Decreasing sensitivity of clinical alcohol screening with the AUDIT-C after repeated negative screens in VA clinics
Introduction
Routine population-based alcohol screening in primary care settings is recommended by the U.S. Preventive Services Task Force based on the effectiveness of brief interventions among patients who screen positive for unhealthy alcohol use (Fleming et al., 2002, Maciosek et al., 2010, Saitz, 2005, Solberg et al., 2008, U.S. Preventive Services Task Force, 2004) and recent U.S. health care reform established alcohol screening as a standard preventive health benefit (HealthCare.gov, 2013a, HealthCare.gov, 2013b). The U.S. Medicare program, which provides health care coverage to adults aged 65 and older, reimburses providers for annual alcohol screening (Centers for Medicare and Medicaid Services, 2011) and the U.S. Veterans Health Administration (VA) has required annual screening with the Alcohol Use Disorders Identification Test-Consumption (AUDIT-C) questionnaire since 2004 (Hawkins et al., 2007). However, the optimal frequency of alcohol screening is unknown.
Recent studies have suggested that alcohol screening identifies few positive screens among patients who repeatedly screen negative on prior annual alcohol screens (Lapham et al., 2014, Lapham et al., 2013). For example, among VA outpatients who have screened negative on 4 screens, only 1.9% screened positive for unhealthy alcohol use on their next screen. Depending on patient characteristics like age, gender and the score on negative screens, the prevalence of positive screens was as low as 0.6% for some patient sub-groups (Lapham et al., 2013). These studies concluded that repeat annual screening for most patients with multiple prior negative screens seemed reasonable but the interval could potentially be extended for some patient sub-groups with very low probabilities of a subsequent positive screen. However, these studies relied on the results of clinical screens and lacked a reference measure for unhealthy alcohol use. Therefore, the validity of repeat screening among these patients remains unknown.
Despite extensive validation in primary care patients, there are several reasons that alcohol screens might perform differently in patients who have had repeated negative screens. In general, when screening is conducted in populations with a very low prevalence of the target condition, the positive predictive value of a screen will be low, regardless of a screen's sensitivity and specificity, meaning, for example, that a large proportion of positive alcohol screens will likely be falsely positive (Brenner and Gefeller, 1997, Hulley and Cummings, 2007, Lachs et al., 1992). In addition, studies of other conditions have shown that the sensitivity of a screen can vary depending on condition prevalence, with decreased performance in populations with a low condition prevalence, such as patients with repeated negative alcohol screens (Lachs et al., 1992). Finally, for alcohol screening in particular, some patients with unhealthy alcohol use may not be willing to accurately report their alcohol use on screens in clinical settings, but may be willing when screening is confidential, which would reduce the sensitivity of clinical screening in these patients (Bradley et al., 2011).
The purpose of this study was to evaluate the performance of repeat annual clinical alcohol screening in 4 samples of VA outpatients with 1–4 prior consecutive negative annual screens. Specifically, we compare the performance of the next annual clinical screen to an alcohol screen on a confidential mailed survey. If the performance of annual clinical alcohol screening is poor in some patient populations with prior negative screens, alternative approaches for assessing alcohol use may be warranted.
Section snippets
Study sample and data
Secondary data from three national VA data sources were obtained for this study. Administrative and clinical data from the electronic medical record (EMR) were extracted from the VA's Corporate Data Warehouse and National Patient Care Databases and included patient-level data on EMR-documented annual alcohol screening performed in VA clinics as well as demographics and clinical diagnoses. Confidential survey data were obtained from the VA Survey of Healthcare Experiences (SHEP), available
Study sample and patient characteristics
Of 462,126 potentially eligible VA outpatients with 2 or more annual clinical alcohol screens during the study period, 20.9% (96,563) were excluded by virtue of screening positive on their first screen (11.9%) or having documented evidence of an alcohol use disorder or addictions treatment during the study (9.0%). Another 74.1% (347,160) were excluded for lack of a survey comparison screen 9 to 15 months after the last prior negative screen, leaving 4% (18,403) of patients with at least 1 to 4
Discussion
As in previous studies, this study found that among patients with repeated negative clinical alcohol screens, the proportion who subsequently screened positive a year later was low and decreased as patients had more prior negative screens (6% decreasing to 1.7% for patients with 1–4 prior negative screens; Lapham et al., 2013, Lapham et al., 2014). In addition, among patients with prior negative screens, the sensitivity of the clinical screen when compared to the survey screen was also low (41%
Role of funding source
The research reported here was funded by the US Department of Veterans Affairs, Veterans Health Administration, Office of Research and Development, Health Services Research and Development (HSR&D) Substance Use Disorders Quality Enhancement Research Initiative (QUERI) RRP 11-021. A preliminary version of this work was presented at both the 36th Annual Research Society for Alcoholism Scientific Meeting in Orlando, FL, June 25, 2013 and the Addiction Health Services Conference in Portland, OR,
Contributors
Authors Bradley and Lapham designed the study and wrote the protocol. Dr. Lapham undertook the statistical analysis and wrote the first draft of the manuscript. All authors contributed to and have approved the final manuscript.
Conflict of interest statement
Dr. Katharine Bradley owns stock for the following pharmaceutical companies, which have no relationship to this study and its results: Johnson and Johnson, Abbot, Pfizer, Proctor and Gamble. No conflict of interests for the other authors.
Disclaimer
The views expressed in this article are those of the authors and do not necessarily reflect the position or policy of the Department of Veterans Affairs, the United States Government, or any of the authors’ institutions.
Acknowledgments
We are grateful to the VA's Office of Analytics and Business Intelligence and Corporate Data Warehouse for making their data available to this study.
References (54)
- et al.
The AUDIT-C: screening for alcohol use disorders and risk drinking in the presence of other psychiatric disorders
Compr. Psychiatry
(2005) - et al.
Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases
J. Clin. Epidemiol.
(1992) - et al.
Estimating risk of alcohol dependence using alcohol screening scores
Drug Alcohol Depend.
(2010) - et al.
AUDIT: The Alcohol Use Disorders Identification Test: Guidelines for Use in Primary Care
(2001) Mode of questionnaire administration can have serious effects on data quality
J. Public Health (Oxf)
(2005)- et al.
Two brief alcohol-screening tests from the Alcohol Use Disorders Identification Test (AUDIT): validation in a female Veterans Affairs patient population
Arch. Intern. Med.
(2003) - et al.
Screening for problem drinking: comparison of CAGE and AUDIT
J. Gen. Intern. Med.
(1998) - et al.
AUDIT-C as a brief screen for alcohol misuse in primary care
Alcohol Clin. Exp. Res.
(2007) - et al.
Brief approaches to alcohol screening: practical alternatives for primary care
J. Gen. Intern. Med.
(2009) - et al.
Using alcohol screening results and treatment history to assess the severity of at-risk drinking in Veterans Affairs primary care patients
Alcohol Clin. Exp. Res.
(2004)