The depression and anxiety screenings physician offices administer when patients come in for an office visit—which are linked to performance metrics—may be ineffective in identifying patients who need help.
Among patients newly diagnosed with depression who had been screened in the 30 days before their diagnosis, the Patient Health Questionnaire (PHQ-2) screening instrument detected depression risk in just 57.7% of patients, according to the findings of a study published in the Annals of Family Medicine.
Similarly, the Generalized Anxiety Disorder 2 (GAD-2) only detected anxiety risk 57.3% of the time among patients newly diagnosed with anxiety who had been recently screened, says the study, which was funded by the AMA’s Electronic Health Record Use Research Grant Program that began in 2019 to identify patterns in EHR use that may detract from patient care. The AMA has awarded more than $2 million in grants to 26 organizations over the years to study a variety of EHR-use topics.
“Focusing on incentivized process measures like intake screening questionnaires leads to repetitive and, we hypothesize inaccurate completion,” wrote the authors of the study.
In turn, the screenings may not make as much of a positive impact on reducing mortality and morbidity from depression and anxiety as previously thought. Further, this time-consuming screening may contribute to physician burnout.
“The implementation of screening questionnaires during clinical encounters was done with good intentions, but in practice, clinicians don’t often trust the results of the screens, and it takes up a lot of our limited cognitive bandwidth,” said Chicago family physician Jeff Panzer, MD, MS, one of the study’s authors. “We are so busy asking screening questions that there’s not enough time to ask patients what they’d like to get out of their limited time in the office.”
These findings build on earlier research from the study authors that deemed one in three patient-screening questions to be repetitively “excessive.” Christine A. Sinsky, MD, vice president of professional satisfaction at the AMA, is a co-author on both studies.
How to boost accuracy
Researchers analyzed EHR data from 24 federally qualified health centers for adult patients that had at least one visit between 2019 and 2021. This included nearly 1.9 million PHQ-2 and nearly 1.6 million GAD-2 screenings that medical assistants typically administered verbally during the intake process.
In addition to evaluating PHQ-2 and GAD-2 scores for patients with recent diagnoses, researchers also compared the rates of positive PHQ-2 and GAD-2 screenings administered within the study population to U.S. Census data and the published literature.
They found that the tests had more than 90% sensitivity in published literature, compared with the less than 60% found in the EHR data in their study.
Variations in how practices administer questionnaires and time constraints may account for underlying factors leading to inaccuracies, but the study's authors wrote that more work is needed to understand what led to the difference in the numbers.
In the meantime, they suggested that screening at predetermined intervals—rather than at every clinical encounter—may help boost detection rates. Also, relying on self-administration methods, either electronic or paper, may increase accuracy and put less of a burden on staff and patients.
“I would separate these questions from the intake of a visit. My take-home was not that these questionnaires aren’t useful tools, but that when they are shoved into the beginning of a visit or more so the beginning of every single visit, they lose value,” said Dr. Panzer, vice president of care transformation at Tapestry 360 Health, a federally qualified health center with 16 locations in the Chicago area serving as a medical home for over 28,000 patients.
Dr. Panzer said a huge step would be eliminating performance measures that use these screening instruments as an incentive, which can have the inadvertent consequence of practices administering them too frequently. For example, practices may screen a patient at a visit in December and again during a visit in January to ensure they get credit for the measure for each calendar year.
“Everybody recognizes that depression and anxiety are huge contributors to morbidity and mortality in our country, but it doesn’t help to have a performance measure that is very distal from the outcome. It creates so much energy and work around making sure those boxes are checked and not the outcomes that actually matter,” Dr. Panzer said.
Dive deeper on this topic by reading this recent JAMA® Viewpoint column, "Improving Health Care Quality Measurement to Combat Clinician Burnout," co-written by AMA President Jesse M. Ehrenfeld, MD, MPH. Among other things, Dr. Ehrenfeld and his colleagues cite research showing that outpatient physicians spend 2.6 hours a week on metric reporting while nonphysician staff spends an additional 12.5 hours weekly.
