Newsbytes

Doctor develops dashboard to address sickle cell disease

August 2022—Transfusion medicine specialists at Phoenix Children’s Hospital may be used to trending laboratory biomarkers for sickle cell disease patients in their heads, but a dashboard supporting red blood cell exchange procedures for such patients eases the burden on their cerebral cortices.

It’s common practice to mentally plot the direction a patient’s biomarkers have been trending relative to the target range, says Ajay Perumbeti, MD, who was medical director, lab informatics, clinical integration, and decision support, at Phoenix Children’s Hospital at CAP TODAY press time. (Dr. Perumbeti is now an informatics fellow at the University of Arizona Medical Center Phoenix and Banner Health System.)

“To be able to do visually what people are doing in their heads is helpful, especially given the fact that for every patient, there may be 10 decisions that you have to make,” Dr. Perumbeti says. “Any tool that will help facilitate making those decisions will allow you to spend more time on nuanced care.”

To this end, Dr. Perumbeti and a handful of colleagues at PCH developed an RBC exchange dashboard, which resides on the hospital’s intranet. Users can access the dashboard through a link on the intranet browser. The dashboard tracks about a dozen biomarkers and, thereby, allows users to view whether RBC exchange is reducing sickle cell levels and can help control other risks associated with the disease. Among the markers it tracks are hemoglobin S, or sickle hemoglobin, as well as such anemia markers as hematocrit, plasma hemoglobin, and hemoglobin F, and markers that can indicate iron loading, such as ferritin in blood serum.

Many of the measures tracked on the dashboard are interrelated and address various risks that affect sickle cell patients, Dr. Perumbeti says. For example, a graph at the top left of the dashboard measures hemoglobin S as a percentage of the patient’s total hemoglobin. The graph tracks how the patient’s hemoglobin S percentage fluctuates following an RBC exchange, and it includes a threshold line at 30 percent to mark the point at which stroke risk increases in this group. Below the hemoglobin S graph, a similar graph tracks hemoglobin levels in grams per deciliter. The latter includes a threshold mark at 9 g/dL to represent a more serious level of anemia that can be dangerous for sickle cell patients.

“There is a kind of balancing act that occurs,” Dr. Perumbeti says, referring to how biomarkers need to be monitored for different reasons. “There is still a little bit of human intuition in terms of looking at the trend lines,” he adds, but this takes out the preprocessing step of determining which direction biomarkers are trending.

Dr. Perumbeti views the dashboard’s ability to help highlight the uniqueness of each patient’s health situation as a key benefit. “Every trend line is specific to a patient, and the trends are different across patients, so it highlights that for this patient we really need to focus on improving the hemoglobin or dropping the sickle cell levels. With each of those trend lines there is an action step that you can take to bring the trend line to about the threshold mark.”

Dr. Perumbeti also believes the dashboard can serve as a training aid for rotating pathology fellows, residents, medical students, and others since the desirable biomarker thresholds are clearly marked. Data visualization also makes it easier for trainees to see the risk factors physicians are trying to control via RBC exchange. Dr. Perumbeti often showed the dashboard to his patients to give them a better understanding of where their sickle cell levels were in relation to the 30 percent threshold for stroke risk and the efforts needed to reduce those levels so they would be closer to the threshold or even below it.

The RBC dashboard went live at Phoenix Children’s Hospital in early 2021. A preliminary study was conducted on post­implementation data from patients who met the criteria for RBC exchange six months before and after implementing the dashboard. The study showed an 18 percent improvement in patients with hemoglobin levels greater than 8 g/dL and a 13 percent decrease in hemoglobin S levels (n=8), with no significant change in RBC exchange intervals in this population. Furthermore, after the dashboard was implemented, six sickle cell disease patients were flagged with ferritin levels greater than 1,000 ng/mL as a measure of refractory iron overload and brought to the attention of the apheresis and hematology teams.

Dr. Perumbeti notes the study’s relatively small sample size and that the study assessed a total of 10 clinicians and nurses as they used the dashboard over a cumulative total of 43 days during the six-month postimplementation period.

The dashboard was built using Microsoft’s Power BI data-visualization software, which PCH uses to graph data. However, it could just as easily be created using other popular data-visualization products, such as Tableau, Dr. Perumbeti says. To create PCH’s dashboard, a data engineer copied relevant data from the hospital’s Allscripts EHR into a research database. The Power BI software was then able to pull information from the database into charts and graphs in near real time.

It typically takes three people to create this type of dashboard, Dr. Perumbeti says. This includes a transfusion medicine specialist with domain expertise, the data engineer who copies the required data elements from the EHR into the research database, and a data scientist or data analyst with expertise in Power BI or other data-visualization software to design and set up the charts.

The most challenging step, Dr. Perumbeti says, is mapping the EHR data into a research or quality database, which can take days to months depending on the resources available in the hospital’s information technology or informatics department. At PCH, this step took about two to three months, and creating the chart visualizations took another month or two, he recalls.

Dr. Perumbeti hopes that increased focus on the lab biomarker trends displayed on the dashboard will help PCH improve sickle cell patient care and reduce the incidence of stroke in this patient group. A subgoal of controlling these patients’ biomarkers via the dashboard is to reduce routine hospital visits and eliminate sickle cell-related hospitalizations, Dr. Perumbeti says.

PCH’s RBC exchange dashboard is not the first dashboard that Dr. Perumbeti has helped build. Before joining PCH, he had helped develop a dashboard at Children’s Hospital of Los Angeles for tracking how patients respond to simple transfusion.

Many of the patients at that hospital who were getting transfused had other medical conditions, he explains. And in some cases, a patient may have needed a transfusion every three or four days, but the clinician ordering the transfusion might have been focusing on other pressing issues affecting the patient’s health. “The whole premise was that we wanted to identify when patients were getting transfused but not responding,” he adds.

Dr. Perumbeti hopes that, in the future, machine-learning algorithms can use data on the effects of transfusion and RBC exchange generated by such dashboards to better predict patient outcomes. “You need dashboards to develop intuition to appropriately label the outcomes that you are interested in,” he says. “Then when you can label outcomes that you are interested in, you can go back and use modeling to make predictions, and those predictions might contain information that is nonintuitive.”—Renee Caruthers

LigoLab adds remote ordering and result reporting to platform

LigoLab has released Web-based remote-ordering and result-reporting functionality for its LigoLab lab information system and revenue cycle-management laboratory operating platform. The new feature, LigoLab Web Connect, allows users to place orders, retrieve lab reports, and track the status of cases in real time from any location with an Internet connection.

LigoLab Web Connect provides such features as:

• a dashboard that displays the status of all orders.
• customizable and automated message notifications to lab clients.
• patient search by patient name or patient ID.
• shortcut buttons for frequently ordered tests.
• configurable ask-at-order-entry questions.
• a free-text field for clinical history.
• a searchable archive based on se­lected fields.

LigoLab, 800-544-6522

Myriad Genetics to offer tests via Epic EHR

Myriad Genetics and Epic have announced a partnership under which Epic will integrate Myriad’s full line of genetic tests into its electronic health record system.

“The integration creates a seamless, end-to-end workflow solution for health care providers to order Myriad tests and review results directly within their everyday Epic platform without additional steps or manual ordering processes,” according to a press release from Myriad. In addition, patients will be able to access their Myriad test results and other health information from their Epic EHR portal.

The undertaking is expected to be completed later this year.

Myriad Genetics, 800-469-7423

Gestalt Diagnostics releases AI requisition engine

Gestalt Diagnostics has introduced its artificial intelligence requisition engine, or AIRE.

The cloud-based system “is an intuitive and continual-learning AI algorithm that is highly scalable and accurately interprets requisition and other form types, including complicated ones,” according to a press release from Gestalt.

AIRE validates patient information against a database and flags data that need to be corrected. It can also be used for onboarding new physicians and clients.

Gestalt Diagnostics, 509-492-4912

Dr. Aller practices clinical informatics in Southern California. He can be reached at raller@usc.edu. Dennis Winsten is founder of Dennis Winsten & Associates, Healthcare Systems Consultants. He can be reached at dwinsten.az@gmail.com.