The technologies available to ensure high-quality care and patient safety are varied, but all depend on data, especially from electronic health record systems, to ensure care providers are making the best decisions during care delivery and have developed safe treatment plans.
But making clinical decision support and advanced analytics models work together optimally is easier said than done.
In this latest installment in Healthcare IT News’ technology optimization special report series, three experts in quality and safety IT – from research firms Black Book Research and KLAS Enterprises and IT vendor Pascal Metrics – describe some best practices to ensure these technologies work in concert toward improved outcomes and reduced medical errors.
Integrating risk, quality and safety
To start, “hospital leaders need to have the systems in place to enable communication across risk, quality and safety to get a comprehensive and timely representation of what is happening in real time,” said Douglas Brown, president and managing partner of Black Book Research, a healthcare technology and services research firm.
“When patient safety and clinical quality data are contained in separate systems that cannot talk to each other, it’s easy to overlook signals that may seem inconsequential alone, but together add up to major troubles.”
By connecting and integrating claims, insurance, legal, financial, safety and clinical data, software platforms will bridge the gap between risk, quality and safety to make sure critical patterns are not missed, Brown added.
“A holistic view simplifies collaboration on overlapping issues and fosters a closer working relationship between the three disciplines to improve patient safety across the continuum of care without labor-intensive analysis and duplication of efforts,” he explained.
Standardized approach to safety
Brown offered another best practice: Strive for a standardized approach in safety (industry-wide) to increase the accuracy of event reporting in conjunction with analytical surveillance systems to capture events automatically.
“Your program should be in a constant state of evaluation and improvement,” he insisted. “This process is one of the keys to making your quality and safety programs lasting and effective.”
Without a baseline of safety and quality event definitions, it proves difficult to correlate drivers and common causes – or interventions with differing outcomes that are useless when sharing data to achieve interoperability goals, he said.
“It is crucial to establish a foundation of standardization in order to benefit from that apples-to-apples comparison, both within the organization and as benchmarking to others,” Brown said.
Use outcomes data to measure
Drew Ladner, chairman and CEO of Pascal Metrics, which works with EHR data to improve quality and safety, said that first and foremost, quality and safety IT optimization is all about outcomes.
“Improving in any domain, and, in any industry, requires knowing what’s going on,” he said. “There is no substitute for outcomes data in knowing the truth. Therefore, truth depends on measurement, and it is outcomes that deliver measurement.
“Without outcomes-derived measurement, providers and health systems lack a lens truly to know what is going on,” he explained. “Indeed, most are not using outcomes but, instead, relying on event reporting or billing/coded data, based on extensive peer-reviewed and real-world evidence.”
To be clear, Ladner stated, voluntary event/incident reporting, on what most providers rely, is a useful source of learning, but not a useful source of measurement, according to extensive peer-reviewed and real-world evidence, for example, capturing approximately 5% of events.
“And during a time in the field when many providers and health systems aspire to become high-reliability organizations, relying on an unreliable method – for example, event reporting – of evaluating progress is problematic,” he said. “Consider replacing an unreliable method with a reliable measure of reliability, namely EHR-based, clinically validated adverse event outcomes.”
For example, one of the major tools that providers use to improve is root cause analysis. But if this analysis relies on unreliable event reporting data, there is a significant opportunity to extend the adoption of outcomes not only to measure, but to provide a timely, continuous stream of data to support quality improvement, Ladner suggested.
“In short, the opportunity is to use EHR-based, clinically validated outcomes data to drive the common cause analysis that extends far beyond the approximate 5%-of-events lens upon which root cause analysis relies,” he said.
“By doing so, providers can move beyond working hard to improve 5% of the problem, avoid missing patterns in 95% of the event data, and deliver more credibility in the organizational culture to quality improvement.”
In sum, outcomes provide scientifically validated and clinically useful measurement that is critical for delivering credible data required to support change management during challenging intervention and improvement, he said.
Outcomes for outcomes’ sake is neither prudent nor the goal – the purpose of outcomes is supporting change with scientific validation and clinical credibility that results in successful intervention and improvement, which is measured by outcomes, he said.
Use outcomes data to predict
Ladner went on with another optimization best practice, saying that outcomes are essential not only to measure, but to train advanced analytics models using machine learning and AI. Accurate and actionable advanced analytics are ever-widening in popularity in safety and quality but elusive to those without EHR-based, clinically validated adverse event outcomes, he added.
“If CIOs are optimizing quality and safety technology that are not driven with EHR-based, clinically validated outcomes, then the first step is to change the technology,” he insisted. “Why? For example, patient deterioration – which many providers laudably seek to predict – is not an outcome. Instead, outcomes in patient safety are EHR-based – versus self-reported or billing/coded – clinically validated adverse events.”
Without clinical validation, outcomes lack credibility, thereby crippling earnest efforts to engage clinicians when it comes to intervention and improvement initiatives, Ladner said.
If a provider wants to predict a medication-related hypoglycemic event, the model used should have been trained with a high volume of EHR-based, clinically validated medication-related glycemic events – not mortality and morbidity data to which most researchers have been historically relegated, he contended.
“For reasons of both accuracy and actionability,” he added, “predicting outcomes using health IT- or EHR-based validated outcomes is far superior to predicting patient condition deterioration – which is what most providers are doing today if they’re using any machine learning or AI to predict safety problems.
It’s all about outcomes
“Whether in detection or prediction, it’s all about outcomes,” he continued. “Whether predicting global harm, for example, a patient will suffer some kind of injury or death as a result of the care versus the disease; specific harm, for example, a patient will suffer from a specific kind of preventable harm; or other safety vulnerability; or simply knowing what’s going on to figure out what to do now, having EHR-based, clinically validated outcomes is essential.”
And this foundation of AI-assisted patient safety – starting with applying machine learning, AI or other technology to adverse events – becomes likewise the foundation for AI-assisted quality, Ladner added.
“The imperative to ‘First, do no harm,’ reminds us that patient safety is foundational and the core of quality,” Ladner said. “All safety problems are quality problems, but not all quality problems are safety problems. Therefore, choose the right evidence-based method and proven technology to address patient safety and extend to quality improvement – versus the other way around.”
Key to clinical transformation
Research firm KLAS looks at quality management in three focus areas: core measures reporting, quality performance improvement, and patient safety and risk. It has started tracking core measures reporting on the ambulatory side. This is because, the firm said, that quality and safety reinforce each other and are key attributes to clinical transformation.
Many providers have focused on two core attributes (there are many more) of a successful and optimized system: A consolidated platform that encompasses both quality and risk solutions, and integration into the EHR, said Ryan Pretnik, director of research, strategy – analytics, at KLAS Enterprises.
“Regarding a consolidated platform that encompasses both quality and risk solutions, in the provider world, having a solution that can check off more boxes on the capabilities end is always very intriguing to providers, even if the solution isn’t the top-performing solution in the space; sometimes good is good enough,” he said.
“Why is that? Having fewer systems to work with is a plus. It’s easier for provider departments to learn and work on the solution, which drives better adoption of the solution, and since solutions that encompass multiple products tend to be integrated nicely, providers don’t have to hop from system to system or try combining data from different solutions to get a holistic view of the data.”
Also, pricing, maintenance of the system and upkeep of multiple systems is taxing for providers, who already manage thousands of applications, he added.
“Having a consolidated, integrated platform that encompasses modules or solutions like claims, peer reviews, credentialing, eCQMs, surveillance, regulations, performance improvement, etc., helps to drive higher provider satisfaction,” Pretnik advised.
“Here is a reference conversation with a provider we recently spoke with: ‘The quality and safety system we currently use is very integrated because we have many, if not all, of their modules/solutions, and the modules/solutions work very nicely together. The system is fully integrated, which makes the solution very intuitive and easy to use, which helps support us on our high-reliability journey.'”
EHR integration is key
And on a final note, integrating quality and safety into the EHR, which is key, he noted. With the EHR being a hub for information, having a quality and risk system systematically work with an EHR is a key driver in decisions by providers, he said.
“Just like other systems, having an optimized workflow and being able to accurately pull or push data to or from the EHR that your quality and risk solutions play nicely with, helps providers feed their EHR or other solution of choice with the accurate information needed for users to interact with,” he said.
“When we look at the scoring of solutions in the quality and risk space, solutions that integrate nicely with the EHR tend to be associated with an increase in overall provider satisfaction.”
This shows how important this core attribute is to providers, he added. For vendors looking to make a jump in this space, having integration with multiple EHRs, with seamless workflow and data transfer, is incredibly valuable to providers, Pretnik said.
“Here,” he concluded, “is a quote from a provider we recently spoke with: ‘I am really happy with the integration we get with our quality and risk solution. The overall product quality is top shelf. The product is basically attached to our EHR, which is fantastic. The solution deals with the EHR by basically reading all the data and helping us produce an output. We started using the solution because of the tight EHR integration and the need for reporting to the governing bodies.”
Source: Read Full Article
