A comparison of two text message-based intervention studies for health promotion in Cape Town, South Africa: The lessons learnt

Abstract

According to the World Bank, there are 128 mobile subscriptions per 100 people in South Africa as of 2011. Cellphones are therefore viewed as having considerable potential for health promotion. However, evaluations of interventions and evidence of effectiveness are still limited in developing countries. Two separate trilingual (isiXhosa, Afrikaans and English) short-message-service (SMS) interventions that took place from 2012 to 2013 in Cape Town are evaluated.

The SMS campaigns aimed to increase health knowledge by disseminating health information via SMS. Specifically, each study involved sending SMSes to participants to provide them with information regarding how to control hypertension and how to be healthy during pregnancy, respectively. The hypertension campaign was conducted with hypertensive patients at a Community Health Centre, while the antenatal campaign was with pregnant women attending a Midwife Obstetric Unit. Facility staff guided the preparation and checked the health promotion content of both campaigns.

There were 223 and 206 participants at baseline in the hypertension and antenatal campaigns respectively. Both studies used mixed research methods of a randomised control trial followed by a focus group. Participants were randomised either to be sent SMSes (experiment group) or not (control group). Intervention in the hypertension campaign involved sending 5 SMSes per week for 16 weeks. The pattern of dissemination was different in the antenatal campaign: SMSes were staggered according to the week of pregnancy at the time of recruitment into the study. The total number of SMSes sent to all those in the experiment group was 101.

34.53% and 46.60% of the participants were lost to follow-up (LTFU) in the hypertension and antenatal campaign respectively. There was no differential LTFU in either study. At the end of the antenatal campaign, there were no significant differences in the level of knowledge (assessed by nine questions) between the experiment and the control group (all p > 0.05). Similar results emerged in the hypertension campaign. Despite no significant improvements in knowledge, both campaigns’ participants (experiment group) reported high levels of behavioural change. Participants in both focus groups remarked that the SMSes acted as reminders or “light bulbs”.

SMSes appear to be more effective at improving motivation than increasing knowledge, although objective measures to verify behavioural change could not be obtained in either campaign. Further research is needed to investigate how health knowledge can be improved via mHealth interventions, particularly in populations of low socioeconomic status where high rates of LTFU are a reality.

There’s an App for That! Promoting Antiretroviral Medication Adherence in Rural Georgia: The Music for Health Project

Consistent high levels of adherence to antiretroviral drug therapy (ART) are needed to sustain undectable viral loads (VL) in persons living with HIV/AIDS (PLWH). The result is improved health and prevention of HIV transmission. Rural dwelling PLWH encounter barriers such as low health care resources, transportation, poverty, stigma, and depression that contribute to adherence challenges. The goal of the Music for Health smartphone app is to use technology to improve adherence to ART. The app consists of a music program called the LIVE Network that includes animated music videos specially developed and tailored for PLWH, a manual with web links, and a pill count survey. The program is designed to educate, motivate, and increase self-confidence in rural PLWH to adhere to ART. We are conducting a randomized controlled clinical trial to study the efficacy of this app compared to an equivalent educational app in 240 rural PLWH in Georgia. Eligibility criteria include: HIV infected, initiating ART for the first time or changing a regimen due to side effects or ART drug resistance, ≥18 years of age, English speaking, and willing to complete study activities. Once randomized at baseline, each participant will receive a smartphone loaded with the appropriate app and will have a supervised listening/viewing session. They will receive regular text message reminders to use the app and for monthly unannounced pill counts. Pill counts will also be collected via smart phone. All participants will be followed up at 3, 6, 9 months using computerized interviews, hair samples for ART drug levels, and lab values extracted from medical records. The app was pre-tested in 3 focus groups conducted in 3 different rural counties. Nine men and 4 women participated; 10 (77%) were African American. All reviewed the app and provided feedback. Only one person did not own a mobile phone and 7 (54%) owned a smart phone. After coaching all were able to use the smart phone and app. All songs and videos were rated ≥ 7 on a scale of 0 to 10 and 5 songs had a median score of 10. Participants found the songs relevant and many thought songs addressed a situation they were currently experiencing: “This is my first year living with HIV, so songs like that I can relate to… dealing with issues and questions that I have in my own head… that I don’t talk to people about.” Videos were edited to incorporate changes requested by the participants. Study recruitment will begin in 2014. If successful this app could transform the delivery of HIV adherence self-management care by overcoming barriers in this vulnerable group.

Virtual Validity: mHealth Simulation Games, Diagnostic Indicators, and Behavior Change

Abstract

Risky decisions (e.g., having unprotected sex; using methamphetamine) are often context dependent, automatic, and affectively based, but the contextual triggers that contribute to those risks may not be well understood by the individual him or herself. Virtual game simulations, designed to capture real-life situations, for example for PTSD sufferers, (Rizzo et al., 2009; McLay et al., 2011) are used by clinicians to diagnose then, personalize therapy. Because Read et al. (2006) have designed interactive virtual date environments (VE) to simulate the risk challenges MSM typically encounter in real life, could MSM’s virtual decisions be used to diagnose men’s risks in everyday life? However, to date, no tests of a simulation game’s virtual validity (link between game and real-world risk-taking) have been conducted: That was this project’s goal. We tested the hypothesis that decisions made within the VE would be correlated with real-life risky choices made in the past 3 months (e.g., unprotected sex, alcohol use, methamphetamine use). High-risk men who have sex with men of color (18-30yr old MSM of Latino & African-American descent) were recruited online via hook-up/social networking sites on both mobile apps and through the Internet. Participants reported past risky behaviors and then made a series of automatically recorded choices in an online VE. We found that those who drank alcohol in the past 90 days were likely to choose to drink alcohol in the VE, p(116)=.332, P=.000. Similarly, those who took methamphetamine in the last 90 days, were also likely to choose to take methamphetamine in the VE, p(116)=.866, P=.000. Those who were the insertive partner most in real-life (over 60% of the time), chose to be the insertive partner in the VE, and those who were the receptive partner most in real-life (over 60% of the time), chose to be the receptive partner in the VE, Anal Sex Position p(90)=.747, P=.000. Finally, those who chose to have unprotected anal intercourse (UAI) in the VE, were more likely to have engaged in risky anal sex in the past 90 days, UAI (Time 1) p (116)=.612, P=.000. These findings suggest the potential value of VE for unobtrusively diagnosing, predicting, and understanding the circumstances under which real-life risk-taking might take place. These findings suggest virtual simulation games could be used in apps and over the web to diagnose MSM’s real-life risky choices, potentially yielding more behavior change, even for those with contextual challenges they don’t understand.

Use of an Automated Telephone Response System for a Women’s Lifestyle Physical Activity Program

Abstract

Physical inactivity is ranked fourth as a global risk factor for mortality. Internationally, physical inactivity disproportionately affects women (33.9%) as compared to men (27.9%). A need exists for innovative physical activity behavioral interventions for women using affordable, accessible technology. The 48-week, community based Women’s Lifestyle Physical Activity Program for African American women strived to meet this need by creating an intervention that includes the use of 6 group visits and a personalized step goal. One of three study conditions also received 11 motivational tips to overcome barriers via an automated telephone response system (ATRS), while another received 11 motivational personal calls and a third received no calls. The purpose of this presentation is to present the development and then examine the use of phone technology in the ATRS condition. The research questions are: 1) What was the preferred means of receiving the automated calls (cell phone, home phone, work phone)? 2) Did the number of automated calls completed differ by the means of receiving the calls? 3) What were the challenges encountered with an ATRS?

A randomized cluster, Latin Squares, clinical trial design was used with the order of the conditions randomly assigned to six community sites. Over three years each of the study sites received each of the three study conditions. Five group visits occurred every five weeks in the first 24 weeks with a booster visit in the second 24 weeks. Women in the ATRS condition received one to two calls between group visits that delivered motivational tips. They could listen to more than one tip.

Ninety-seven women of the total 288 enrolled were in the ATRS condition. Of these women, 68% provided cell numbers, 69% provided home numbers, and 27% provided work numbers. The majority of women chose to have messages delivered to their cell phones (73%), followed by home phones (22%), or work phones (5%). The women listened to an average of 9.5 motivational tips. Specific challenges encountered were phone numbers that were not in service throughout the study, which was encountered with 10.8% of the phones at least one time during the study as well as unanswered calls (8.9%). Selected solutions designed to address challenges included closely tracking weekly call reports, using participants’ alternative contacts and email to reach them, and always updating participant phone numbers during data collections.

This study has demonstrated that automated telephone messages can be successfully delivered via cell phone technology.

Engaging Individuals to Promote Diabetes Risk Awareness with txt4health: A Program Evaluation of Beacon Community txt4health Pilots in Southeast Michigan and Greater Cincinnati

Abstract

There are an estimated 25.8 million American children and adults, roughly 8.3% of the U.S. population, living with diabetes. Diabetes is particularly burdensome on minority populations where health disparities persist. Mobile technologies are an attractive method for reaching broad populations given the high penetration of cell phones across diverse groups, and may be a useful strategy for raising awareness of type 2 diabetes. To raise awareness of the risks associated with type 2 diabetes, pilots of txt4health, an automated 14-week text message program, were launched by the Southeast Michigan and Greater Cincinnati Beacon Communities. This investigation sought to evaluate the txt4health program pilot in Southeast Michigan and Greater Cincinnati through documenting participant usage of txt4health, as well as user perceptions of the program.

In this two part evaluation, we conducted a retrospective records analysis of individual-level txt4health system usage data from participants in Southeast Michigan and Greater Cincinnati to determine usage of the program. We also conducted a multimodal user survey with 161 txt4health users recruited through the program to understand participant perceptions of program satisfaction, participant use, and self-reported perceptions of behavior change. Preliminary results from the retrospective records analysis reveal that across both pilots, 5,570 participants initiated enrollment in txt4health, of which 33% completed the two-step enrollment process. Once enrolled, the majority of participants set a weight loss goal (74%), and tracked their weight (89%) and physical activity (65%) at least once during the program; however 56% dropped out before the end of the program, with 70% of dropouts occurring before the end of the fourth week. Among program completers, rates of weekly weight tracking were low with 22% of participants logging weekly weights at least five times, yet rates of weekly physical activity tracker were greater with 49% logging weekly physical activity at least five times. Despite high attrition across the pilots, surveyed txt4health users report high levels of program satisfaction, with 67% reporting satisfaction scores of eight or higher on a ten point scale (10 =most satisfied; M=8.2, SD=1.6). The majority of participants report that txt4health helped them make lifestyle and behavior changes related to diet and physical activity. While broadly focused public health text message interventions may have a great reach, individual engagement among participants widely varies, suggesting that this type of approach may not be appropriate for all, but is a feasible and acceptable delivery modality for a large subset of people.

Air Aware: A smartphone app to increase awareness of environmental factors triggered for symptoms of breathing-related illnesses

Abstract

Breathing-related chronic illnesses such as asthma and chronic obstructive pulmonary disorder (COPD) affect millions of people. People with these conditions are especially impacted by environmental and weather-related factors that contribute to symptom exacerbation. This includes worsening dyspnea, anxiety, fatigue and depression. Self-monitoring of symptoms is becoming an increasingly popular tool in treatment for these illnesses. One such tool is being developed by the Department of Medical Social Sciences (MSS) at Northwestern University Feinberg School of Medicine. This tool is an iOS based mobile application that incorporates validated measures from the NIH PROMIS system with publicly available environmental and weather-related data through RSS and API interfaces. The application associates a person’s dyspnea, anxiety and fatigue symptoms with environmental variables such pollution and pollen levels and weather related data such as temperature and humidity levels for a given GPS location. These variables are obtained by calling publicly available web services that take GPS data as inputs. The GPS readings are obtained through the GPS API available in the iOS devices. Since a person’s sensitivity to these external factors is highly individualized, the application first prompts users to complete a short battery of PROMIS computer adaptive tests (CAT) under a variety of environmental conditions and then calculates correlation coefficients. These prompts are enabled through the notification system available in the iOS platform. CATs reduce patient burden by administering highly targeted items to the user and stops when enough information is received. Once the application has collected enough data to determine if a correlation(s) exists, it will alert the user when environmental or weather conditions are at a level that would indicate a symptom is likely to occur. Since correlation does not equate causation, the application does not provide specific treatment recommendations for the symptoms; instead it aims to increase awareness of the existing conditions and how it may affect one’s immediate well-being. User acceptance and effectiveness will be evaluated in a future feasibility study. This will aid in the refinement of criteria for determining the alert mechanism based on historical stored data.