Telephone self-monitoring among alcohol use disorder patients in early recovery: a randomized study of feasibility and measurement reactivity

Abstract

Frequent symptom self-monitoring protocols have become popular tools in the addiction field. Interactive Voice Response (IVR) is a telephone monitoring system that has been shown to be feasible for collecting frequent self-reports from a variety of research populations. Little is known, however, about the feasibility of using IVR monitoring in clinical samples, and few controlled trials exist assessing the impact of any type of frequent self-report monitoring on the behaviors monitored. This pilot study with patients in early recovery from an alcohol use disorder (n = 98) evaluated compliance with two IVR monitoring protocols, subjective experiences with monitoring, and change in symptoms associated with monitoring (i.e., measurement reactivity). Participants were randomly assigned to call an IVR system daily for 28 days, once per week for 4 weeks, or only to complete 28-day follow-up assessment including retrospective drinking reports. Monitoring calls assessed alcohol craving, substance use, emotional well-being, and PTSD symptoms. Most monitoring participants completed calls on at least 75% of scheduled days (72.2% and 59.2% for daily and weekly, respectively). Including reconstructed data from follow-up of missed calls yielded 77.8% and 74.1% of maximum data points, respectively. Most monitoring participants indicated the protocol was manageable and reported positive or no effects of monitoring on urges to use alcohol, actual drinking, and PTSD symptoms. Analyses of measurement reactivity based on assessment one month after randomization found no significant group differences on drinking, craving for alcohol, or PTSD-related symptoms. Results suggest that IVR technology is feasible and appropriate for telephone symptom monitoring in similar clinical samples.

Introduction

There is a long tradition of using frequent self-monitoring to study various aspects of drinking behavior, including consumption patterns (Carney et al., 1998), the validity of retrospective consumption measures such as the Timeline Follow-Back (Searles et al., 2002), and relapse (Kranzler et al., 2004, Litt et al., 2000; see also Leigh, 2000 for an overview of this literature). Self-monitoring is defined here as any behavior that entails reflecting on and providing self-reports about a specific behavior, such as drinking, or emotional state regardless of the mode of reporting (e.g., via paper-and-pencil diaries, electronic pagers or palm pilots, telephone data capture systems, etc.). While such monitoring studies have proliferated in the addiction literature over the past 5–10 years, relatively little emphasis has been placed on the feasibility of these methodologies for clinical samples and how the monitoring protocols themselves may influence the behavior of interest (i.e., measurement reactivity).

Most studies utilizing frequent self-monitoring for substance use report that subject compliance is very good across a variety of monitoring methodologies when incentives are offered (paper-and-pencil logs mailed daily: Carney et al., 1998; reports made via cell phone: Collins et al., 2003; calls made to an Interactive Voice Response (IVR) system: Searles et al., 2002, Searles et al., 1995). However, these studies excluded those with moderate or severe alcohol dependence (e.g., Carney et al., 1998, Searles et al., 2002) or alcoholics without stable housing (e.g., Litt et al., 2000). Studies of frequent self-monitoring with treated alcoholics (Filstead, 1988, Ogborne and Annis, 1988) using daily logs that were mailed biweekly or monthly have found extremely poor compliance, though in addition to the long lags in submitting reports, these studies do not appear to have included any payment incentives for monitoring. Thus, it remains unknown whether a more representative sample of people in treatment for alcohol use disorders will comply adequately with a frequent telephone monitoring protocol that includes both monetary incentives and a way to track day-to-day protocol compliance.

A second critical methodological issue is that self-monitoring itself, in any form, may affect the behaviors being monitored (measurement reactivity). Indeed, self-monitoring of craving and drinking has been found to be an active component of some cognitive-behavioral interventions for substance use disorders (Garvin et al., 1990, Miller et al., 1995, Mullen et al., 1997), but not others (Harris and Miller, 1990). Thus, monitoring craving, alcohol use, and co-occurring symptoms could serve to suppress or moderate these symptoms for some, while for others paying attention to their symptoms could increase awareness and bring about a symptom exacerbation.

Only two experimental studies of measurement reactivity in the addiction literature (Ogborne and Annis, 1988, Sobell et al., 1989) randomly assigned participants to self-monitoring conditions or to no-monitoring conditions with follow-up assessments. Neither study found significant group differences in drinking behavior, though participants in both studies had fairly poor compliance with the paper-and-pencil monitoring protocols, possibly confounding the measurement reactivity results. Other, non-experimental studies have yielded mixed results. Using a paper-and-pencil log that was submitted weekly, Litt et al. (1998) found that alcoholics in a self-monitoring study reported that the monitoring helped them drink less but showed no differences in self-reported abstinence rates, alcohol intake, or days drinking when compared with a non-randomized control group of patients who did not self-monitor, but who were involved in treatment outcome assessments. However, self-monitoring compliance was also poor in this study with 70% of those who were debriefed admitting to having recorded the correct time and date but delaying their actual recordings until later. A study of problem-drinking college students using ecological momentary assessment via hand held computer (Hufford et al., 2002) found no change in drinking compared to baseline over the course of the 2-week monitoring period. However, the participants reported significantly reduced motivation to change their drinking behavior at the end of the study. In contrast, a much longer study with a community sample of drinkers involving daily IVR reports (Helzer et al., 2002) found a 19% overall reduction in drinking during the second year of monitoring relative to the first year of monitoring, with 45% of the sample of white males significantly reducing their alcohol intake.

Most research in the addiction field comparing frequent monitoring with retrospective reports of drinking behavior finds good aggregate correspondence between the two methodologies (Sobell and Sobell, 2003), and therefore the more demanding and expensive frequent monitoring methodology has been recommended only for research evaluating the temporal covariation between phenomena of interest (Collins and Graham, 2002, Leigh, 2000). Given our ultimate interest in the temporal covariation between alcohol use, craving, and Posttraumatic Stress Disorder (PTSD) symptoms among patients in early recovery from an alcohol use disorder, we conducted a preliminary study to evaluate whether daily or weekly symptom self-monitoring using an IVR system is feasible, acceptable, and associated with symptom changes (i.e., introduces measurement reactivity). Our interest in patients with comorbid alcohol use disorders and PTSD stems from the findings that the two disorders frequently co-occur (Brady, 2001, Ouimette et al., 2000, Stewart, 1996), and the combination of the two is associated with poorer treatment outcomes (Benda, 2001, Brown et al., 1996, Dickey and Azeni, 1996, Ouimette et al., 2000, RachBeisel et al., 1999, Siegfried, 1998, Steindl et al., 2003), more severe clinical profiles (McFall et al., 1991, Zoricic et al., 2003), as well as higher treatment costs (Brown et al., 1999). There is also a paucity of information about how the two disorders may affect one another in real time (see Stewart, 1996) and a better understanding of the day-to-day relationship between them could lead to improved interventions for those with both disorders. As noted, such temporally sensitive questions are best addressed with closely spaced assessment intervals rather than through infrequent assessments that rely too heavily on participant memory and impressions (Collins and Graham, 2002, McKay, 1999; see also Brown et al., 1998).

Although our primary interest is in the basic feasibility and measurement reactivity associated with frequent self-monitoring among dually diagnosed individuals in substance abuse treatment in early recovery, we also hoped to evaluate the feasibility and acceptability of IVR technology for this population. We selected IVR over the other alternatives for the following reasons: (1) the monitoring materials cannot be lost, sold, or destroyed because they are remote to the participant and must be accessed via a toll-free telephone number; (2) compliance can be easily tracked via daily monitoring of a linked study internet website and problems with compliance can therefore be quickly addressed; (3) the IVR system gives participants daily updates about the amount of money they have earned; (4) previous research with samples that included drinkers who met criteria for alcohol use disorders has shown very good compliance (Searles et al., 1995; Searles et al., 2000).

Subsequently, we identified the following research questions. (1) What proportion of eligible participants will consent to daily IVR monitoring with monetary compliance incentives? (2) What level of monitoring compliance is achieved with and without reconstruction of data by follow-up for missed calls and do monitoring compliance rates differ between daily and weekly conditions? (3) Do participants in the daily and weekly conditions report different subjective experiences related to protocol participation (convenience, willingness to continue, effects on symptom levels)? and (4) Compared to a no-monitoring control condition, are changes in symptom levels at 28 days post-randomization associated with monitoring condition or co-morbidity status at baseline (e.g., PTSD, other drug use)?