Phoenix Children's makes big gains with homegrown data warehouse and apps

Photo: Phoenix Children’s

Hospitals and health systems have worked hard over the past decade-plus to roll out and optimize electronic health records. But EHRs by themselves have not automatically resulted in widespread improvements in clinical care and patient outcomes.


“During the early phase of EHR implementation, when a large majority of providers first started converting from paper to electronic, we were promised meaningful benefits that simply did not pan out,” said Dr. Vinay Vaidya, chief medical information officer at Phoenix Children’s. “Instead of providing efficiencies for physicians, EHRs created additional work by adding to the burden of documentation, significantly contributing to provider burnout across the industry.

“Trying to develop ways to fix and perfect the EHR process quickly became the focus, and while that was the right approach at the time, it became the undue focus of healthcare IT interventions, and we lost track of why EHRs were developed in the first place,” he added.

EHRs were meant to be a means to an end, a tool to capture and generate electronic data that could be harnessed for coordinated patient care. But in many organizations, clinicians and IT became so focused and bogged down in what seemed like a cumbersome process, they lost sight of the benefits EHRs could provide.

“About six years ago, Phoenix Children’s decided to take a balanced approach to EHR utilization and implementation,” Vaidya recalled. “While we continued the ongoing optimization and maintenance of the EHR, we did not allow it to consume our focus, resources and effort.

“Our goal was to place patient outcomes back on center stage where it belongs,” he continued. “As a result, we challenged ourselves in a deliberate and methodical manner to start using EHR data to streamline patient care and thereby achieve meaningful clinical outcomes.”

The clinicians were spending valuable time entering data into the EHR. Phoenix Children’s decided it was time to recognize and reward their dedication by providing them valuable insights on their patients that were easily digestible, actionable and clinically meaningful.

“Insights that would never have been possible without the power of data,” Vaidya explained. “It was time to produce the benefits that have been reaped time and again in every industry where data was used to its fullest extent. It was time to pull healthcare into the 21st century and prevent it from becoming lost in endless cycles of EHR optimization.”


Phoenix Children’s was fortunate to own a robust, internally developed data warehouse using Microsoft technologies. Virtually every electronic system used within the organization contributed to this data warehouse, resulting in the rapid availability and near-real-time accessibility of interconnected system data.

Notably, data from the remotely hosted Altera (Allscripts) EHR arrived at the data center in real time, forming a sturdy foundation and infrastructure for forthcoming data-driven solutions, Vaidya said.

“We recognized the sheer abundance of data alone would not solve the problem, and understood the potential for this to exacerbate rather than solve issues,” he noted. “It was clear we needed to extract, analyze and synthesize the data meaningfully, while presenting it in a clinically intuitive manner to offer actionable insights.

“To ensure success, our data team became integral to the clinical workflow, collaborating closely with clinicians to foster a robust clinical-IT partnership,” he continued.

Staff members were confident that, when executed correctly, these systems could gain rapid acceptance among clinical users, providing a comprehensive patient-by-patient view of clinical conditions as well as population-wide insights.

“These tools would assist providers in identifying care gaps, stratifying patients by severity and complementing their clinical expertise,” Vaidya said. “Although our ultimate vision encompassed widespread deployment across inpatient, ambulatory, emergency room and surgery domains, we initially directed our efforts toward chronic disease management.

“This focus stemmed from the understanding that this patient group required the most care coordination and stood to benefit significantly,” he continued.

In 2017, Phoenix Children’s launched its first dashboard for chronic disease management, targeting juvenile idiopathic arthritis (rheumatoid arthritis).

“Its resounding success led to rapid organizational adoption, sparking interest among physicians from diverse specialties who sought similar clinical tools,” Vaidya noted.


Fast forward to today, six years later, and the health system’s journey has culminated in more than 50 custom-built clinical dashboards that play a pivotal role in driving care excellence throughout Phoenix Children’s. These dashboards have found their applications across the ambulatory, inpatient, emergency room, operating room and intensive care units.

“During the COVID-19 pandemic, these dashboards proved invaluable in rapidly addressing the backlog of delayed elective surgeries, and today these tools continue to help us manage complex cases of more than 5,000 children with epilepsy, scoliosis patients undergoing surgery, chemotherapy oversight, fetal care, suicide prevention, social determinants of health, cystic fibrosis and many others,” Vaidya explained.

“The driving force behind these dashboards is the data sourced from our robust data warehouse, and they are built using Microsoft’s Power BI as the visualization tool,” he continued. “With time, our expertise in creating these innovative dashboards has accelerated, enabling me and my small, dedicated team of four to five clinical analysts to respond to critical clinical needs with agility.”

When a new high-priority clinical need is identified, the team can conceptualize and deploy an initial, fully functional draft dashboard within one to two weeks. This turnaround time has revolutionized the approach, fostering provider engagement, buy-in and enthusiasm for data-driven solutions.

“Our focus doesn’t revolve solely around technology or the flashy interactive dashboard visualizations; rather, it centers on the rapid, efficient and effective resolution of clinical challenges, ultimately enhancing quality and safety,” Vaidya said. “Each dashboard is carefully crafted with the intention of providing actionable insights, and as a result, these insights guide actions that are meticulously tracked through progress indicators and outcome measures displayed on the dashboards.

“The adoption of the dashboards has been truly remarkable, as these solutions have been embraced wholeheartedly by various clinical users,” he continued. “Nurses incorporate them into safety huddle rounds, pharmacists employ them to monitor medication safety, physicians rely on them to monitor high-acuity patients in intensive care units, care coordinators leverage them to guide patients through the health system, schedulers use them to track patients with follow-up delays, and senior leaders depend on them to oversee key metrics.”

While the usage of enterprise wide healthcare data has been impactful, the recent focus has expanded to encompass the collection of patient-generated data from home using remote patient monitoring. This has led to the establishment of five robust home monitoring programs, offering invaluable clinical insights from the patient’s home across a variety of clinical conditions.


“Projects or solutions implemented without clearly defined metrics are like ships without a compass,” Vaidya said. “Recognizing this, we made the strategic decision to ensure all our solutions encompass integrated metrics within the dashboards.

“This thoughtful inclusion serves as a valuable tool, enabling us to effectively track both process measures (actions taken) and outcome measures (results obtained),” he continued. “This approach not only provides an objective measurement of results but also empowers us to evaluate progress, make informed decisions and refine strategies throughout the journey.”

The results obtained validate the carefully chosen strategy and approach. The construction of apps and dashboards that seamlessly deliver real-time data to clinicians has yielded a remarkable impact on patient safety and health outcomes throughout the health system, Vaidya said.

“Our WATCHER program has been highly effective in mitigating acute code events by proactively guiding clinicians to address medical concerns before they escalate into emergencies,” he explained. “Since its implementation, it has identified more than 100 children requiring elevated care levels, resulting in immediate ICU transfers. Clinical teams intervene in response to alerts more than 50% of the time, and since the inception of the program, Phoenix Children’s acute care units have witnessed zero preventable code events.

“The impact of our Malnutrition App is evident as it consistently identifies more than 10 patients each week with potential malnutrition who are then referred to nutritionists for further comprehensive evaluation,” he continued. “Among the algorithm-identified patients, as many as 60-80% are confirmed to have a positive diagnosis of malnutrition and are promptly started on a malnutrition treatment plan.”

Introduced in the fall of 2022, the Cleft Palate App has already been embraced by more than 40 patient families, facilitating the monitoring of their child’s weight and growth. This helped Phoenix Children’s to reduce in-person visits by 20%. In addition, by carefully monitoring the feeding and growth, the health system has been able to reduce the incidence of malnutrition in this patient population by 26%.

“Early results in the first 20 patients using the Leukemia Home Monitoring App showed promising results,” he noted. “This population has a high incidence of readmission in the first month since diagnosis. We were able to successfully reduce the total days spent in the hospital during the first month, a decrease from 19% to 6%.”


Getting started on the data analytics journey is more important than endless planning for perfection, Vaidya advised.

“Don’t wait until you have every bit of data you think you will need or until it is perfectly accurate, because such a state does not exist,” he stated. “Instead, start with the least amount of data that is needed to generate initial actionable insights. Date accuracy and perfection are iterative processes that will naturally follow.

“While it is important to have a smooth-functioning EHR that does not hinder clinical workflows, remember your true goal is to achieve meaningful patient outcomes, not create the perfect EHR software,” he continued. “Do optimize the EHR to meet your needs, but don’t indulge in endless cycles of EHR optimization that add very little clinical value.”

Instead, focus heavily on extracting and analyzing EHR data and delivering it back to the clinical team in real time with actionable insights that would otherwise have been difficult to uncover, he added.

“Remember, healthcare data is not limited just to clinical encounters during inpatient or ambulatory visits,” he said. “A huge amount of untapped, valuable clinical data can be unleashed with remote patient monitoring, which should be an important segment of your healthcare analytics portfolio.

“Combining home monitoring data with clinical encounter data provides richer and more comprehensive clinical information and requires patients to submit their own data from home, strengthening patient engagement and satisfaction,” he added.

As one thinks through the best approach for a hospital or health system, one needs to decide if the organization is better suited to adopt off-the-shelf, vendor-based analytics systems or, if able, develop systems that may be faster, more flexible and cost-effective, Vaidya advised.

“We chose to develop custom solutions in-house, but you should evaluate the pros and cons of each approach,” he added.

Finally, artificial intelligence and machine learning have tremendous potential to positively impact all aspects of healthcare delivery, so it’s time to start investing in the effort and ensure it’s guided by a thoughtful strategy, he said.

“This transformation will not happen overnight, so don’t neglect your existing traditional business intelligence projects while chasing AI,” he concluded. “There is tremendous value in your existing data and not all of it must be predictive in nature.”

Follow Bill’s HIT coverage on LinkedIn: Bill Siwicki
Email him: [email protected]
Healthcare IT News is a HIMSS Media publication.

Source: Read Full Article