In recent years, there have been advances in the development of various wearable transdermal alcohol sensor (TAS) devices. These devices measure alcohol consumption from vapors off the skin via sweat, known as transdermal alcohol concentration. The advantage of these devices is the possibility to wear them all day, allowing for long-term continuous data capture [1] and recording drinking episodes accurately in near real time. These are features that overcome some of the limitations of other methods. Breath and urine analysis is limited by the short half-life of ethanol [2]. Blood markers of heavy alcohol consumption (eg, gamma-glutamyl transferase, liver function tests, mean cell volume, carbohydrate-deficient transferrin) have limitations in terms of sensitivity or specificity [2,3]. Although other alcohol biomarkers such as phosphatidyl ethanol found in blood and ethyl glucuronide and ethyl sulfate found in urine have been found to have high sensitivity and specificity, these tests can only be used to detect if alcohol has been consumed within the last number of days [4-6]. Further, all these tests require administration by a trained health professional. In contrast, TAS devices are noninvasive, objective, easy-to-use, low-maintenance, and allow for the study of behaviors in real-world contexts for potentially weeks or months at a time [7].

The newer TAS devices have the potential to communicate wirelessly through a sim card or with a smartphone [7]. Using this as a method of data collection could reduce the time and resources required for data capture. There is also the possibility for information to be uploaded over a mobile network and delivered to the patient, health professional, or researcher in near real time. Evidence suggests there is a slight time lag between drinking and peak transdermal alcohol concentration [7-10], but preliminary studies with newer generation devices have demonstrated a reduction in this time lag [7,9]. SCRAM is a chunkier device with various types worn on the ankle and is similar in appearance to a house arrest monitor, whereas WrisTAS, BACtrack, ION, and Quantac Tally are worn exclusively on the wrist and are smaller in style, closer in appearance to a FitBit, health watch, or pedometer, and are approximately the same size as a watch (information along with an image of each device is summarized in Multimedia Appendix 1).

The acceptability and feasibility of health care interventions are important issues to consider in their development, evaluation, and implementation [22]. Although previous studies [7,12,23-25] have alluded to the acceptability and feasibility of TASs to objectively monitor alcohol consumption, there are a limited number of studies addressing this, and to our knowledge, there are no systematic reviews specifically investigating this. In this review, we consider acceptability as the device being perceived as appropriate, which is based on both cognitive and emotional responses to the devices and that this acceptability can be assessed before, during, or after wearing the device [22]. We consider feasibility as the extent to which this device could be implemented practically within the identified setting. This systematic review investigates the current knowledge by systematically identifying and evaluating the existing literature on the use of TAS devices in clinical and nonclinical populations, alone or in conjunction with a psychosocial intervention. The objective of this review was to assess the acceptability and feasibility of TAS devices with an overarching objective of establishing the barriers and facilitators to implementing these devices.

This systematic review has been written according to the PRISMA-P (Preferred Reporting Items for Systematic review and Meta-Analysis for Protocols) guidelines [26]. This protocol has been registered on the Prospective Register of Systematic Reviews (PROSPERO CRD42021231027). On review of the results, it was decided that the findings of this systematic review should be reported in 2 papers: this paper focusing on acceptability and feasibility outcomes and another paper on accuracy outcomes [27].

Studies meeting all the following criteria were included: full text, original studies, published in peer-reviewed journals, written in English, using a wearable transdermal sensor device(s) or investigating attitudes and experience of TAS use reporting acceptability or feasibility outcomes. There were no restrictions on publication year or participant clinical characteristics. Data based on conference abstracts, dissertations, and grey literature were not included.

Bibliographic databases included CINAHL, EMBASE, Google Scholar, MEDLINE, PsycINFO, PubMed, and Scopus. Searches were carried out February 2021 (Textbox 1 and Multimedia Appendix 2 and Multimedia Appendix 3). The searches were supplemented by cross-checking the reference lists of key publications, related systematic reviews, and all included papers.

All identified titles and abstracts were screened in Covidence. From this list, the full text was retrieved and assessed for eligibility (EB). Any queries were discussed with a second reviewer (SH). Any disagreement was resolved through discussion with a third reviewer (PD). A data extraction form was created, piloted, and refined as necessary. EB extracted the data independently (Multimedia Appendix 4) and the second reviewer completed entries check for accuracy.

All outcome measures reported in the included studies were extracted, including both objective and self-reported measures. The definition of acceptability outcomes for TASs in the context of this study are factors that affect participant willingness to use the device in treatment or rehabilitation. The definition of feasibility outcomes for TASs in the context of this study are factors that would impact the introduction, including the operational capability of these devices in alcohol treatment services as part of treatment and individual skills required by wearers.

We used the Mixed Methods Appraisal Tool (MMAT), as it was designed for the appraisal in reviews that include a range of designs (qualitative, quantitative, and mixed methods) [28]. For each included study, we determined the study design and then applied the appropriate screening criteria; this provided an overall quality score for the study [28]. EB independently completed the appraisal, and queries were discussed with a second reviewer (SH). Any disagreements were resolved by discussion with a third reviewer (PD).

The data are summarized using a structured narrative description. As there are no standard outcomes for acceptability or feasibility measures for TASs, we include any acceptability and feasibility measures they report. We found that outcomes fell into one of the following categories and thus, these are used as subheadings in the results narrative: comfort, appearance, ease of use, social perceptions, perceptions of alcohol use, barriers/suggestions, the criminal justice system, device tampering, and compliance.

After removing duplicates, 125 papers were screened, 31 papers were excluded at the title and abstract screening, and 94 full-text papers were assessed for eligibility. A total of 76 papers were then excluded. There were 8 additional papers identified by citation searching; of these, 5 were included. The final sample included 22 publications (Figure 1).

Sixteen studies used a version of SCRAM [8,12,14,23-25,29-38], 4 used a version of WrisTAS [39-42], 3 used BACtrack Skyn [7,38,43], and 1 used Quantac Tally [7]. Some studies used more than 1 version or brand of device. Almost half of the included studies (10/22, 45%) aimed to assess how TASs can be used to measure alcohol consumption. Out of the 22 studies, 8 (36%) aimed to assess acceptability, adherence, and feasibility with TASs, and 5 (23%) used TASs to explore their effectiveness in implementing CM for alcohol reduction treatment and evaluating the efficacy of CM for alcohol use reduction. One investigated nonalcoholic energy drinks and one investigated alcohol-related positive mode enhancement and negative mood reduction.

Most studies (18/22, 82%) included participants who were adults in good health. Only 4 out of 22 (18%) studies included participants who were diagnosed with alcohol dependence. These 4 studies recruited participants from community-based clinics receiving alcohol treatment [8,23,24] or were recruited on admission to a general hospital substance abuse unit [42]. Most were conducted in the United States (17/22, 77%). The earliest paper included is from 1992 but the majority (17/22, 77%) of the studies were published since 2015.

There were 821 participants enrolled in total across the studies, with 793 included in the procedure or analysis. Therefore, 28 participants that were enrolled were not included in the results (28/821, 3.4%) (for reasons such as withdrawing or missing data). One paper [7] was still in the early stages of data collection for one of their studies and so these participant numbers were unknown. Not all studies included detailed information on the participants’ age, gender, and ethnicity. Where information was provided, it could be seen that participants’ ages ranged from 18 to 57 years, the majority included both females and males, and for most, Caucasian participants represented a large majority of the sample (see Table 1).

We found that no study defined acceptability and feasibility in terms of their research. One described how they measured protocol feasibility [40] and one described how they measured acceptability and feasibility of the device [43]. Luczak et al [40] report protocol feasibility as the reliability of each component of the protocol, the validity of TAS data, participant compliance, and reactivity. Rosenberg et al [43] report measuring acceptability using the Acceptability of Intervention Measure scale and feasibility with the amount of alcohol monitor data produced and the correlation between device-reported drinking events and drinking events reported by participants. They also used the Feasibility of Intervention Measure scale.

All studies, except for 2 studies [7,24], met a minimum of 4 out 5 MMAT criteria (>80%) (all scores reported in Table 1 and Multimedia Appendix 5). This is due to Alessi et al [24] not providing details about randomization and participant information. Alessi et al [24] included alcohol treatment outpatients who were randomized to usual care in 2 previous studies; these 2 previous studies are not clearly stated. Wang et al’s study [7] was difficult to score due to incomplete data collection, as their study was still ongoing at the time of publication. With the MMAT, exclusion of low methodological quality studies is discouraged [28]. Due to the nature of many studies, blinding of participants or staff was not possible; in some, there were clear differences between groups [34] or CM incentives were provided [25], where the staff were required to know participant allocation. In many studies, there was only 1 group of participants who all completed the same task and so randomization was not required [7,8,12,23,29-33,35-43]. Not all studies provided clear information on participant information, randomization, incomplete outcome data, and selective reporting; therefore, there was potential bias due to limited information [7,12,14,23,24,29,34,35,38,40-43].