Antidepressant Overtreatment Associated With Brief Screening
Background: Use of brief depression symptom measures for identifying or screening cases may help to address depression undertreatment, but whether it also leads to diagnosis and treatment of patients with few or no symptoms—a group unlikely to have major depression or benefit from antidepressants—is unknown. We examined the associations of use of a brief depression symptom measure with depression diagnosis and antidepressant recommendation and prescription among patients with few or no depression symptoms.
Methods: We conducted exploratory observational analyses of data from a randomized trial of depression engagement interventions conducted in primary care offices in California. Analyses focused on participants scoring <10 on a study-administered 9-item Patient Health Questionnaire (PHQ-9) (completed immediately before an office visit and not disclosed to the provider) with complete chart review data (n = 595). We reviewed visit notes for evidence of practice administration of a brief symptom measure (independent of the trial) and whether the provider (1) diagnosed depression or (2) recommended and/or prescribed an antidepressant.
Results: Among the 545 patients without a practice-administered measure, 57 (10.5%) had a visit diagnosis of depression; 9 (1.6%) were recommended and another 21 (3.8%) prescribed an antidepressant. Among the 50 patients (8.4% of total sample) with a practice-administered measure, 10 (20%) had a visit diagnosis of depression; 6 (12%) were recommended and another 6 (12%) prescribed an antidepressant. Adjusting for nesting within providers, trial intervention, stratification variables, and sample weighting, use of a brief symptom measure was associated with depression diagnosis (adjusted odds ratio, 3.2; 95% confidence interval, 1.1–9.2) and antidepressant recommendation and/or prescription (adjusted odds ratio, 3.80; 95% confidence interval, 1.0–13.9). Analyses using progressively lower PHQ-9 thresholds (<9 to <5) and examining antidepressant prescription alone yielded consistent findings. Analyses by practice-administered measure (PHQ-9 vs PHQ-2) indicated the study findings were largely associated with PHQ-9 use.
Conclusions: These exploratory findings suggest administration of brief depression symptom measures, particularly the PHQ-9, may be associated with depression diagnosis and antidepressant recommendation and prescription among patients unlikely to have major depression. If these findings are confirmed, researchers should investigate the balance of benefits and risks (eg, overdiagnosis of depression and overtreatment with antidepressants) associated with use of a brief symptom measure.
Underidentification of depression is prevalent and multifactorial in origin. Consequently, many practices use brief depression symptom measures, such as the 2- or 9-item versions of the Patient Health Questionnaire (PHQ-2 or PHQ-9), to aid in identifying or screening cases. The US Preventive Services Task Force (USPSTF) and others have endorsed the use of these measures in practices with appropriate diagnostic, treatment, and follow-up systems.
However, no studies have examined whether practice administration of a brief depression symptom measures also leads to antidepressant prescribing among people with few or no depression symptoms. Both the USPSTF and the Canadian Task Force on Preventive Health Care (CTFPHC) emphasize the need to examine this question. The issue is important to address, given that brief symptom measure validation studies using expert structured diagnostic interviews as the reference indicate that individuals with few or no depression symptoms are unlikely to benefit from antidepressants. In PHQ-9 validation studies, those with scores <10 largely comprised nondepressed people and those with minor depression. Such individuals may benefit from behavioral therapies, but evidence from randomized controlled trials (RCTs) indicates most do not benefit from antidepressants. In other words, antidepressant treatment in such patients suggests potential overtreatment.
Despite trial evidence of little benefit, antidepressants are frequently prescribed to patients with few or no depression symptoms, resulting in unnecessary costs and potential detrimental effects (eg, labeling, medication toxicity). Discussions about antidepressants with such patients may also burden office visits, distracting from more salient issues. In contrast with the USPSTF, the CTFPHC recommended against using brief symptom measures, in part because of concerns about antidepressant overtreatment, but it noted a lack of RCTs examining how brief measures contribute both to appropriate depression treatment and overtreatment.
We examined these issues in exploratory observational analyses of data from an RCT. We focused on the subgroup of participants with few or no depression symptoms, defined by a score of <10 on a study-administered PHQ-9. We examined whether a practice-administered brief depression symptom measure was associated with (1) increased diagnosis of depression and (2) increased recommendation and/or prescription of antidepressants. Providers were not informed about the study-administered PHQ-9 or the results. Depression diagnosis and antidepressant recommendation and/or prescription were considered in tandem to help gauge whether antidepressants were intended to address depression or other conditions (eg, insomnia, chronic pain). We examined the composite outcome of antidepressant recommendation and/or prescription because both components reflect clinicians' assessments of treatment need; prescriptions also reflect patient preferences.23– We ascertained use of brief symptom measures, diagnosis of depression, recommendation and prescription of antidepressants from visit medical records.
Given evidence that case finding and screening to identify other conditions lead to overdiagnosis and overtreatment, we hypothesized that use of a brief depression symptom measure would be associated with both increased depression diagnosis and increased antidepressant recommendation and prescription. While a PHQ-9 score <10 helps identify patients unlikely to have major depression, any cut point results in misclassification; lower cut points have increasing specificity in ruling out major depression. Thus, we also examined depression diagnosis and antidepressant recommendation and/or prescription associated with use of a brief symptom measure at progressively lower study-administered PHQ-9 thresholds (<9, <8, <7, <6, <5). Finally, we examined how the associations of use of a brief symptom measure with the study outcomes varied by the specific measure used.
Abstract and Introduction
Abstract
Background: Use of brief depression symptom measures for identifying or screening cases may help to address depression undertreatment, but whether it also leads to diagnosis and treatment of patients with few or no symptoms—a group unlikely to have major depression or benefit from antidepressants—is unknown. We examined the associations of use of a brief depression symptom measure with depression diagnosis and antidepressant recommendation and prescription among patients with few or no depression symptoms.
Methods: We conducted exploratory observational analyses of data from a randomized trial of depression engagement interventions conducted in primary care offices in California. Analyses focused on participants scoring <10 on a study-administered 9-item Patient Health Questionnaire (PHQ-9) (completed immediately before an office visit and not disclosed to the provider) with complete chart review data (n = 595). We reviewed visit notes for evidence of practice administration of a brief symptom measure (independent of the trial) and whether the provider (1) diagnosed depression or (2) recommended and/or prescribed an antidepressant.
Results: Among the 545 patients without a practice-administered measure, 57 (10.5%) had a visit diagnosis of depression; 9 (1.6%) were recommended and another 21 (3.8%) prescribed an antidepressant. Among the 50 patients (8.4% of total sample) with a practice-administered measure, 10 (20%) had a visit diagnosis of depression; 6 (12%) were recommended and another 6 (12%) prescribed an antidepressant. Adjusting for nesting within providers, trial intervention, stratification variables, and sample weighting, use of a brief symptom measure was associated with depression diagnosis (adjusted odds ratio, 3.2; 95% confidence interval, 1.1–9.2) and antidepressant recommendation and/or prescription (adjusted odds ratio, 3.80; 95% confidence interval, 1.0–13.9). Analyses using progressively lower PHQ-9 thresholds (<9 to <5) and examining antidepressant prescription alone yielded consistent findings. Analyses by practice-administered measure (PHQ-9 vs PHQ-2) indicated the study findings were largely associated with PHQ-9 use.
Conclusions: These exploratory findings suggest administration of brief depression symptom measures, particularly the PHQ-9, may be associated with depression diagnosis and antidepressant recommendation and prescription among patients unlikely to have major depression. If these findings are confirmed, researchers should investigate the balance of benefits and risks (eg, overdiagnosis of depression and overtreatment with antidepressants) associated with use of a brief symptom measure.
Introduction
Underidentification of depression is prevalent and multifactorial in origin. Consequently, many practices use brief depression symptom measures, such as the 2- or 9-item versions of the Patient Health Questionnaire (PHQ-2 or PHQ-9), to aid in identifying or screening cases. The US Preventive Services Task Force (USPSTF) and others have endorsed the use of these measures in practices with appropriate diagnostic, treatment, and follow-up systems.
However, no studies have examined whether practice administration of a brief depression symptom measures also leads to antidepressant prescribing among people with few or no depression symptoms. Both the USPSTF and the Canadian Task Force on Preventive Health Care (CTFPHC) emphasize the need to examine this question. The issue is important to address, given that brief symptom measure validation studies using expert structured diagnostic interviews as the reference indicate that individuals with few or no depression symptoms are unlikely to benefit from antidepressants. In PHQ-9 validation studies, those with scores <10 largely comprised nondepressed people and those with minor depression. Such individuals may benefit from behavioral therapies, but evidence from randomized controlled trials (RCTs) indicates most do not benefit from antidepressants. In other words, antidepressant treatment in such patients suggests potential overtreatment.
Despite trial evidence of little benefit, antidepressants are frequently prescribed to patients with few or no depression symptoms, resulting in unnecessary costs and potential detrimental effects (eg, labeling, medication toxicity). Discussions about antidepressants with such patients may also burden office visits, distracting from more salient issues. In contrast with the USPSTF, the CTFPHC recommended against using brief symptom measures, in part because of concerns about antidepressant overtreatment, but it noted a lack of RCTs examining how brief measures contribute both to appropriate depression treatment and overtreatment.
We examined these issues in exploratory observational analyses of data from an RCT. We focused on the subgroup of participants with few or no depression symptoms, defined by a score of <10 on a study-administered PHQ-9. We examined whether a practice-administered brief depression symptom measure was associated with (1) increased diagnosis of depression and (2) increased recommendation and/or prescription of antidepressants. Providers were not informed about the study-administered PHQ-9 or the results. Depression diagnosis and antidepressant recommendation and/or prescription were considered in tandem to help gauge whether antidepressants were intended to address depression or other conditions (eg, insomnia, chronic pain). We examined the composite outcome of antidepressant recommendation and/or prescription because both components reflect clinicians' assessments of treatment need; prescriptions also reflect patient preferences.23– We ascertained use of brief symptom measures, diagnosis of depression, recommendation and prescription of antidepressants from visit medical records.
Given evidence that case finding and screening to identify other conditions lead to overdiagnosis and overtreatment, we hypothesized that use of a brief depression symptom measure would be associated with both increased depression diagnosis and increased antidepressant recommendation and prescription. While a PHQ-9 score <10 helps identify patients unlikely to have major depression, any cut point results in misclassification; lower cut points have increasing specificity in ruling out major depression. Thus, we also examined depression diagnosis and antidepressant recommendation and/or prescription associated with use of a brief symptom measure at progressively lower study-administered PHQ-9 thresholds (<9, <8, <7, <6, <5). Finally, we examined how the associations of use of a brief symptom measure with the study outcomes varied by the specific measure used.
SHARE
