The Brief Addiction Monitor – Continuous (BAM) was developed within the United States Department of Veterans Affairs (VA) by Cacciola and colleagues (2013) to routinely track a client’s status across the course of treatment. Most items ask the client to reflect on the past 30 days, covering recent alcohol and drug use alongside the everyday circumstances that raise or lower the likelihood of continued use, such as cravings, mood, sleep, risky situations, self-help attendance, and social support; one item asks about the coming 30 days. The version implemented on NovoPsych is the continuous form (sometimes referred to as the BAM-R), in which frequency items record the actual number of days over the past 30, and the remaining items are scaled to the same 0-30 range.

BAM Domains

The BAM is scored as three domains that together span substance use and the risk and protective circumstances surrounding recovery:

• Substance Use – the frequency of any recent alcohol (including heavy use) and drug use.
• Risk Factors – circumstances and states that increase the likelihood of continued or resumed use, including low mood, poor sleep, cravings, risky situations, family/social problems, and poor physical health.
• Protective Factors – supports and behaviours that sustain recovery, including confidence in staying abstinent, self-help attendance, religion/spirituality, meaningful activity (e.g., work, school, volunteer), adequate income, and supportive relationships.

Clinical Applications of the BAM

A progress-monitoring tool, not a screener. The BAM is designed for people who are already in, or entering, treatment for substance use. It is not a screening tool for identifying substance use problems, and is not intended to establish a diagnosis. Some item wording refers to “recovery,” which may not sit well with clients who are pre-contemplative or not yet seeking treatment (U.S. Department of Veterans Affairs, n.d.-a). It can be completed by the client as a self-report questionnaire or administered by a clinician in person or by telephone (U.S. Department of Veterans Affairs, n.d.-a).

Settings. The BAM is designed for repeated administration during substance use treatment, and is used across outpatient, residential, and community settings. Published use spans VA specialty and residential programmes (Blanchard et al., 2023; Dams et al., 2023), a non-veteran outpatient service (Schumm et al., 2022), a US telehealth service for opioid use disorder (Burke et al., 2024), and a community health centre (Phillips et al., 2026).

Considerations outside veteran settings. The BAM was developed within the US Department of Veterans Affairs (Cacciola et al., 2013), and much of its psychometric evidence comes from veteran samples that are primarily male. However, it has also been evaluated in a large non-veteran outpatient SUD sample (Schumm et al., 2022) and used to track outcomes in other non-VA treatment settings (Burke et al., 2024; Phillips et al., 2026).

Reviewing results with the client. The three domains provide a simple, shared picture of where a client currently sits and what is changing, which supports collaborative goal-setting and helps direct attention to the areas most in need of work (U.S. Department of Veterans Affairs, n.d.-a). The developers intend for the BAM to be used by feeding results back (particularly as graphs of change), and group members in focus groups have reported finding the graphs helpful for examining change over time (Drapkin et al., 2010, as cited in DePhilippis et al., 2014).

Identifying treatment targets. Because each item reflects a concrete, clinically meaningful behaviour or state, individual item responses can also be reviewed to identify specific targets for intervention as well as strengths and resources that can support recovery. The VA consistently advises attending to the items more closely than to the domain scores, because the items have stronger psychometric support and richer implications for treatment planning (U.S. Department of Veterans Affairs, n.d.-a.; DePhilippis et al., 2014). Treatment typically focuses on minimising Risk and maximising Protective Factors, with the goal of reducing Substance Use. Two clients can arrive at the same Risk Factors score by entirely different routes: one through poor physical health, sleep, and mood, another through craving, risky situations, and interpersonal conflict, and the domain score alone does not distinguish them.

Tracking change over time. Administering the BAM at intake and at regular intervals lets the clinician and client see how substance use, risk, and protective factors shift over the course of care and aid in treatment planning. Specific BAM items can also be linked to treatment objectives, allowing subsequent administrations to provide measurable evidence of progress toward those goals. This is the BAM’s primary purpose (U.S. Department of Veterans Affairs, n.d.-a) and where its evidence is strongest relative to other applications (Blanchard et al., 2023; Dams et al., 2023; Burke et al., 2024; Phillips et al., 2026).

BAM Scoring

Item-level scoring. Each BAM item is scored from 0 to 30. Items asking for a number of days or nights over the past month (2, 3, 4, 5, 6, 7a to 7g, 10, 11, 13, and 16) are scored as the actual number reported. Items with five rated response options (1, 8, 9, 12, 15, and 17) are scored 0, 8, 15, 22, or 30. Item 14, which asks whether the respondent has adequate legal income, is scored No = 0 and Yes = 30.

Contingent items. Two conditional blocks are displayed only if the respondent reported any use on the preceding item, covering eight items in total. If the respondent reports no alcohol use in the past 30 days (i.e., item 4 = 0 days), the heavy-drinking item (item 5) is skipped; if they report no drug use (i.e., item 6 = 0 days), the seven specific-drug items (7a to 7g) are skipped. Skipped items are scored zero, so the domain totals remain on the same scale for every client. Note that a skipped item still appears in the Client Responses table as “Not Completed”: item 5 whenever item 4 is zero, and items 7a to 7g whenever item 6 is zero. A “Not Completed” entry against one of these items means the question was not asked because the preceding answer made it redundant, not that data are missing, and the Substance Use total is complete.

BAM Results Table

Domains. The BAM is scored as three domain totals by summing the relevant items. Each item on the continuous form contributes a value from 0 to 30. The specific-drug items (questions 7a to 7g) and the recovery-satisfaction item (question 17) are recorded for clinical review but are not included in the three domain totals.

• Substance Use (3 items: 4, 5, 6; total ranging from 0 to 90), capturing the frequency of recent alcohol and drug use (days of drinking, days of heavy drinking, and days of illicit drug use or misuse of prescription medication). Note: Item 6 in the NovoPsych version of the BAM reads “…or misuse any prescription medications?” whereas the original source reads “abuse”. The change follows current terminology guidance. The National Institute on Drug Abuse lists “abuse” among the terms to avoid, giving “misuse” or “used other than prescribed” as the preferred alternatives for prescription medications, on the basis that “abuse” carries a high association with negative judgement and punishment (National Institute on Drug Abuse, n.d.). The VA’s own content style guidance advises the same (U.S. Department of Veterans Affairs, n.d.-b).
• Risk Factors (6 items: 1, 2, 3, 8, 11, and 15; total ranging from 0 to 180), capturing states and circumstances that raise the likelihood of continued use (physical health, sleep problems, mood problems, cravings, risky situations, and conflict with family or friends).
• Protective Factors (6 items: 9, 10, 12, 13, 14, and 16; total ranging from 0 to 180), capturing supports and behaviours that sustain recovery (confidence in remaining abstinent, self-help attendance, spirituality, work or study, adequate income, and contact with supportive people).

BAM Domain Scores Graph

BAM Domain Score Interpretation

The BAM is interpreted using raw domain scores and interpretation is directional. On the Substance Use and Risk Factor domains, higher scores indicate more substance use and more risk-related circumstances. On the Protective Factors domain, higher scores indicate more recovery support. A single administration gives a snapshot of where the client currently sits across the three domains; the greatest clinical value comes from comparing scores across administrations to see what is improving and what is not.

Because each BAM item reflects a distinct and concrete behaviour or state, reviewing individual item responses is often as informative as the domain totals. Items carry richer implications for treatment planning than the composite scores. A high number of days of drug use, a rise in risky situations, or a fall in self-help attendance each points to a specific and actionable focus for the next stage of treatment. Two clients with the same Risk score may present entirely differently: one with physical health, sleep and mood difficulties, the other with craving, risky situations and interpersonal conflict; so the total domain score alone cannot guide the plan.

The items the report surfaces in its interpretive text are selected and ordered by score on the shared 0 to 30 metric, not by clinical importance, so that list is a starting point for review rather than a priority order. The full set of responses is set out at the end of the report, grouped by domain and numbered as they appear on the questionnaire (1 to 17, with 7a to 7g), so each answer can be reviewed directly alongside the domain totals.

There are no norms, cut-scores, or percentile ranks for the BAM domain scores, and the measure does not produce a total score. Results are compared against the same client’s earlier administrations rather than against a reference population. There is no threshold at which a domain score becomes “high” or “clinically significant.”

BAM Alcohol and Other Drug Use Graph Over Time

BAM Change Score Interpretation

The BAM is designed for repeated administration as part of measurement-based care (MBC), and change across administrations is the primary interpretive output.

On a repeat administration the report describes change against two reference points: the most recent prior administration, and the client’s initial administration. Reporting both separates a recent shift from the overall trajectory, which can move in opposite directions. On a second administration the two reference points are the same, so only one comparison is given.

A reduction in the Substance Use or Risk Factors domain, or an increase in the Protective Factors domain, reflects movement in the favourable direction. Movement on Protective Factors should be read with particular caution: the internal consistency of that domain is very low (Cronbach’s alpha .32 and .22; Hallinan et al., 2021), so a change in the total should always be checked against which of its six items actually moved. Change is reported as a raw point difference. No threshold defines how large a change must be before it counts as meaningful. Change is therefore read directionally and in the context of the client’s own trajectory.

The evidence base indicates that BAM scores move meaningfully during treatment, with the heavy alcohol use, drug use, craving, mood, and self-help items showing the greatest responsiveness (Blanchard et al., 2023).

BAM Recovery Engagement and Confidence Graph Over Time

BAM Graphs

Single administration. The report presents three horizontal bar plots, one per domain, each on its own raw-score axis: Substance Use 0 to 90, Risk Factors and Protective Factors 0 to 180.

Repeat administrations, item level. Because attending to the items more closely than to the domain scores is important, the report plots individual items over time and presents these first, grouped so that items which are read together are seen together:

• Substance Use: days of drinking, heavy drinking and drug use (items 4, 5 and 6), plus a separate plot of the specific substances (7a to 7g). Both plots count days, and the section carries a note saying so, since the shared 0 to 30 axis means something different elsewhere in the report. The second plot draws a line only for a substance the client has reported using at some point in the window, so a drug that has reappeared is shown with its preceding run of zeros intact and a drug never used is left off entirely.
• Risk Factors: physical health, sleep, mood and interpersonal conflict (items 1, 2, 3 and 15); and craving and exposure to risky situations (items 8 and 11).
• Protective Factors: confidence, self-help attendance and spiritual support (items 9, 10 and 12); and work or study, income and supportive contact (items 13, 14 and 16).
• Recovery satisfaction: item 17 on its own, carrying a note that this score forms no part of any domain.

BAM Life Stability and Social Support Graph Over Time

Repeat administrations, domain level. The report presents a trajectory plot of each domain across administrations, so the movement of Substance Use, Risk Factors and Protective Factors over the course of treatment can be seen together. The three are plotted separately rather than combined, because Substance Use runs to 90 while the other two run to 180, and because Protective Factors reads in the opposite direction.

BAM Substance Use Domain Score Graph Over Time

Background gradient. All plots carry a background gradient whose colour deepens as more of the construct is present: red as Substance Use and Risk Factors scores rise, green as Protective Factors scores rise. On the trajectory plots the gradient is a wash behind the line; on the single-administration plots it is a strip along the base of each chart. Only the upper portion of each scale is shaded, and the shading fades out gradually rather than beginning at a fixed point. The lower end is left unshaded deliberately: a low Substance Use or Risk Factors score reflects the absence of a reported problem rather than a positive finding. The shading is a visual guide only. The BAM has no validated cut scores, norms, or severity bands, so no score should be read as having crossed a threshold on the basis of where it falls within the gradient.

BAM Development

The BAM was developed by Cacciola and colleagues (2013) as a brief instrument for routine progress monitoring in substance use treatment. Items were assembled in three steps: a review of existing assessment instruments and outcome monitoring systems, consultation with VA clinicians and a national scientific advisory group, and pilot testing of a 25-item draft with 25 outpatients. Items were retained on the basis of that consultation, their empirical links to substance use and treatment outcomes, and the goals of the monitoring system (Cacciola et al., 2013). The original development study was conducted with 150 patients entering outpatient substance use treatment at a United States Department of Veterans Affairs medical centre, and the instrument has since been examined in large national veteran samples and non-veteran treatment settings.

Since its development, the BAM has been administered in several versions that differ in response format and recall period. The BAM version implemented in NovoPsych is the continuous form (sometimes referred to as the BAM-R), in which frequency items record the actual number of days over the past 30. In clinical use, the instrument is scored as three domains, Substance Use, Risk Factors, and Protective Factors, following the scoring convention adopted by the United States Department of Veterans Affairs and used in subsequent applied studies (Burke et al., 2024; Phillips et al., 2026).

BAM Reliability

Internal consistency has been reported for several different groupings of BAM items, on several different versions of the instrument, and the figures are not interchangeable.

For the three domains NovoPsych scores, on the version NovoPsych implements, only one study was identified. In a national veteran sample completing the continuous form, Cronbach’s alpha was .73 at both administrations for Risk Factors, .66 and .68 for Substance Use, and .32 and .22 for Protective Factors (Hallinan et al., 2021, N = 22,453). Risk Factors is adequate, Substance Use is questionable, and Protective Factors is too low to be read as a scale. The Protective Factors domain is therefore best reviewed at the level of its individual items rather than as a domain total.

Test-retest reliability was examined in the development study, on the continuous form, but the design does not isolate temporal stability. Two administrations were conducted at the three-month follow-up, one by the treating clinician and one by a research technician, planned to fall within one week of each other, in counterbalanced order (Cacciola et al., 2013, n = 74). Agreement was acceptable for 13 of the 15 items and excellent for all three of that study’s factors, with poor agreement for days of drug use (ICC = .24) and days in risky situations (ICC = .25). Because the two administrations differed in who asked the questions, the result confounds retest interval with rater and setting. Participants reported more self-help attendance, less craving, and more recovery satisfaction to clinicians than to researchers.

BAM Factor Structure

The BAM produces three scores: Use, Risk, and Protective. These groupings come from the Department of Veterans Affairs’ scoring guidelines and reflect clinical reasoning about which items belong together rather than a statistical analysis of how the items behave. The guidelines say so themselves, describing the subscale scoring as “supplementary and very preliminary, based on clinical judgment rather than empirical data”, and directing clinicians to the individual items as “the most clinically relevant use of this measure” (U.S. Department of Veterans Affairs, 2009). The item groupings are the same on every version of the BAM; only the item metric differs (U.S. Department of Veterans Affairs, n.d.-a).

Factor-analytic work has been mixed, and the studies do not all use the same version. The original development study, on the continuous form, proposed a different three-factor structure comprising Recovery Protection, Physical and Psychological Problems, and Substance Use and Risk, formed from 11 of the 15 items analysed (Cacciola et al., 2013). Larger veteran samples have supported an alternative four-factor structure of Alcohol Use, Stressors, Risk and Stability, with better model fit (Gaddy et al., 2018, N = 4,955, categorical form; Hallinan et al., 2021, N = 22,453, continuous form). In a non-veteran outpatient sample, which also used the categorical form, neither structure fit well (Schumm et al., 2022, N = 2,227).

Given this, NovoPsych reports the three operational domains as summed monitoring scores but places the emphasis on item-level review. This is not simply a NovoPsych preference. Gaddy and colleagues (2018) concluded that it is not advised to rely heavily on the BAM subscale scores for treatment planning, and that tracking change on individual items may be of greater clinical utility, and the VA’s own guidance has taken the same position since the scoring guidelines were first drafted, describing item-level examination as “the most clinically relevant use of this measure” (U.S. Department of Veterans Affairs, 2009) and, in the current instructions, directing clinicians to attend to the items “to a greater degree than the composite scores” (U.S. Department of Veterans Affairs, n.d.-a).

BAM Validity

Content for the BAM was drawn from behaviours and states with established links to substance use outcomes (Cacciola et al., 2013). In the development study, using the continuous form, two of the exploratory factors predicted dropout from the first stage of outpatient treatment (n = 121). Fewer protective factors at intake and higher substance use and risk at intake each raised the odds of dropping out by about 5% per point, with the model accounting for 16.9% of the variance (Cacciola et al., 2013). Those odds ratios are expressed per point of a standardised T-score on that study’s exploratory factors, not per raw point on a NovoPsych domain, and cannot be applied to the 0 to 180 Protective Factors or 0 to 90 Substance Use totals. Note that the Recovery Protection factor was reverse-scored in that analysis, so its direction is opposite to the NovoPsych Protective Factors domain, where higher scores indicate more recovery support.

However, evidence for prediction is weak and does not replicate. In a national veteran sample, no intake item, no combination of items, and none of the three domain scores predicted 90-day treatment retention or 12-month mortality, with all areas under the curve below 0.70 (Blanchard et al., 2023, N = 32,002). Based on the evidence available, the BAM’s value lies in describing current status and detecting change rather than in forecasting outcomes at intake.

BAM Sensitivity to Change

The strongest and most consistent evidence for the BAM concerns its responsiveness to change during treatment. In a national veteran sample completing the continuous form, item scores changed from intake to follow-up with effect sizes ranging from small to large. The largest changes were on heavy alcohol use, self-help attendance, drug use, craving, and mood, and the Substance Use domain showed a large effect (Cohen’s d = 0.95; Blanchard et al., 2023, n = 7,523 with repeat administrations). It is the only study identified that reports effect sizes for the three domains as NovoPsych scores them.

In a veteran residential treatment sample, also using the continuous form, admission-to-discharge change produced robust effect sizes from 0.76 to 1.60 (Dams et al., 2023, n = 2,668). These were calculated on the three exploratory factors from the original development study rather than on the Use, Risk and Protective domains, so they show that the instrument detects change but say nothing specific about the domain scores NovoPsych reports.

Improvement in the same direction has been reported in non-veteran settings, including a national telehealth programme for opioid use disorder (Burke et al., 2024, n = 1,447) and a community health centre programme (Phillips et al., 2026, n = 982 with repeat administrations), where substance use and risk scores fell, and protective scores rose. Both of these studies used the categorical form rather than the continuous form; neither reported effect sizes, and Phillips and colleagues ran no inferential tests, so the direction of change is supported but its magnitude on the continuous form is not. The protective gain in the telehealth sample was small, was not statistically significant in the youngest or the oldest age bands, and was judged by the authors to be of doubtful clinical relevance.

Assess your patients with the Brief Addiction Monitor – Continuous (BAM) using a NovoPsych account

• For Psychologists & Mental Health Clinicians
• Send Assessments to Patient’s Phone
• Receive Comprehensive Reports
• Access Over 100 Validated Assessments​
• Instant Psychometric Scoring
• Track Symptoms​
• Inform Treatment​