program, a two-pronged strategy previously unexplored, and determined that patients in the combined program had greater weight loss over 12 months than patients in the other two groups. Findings are published in the JAMA.

“Population health managers are already doing outreach to people who would benefit from weight loss, such as patients with hypertension or type 2 diabetes,” said corresponding author Heather Baer, ScD, an associate epidemiologist in the Division of General Internal Medicine and Primary Care at the Brigham. “The idea was to leverage population health management, which is a strategy that’s been implemented at the Brigham and in other institutions, to see if it could be integrated with an online weight loss program and be sustainable over time.”

The study followed 840 patients in the Brigham’s Primary Care Center of Excellence, a network of 15 practices, between 2016 and 2019. Participants, all of whom either had overweight or obesity and hypertension or type 2 diabetes, were divided into three study groups. The “usual care” group received general information about weight management in the mail; the “online only” group participated in an online weight-loss program, which included meal plans, activity trackers, and progress reporting features; and the “combined intervention” group participated in the online program while receiving additional outreach and support from non-clinical staff members monitoring their progress.

Over the course of 12 months, the usual care, online only, and combined intervention groups lost an average of 2.7 pounds, 4.1 pounds, and 6.9 pounds, respectively. The researchers found that the combined intervention produced a small but statistically significant increase in weight loss compared to the other two methods. Moreover, participants in the combined intervention group decreased their body weight by an average of 3 percent, a statistically significant difference from the 1.4 and 1.9 percent decreases found in the usual care and online only groups.

The researchers acknowledge that while the amount of weight loss is fairly low, clinically significant health benefits are associated with a 3-5 percent decrease in weight among individuals with overweight or obesity. Approximately one third of participants in the combined intervention group lost 5 percent or more of their weight, compared to only 14.9 and 20.8 percent of participants in the usual care and online only groups, respectively. Moreover, the weight loss in the combined intervention group was also found to be maintained at 18 months, in spite of the fact that active interventions ceased after 12 months.

“Our findings suggest that the combined intervention facilitates potentially sustainable weight loss in the long term,” Baer said. She notes that the generalizability of the team’s findings is limited, given that the the majority of participants were white, well-educated, and English-speaking. Nevertheless, the combined intervention approach may offer an effective form of weight management that is more easily available to some patients who may face barriers in accessing traditional in-person care.

The results are especially notable as health care increasingly shifts to virtual forms amidst the COVID-19 pandemic. “We knew that online weight loss programs can be more convenient, accessible, and cheaper for patients, but now both physicians and patients have become more comfortable with the idea of care being delivered virtually,” Baer said. “Virtual care is going to stay with us post-pandemic, and this study provides more evidence that even pre-pandemic, virtual care for some conditions is working.”

This study was supported by a Patient-Centered Outcomes Research Institute (PCORI) Award (14-092-1758).

Paper cited: Baer, HJ et al. “Effect of an Online Weight Management Program Integrated with Population Health Management On Weight Change: A Randomized Clinical Trial” Journal of the American Medical Association DOI: 10.1001/jama.2020.18977


Study Dives into Genetic Risk of Alzheimer’s and Dementia for Diverse Latinx Groups

Investigators find evidence that the genetic basis for cognitive decline may vary by genetic ancestry 

Alzheimer’s disease and related dementias (ADRD) are degenerative brain diseases that affect over 5.8 million people in the U.S. Rates of ADRD are higher among Black and Latinx individuals compared to non-Latinx white individuals. APOE-ε4 is the strongest known genetic risk factor for ADRD, however this finding is largely based on studies of individuals of European descent. To better understand the association of the APOE gene with cognitive decline in Latinx populations, researchers from Brigham and Women’s Hospital and collaborators analyzed metrics of cognitive decline in six diverse Latinx populations: those of Cuban, Central American, Dominican, Mexican, Puerto Rican, and South American backgrounds. They found that the APOE-ε4 genetic variant was associated with risk of cognitive decline in Latinx populations, with the strongest effect among those of Cuban backgrounds. Findings are published in Alzheimer’s and Dementia.

“Latinx populations suffer more from Alzheimer’s and dementia than other populations,” said Einat Granot-Hershkovitz, PhD, first author on the paper and a research fellow in the Division of Sleep and Circadian Disorders at the Brigham. “Historically, Latinx populations have been underrepresented in research, especially genetic research. But our findings highlight how important it is to look beyond European ancestry and European genetic risk factors to understand in what ways genetics may or may not contribute to their risk.”

The six Latinx subgroups differ based on what proportion of their genes trace back to African, European, and Amerindian (indigenous American) ancestries. Cuban individuals, for example, have the lowest proportion of Amerindian ancestry among the six groups studied and a relatively higher proportion of European ancestry. When the researchers looked at whether the effect of APOE-ε4 is modified by genetic ancestry proportions, they observed that increased proportion of genetic Amerindian ancestry protects from the risk conferred by APOE-ε4 on cognitive decline.

“We think proportion of ancestry is a crude measure of genetics, but there are probably specific genetic factors that may be specific to Amerindian ancestry that drive this difference in the effect of APOE,” said Tamar Sofer, PhD, director of the biostatistics core in the Division of Sleep and Circadian Disorders at the Brigham and principal investigator of the paper. She acknowledged that non-genetic factors — like nutrition, sleep, physical activity, or exposure to toxins — could also play a role.

The researchers based their findings on 4,183 Latinx individuals participating in the Hispanic Community Health Study/Study of Latinos and the ancillary Study of Latinos-Investigation of Neurocognitive Aging. On average, cognitive tests were administered to the participants seven years apart. Since many of the participants are still too young to have developed ADRD, the researchers looked at associated conditions like significant cognitive decline and mild cognitive impairment, which often precede more serious states of dementia.

“Ideally there will be another study so we can see what happens long-term as the population becomes older,” Sofer said. “We can verify whether the pattern we see is related to Alzheimer’s or a different type of dementia or cognitive decline. We still want to verify the long-term patterns before we communicate risks to people in clinic.”

Still, the study is important as a large-scale investigation into an under-researched topic.

“Latinx populations should know that they are represented in this area of genetic research now,” Granot-Hershkovitz said. “It’s not easy to participate in a research study, but there are revelatory findings from this amazing dataset, and hopefully there will be more in the future. This brings us a step further in addressing Latinx health disparities.”

This work was supported by the National Institute on Aging (R01AG048642, RF1AG054548, RF1AG061022, and R21AG056952). Hector González, PhD, of the University of California, San Diego, receives additional support from P30AG062429 and P30AG059299. The Hispanic Community Health Study/Study of Latinos is supported by contracts from numerous organizations and institutions listed in the paper.

Paper cited: Granot-Hershkovitz, E et al. “APOE alleles association with cognitive function differs across Hispanic/Latino groups and genetic ancestry in the Study of Latinos-Investigation of Neurocognitive Aging (HCHS/SOL)” Alzheimer’s and Dementia DOI: 10.1002/alz.12205