Addressing bias in artificial intelligence for public health surveillance.

Components of artificial intelligence (AI) for analysing social big data, such as natural language processing (NLP) algorithms, have improved the timeliness and robustness of health data. NLP techniques have been implemented to analyse large volumes of text from social media platforms to gain insights on disease symptoms, understand barriers to care and predict disease outbreaks. However, AI-based decisions may contain biases that could misrepresent populations, skew results or lead to errors. Bias, within the scope of this paper, is described as the difference between the predictive values and true values within the modelling of an algorithm. Bias within algorithms may lead to inaccurate healthcare outcomes and exacerbate health disparities when results derived from these biased algorithms are applied to health interventions. Researchers who implement these algorithms must consider when and how bias may arise. This paper explores algorithmic biases as a result of data collection, labelling and modelling of NLP algorithms. Researchers have a role in ensuring that efforts towards combating bias are enforced, especially when drawing health conclusions derived from social media posts that are linguistically diverse. Through the implementation of open collaboration, auditing processes and the development of guidelines, researchers may be able to reduce bias and improve NLP algorithms that improve health surveillance.

An Evaluation of the User Experience and Privacy Concerns of Individuals Misusing Opioids Using a Location Tracking Mobile Application.

Opioid use disorder is a growing public health concern in the United States, causing economic burden and hindered by stigma. New forms of data, including location data, may improve the effectiveness of interventions for preventing and treating opioid use disorder and/or misuse, increase access to treatment and address racial and ethnic disparities. This qualitative study aimed to identify factors that contribute to users' experience with a publicly available location-tracking mobile app - and investigate their privacy and ethical concerns. The study was conducted through two 15-minute interviews within a 48-h time frame. Participants were recruited from a pool of past research participants, Facebook ads, and referrals, and had to meet certain inclusion criteria related to opioid use disorder and/or misuse. The study had a final sample of 30 participants, 15 male and 15 female. The study suggests that a simple onboarding process and convenient experience can enhance participant adherence to the study app and other similar location-based research apps. However, the study also found that participants had concerns about privacy and transparency about locational privacy when sharing their location data. To improve the app, researchers suggest incorporating user behavior earlier in the app development stage. The study also highlights the importance of addressing ethical and privacy concerns such as limiting the types of collected data, incorporating data encryption and retention strategies, giving access to research staff only, and not sharing the data with third-party companies or law enforcement agencies to increase user satisfaction.

Direct and Indirect Effects of Cannabis Risk Perceptions on Cannabis Use Frequency.

Background: Perceived risk of harm associated with cannabis use has decreased in recent decades, particularly among emerging adults who show the highest prevalence of use. Cannabis-related protective behavioral strategies (PBS) are associated with lower cannabis use and fewer consequences; however, individuals who perceive using cannabis as low risk may use cannabis PBS less often. Therefore, using cross-sectional data, we examined the associations between perceived risk of harm associated with cannabis use, cannabis PBS, and cannabis use frequency. Method: Participants were 146 emerging adults between the ages of 18-25 (56.2% female) who reported consuming cannabis at least 3 times/week and completed measures of past-month cannabis use, past three-month use of cannabis PBS, and perceived risk of harm associated with cannabis use. Path analyses examined direct and indirect effects of perceived risk of cannabis-related harm on cannabis frequency through cannabis PBS. Results: Most (66.4%) participants reported no perceived risk of harm associated with occasional cannabis use, whereas 30.1% reported no perceived risk of harm associated with regular cannabis use. Findings indicated a significant indirect effect between perceived risk of harm and cannabis use frequency through cannabis PBS, b = -10.23, SE = 3.80, 95% CI [-17.67, -2.80], p = .007. Conclusions: Among emerging adults who consume cannabis regularly, findings suggest that a greater perceived risk of cannabis-related harm is associated with decreased cannabis use frequency via increased use of cannabis PBS. Although future analyses evaluating causal mechanisms are needed, these findings have clinical implications for harm reduction interventions focused on cannabis use.

Ethical Perspectives in Using Technology-Enabled Research for Key HIV Populations in Rights-Constrained Settings.

PURPOSE OF REVIEW: People who use illicit drug substances (e.g., heroin) and sex workers are vulnerable to acquiring HIV. Due to the criminalization of illicit drug substances and sex work in many countries, these populations often reside in rights-constrained settings where their well-being, freedom, and access to HIV prevention and care services may be compromised due to legal prosecutions and societal stigma. RECENT FINDINGS: This study conducted a literature review on papers that evaluated a combination of the following three components: ethics, technology-based research, and populations who use drug substances and/or sex workers. We explored research on these ethical perspectives from both key populations and researchers. Findings revealed potential risks in data security and possible harmful implications of compromised data within these rights-constrained settings. Best practices were explored within the literature to identify potential methods for addressing these ethical concerns and improving HIV prevention and care.

Potential application of conversational agents in HIV testing uptake among high-risk populations.

Human Immunodeficiency Virus (HIV) continues to be a significant public health problem, with ~1.2 million Americans living with HIV and ~14% unaware of their infection. The Centers for Disease Control and Prevention recommends that patients 13 to 64 years of age get screened for HIV at least once, and those with higher risk profiles screen at least annually. Unfortunately, screening rates are below recommendations for high-risk populations, leading to problems of delayed diagnosis. Novel technologies have been applied in HIV research to increase prevention, testing and treatment. Conversational agents, with potential for integrating artificial intelligence and natural language processing, may offer an opportunity to improve outreach to these high-risk populations. The feasibility, accessibility and acceptance of using conversational agents for HIV testing outreach is important to evaluate, especially amidst a global coronavirus disease 2019 pandemic when clinical services have been drastically affected. This viewpoint explores the application of a conversational agent in increasing HIV testing among high-risk populations.

Network interventions for managing the COVID-19 pandemic and sustaining economy.

Sustaining economic activities while curbing the number of new coronavirus disease 2019 (COVID-19) cases until effective vaccines or treatments become available is a major public health and policy challenge. In this paper, we use agent-based simulations of a network-based susceptible-exposed-infectious-recovered (SEIR) model to investigate two network intervention strategies for mitigating the spread of transmission while maintaining economic activities. In the simulations, we assume that people engage in group activities in multiple sectors (e.g., going to work, going to a local grocery store), where they interact with others in the same group and potentially become infected. In the first strategy, each group is divided into two subgroups (e.g., a group of customers can only go to the grocery store in the morning, while another separate group of customers can only go in the afternoon). In the second strategy, we balance the number of group members across different groups within the same sector (e.g., every grocery store has the same number of customers). The simulation results show that the dividing groups strategy substantially reduces transmission, and the joint implementation of the two strategies could effectively bring the spread of transmission under control (i.e., effective reproduction number ≈ 1.0).

Risky Drinking in Adolescents and Emerging Adults: Differences between Individuals Using Alcohol Only versus Polysubstances.

BACKGROUND: Factors related to risky drinking (e.g., motives, protective behavioral strategies [PBS]) may vary between youth who engage in polysubstance use compared to those who consume alcohol only. We examined differences in factors among youth who consume alcohol only compared to alcohol with other substances (i.e., polysubstance use), and correlates associated with risky drinking between the groups. METHODS: Participants (N = 955; ages 16-24; 54.5% female) who reported recent risky drinking completed measures of alcohol/substance use, alcohol-related consequences, drinking motives, alcohol PBS, mental health symptoms, and emotion dysregulation. Participants were in the polysubstance group if they reported using at least one other substance (e.g., cannabis, stimulants) in addition to alcohol in the past three months. Chi-square and t-tests examined differences between the two groups and multiple regression analyses examined correlates of risky drinking. RESULTS: Most participants (70.4%, n = 672) reported polysubstance use; these individuals engaged in riskier patterns of drinking, experienced more alcohol-related consequences, used fewer PBS, had stronger drinking motives (enhancement, social, coping), endorsed more mental health symptoms, and reported more emotion dysregulation. Regression models showed that emotion dysregulation significantly associated with risky drinking in the alcohol-only group; conformity and coping motives, alcohol PBS, and anxiety symptoms significantly associated with risky drinking in the polysubstance group. CONCLUSIONS: Among risky drinking youth, results indicated youth engaging in polysubstance use have greater comorbidities and individual-level factors associated with risky drinking than youth who consume alcohol only. These findings may inform the tailoring of interventions for individuals who engage in risky drinking and polysubstance use.

A Mobile Health Application Using Geolocation for Behavioral Activity Tracking.

The increasing popularity of mHealth presents an opportunity for collecting rich datasets using mobile phone applications (apps). Our health-monitoring mobile application uses motion detection to track an individual's physical activity and location. The data collected are used to improve health outcomes, such as reducing the risk of chronic diseases and promoting healthier lifestyles through analyzing physical activity patterns. Using smartphone motion detection sensors and GPS receivers, we implemented an energy-efficient tracking algorithm that captures user locations whenever they are in motion. To ensure security and efficiency in data collection and storage, encryption algorithms are used with serverless and scalable cloud storage design. The database schema is designed around Mobile Advertising ID (MAID) as a unique identifier for each device, allowing for accurate tracking and high data quality. Our application uses Google's Activity Recognition Application Programming Interface (API) on Android OS or geofencing and motion sensors on iOS to track most smartphones available. In addition, our app leverages blockchain and traditional payments to streamline the compensations and has an intuitive user interface to encourage participation in research. The mobile tracking app was tested for 20 days on an iPhone 14 Pro Max, finding that it accurately captured location during movement and promptly resumed tracking after inactivity periods, while consuming a low percentage of battery life while running in the background.

Role of Peer Coaches in Digital Interventions for MOUD Initiation and Maintenance.

Millions of Americans suffer from opioid use disorder (OUD) in the United States, resulting in thousands of deaths. The ongoing opioid crisis necessitates novel approaches to reduce the morbidity and mortality associated with misusing opioids. Studies using peer support models show success in engaging persons living with OUD in initiating treatment and decreasing relapse. Although most studies have focused on patients in clinical settings, community studies integrating peer community leaders also show promise. This viewpoint paper explores the use of peer coaches in online interventions in the community setting.

A Qualitative Exploration of the Potential Role of Using Online Social Media Support Communities to Increase Initiation of Medications for Opioid Use Disorder (MOUD).

Objectives: This study sought to explore the potential role of peer-led online communities to increase use of medications for opioid use disorder. Methods: From January through March 2020, participants with opioid use disorder and their family members/friends were recruited from paid Facebook ads; public health key stakeholders were recruited from referrals from the study team and opioid experts. Thirty participants from California were interviewed; 23 persons reporting opioid misuse, 3 family members/friends of persons misusing opioids, and 4 public health key stakeholders. We conducted semi-structured interviews asking about preferences, barriers and facilitators of treatment options for opioid use disorder, and perspectives around the use of digital/online communities. The categories of participants interviewed were each asked slightly different questions depending upon their role. Results: Results suggest that participants who misuse opioids (1) may prefer to engage in online communities rather than in-person meetings to discuss their opioid use, (2) generally prefer to receive opioid-related information from other patients with opioid use disorder and/or those in recovery rather than from health providers or other individuals, and (3) thought that an online community could be beneficial for helping address their opioid use. Conclusion: Results suggest an openness and interest in a peer-led online community to discuss opioid use and treatment among people who misuse opioids.

Social Media Images as an Emerging Tool to Monitor Adherence to COVID-19 Public Health Guidelines: Content Analysis.

BACKGROUND: Innovative surveillance methods are needed to assess adherence to COVID-19 recommendations, especially methods that can provide near real-time or highly geographically targeted data. Use of location-based social media image data (eg, Instagram images) is one possible approach that could be explored to address this problem. OBJECTIVE: We seek to evaluate whether publicly available near real-time social media images might be used to monitor COVID-19 health policy adherence. METHODS: We collected a sample of 43,487 Instagram images in New York from February 7 to April 11, 2020, from the following location hashtags: #Centralpark (n=20,937), #Brooklyn Bridge (n=14,875), and #Timesquare (n=7675). After manually reviewing images for accuracy, we counted and recorded the frequency of valid daily posts at each of these hashtag locations over time, as well as rated and counted whether the individuals in the pictures at these location hashtags were social distancing (ie, whether the individuals in the images appeared to be distanced from others vs next to or touching each other). We analyzed the number of images posted over time and the correlation between trends among hashtag locations. RESULTS: We found a statistically significant decline in the number of posts over time across all regions, with an approximate decline of 17% across each site (P<.001). We found a positive correlation between hashtags (#Centralpark and #Brooklynbridge: r=0.40; #BrooklynBridge and #Timesquare: r=0.41; and #Timesquare and #Centralpark: r=0.33; P<.001 for all correlations). The logistic regression analysis showed a mild statistically significant increase in the proportion of posts over time with people appearing to be social distancing at Central Park (P=.004) and Brooklyn Bridge (P=.02) but not for Times Square (P=.16). CONCLUSIONS: Results suggest the potential of using location-based social media image data as a method for surveillance of COVID-19 health policy adherence. Future studies should further explore the implementation and ethical issues associated with this approach.

Online personal ads and its implication for sexual health risks: an observational analysis of posts.

BACKGROUND: People use online personal advertisements (ads) to solicit sexual partners. Data from online ads are often publicly available, allowing researchers to better understand the topics discussed and potential role of online ads in sexual health and risk behaviours. This study aims to examine those research questions. METHODS: We collected posts on ClassifiedAds.com from the US (August 2019- February 2020). Gender identity was abstracted from post title, when available. HIV risk-related terms were abstracted from post title and content and were used to dichotomise the following variables: (1) ads describing sexual risk behaviours; (2) ads mentioning concerns over substance use or sexually transmitted infections/HIV; and (3) ads mentioning an interest in substance use. Descriptive analysis and logistic regression were conducted. RESULTS: A total of n =12866 unique ads were posted; 72.4% posted by men. Most posts came from the southern US (34.7%) and from urban areas (89.6%). Few ads contained images (21.9%) or words associated with substance use (1%), while 26% mentioned substance use or STI/HIV concern. Logistic regression models indicated that ads in the South were less likely to contain substance use or STI/HIV concern than ads from the Midwest. Also, women were less likely than men to post about substance use or mention substance use or STI/HIV concern. CONCLUSIONS: Personal ads offer a glimpse into Internet-mediated sex encounters and sexual health risks among those who seek sex online. The real-time data collected has potential to highlight elevated prevalence of sexual risk behaviours that may put individuals at risk for STI/HIV or substance use.

Accounting for Context in Randomized Trials after Assignment.

Many preventive trials randomize individuals to intervention condition which is then delivered in a group setting. Other trials randomize higher levels, say organizations, and then use learning collaboratives comprised of multiple organizations to support improved implementation or sustainment. Other trials randomize or expand existing social networks and use key opinion leaders to deliver interventions through these networks. We use the term contextually driven to refer generally to such trials (traditionally referred to as clustering, where groups are formed either pre-randomization or post-randomization - i.e., a cluster-randomized trial), as these groupings or networks provide fixed or time-varying contexts that matter both theoretically and practically in the delivery of interventions. While such contextually driven trials can provide efficient and effective ways to deliver and evaluate prevention programs, they all require analytical procedures that take appropriate account of non-independence, something not always appreciated. Published analyses of many prevention trials have failed to take this into account. We discuss different types of contextually driven designs and then show that even small amounts of non-independence can inflate actual Type I error rates. This inflation leads to rejecting the null hypotheses too often, and erroneously leading us to conclude that there are significant differences between interventions when they do not exist. We describe a procedure to account for non-independence in the important case of a two-arm trial that randomizes units of individuals or organizations in both arms and then provides the active treatment in one arm through groups formed after assignment. We provide sample code in multiple programming languages to guide the analyst, distinguish diverse contextually driven designs, and summarize implications for multiple audiences.

The Role of Misinformation and Stigma in Opioid Use Disorder Treatment Uptake.

Background: Deaths due to opioid overdose continue to rise in the United States. Despite availability of effective treatment for opioid use disorder, uptake is low among those who misuse opioids. Methods: This paper explores the role of misconception, stigma, and misinformation in influencing decisions to initiate medications for opioid use disorder among patients and providers. Conclusion: Misinformation about opioids has been prevalent among future healthcare providers and first responders as well as pharmaceutical companies, which may have implications for treatment. Among individuals with opioid use disorder, treatment uptake and adherence have been negatively affected by misconceptions about treatment efficacy and side effects, as well as stigma. We discuss the role of social media, education, and the community, in mitigating misinformation and addressing misconceptions about opioids and treatment options.

High-intensity drinking among adolescent and emerging adult risky drinkers.

Background: High-intensity drinking (HID; 8+ U.S. standard drinks for women, 10+ men) is initiated during adolescence/emerging adulthood, increasing risk for negative outcomes, including blackouts. We examined baseline data from a study of risky drinking youth to identify factors associated with HID. Methods: Risky drinkers (ages 16-24) were recruited online (positive 3-month AUDIT-C score) as part of a larger study to examine social media interventions for risky drinking. We used baseline survey data to examine HID in relation to demographics, substance use-related variables, and individual and social factors. Results: Among 931 risky drinkers, 29.8% reported past-month HID, and those with HID reported greater substance use and consequences. In multivariable analysis, HID was associated with male sex; greater social motives, impulsivity, and motivation; lower self-efficacy; and greater likelihood of not living with parents, drinking with important peers, and parental disapproval of posting drinking pictures. When examining age group interactions (16-20; 21-24), underage drinkers with high sensation-seeking scores and lower parental disapproval of posting drinking pictures on social media reported greater HID. Conclusions: Among risky drinking youth, male sex, social motives, impulsivity, higher motivation to and lower-self-efficacy to reduce drinking, living away from parents, more frequent drinking with important peers, and lower parental disapproval of posting drinking pictures on social media were positively associated with HID. Further, HID was associated with greater health consequences, underscoring the need for HID interventions. Such interventions may benefit from enhancing motivation and self-efficacy, particularly in social contexts, as well as increasing positive peer and leisure activities to reduce HID.

Mental Health among African American and Latinx Men who have sex with men after the COVID-19 Lockdown in Los Angeles - Findings from the HOPE cohort.

This study aimed to examine the depression and anxiety among men of color (primarily African American and Latinx) who have sex with men after the lockdown due to the COVID-19 pandemic. Outcomes included 21-item Beck Depression Inventory (BDI), 7-item Generalized Anxiety Disorder (GAD), and a 10-item COVID-related anxiety measure using a modified H1N1-related anxiety question. Independent variables were food insecurity and belief in government efficiency. Data were analyzed by Regression models with random cluster effects. Food insecurity experiences were significantly associated with higher depression (p < 0.001), higher anxiety (p < 0.001), and higher pandemic-related anxiety (p < 0.001). Higher levels of belief in government efficiency were significantly associated with lower depression (p < 0.05), less anxiety (p < 0.05), and less pandemic-related anxiety (p-value < 0.001). These findings emphasize the importance of establishing trust between government and at-risk communities when issuing public health policies, especially during unforeseen circumstances, as well as to ensure basic human rights, such as food security.

An online advertising intervention to increase adherence to stay-at-home-orders during the COVID-19 pandemic: An efficacy trial monitoring individual-level mobility data.

The COVID-19 pandemic has led public health departments to issue several orders and recommendations to reduce COVID-19-related morbidity and mortality. However, for various reasons, including lack of ability to sufficiently monitor and influence behavior change, adherence to these health orders and recommendations has been suboptimal. Starting April 29, 2020, during the initial stay-at-home orders issued by various state governors, we conducted an intervention that sent online website and mobile application advertisements to people's mobile phones to encourage them to adhere to stay-at-home orders. Adherence to stay-at-home orders was monitored using individual-level cell phone mobility data, from April 29, 2020 through May 10, 2020. Mobile devices across 5 regions in the United States were randomly-assigned to either receive advertisements from our research team advising them to stay at home to stay safe (intervention group) or standard advertisements from other advertisers (control group). Compared to control group devices that received only standard corporate advertisements (i.e., did not receive public health advertisements to stay at home), the (intervention group) devices that received public health advertisements to stay at home demonstrated objectively-measured increased adherence to stay at home (i.e., smaller radius of gyration, average travel distance, and larger stay-at-home ratios). Results suggest that 1) it is feasible to use mobility data to assess efficacy of an online advertising intervention, and 2) online advertisements are a potentially effective method for increasing adherence to government/public health stay-at-home orders.

Patterns of same-day alcohol and cannabis use in adolescents and young adults with risky alcohol use.

Same-day alcohol and cannabis use is relatively common in adolescents and young adults, constituting a higher-risk behavior relative to single-substance use. However, the association between quantity of alcohol and cannabis use on co-use days is understudied. We examined the association between the quantity of alcohol and same-day cannabis use with a multilevel regression analysis in a sample of youth (16-24 years old) with risky alcohol use. Participants reported one or more days of alcohol and cannabis over the past month (N = 468). Quantity of cannabis use was highest on heavy drinking days [M = 0.91 grams, SD = 0.68] followed by moderate drinking days (M = 0.78 grams, SD = 0.63), and lowest on days without alcohol use (M = 0.74 grams, SD = 0.64, p < 0.001). In multilevel modeling analyses, adjusted for clustering within individuals, greater quantity of drinking on a given day was associated with greater cannabis use (estimate = 0.03, p < 0.001). When using alcohol and cannabis on the same day, greater alcohol use was associated with greater cannabis use. Preventing days of heavy use of multiple substances, particularly among at-risk drinkers, may complement interventions addressing co-use generally to prevent substance-related consequences.

Technology-Delivered Intervention Strategies to Bolster HIV Testing.

Since the beginning of the HIV epidemic, there have been more than 75 million cases. Currently, there about 1.2 million living with HIV in the USA. Despite current testing recommendations, test rates continue to be suboptimal. Investigators have studied the use of digital technology to promote HIV testing, especially among high-risk populations. PURPOSE OF REVIEW: This non-systematic review provides an overview of the scientific research between 2015 and 2020 focused on the use of digital technology to bolster HIV testing and suggests novel technologies for exploration. RECENT FINDINGS: A total of 40 studies were included in the review that span a wide range of available technology. Studies effectively increased HIV testing among study participants. Generally, participants in the intervention/exposure groups had significantly higher rates of HIV test uptake compared to participants in the comparison groups at study follow-up. For a variety of reasons (e.g., differences in ways the technologies were used and study design), no digital tool clearly performed better than others, but each have the capacity to increase outreach and self-testing. An exploration of the potential use of nascent technologies is also discussed, as well as the authors' experiences using a number of these technologies in our research.

Applying social norms interventions to increase adherence to COVID-19 prevention and control guidelines.

Despite widespread national, state, and local guidelines for COVID-19 prevention, including social distancing and mask orders, many people continue to not adhere to recommendations, including congregating in groups for non-essential activities, putting themselves and others at risk. A social psychological perspective can be used to understand reasons for lack of adherence to policies and methods for increasing adherence based on successes from other behavior change campaigns. This manuscript seeks to describe some of the social psychological research that may be relevant to COVID-19 prevention and behavior change, describe how these theories have been previously applied in various domains to change behavior, and provide examples of how these approaches might be similarly applied to control the pandemic. We provide concrete examples of actions that can be taken based on social psychological research that might help to increase adherence to COVID-19 recommendations and improve prevention and control of the virus.