BMJ: EHRs could help flag patients for domestic abuse risk
EHRs could be used to predict a patient’s risk of suffering domestic abuse, according to research published Sept. 29 in the online version of the British Medical Journal.
Lead author Ben Y. Reis, PhD, from Boston Children’s Hospital informatics program and Harvard Medical School in Boston, and colleagues examined whether historical electronic data could be used to target patients at risk for domestic abuse.
They analyzed medical records from more than 560,000 non-identifiable patients over 18 years of age with at least four years of health data available from hospital or emergency department admissions. These patients had more than 16 million diagnoses among them and cases of abuse were recorded according to established recordkeeping codes.
Using EHRs and risk factors, the authors developed a scoring system to predict which patients were likely to eventually receive a domestic abuse diagnosis. According to the authors, they were able to use the system to predict future diagnoses of abuse 10 to 30 months in advance.
The authors identified certain risk factors associated with a future diagnosis of abuse. For women this included being seen in hospitals or emergency rooms for injuries, poisoning or alcoholism. And men seen for conditions such as depression and psychosis were most at risk for a diagnosis of future abuse.
According to the researchers, 5,829 of the patients met the narrow case definition for abuse, while 19,303 met the broader case definition for abuse.
Since domestic abuse is the single most common cause of injury to women in the United States, the authors said, “it is critical that at–risk patients be identified as early as possible.” And, according to Reis, the risk profiles developed by the authors should be able to help doctors with early diagnosis, “perhaps many years in advance."
"With increasing amounts of data becoming available,” said Reis, “this work has the potential to bring closer the vision of predictive medicine, where vast quantities of information are used to predict individuals' future medical risks in order to improve medical care and diagnosis."
Lead author Ben Y. Reis, PhD, from Boston Children’s Hospital informatics program and Harvard Medical School in Boston, and colleagues examined whether historical electronic data could be used to target patients at risk for domestic abuse.
They analyzed medical records from more than 560,000 non-identifiable patients over 18 years of age with at least four years of health data available from hospital or emergency department admissions. These patients had more than 16 million diagnoses among them and cases of abuse were recorded according to established recordkeeping codes.
Using EHRs and risk factors, the authors developed a scoring system to predict which patients were likely to eventually receive a domestic abuse diagnosis. According to the authors, they were able to use the system to predict future diagnoses of abuse 10 to 30 months in advance.
The authors identified certain risk factors associated with a future diagnosis of abuse. For women this included being seen in hospitals or emergency rooms for injuries, poisoning or alcoholism. And men seen for conditions such as depression and psychosis were most at risk for a diagnosis of future abuse.
According to the researchers, 5,829 of the patients met the narrow case definition for abuse, while 19,303 met the broader case definition for abuse.
Since domestic abuse is the single most common cause of injury to women in the United States, the authors said, “it is critical that at–risk patients be identified as early as possible.” And, according to Reis, the risk profiles developed by the authors should be able to help doctors with early diagnosis, “perhaps many years in advance."
"With increasing amounts of data becoming available,” said Reis, “this work has the potential to bring closer the vision of predictive medicine, where vast quantities of information are used to predict individuals' future medical risks in order to improve medical care and diagnosis."