Socially situated artificial intelligence enables learning from human interaction.

Regardless of how much data artificial intelligence agents have available, agents will inevitably encounter previously unseen situations in real-world deployments. Reacting to novel situations by acquiring new information from other people-socially situated learning-is a core faculty of human development. Unfortunately, socially situated learning remains an open challenge for artificial intelligence agents because they must learn how to interact with people to seek out the information that they lack. In this article, we formalize the task of socially situated artificial intelligence-agents that seek out new information through social interactions with people-as a reinforcement learning problem where the agent learns to identify meaningful and informative questions via rewards observed through social interaction. We manifest our framework as an interactive agent that learns how to ask natural language questions about photos as it broadens its visual intelligence on a large photo-sharing social network. Unlike active-learning methods, which implicitly assume that humans are oracles willing to answer any question, our agent adapts its behavior based on observed norms of which questions people are or are not interested to answer. Through an 8-mo deployment where our agent interacted with 236,000 social media users, our agent improved its performance at recognizing new visual information by 112%. A controlled field experiment confirmed that our agent outperformed an active-learning baseline by 25.6%. This work advances opportunities for continuously improving artificial intelligence (AI) agents that better respect norms in open social environments.

Ethics and society review: Ethics reflection as a precondition to research funding.

Researchers in areas as diverse as computer science and political science must increasingly navigate the possible risks of their research to society. However, the history of medical experiments on vulnerable individuals influenced many research ethics reviews to focus exclusively on risks to human subjects rather than risks to human society. We describe an Ethics and Society Review board (ESR), which fills this moral gap by facilitating ethical and societal reflection as a requirement to access grant funding: Researchers cannot receive grant funding from participating programs until the researchers complete the ESR process for their proposal. Researchers author an initial statement describing their proposed research's risks to society, subgroups within society, and globally and commit to mitigation strategies for these risks. An interdisciplinary faculty panel iterates with the researchers to refine these risks and mitigation strategies. We describe a mixed-method evaluation of the ESR over 1 y, in partnership with a large artificial intelligence grant program at our university. Surveys and interviews of researchers who interacted with the ESR found 100% (95% CI: 87 to 100%) were willing to continue submitting future projects to the ESR, and 58% (95% CI: 37 to 77%) felt that it had influenced the design of their research project. The ESR panel most commonly identified issues of harms to minority groups, inclusion of diverse stakeholders in the research plan, dual use, and representation in datasets. These principles, paired with possible mitigation strategies, offer scaffolding for future research designs.

Metaplastic carcinoma of the breast: matched cohort analysis of recurrence and survival.

PURPOSE: Metaplastic breast carcinoma (MBC) is a rare subtype of breast cancer, defined as mammary carcinoma with squamous or mesenchymal differentiation, that may include spindle cell, chondroid, osseous, or rhabdomyoid differentiation patterns. The implications of MBC recurrence and survival outcomes remains unclear. METHODS: Cases were ascertained from a prospectively maintained institutional database of patients treated from 1998 to 2015. Patients with MBC were matched 1:1 to non-MBC cases. Cox proportional-hazards models and Kaplan-Meier estimates were used to evaluate outcome differences between cohorts. RESULTS: 111 patients with MBC were matched 1:1 with non-MBC patients from an initial set of 2400 patients. Median follow-up time was 8 years. Most patients with MBC received chemotherapy (88%) and radiotherapy (71%). On univariate competing risk regression, MBC was not associated with locoregional recurrence (HR = 1.08; p = 0.8), distant recurrence (HR = 1.65; p = 0.092); disease-free survival (HR = 1.52; p = 0.065), or overall survival (HR = 1.56; p = 0.1). Absolute differences were noted in 8-year disease-free survival (49.6% MBC vs 66.4% non-MBC) and overall survival (61.3% MBC vs 74.4% non-MBC), though neither of these reached statistical significance (p = 0.07 and 0.11, respectively). CONCLUSION: Appropriately-treated MBC may exhibit recurrence and survival outcomes that are difficult to distinguish from those of non-MBC. While prior studies suggest that MBC has a worse natural history than non-MBC triple-negative breast cancer, prudent use of chemotherapy and radiotherapy may narrow these differences, although studies with more power will be required to inform clinical management. Longer follow-up among larger populations may further elucidate the clinical and therapeutic implications of MBC.

Can incorrect artificial intelligence (AI) results impact radiologists, and if so, what can we do about it? A multi-reader pilot study of lung cancer detection with chest radiography.

OBJECTIVE: To examine whether incorrect AI results impact radiologist performance, and if so, whether human factors can be optimized to reduce error. METHODS: Multi-reader design, 6 radiologists interpreted 90 identical chest radiographs (follow-up CT needed: yes/no) on four occasions (09/20-01/22). No AI result was provided for session 1. Sham AI results were provided for sessions 2-4, and AI for 12 cases were manipulated to be incorrect (8 false positives (FP), 4 false negatives (FN)) (0.87 ROC-AUC). In the Delete AI (No Box) condition, radiologists were told AI results would not be saved for the evaluation. In Keep AI (No Box) and Keep AI (Box), radiologists were told results would be saved. In Keep AI (Box), the ostensible AI program visually outlined the region of suspicion. AI results were constant between conditions. RESULTS: Relative to the No AI condition (FN = 2.7%, FP = 51.4%), FN and FPs were higher in the Keep AI (No Box) (FN = 33.0%, FP = 86.0%), Delete AI (No Box) (FN = 26.7%, FP = 80.5%), and Keep AI (Box) (FN = to 20.7%, FP = 80.5%) conditions (all ps < 0.05). FNs were higher in the Keep AI (No Box) condition (33.0%) than in the Keep AI (Box) condition (20.7%) (p = 0.04). FPs were higher in the Keep AI (No Box) (86.0%) condition than in the Delete AI (No Box) condition (80.5%) (p = 0.03). CONCLUSION: Incorrect AI causes radiologists to make incorrect follow-up decisions when they were correct without AI. This effect is mitigated when radiologists believe AI will be deleted from the patient's file or a box is provided around the region of interest. CLINICAL RELEVANCE STATEMENT: When AI is wrong, radiologists make more errors than they would have without AI. Based on human factors psychology, our manuscript provides evidence for two AI implementation strategies that reduce the deleterious effects of incorrect AI. KEY POINTS: • When AI provided incorrect results, false negative and false positive rates among the radiologists increased. • False positives decreased when AI results were deleted, versus kept, in the patient's record. • False negatives and false positives decreased when AI visually outlined the region of suspicion.

A Pilot Study of Accelerometer-Based Biometric Data Collection among Patients Undergoing Locoregional Therapies.

This study evaluated the feasibility of measuring patient recovery after locoregional therapies (LRTs) using a wearable activity tracker (WAT). Twenty adult patients with cancer were provided with a WAT device to wear for a minimum of 7 days prior to their procedure (baseline) and for up to 30 days after their procedure (recovery). Daily step counts were continuously recorded. Patient responses to the Short Form 36-Item Health Survey (SF-36) were also collected before and after LRT. Analysis of WAT data demonstrated a mean of 4,850 daily steps taken at baseline, which decreased to 2,000 immediately after LRT and then rapidly increased to approximately 4,300 daily steps over an average of 10 days (P < .001). No significant changes were observed in SF-36 responses between baseline and follow-up assessments (P > .10). These results suggest that WAT devices capture dynamic periprocedural data not reflected in survey-based assessments and may be used to monitor patient recovery after interventional oncologic procedures.

Black individuals' suspicion of Whites' motives negatively predicts expected efficacy with White social partners.

OBJECTIVES: Interracial efficacy is critical to intergroup interactions. However, the antecedents of interracial efficacy are unclear and rarely studied from the perspectives of Black individuals. The current work addresses these empirical gaps by testing whether individual differences in suspicion of Whites' motives negatively relate to expected efficacy in interracial interactions. Suspicion was operationalized as beliefs that Whites' positivity toward people of color (POC) is primarily motivated by concerns with appearing prejudiced. METHOD: Four studies employing correlational and experimental vignette designs with Black adult participants (Ntotal = 2,295; ∼60% female) tested suspicion's hypothesized negative relation with three conceptions of interracial efficacy (general efficacy, liking-based efficacy, and respect-based efficacy). RESULTS: Four studies provided consistent evidence that suspicion of Whites' motives negatively related to expected efficacy in interactions with White social partners. Moreover, this relationship was unique to contexts with White partners and did not extend to imagined interactions with Black partners or members of other outgroup members (e.g., Latine partners). CONCLUSIONS: Results further suggest suspicion increases expected threat (i.e., anticipated uncertainty/anxiety), which undermines Black individuals' confidence in interactions with White partners. (PsycInfo Database Record (c) 2023 APA, all rights reserved).

Digital Breast Tomosynthesis and Digital Mammography Recall and False-Positive Rates by Time of Day and Reader Experience.

Background Digital breast tomosynthesis (DBT) image interpretation might be more cognitively demanding than interpretation of digital mammography (DM) images. The time of day of interpretation might affect recall and false-positive (FP) rates, especially for DBT. Purpose To determine whether recall and FP rates vary by time of day of interpretation for screening mammography for breast cancer performed with DM and DBT. Materials and Methods This is a retrospective study examining 97 671 screening mammograms interpreted by 18 radiologists between January 2018 and December 2019 at one of 12 community radiology sites. The association between the time of day of interpretation, the type of image interpreted (DM vs DBT), and radiologist experience (≤5 posttraining years vs >5 posttraining years) and the likelihood of a patient being recalled from screening mammography and the likelihood of whether the interpretation was FP or true positive were analyzed. Analyses were conducted using generalized linear mixed modeling with a binary distribution and sandwich estimation where observations were nested by radiologist. Results Screening mammograms interpreted by 18 radiologists were reviewed (40 220 DBTs, 57 451 DMs). Nine radiologists had 5 or fewer posttraining years of experience, and nine had more than 5 posttraining years of experience. The overall recall rates for DM (10.2%) and DBT (9.0%) were different (P = .006); FP rate also differed (9.8% DM, 8.6% DBT; P = .004). For radiologists with 5 or fewer posttraining years of experience, odds of recall increased 11.5% (odds ratio [OR] = 1.12, P = .01) with every hour when using DBT, but this was not found for DM (OR = 1.09, P = .06); DBT and DM were different (OR = 1.12 vs 1.09, P = .02). For radiologists with more than 5 posttraining years of experience, no evidence of increase in recall was observed for DBT (OR = 1.02, P = .27) or DM (OR = 1.0, P = .80), and there was no evidence that these were different (OR = 1.02 vs 1.0, P = .13). Conclusion Patients were more likely to be recalled when their screening digital breast tomosynthesis images were interpreted later in the day by less-experienced radiologists. © RSNA, 2022 Online supplemental material is available for this article.

How do US orthopaedic surgeons view placebo-controlled surgical trials? A pilot online survey study.

Randomised placebo-controlled trials (RPCTs) are the gold standard for evaluating novel treatments. However, this design is rarely used in the context of orthopaedic interventions where participants are assigned to a real or placebo surgery. The present study examines attitudes towards RPCTs for orthopaedic surgery among 687 orthopaedic surgeons across the USA. When presented with a vignette describing an RPCT for orthopaedic surgery, 52.3% of participants viewed it as 'completely' or 'mostly' unethical. Participants were also asked to rank-order the value of five different types of evidence supporting the efficacy of a surgery, ranging from RPCT to an anecdotal report. Responses regarding RPCTs were polarised with 26.4% viewing it as the least valuable (even less valuable than an anecdote) and 35.7 .% viewing it as the most valuable. Where equipoise exists, if we want to subject orthopaedic surgeries to the highest standard of evidence (RPCTs) before they are implemented in clinical practice, it will be necessary to educate physicians on the value and ethics of placebo surgery control conditions. Otherwise, invasive procedures may be performed without any benefits beyond possible placebo effects.

Cheek Donor Site for Full-Thickness Skin Graft Repair of the Nasal Ala: Outcomes of a Retrospective Cohort Study.

BACKGROUND: Full-thickness skin grafts (FTSGs) are useful repairs for reconstructing nasal alar defects. Traditional donor sites include the preauricular, postauricular, and supraclavicular skin. OBJECTIVE: To evaluate esthetic outcomes and complications of nasal alar defects repaired with FTSGs from the medial cheek. MATERIALS AND METHODS: A retrospective chart review of Mohs surgery patients who had FTSG repair of the nasal ala between January 2015 and August 2020 was performed. Demographic, surgery, and follow-up visit data were reviewed. Cosmesis was rated by a facial plastic surgeon, a Mohs surgeon, and a plastic surgeon using baseline, defect, and follow-up visit photographs. RESULTS: Sixty-nine patients with FTSG repairs of nasal alar defects were identified. 51 of 69 patients (73.9%) had the cheek donor site, and 18 of 69 patients (26.1%) had a noncheek donor site. The mean (SD) rater visual analog score for both cohorts was good with no significant difference (cheek: 65.9 [13.8]; noncheek: 66.1 [15.3]; p = .96). A notable difference in the complication rate by donor site was observed (cheek: 6.9%, noncheek: 16.7%; p = .13), although it did not reach significance. CONCLUSION: The cheek is a reliable FTSG donor site for nasal alar defects after Mohs micrographic surgery, with a trend toward fewer complications.

Facing social exclusion: a facial EMG examination of the reaffiliative function of smiling.

Social exclusion influences how expressions are perceived and the tendency of the perceiver to mimic them. However, less is known about social exclusion's effect on one's own facial expressions. The aim of the present study was to identify the effects of social exclusion on Duchenne smiling behaviour, defined as activity of both zygomaticus major and the orbicularis oculi muscles. Utilising a within-subject's design, participants took part in the Cyberball Task in which they were both included and excluded while facial electromyography was measured. We found that during the active experience of social exclusion, participants showed greater orbicularis oculi activation when compared to the social inclusion condition. Further, we found that across both conditions, participants showed greater zygomaticus major muscle activation the longer they engaged in the Cyberball Task. Order of condition also mattered, with those who experienced social exclusion before social inclusion showing the greatest overall muscle activation. These results are consistent with an affiliative function of smiling, particularly as social exclusion engaged activation of muscles associated with a Duchenne smile.

Motives matter: White instructors' external race-based motives undermine trust and belonging for Black college students.

OBJECTIVE: Trust is fundamental to successful educational relationships. Yet, numerous barriers inhibit the development of trust between students of color (SOC) and White instructors. The current research examined a metacognitive obstacle to the development of cross-race classroom trust: Primarily External Race Motives (PERM). PERM was defined as the experience that instructors were more concerned with avoiding the appearance of prejudice than having self-directed egalitarian motives. METHOD: Using within-subjects vignettes (n = 313; 74.8% female), between-subjects cross-sectional designs (n = 386; 70.5% female), and longitudinal methods (n = 135; 45.2% female), the current work tested the primary hypothesis that PERM would undermine instructor trust and classroom belonging. Hypotheses were tested with Black adults (Study 1) and college students (Studies 2 and 3). RESULTS AND CONCLUSIONS: Whether with hypothetical, past, or present White educators, feeling that instructors have primarily external race-based motives undermined instructor trust and classroom belonging. In all studies, the relationship between PERM and classroom belonging was mediated by instructor (mis)trust. The results provide evidence that motives viewed to be primarily external undermine instructional relationships for SOC. (PsycInfo Database Record (c) 2022 APA, all rights reserved).

The Societal Readiness Thinking Tool: A Practical Resource for Maturing the Societal Readiness of Research Projects.

In this paper, we introduce the Societal Readiness (SR) Thinking Tool to aid researchers and innovators in developing research projects with greater responsiveness to societal values, needs, and expectations. The need for societally-focused approaches to research and innovation-complementary to Technology Readiness (TR) frameworks-is presented. Insights from responsible research and innovation (RRI) concepts and practice, organized across critical stages of project-life cycles are discussed with reference to the development of the SR Thinking Tool. The tool is designed to complement not only shortfalls in TR approaches, but also improve upon other efforts to integrate RRI, sustainability, and design thinking in research and innovation cycles. Operationalization and early-stage user tests of the Tool are reported, along with discussion of potential future iterations and applications.

Healthcare disparities among anticoagulation therapies for severe COVID-19 patients in the multi-site VIRUS registry.

Here we analyze hospitalized andintensive care unit coronavirus disease 2019 (COVID-19) patient outcomes from the international VIRUS registry (https://clinicaltrials.gov/ct2/show/NCT04323787). We find that COVID-19 patients administered unfractionated heparin but not enoxaparin have a higher mortality-rate (390 of 1012 = 39%) compared to patients administered enoxaparin but not unfractionated heparin (270 of 1939 = 14%), presenting a risk ratio of 2.79 (95% confidence interval [CI]: [2.42, 3.16]; p = 4.45e-52). This difference persists even after balancing on a number of covariates including demographics, comorbidities, admission diagnoses, and method of oxygenation, with an increased mortality rate on discharge from the hospital of 37% (268 of 733) for unfractionated heparin versus 22% (154 of 711) for enoxaparin, presenting a risk ratio of 1.69 (95% CI: [1.42, 2.00]; p = 1.5e-8). In these balanced cohorts, a number of complications occurred at an elevated rate for patients administered unfractionated heparin compared to patients administered enoxaparin, including acute kidney injury, acute cardiac injury, septic shock, and anemia. Furthermore, a higher percentage of Black/African American COVID patients (414 of 1294 [32%]) were noted to receive unfractionated heparin compared to White/Caucasian COVID patients (671 of 2644 [25%]), risk ratio 1.26 (95% CI: [1.14, 1.40]; p = 7.5e-5). After balancing upon available clinical covariates, this difference in anticoagulant use remained statistically significant (311 of 1047 [30%] for Black/African American vs. 263 of 1047 [25%] for White/Caucasian, p = .02, risk ratio 1.18; 95% CI: [1.03, 1.36]). While retrospective studies cannot suggest any causality, these findings motivate the need for follow-up prospective research into the observed racial disparity in anticoagulant use and outcomes for severe COVID-19 patients.

Identification and Description of Potentially Influential Social Network Members using the Strategic Player Approach.

Background: Diffusion of innovations theory posits that ideas and behaviors can be spread through social network ties. In intervention work, intervening upon certain network members may lead to intervention effects "diffusing" into the network to affect the behavior of network members who did not receive the intervention. The strategic players (SP) method, an extension of Borgatti's Key Players approach, is used to balance the (sometimes) opposing goals of spreading the intervention to as many members of the target group as possible, while preventing the spread of the intervention to others. Objectives: We sought to test whether members of the SP set have network position and non-network differences (such as demographic, attitudinal, or behavioral differences) compared to the remaining members of the target group (non-SPs). Methods: A first-year class at a private residential university (N = 1342) completed network and non-network measures. Analyses were restricted only to heavy drinkers, leading to a final analytic sample of 529. Results: SPs and non-SPs differed on multiple network variables, but did not differ on most demographic, attitudinal, and behavior variables. Conclusions: As designed, the SP program identified participants who were distinguished by their network position. The fact that they did not also differ on other characteristics shows the SPs are not significantly different than heavy drinkers who were not selected.

Technology-Delivered Cognitive-Behavioral Interventions for Alcohol Use: A Meta-Analysis.

BACKGROUND: Cognitive-behavioral therapy (CBT) has long-standing evidence for efficacy in the treatment of alcohol use, yet implementation in clinical practice has been challenging. Delivery of CBT through technology-based platforms, such as web-based programs and mobile applications, has the potential to provide widespread access to this evidence-based intervention. While there have been reviews indicating the efficacy of technology-based delivery of CBT for various psychiatric conditions, none have focused on efficacy for alcohol use. The current meta-analysis was conducted to fill this research gap. METHODS: Descriptive data were used to characterize the nature of the literature on technology-delivered, CBT-based interventions for alcohol use ("CBT Tech"). Inverse-variance-weighted effect sizes were calculated, and random effects, effect sizes were pooled in 4 subgroups. RESULTS: Fifteen published trials conducted primarily with at-risk or heavy drinkers were identified. Of these studies, 60% explicitly targeted alcohol use moderation. The content of CBT Tech programs varied, ranging from 4 to 62 sessions/exercises, with many programs combining elements of motivational interviewing (47%). With respect to efficacy, CBT Tech as a stand-alone treatment in contrast to a minimal treatment control showed a positive and statistically significant, albeit small effect (g = 0.20: 95% CI = 0.22, 0.38, kes  = 5). When CBT Tech was compared to treatment as usual (TAU), effects were nonsignificant. However, when CBT Tech was tested as an addition to TAU, in contrast to TAU only, the effect size was positive, significant (g = 0.30: 95% CI = 0.10, 0.50, kes  = 7), and stable over 12-month follow-up. Only 2 studies compared CBT Tech to in-person CBT, and this pooled effect size did not suggest superior efficacy. CONCLUSIONS: These results show a benefit for technology-delivered, CBT-based interventions as a stand-alone therapy for heavy drinking or as an addition to usual care in specialty substance use settings.

A giant left atrial mass-Clinical treatment dilemma.

A 74-year-old female current 75 pack-year smoker presented with shortness of breath and mild hemoptysis. Chest computed tomography showed a large right upper lobe mass compressing the superior vena cava, invading the right pulmonary veins, and occupying the majority of the left atrium. Brain magnetic resonance imaging revealed a 13 mm right parietal lesion with surrounding edema consistent with metastasis. A 3D TEE showed a large mobile mass in the left atrium. Bronchoscopy confirmed that the tumor mass was consistent with a moderately to poorly differentiated squamous cell carcinoma. She underwent chemotherapy, radiation, and immune therapy. She was also started on warfarin for anticoagulation after the initial chemotherapy with resolution of the left atrial mass. We feel that the patient most likely had carcinogenic thrombus in the pulmonary veins and left atrium.

Artificial Intelligence and the Future of Primary Care: Exploratory Qualitative Study of UK General Practitioners' Views.

BACKGROUND: The potential for machine learning to disrupt the medical profession is the subject of ongoing debate within biomedical informatics and related fields. OBJECTIVE: This study aimed to explore general practitioners' (GPs') opinions about the potential impact of future technology on key tasks in primary care. METHODS: In June 2018, we conducted a Web-based survey of 720 UK GPs' opinions about the likelihood of future technology to fully replace GPs in performing 6 key primary care tasks, and, if respondents considered replacement for a particular task likely, to estimate how soon the technological capacity might emerge. This study involved qualitative descriptive analysis of written responses ("comments") to an open-ended question in the survey. RESULTS: Comments were classified into 3 major categories in relation to primary care: (1) limitations of future technology, (2) potential benefits of future technology, and (3) social and ethical concerns. Perceived limitations included the beliefs that communication and empathy are exclusively human competencies; many GPs also considered clinical reasoning and the ability to provide value-based care as necessitating physicians' judgments. Perceived benefits of technology included expectations about improved efficiencies, in particular with respect to the reduction of administrative burdens on physicians. Social and ethical concerns encompassed multiple, divergent themes including the need to train more doctors to overcome workforce shortfalls and misgivings about the acceptability of future technology to patients. However, some GPs believed that the failure to adopt technological innovations could incur harms to both patients and physicians. CONCLUSIONS: This study presents timely information on physicians' views about the scope of artificial intelligence (AI) in primary care. Overwhelmingly, GPs considered the potential of AI to be limited. These views differ from the predictions of biomedical informaticians. More extensive, stand-alone qualitative work would provide a more in-depth understanding of GPs' views.

Chronic dietary creatine enhances hippocampal-dependent spatial memory, bioenergetics, and levels of plasticity-related proteins associated with NF-κB.

The brain has a high demand for energy, of which creatine (Cr) is an important regulator. Studies document neurocognitive benefits of oral Cr in mammals, yet little is known regarding their physiological basis. This study investigated the effects of Cr supplementation (3%, w/w) on hippocampal function in male C57BL/6 mice, including spatial learning and memory in the Morris water maze and oxygen consumption rates from isolated mitochondria in real time. Levels of transcription factors and related proteins (CREB, Egr1, and IκB to indicate NF-κB activity), proteins implicated in cognition (CaMKII, PSD-95, and Egr2), and mitochondrial proteins (electron transport chain Complex I, mitochondrial fission protein Drp1) were probed with Western blotting. Dietary Cr decreased escape latency/time to locate the platform (P < 0.05) and increased the time spent in the target quadrant (P < 0.01) in the Morris water maze. This was accompanied by increased coupled respiration (P < 0.05) in isolated hippocampal mitochondria. Protein levels of CaMKII, PSD-95, and Complex 1 were increased in Cr-fed mice, whereas IκB was decreased. These data demonstrate that dietary supplementation with Cr can improve learning, memory, and mitochondrial function and have important implications for the treatment of diseases affecting memory and energy homeostasis.

Do what you say and say what you are going to do: A preliminary meta-analysis of client change and sustain talk subtypes in motivational interviewing.

Objective: This meta-analysis examines the predictive validity of client change language subtypes in motivational interviewing (MI) sessions addressing addictive behavior change. Method: A systematic review identified k = 13 primary studies, contributing 16 MI conditions (N = 1556). The pooled correlation coefficient was used to assess the significance, direction, and strength of seven language subtypes (i.e., reason, desire, need, ability, commitment, taking steps, and other) by three valences (i.e., frequency positive or change talk, frequency negative or sustain talk, and proportion change talk) and their relationship to subsequent engagement in addictive behavior. Results: For frequency measures, more sustain talk related to reason, desire, ability, and other were associated with more addictive behavior at follow up. Other change talk was associated with MI outcomes but in an unexpected direction (i.e., more addictive behavior). Proportion measures showed more proportion change talk-reason and -other statements were associated with less addictive behavior at follow up. Sensitivity analyses indicated some heterogeneity and instability of effect sizes, but no evidence of publication bias. Conclusions: This preliminary meta-analysis suggests that aggregate measures of change and sustain talk are comprised of statement subtypes that are not equally meaningful in predicting outcome following MI for addictive behavior change.