2019 Top 10 Most Prevalent Patient Harms Happening in U.S. Hospitals
Pascal Metrics is pleased to release the first-ever list of its kind on the prevalence of patient harm using EHR-based clinically validated adverse event outcomes data.
Leading health systems from Pascal’s U.S. Community Collaborative have already deployed this solution, and it’s from their June 1, 2018-May 30, 2019 data that Pascal was able to develop its inaugural list of Top 10 Patient Harms. The harms listed below are ones that, according to Levels F-I of The National Coordinating Council for Medication Error Reporting and Prevention’s (NCC MERP) Harm Scale, result in increased hospitalization, irreversible damage or death:
Oversedation, which occurs when patients experience hypotension or respiratory failure because they’ve received too high a dose of analgesics, anti-anxiety drugs or sedatives.
Sepsis, which is a potentially fatal inflammatory reaction to a bloodstream infection.
Glycemic Events, which occur when a patient’s blood sugar levels are either too high or too low.
Abnormal Surgical Bleeding, which is defined as any post-invasive procedure unexpected blood loss.
Respiratory Complications, through which it becomes difficult for patients to breathe.
Organ Injury/Repair/Removal, in which case a provider mistakenly impacts an otherwise healthy organ.
VTEs/Clots/Occlusions, which stem from difficulties with coagulation.
Neonatal Injury, which covers any injury that takes place before, during or after birth.
Fall with Injury, which can occur either because there is a safety hazard present in the facility, or a patient suffers another harm and loses consciousness as a result.
Acute Kidney Injury, which can occur when toxic chemicals build up in the body either through illness or inappropriate prescription management.
Why is this Top 10 list important and even revolutionary?
Let’s start with its importance. This Top 10 list is the first ever not only to provide a validated view of which patient harms representative hospitals and health systems in the United States are finding based on EHR data but also a high-level “heads up” guide for those hospitals and health systems which have not deployed all-cause harm technology systems enterprise-wide.
Dr. Don Berwick, in a recent piece in Annals of Internal Medicine co-authored with Dr. David Classen and Fran Griffin RN, called on boards and all hospitals to find patient safety vulnerabilities using EHR data. This Top 10 list is example of what doing that looks like at the nationwide level.
Let’s dive deeper. Clinical teams cannot efficiently and effectively prevent patient harm (used interchangeably here with “adverse event”) without an accurate understanding of what harms have happened, are happening, and may happen right before them.
If clinicians are going to answer each of those questions – retrospective, concurrent, and prospective – they must use EHR-based clinically validated adverse event outcomes (“AE Outcomes”) data. Most to date have not, largely because they have not had access to these data sets. Consequently, the entire field and industry has relied on proxy data, such as less accurate retrospective claims data, or proxy methods such as voluntary event reporting, which captures approximately 5% of all adverse events (claims do only a little better at 10%).
Therefore, AE Outcomes are important, as they alone enable a health system to measure patient safety with accuracy. They alone enable a board and its clinical leaders, managers, and front-line caregivers to answer the question, “Are our patients safe?”
Without AE Outcomes data, a scientifically valid and clinically credible baseline of patient harm cannot be established and, without that, a health system cannot determine on an “all-cause harm” basis whether they are improving or not. In short, without AE Outcomes, a health system is – when it comes to patient safety – “flying blind.”
And while we’re on the subject of preventing harm, it’s worth pointing out that, amidst extraordinary hype around predictive analytics and AI, providers and health systems cannot predict their way to performance. The whole analytic continuum is essential for use in performance improvement.
Indeed, performance does not start and end with prediction or AI, it starts and ends with EHR-based clinically validated adverse event outcomes (“AE Outcomes”). Further, without AE Outcomes, there is no useful AI in patient safety. AE Outcomes are essential for building accurate, useful advanced analytics, whether in machine learning, AI, or else.
For example, if we want to predict medication-related bleeding, we need medication-related bleeding AE Outcomes data in sufficient volume to train our models. The primary reason researchers have been relying on coarse-grained mortality and morbidity data in recent decades is that they haven’t had access to AE Outcomes.
Until now. Pascal Metrics, which pioneered the use of EHR data to identify AE Outcomes both through detection and prediction, has been accumulating EHR-based clinically validated adverse event outcomes data for over 10 years and holds the largest such data set worldwide. Pascal was the first to use its AE Outcomes data with machine learning and advanced AI techniques to predict “all-cause harm” and has been pioneering how to move well beyond.
Why is this Top 10 list “even revolutionary”? While there are health systems in the United States that have decided to embrace the peer-reviewed published evidence, regulatory direction, and value of identifying and reducing all-cause harm, the state of the field is this: most hospitals and health systems – even those institutions that are considered in the global elite of healthcare – still too often rely on the traditional method of “See something, say something”.[/vc_column_text][vc_empty_space][vc_column_text]
Would we be okay if our families only flew on planes using visual flight rules?
To be crystal clear, Pascal Metrics embraces the use of voluntary event reporting. However, voluntary event reporting is a source of learning – particularly about those safety vulnerabilities that are not captured by or in electronic data – but it is not a source of measurement. There is a better way, it has been validated, and it works to avoid injury and save lives; it’s called health IT-enabled patient safety.
Therefore, the revolutionary aspect of this Top 10 is that it’s disruptive, highly disruptive. Ten years ago, many thought this approach was straight from Buck Rogers. Five years ago, it was viewed as advanced, and perhaps even impressive, but “a bridge too far.” Today – with one out of three U.S. patients suffering patient harm – the hospitals and health systems who care about patient safety and who care about value are embracing this approach.
However, culture in too many hospitals and health systems – which have been so accustomed to relying on event reporting data and drawing on highly retrospective claims data – are slow to embrace the entirely new and unfamiliar approach to finding safety vulnerabilities with the implications of transforming quality improvement.
Moreover, at the field-wide level, the reason why this is revolutionary is that, unlike recent decades, the healthcare organizations who will achieve excellence and well-deserved reputational capital as a result will do so on the basis of data. In the past, data has not been the primary driver of reputational capital, because it’s been prohibitively high-cost to access or simply unavailable.
And because ultimately patients care most about avoiding preventable injury and death, the data that will build the foundations of the highest performing healthcare delivery systems and brands of tomorrow will be AE Outcomes. The clock is ticking.