Reliability of a Timeline Followback measure of caffeinated coffee consumption.

BACKGROUND AND AIMS: The Timeline Followback (TLFB) was initially developed to collect retrospective self- reports of alcohol and drug use. Since its development, several peer-reviewed papers have supported it as a sound psychometric measure for substance use and for several other behaviors. Worldwide, coffee is one of the most widely consumed beverages. Although early epidemiological research suggested that drinking coffee is associated with some health problems, several recent studies have found moderate coffee consumption to have an inverse association with mortality. Because of its widespread use, a psychometrically sound measure of coffee consumption would help better inform research and public health policies. DESIGN: This study investigated the test-retest reliability of a version of the TLFB modified to assess coffee consumption. SETTING AND PARTICIPANTS: 98 clinical psychology doctoral students completed a 30-day coffee TLFB on two occasions separated by 14 days. MEASUREMENTS: Participants used a computerized Survey Monkey™ version of the coffee TLFB to report their caffeinated coffee consumption for the same 30-day interval. For the first administration participants reported their consumption for the preceding 30 days, and for the second administration they reported on the preceding 45 days which included the earlier reporting interval. FINDINGS: Reliability coefficients were found to be very high (0.90-0.97). CONCLUSIONS: As with other substance use versions of the TLFB, these results suggest the TLFB is a good method for measuring coffee consumption in research studies and in primary health care settings where such data may be important.

Two Brief Measures of Alcohol Use Produce Different Results: AUDIT-C and Quick Drinking Screen.

BACKGROUND: Several psychometrically sound measures of alcohol use have been developed to assess drinking. The Alcohol Use Disorders Identification Test (AUDIT) and its shorter counterpart the AUDIT-C, which contains the first 3 AUDIT questions, were developed by the World Health Organization and have become the preferred brief measures for screening and evaluating problem severity. This study compared the first 3 questions on the AUDIT with another psychometrically sound brief measure of alcohol use, the Quick Drinking Screen (QDS). METHODS: Data were obtained from a randomized controlled trial of a mail-based intervention promoting self-change with 472 alcohol abusers (n = 280, no prior alcohol treatment; n = 192, prior alcohol treatment). Participants' retrospective self-reports of alcohol consumption were collected using the QDS and the 3 AUDIT-C questions and compared. Although both measures contain similar questions (2 quantity-frequency and 1 binge drinking), they differ in 2 important ways: (i) temporal interval over which data are collected, and (ii) formatting of response options (i.e., a continuous number vs. categorical). RESULTS: Intraclass correlations for drinking variables were moderate to moderately high. A repeated-measures MANOVA using treatment condition and gender as variables revealed significant differences in absolute values between the 2 drinking measures with the QDS showing greater consumption on almost all variables. Participants' numerical answers on the QDS were compared to their categorical answers to the similar alcohol use questions on the AUDIT-C. The comparison revealed that participants' answers on the AUDIT-C were associated with extreme variability compared to their QDS answers. This variability suggests the AUDIT-C would be unreliable as a quantitative measure of alcohol consumption. CONCLUSIONS: The differences between the 3 alcohol use questions on the AUDIT-C and the same questions on the QDS may reflect the imprecision of the AUDIT-C's drinking response categories. Results suggest that the QDS can be used to identify risky drinking and to provide a more informative characterization of a drinker's alcohol consumption than that provided by the AUDIT-C.

Preventing alcohol-exposed pregnancies among Hispanic women.

Project Healthy CHOICES, a self-administered, mail-based prevention intervention, was developed for women at risk of an alcohol-exposed pregnancy (AEP). Participants were sent their assessment and study materials through the United States Postal Service. This article uses data from a larger study (N = 354) and focuses on the 89 women who identified as Hispanic. Potential participants who called in response to English and Spanish ads and who said they could read and write Spanish were given a choice of receiving the intervention materials in English or Spanish. The main objective of the present study was to evaluate differences in outcomes as a function of (a) the language in which the intervention materials were received, and (b) the participants' acculturation levels. Prior to the study, all women were at risk of an AEP. At the 6-month follow-up, two thirds (66%) of all Hispanic women had reduced their overall risk of an AEP, primarily by practicing effective birth control. These outcomes are similar to those reported for previous Project CHOICES studies. Significantly more women who requested the intervention materials in English (75%) compared to Spanish (41%) reduced their overall risk of an AEP. Women with high English cultural domain scores were at significantly less risk of an AEP due to effective contraception and a reduced overall risk of an AEP. Compared to other Project CHOICES studies, Project Healthy CHOICES is less intensive; it is self-administered, freely available, and can be completed without visiting a health care practitioner or clinic.