JAMIA: Problem lists benefit from inference rules
Researchers implementing a clinical alerting system using inference rules to notify providers of undocumented problems found that clinics using such a system added nearly three times as many problem documentation notes to EHRs than clinics that did not, according to a report published Jan. 3 by the Journal of the American Medical Informatics Association.
“In this study, we performed a cluster randomized, controlled trial of a clinical alerting system that used inference rules to detect and notify providers of undocumented problems, giving them the opportunity to correct these gaps and increase problem list completeness,” wrote lead author Adam Wright, MD, of Brigham and Women’s Hospital in Boston.
“Our goal was to assess whether or not this system would improve problem notation for a broad array of patient conditions.”
Using inference rules developed for 17 conditions, an EHR-based intervention was conducted in 14 primary care practices, with an additional 14 primary care practices affiliated with Brigham and Women’s Hospital acting as controls.
The 17 conditions included attention deficit hyperactivity disorder, asthma/COPD, breast cancer, coronary artery disease, congenital coagulopathy, congestive heart failure, diabetes mellitus, glaucoma, hypertension, hyperthyroidism, hypothyroidism, myasthenia gravis, osteoporosis/osteopenia, rheumatoid arthritis, renal failure, sickle cell disease and stroke.
Wright wrote that inference rules were developed along the following steps: identification of problem associations with structured data, selection of specific problems, development of preliminary rules, characterization of preliminary rules and alternatives, selection of a final rule and validation of the final rule.
Pre-intervention data were collected from November 2009 to May 2010 and intervention data were collected from May 2010 to November 2010.
In primary care practices receiving the intervention, EHRs were programmed to analyze providers’ typed notes and patients’ medications, laboratory results, billing codes and vital signs to detect for one of the 17 conditions studied. If a condition was detected in a patient, but was undocumented in the EHR, an actionable alert would appear onscreen, allowing a provider to accept the alert, ignore the alert once or ignore the alert forever.
Throughout the entire study period, 38,025 patients were seen in the clinics receiving the intervention, 41,039 were seen in the control clinics and 3,894, or 5 percent, were seen in one of each.
According to the report, 41.1 percent of the 17,043 alerts that presented in intervention clinics were accepted, and intervention clinics documented a total of 10,016 study problems during the intervention period compared to control clinics’ total of 3,739.
Researchers concluded that problem inference rules could be used to help providers develop more complete EHRs with up-to-date and accurate problem lists.
“Our results suggest that problem inference rules such as these are a valuable tool for improving problem list completeness and thus may be beneficial for improving patient care,” Wright wrote.
“The rate of notation of study problems increased dramatically during the intervention period as a result of this simple alert-based intervention,” he concluded. “Overall, study problems were approximately three times more likely to be documented when alerts were shown. This increase is clinically important, since many of these problems are used for quality improvement and clinical decision support.”
“In this study, we performed a cluster randomized, controlled trial of a clinical alerting system that used inference rules to detect and notify providers of undocumented problems, giving them the opportunity to correct these gaps and increase problem list completeness,” wrote lead author Adam Wright, MD, of Brigham and Women’s Hospital in Boston.
“Our goal was to assess whether or not this system would improve problem notation for a broad array of patient conditions.”
Using inference rules developed for 17 conditions, an EHR-based intervention was conducted in 14 primary care practices, with an additional 14 primary care practices affiliated with Brigham and Women’s Hospital acting as controls.
The 17 conditions included attention deficit hyperactivity disorder, asthma/COPD, breast cancer, coronary artery disease, congenital coagulopathy, congestive heart failure, diabetes mellitus, glaucoma, hypertension, hyperthyroidism, hypothyroidism, myasthenia gravis, osteoporosis/osteopenia, rheumatoid arthritis, renal failure, sickle cell disease and stroke.
Wright wrote that inference rules were developed along the following steps: identification of problem associations with structured data, selection of specific problems, development of preliminary rules, characterization of preliminary rules and alternatives, selection of a final rule and validation of the final rule.
Pre-intervention data were collected from November 2009 to May 2010 and intervention data were collected from May 2010 to November 2010.
In primary care practices receiving the intervention, EHRs were programmed to analyze providers’ typed notes and patients’ medications, laboratory results, billing codes and vital signs to detect for one of the 17 conditions studied. If a condition was detected in a patient, but was undocumented in the EHR, an actionable alert would appear onscreen, allowing a provider to accept the alert, ignore the alert once or ignore the alert forever.
Throughout the entire study period, 38,025 patients were seen in the clinics receiving the intervention, 41,039 were seen in the control clinics and 3,894, or 5 percent, were seen in one of each.
According to the report, 41.1 percent of the 17,043 alerts that presented in intervention clinics were accepted, and intervention clinics documented a total of 10,016 study problems during the intervention period compared to control clinics’ total of 3,739.
Researchers concluded that problem inference rules could be used to help providers develop more complete EHRs with up-to-date and accurate problem lists.
“Our results suggest that problem inference rules such as these are a valuable tool for improving problem list completeness and thus may be beneficial for improving patient care,” Wright wrote.
“The rate of notation of study problems increased dramatically during the intervention period as a result of this simple alert-based intervention,” he concluded. “Overall, study problems were approximately three times more likely to be documented when alerts were shown. This increase is clinically important, since many of these problems are used for quality improvement and clinical decision support.”