BACKGROUND/RATIONALE:
VA screens for alcohol misuse annually, but the optimal frequency of alcohol screening is unknown. The US Preventive Services Task Force recommends "routine" alcohol screening without specifying a frequency. VA chose annual alcohol screening as a common interval for other preventive care. However, little is known about the yield of repeat annual screening for patients who screen negative on the AUDIT-C (e.g. nondrinkers and low-risk drinkers). Rescreening patients who screen negative on the AUDIT-C may not be useful if those patients rarely convert to positive screens.

OBJECTIVE(S):
To determine whether some subgroups of VA outpatients who screen negative on the AUDIT-C are at such low risk for having a later positive AUDIT-C screen that they may not need to be rescreened annually.

SPECIFIC AIMS
Aim 1 estimated, among male outpatients who initially screened negative for alcohol misuse and had no past-year substance use disorder (SUD) diagnosis or treatment, the probability of remaining a negative screen for alcohol misuse at follow-up alcohol screening. Specifically we: a) evaluated demographic and clinical factors associated with remaining a negative screen at follow-up annual screening; b) estimated the probability of remaining a negative screen at follow-up screening among patients who screened negative on 1, 2, 3, or 4 consecutive prior clinical screens; and c) also estimated the probability of screening negative on a mailed survey screen completed within 90 days of the last negative clinical screen.

Aim 2 estimated the probability of remaining a nondrinker on follow-up screens in parallel analyses to Aims 1a-c above.

Secondary aims included parallel analyses to Aims 1 & 2 1) stratified by any demographic and clinical covariates associated with the probability of remaining a negative screen or nondrinker and 2) for women.

METHODS:
Data sources: Data sources for the study included the Corporate Data Warehouse (CDW), Survey of Healthcare Experiences of Patients, and the National Patient Care Databases.

Population: VA outpatients with no past-year documented substance use disorders (SUD) diagnoses who received care in VA Region 1 between 2004-2008 and had two or more (up to 6) consecutive clinical AUDIT-C screens a year (9-15 months) apart in the CDW. Eligibility was limited to those whose first screen was negative.

Main measures: For Aim 1 analyses, men were classified as negative or positive screens based on AUDIT-C scores 0-3 and 4-12 points, respectively (scores of 0-2 and 3-12 for women, respectively). For Aim 2 analyses, they were classified as nondrinkers or drinkers based on AUDIT-C scores 0 and 1-12, respectively. Results were evaluated as the probability of a positive screen at follow-up (i.e. 1-probability of a negative follow-up screen).

FINDINGS/RESULTS:
Overall 338,531 men and 21,853 women were eligible with at least 1 pair of clinical alcohol screens in which the initial screen was negative for misuse. Because there were adequate numbers of women for most analyses, and the threshold for a positive screen differs for men and women, results for men and women are presented in parallel.

Aims 1a and 2a. The probability of a positive screen at follow-up after an initial negative screen was 5.9% in men and 5.2% in women. The factors most strongly associated with remaining a negative screen at follow-up among men and women were age and the score on the initial negative AUDIT-C. For both samples, younger patients and those with higher initial negative screen scores were most likely to have a positive screen at follow-up. Among men who initially screened negative, the adjusted probability of a positive screen at follow-up ranged from 2.1% to 37.7% depending on age and initial negative screen score. For women, comparable results were 2.3% to 24.5% depending on age and initial negative screen score.

Aims 1b and 2b. The number of prior negative screens had relatively minor impact on the adjusted probability of a positive follow-up screen: decreased from 4.8% to 2.0% as the number of prior negative screens increased from 1 to 4. Despite the relatively minor variation in the adjusted probability of a positive follow-up screen by number of prior negative screens, when the number of prior negative screens and patient age, gender and their prior negative screen scores were considered together, the variation ranged from 0.6% - 39.3%.

Aims 1c and 2c. The adjusted probability of a positive screen at follow-up was greater on surveys than on the last clinical screen: decreased from 8.6% to 4.5% as the number of prior negative screens increased from 1 to 4. Depending on the number of prior negative screens (1 to 3), age, and prior negative screen scores, the adjusted probability of a positive survey screen at follow-up ranged from 1.7% to 74.0%. These results suggest that false negative clinical screens may be common and may be an important reason to continue screening many if not all patients annually.

IMPACT:
The number of patients with an initial negative alcohol screen who need to be rescreened to identify one patient with a positive screen at follow-up ranged from 3 to 159. This study suggests that decreasing the frequency of rescreening older patients who have low negative scores on the AUDIT-C might be reasonable.

---

External Links for this Project

---

PUBLICATIONS:

Journal Articles

1. Lapham GT, Rubinsky AD, Heagerty PJ, Williams EC, Hawkins EJ, Maynard C, Kivlahan DR, Bradley KA. Annual rescreening for alcohol misuse: diminishing returns for some patient subgroups. Medical care. 2013 Oct 1; 51(10):914-21. [view]
2. Lapham GT, Rubinsky AD, Heagerty PJ, Achtmeyer C, Williams EC, Hawkins EJ, Maynard C, Kivlahan DR, Au D, Bradley KA. Probability and predictors of patients converting from negative to positive screens for alcohol misuse. Alcoholism, clinical and experimental research. 2014 Feb 1; 38(2):564-71. [view]

Conference Presentations

1. Lapham GT, Rubinsky AF, Williams EC, Maynard CC, Heagerty P, Kivlahan DR, Au DH, Bradley KA. Factors Associated with Repeated Negative Alcohol Screens: Is Annual Alcohol Screening Unnecessary for Some VA Patients? Poster session presented at: Research Society on Alcoholism Annual Scientific Meeting; 2012 Jun 26; San Francisco, CA. [view]
2. Lapham GT, Rubinsky AD, Williams EC, Hawkins EJ, Kivlahan DR, Maynard C, Bradley KA, Heagerty PJ. The Association between a Positive Screen for Alcohol Misuse and Prior Negative Screen Scores. Paper presented at: Addiction Health Services Research Conference; 2012 Oct 18; New York, NY. [view]

DRA: Mental, Cognitive and Behavioral Disorders, Substance Use Disorders
DRE: Prognosis
Keywords: none
MeSH Terms: none