Feasibility of using text messaging for unhealthy behaviors screening in a clinical setting: a case study on adolescent hazardous alcohol use

Abstract

Underage alcohol use is the leading cause of preventable mortality among adolescents in the USA. Moreover, the average age of onset of underage drinking is 13 years. This study examined the feasibility of using a text messaging survey (TMS) to assess adolescent alcohol use. A sample of 29 adolescents, aged 13–17 years, was recruited from two primary care clinics. They completed a 16 question TMS while in the waiting room and a two-question exit TMS. The participation rate was 87%. Two out of 25 (8%) adolescents met the criteria for hazardous drinking and 28% reported alcohol use. It was found that 38% and 25% of adolescents who completed the exit TMS were asked or advised about drinking, respectively. Text messaging to assess adolescent alcohol use in this setting seems feasible, does not disrupt patient workflow, and can assess many health behaviors before a clinical encounter.

Cell phone, text-message, primary care, screening, adolescents, alcohol

Introduction

Underage alcohol use is the leading cause of preventable mortality among adolescents in the USA.1 Alcohol use often starts during early adolescence; the average age of onset for alcohol use among adolescents in the USA is 13.2 years.2 Furthermore, motor vehicle injuries are the leading cause of death for 15–20-year-olds and 25% of those deaths involved an alcohol-impaired driver.3

Recognizing and assessing underage alcohol use is a key component in preventing and reducing teenage morbidity and mortality. Since most adolescents have health insurance, either public or private, and receive at least one medical check-up annually,4,5 primary care physicians (PCPs) are in a unique position to have an effect on teenage alcohol use by providing early identification, assessment, intervention, and referrals. The American Academy of Pediatrics and the American Medical Association recommend adolescent alcohol screening, identification, advice, and discussion at every visit.6–8 Despite these guidelines, several studies have found that PCP compliance with American Academy of Pediatric/American Medical Association guidelines ranges from 23% to 45%.8–10

The most common reason cited by healthcare professionals for not discussing sensitive topics such as adolescent substance abuse is time.11–13 Identification of health problems before the clinical encounter can guide PCPs in dealing with specific health concerns making the clinical encounter more efficient and possibly even shorter. Only 35% of adolescents reported discussing substance use with their PCP, although 65% said that they wanted to do so.14 When given a computerized option for substance use screening, adolescents preferred it to verbal inquiry from a nurse or physician, and did not differ in what they were likely to reveal.15 Confidentiality is another concern for adolescents. Over half (58%) of high school students have health concerns that they wished to keep private from their parents and from friends and classmates (69%); similarly, 25% reported that they would forego healthcare in some situations if their parents might find out.16

Wireless technology has been used to address PCP and adolescent concerns. Personal digital assistants (PDAs) have been used in clinics to enhance adolescent screening and counseling in primary care. A recent study in five primary care practices in the USA found that a PDA-based screening tool enhances physician counseling and improves adolescents' perceptions of the well visit.17 Another study examining the use of PDAs on alcohol screening reported that with PDA usage more adolescents rated their visit as confidential, more thought they were listened to carefully, and more were very satisfied (all statistically significant differences).18 However, PDAs are expensive, learning to use them requires training and thus uses valuable time in a clinic, and demand a support staff to solve technical problems.

Cell phone texting provides an innovative technological approach to deal with this important public health problem. Its widespread use and fast adoption have created innovative ways for patients and healthcare providers to interact about a variety of healthcare problems.16 Moreover, mobile phones are an effective mode to deliver information owing to familiarity, ubiquity, and affordability.18 Texting may also provide a method of interaction that is, simple, confidential, and preferable to adolescents. Research shows that most adolescents own a cell phone and that texting is their preferred means of communication19,20; 50% of children aged 8–12 and 75% of 17-year-olds own mobile phones.19 Minimal research has been done on the feasibility of using text messaging technology in a clinic.

It is impossible to predict the future of wireless technologies. However, a 2011 report found that 35% of adults in the USA own a smartphone.21 Furthermore, by the end of 2011 smartphones will be owned by 50% of cell phone users.22 Clearly, the trend seems to be the acquisition of more powerful mobile devices. These devices will allow owners short messages service (SMS) capabilities and also internet connection. Once this transition occurs, more diverse and interactive types of surveys and interventions can be developed and tested.

The aim of this paper is to examine the feasibility of using text messaging survey (TMS) technology to assess adolescent alcohol use in a clinic. We use alcohol screening in primary care as a case example.

Methods

Participants and setting

Participants were recruited while attending a pediatric or a family medicine clinic. Eligibility criteria included being 13–17 years old, having a cell phone with text message capabilities, and being mentally and physically able to send and receive text messages. Parental and adolescent consent were obtained before any assessment.

Procedures

A research assistant (RA) approached consecutive adolescents and their parents as they arrived for their clinical appointments. The RA explained the purpose of the study and determined eligibility. Thereafter, adolescents were given a printed card with a text number and survey initiation code and instructions on how to access the survey. Adolescents were instructed to return to the RA after their scheduled appointment to receive instructions on accessing a separate two-question exit survey and to receive a $25 debit card to cover any text message charges and as a token of appreciation for participating.

Materials

This pilot study employed a low-cost system using mobile phones to receive and send SMS survey questions. The application uses text-to-email conversion capabilities from USA mobile phone companies, allowing teens to use their personal phones. This feature is key to ensure adoption and use as it prevents patient workflow disruptions that would occur if teens had to learn how to text on a new phone, be trained to use a PDA or create a username and password to login to a website.

An RA gave participants a printed card with a text number and survey initiation code and instructions on how to access the survey. As soon as adolescents send a text message with the survey initiation code to a study-specific number the automated texting system sends the first questions to cell phones and waits for a response. After each response, the system sends the next question until the last question is reached. This concludes the first stage of the TMS. After the medical assessment, an RA instructs participants to text a different survey initiation code to the same number. The automated system sends the exit questions to cell phones following the same protocol described above. Only two questions are sent and that concludes the TMS. Adolescent responses are stored as a spreadsheet data file where each row and column corresponds to a participant and TMS question, respectively. The participants' phone numbers are then automatically deleted, the file saved, and delivered to the researcher. Once the researcher receives the spreadsheet data file it is converted to Stata version 1123 format (Stata, College Station, Texas, USA). Statistical analyses and reports were also generated using Stata.23

The TMS consisted of 18 questions that were divided in two parts (see the online supplementary appendix). The first part included 16 questions and the second part consisted of two questions sent soon after the adolescent left the clinical encounter. The TMS gathers information about demographics, diet, physical activity, smoking, alcohol, and PCP's assessment and advice on alcohol use. The CRAFFT instrument was used to assess alcohol use. CRAFFT is a mnemonic based on key words within the individual questions: “Have you ever ridden in a CAR driven by someone (including yourself) who had been using alcohol?” “Do you ever use alcohol to RELAX, feel better about yourself, or fit in?” “Do you ever use alcohol while you are by yourself or ALONE?” “Do you ever FORGET things that you did while using alcohol?” “Do your family or FRIENDS ever tell you that you should cut down on your drinking?” “Have you ever gotten into TROUBLE while you were using alcohol?” Two or more “yes” answers suggest that the respondent may have a serious problem with alcohol abuse and an additional assessment is warranted. CRAFFT has been validated and tested on adolescents in a clinic24 and is not affected by age, sex, or race.25

Analysis

Descriptive statistics such as frequencies and percentages were calculated using Stata V.11.23

Results

A total of 55 adolescents were approached to participate. Of these, 18 were ineligible for two reasons (ie, physical or mental disability (n=3), or did not bring their cell phones (n=15)). Eligible participants comprised 37 teens but parental or participant refusals lowered the sample to 32. The study sample included 32 adolescents ages 13–17 for a response rate of 86%. However, three had faulty cell phones so the final sample was 29 adolescents (see figure 1).

Figure 1

Study recruitment flow chart.

Table 1 presents selected characteristics of the study participants.

Table 1

Selected characteristics of the study sample (n=29)

Characteristic	%
Age (years)
13	21.4
14	25.0
15	25.0
16	17.9
17	10.7
Mean (±SD*)	14.7 (1.3)
Sex
Male	48.2
Clinic
Pediatric	79.3
Family medicine	20.7
Length of survey (min:sec)
Mean (±SD*)	15:04 (5:10)

Participants had a mean age of 14.7 years (SD±1.3 years) and were almost equally divided by sex. The average length to complete the TMS was 15 min (SD ±5 min). Table 2 presents results for the CRAFFT questions.

Table 2

Percentage distribution of CRAFFT questions in the study sample (n=29)

Questions	%
Have you ever ridden in a CAR driven by someone (including yourself) who had been using alcohol?	20.7
Do you ever use alcohol to RELAX, feel better about yourself, or fit in?*	3.6
Do you ever use alcohol while you are by yourself or ALONE?	0.0
Do you ever FORGET things that you did while using alcohol?*	7.1
Do your family or FRIENDS ever tell you that you should cut down on your drinking?†	3.9
Have you ever gotten into TROUBLE while you were using alcohol?	3.5

*

One missing value.

†

Three missing values.

For example, the most commonly reported risk behavior was riding in a CAR driven by someone (including yourself) after using alcohol (21%). Two adolescents met the CRAFFT criteria for hazardous drinking but their PCPs neither asked nor advised them on the hazards of alcohol consumption. The exit survey was completed by 55% (16/29) of the participants. Thirty eight per cent (6/16) of PCPs asked adolescents if they were using alcohol while 25% (4/16) advised about the health hazards of underage drinking.

Discussion

The aim of this pilot study was to examine the feasibility of using a TMS in primary care taking as a case example screening adolescents' alcohol use. A major concern was participation rates because both parents and adolescents had to consent. However, this concern proved unfounded since 86% of eligible adolescents completed the first and longest part of the TMS. This response rate is comparable to school-based surveys26 and higher than population-based computer-assisted interviews.2 The high participation rate shows that using cell phone text message in primary care is well accepted by parents and adolescents to gather sensitive data and can facilitate screening and assessment of unhealthy behaviors.

Another concern was possible waiting room delays, which would compromise the use of this technology in clinics. However, this pilot study also showed that this technology did not increase waiting room time. Furthermore, PCPs rarely noticed that their adolescent patients were participating in a research study.

Security of sensitive data is also a factor that needs to be considered. However, none of the participants expressed any concern. Nevertheless, encryption data protocols for text messaging authentication and confidentiality are currently available (eg, SecureSMS) (CellTrust Corporation, Scottsdale, Arizona, USA) and should be used when electronically collecting and transmitting sensitive data. Security encryption protocols must become the norm when using wireless technology for data gathering.

This pilot study also has some limitations. The first is the small sample typical of this kind of study. Another limitation is the lower response rate for the exit survey. A different approach is required to increase participation rates. Possible approaches include sending the TMS later in the day, providing adolescents with an additional incentive to dial and answer the TMS at a later time, or incorporating the TMS after the PCP clinical evaluation before leaving the examining room.

Notwithstanding these limitations, this study also has several strengths. First, we obtained a high response rate (86%). This is higher than similar surveys.2,26 Second, 38% of adolescents ≥15 years of age reported smoking, higher than other school-based (33%)27 or population-based surveys (29%).2 This confirms previous findings where adolescents reported predilection for, and feeling more comfortable with, mobile technology to answer sensitive questions about illegal and socially undesirable behaviors more honestly.28–30 Third, even within this small sample, two adolescents met the CRAFFT criteria for hazardous alcohol use but neither of them were identified by their PCPs. Of additional concern, is the low proportion of PCPs meeting the recommended clinical guidelines. Fourth, the technology proved reliable and data capturing and storage was done immediately thus reducing data entry errors, costs, time, and quality control needs. Finally, there were no waiting room delays and the time needed to complete the TMS was small.

Earlier studies have shown the feasibility and usefulness of electronic means for patients' health data.31–32 However, these studies have some limitations (eg, cost, patient training, length of assessment, and computer and internet literacy) and are prone to create a social divide with respect to access and literacy. Our study builds upon this previous work but uses a technology that overcomes past limitations and all socioeconomic groups have access to cell phones and know how to text.33 This study provides preliminary evidence for the feasibility of using SMS technology to survey adolescents in a clinic before their evaluation. This technology can screen adolescents for unhealthy or hazardous behaviors before they are evaluated by their PCPs. Information on these behaviors can be relayed to PCPs before the visit and therefore prompt them to initiate a conversation on identified hazardous behaviors and make more effective use of their limited time. Moreover, SMS shortcomings on when and where to reach participants include time zones, improper hours, carrier coverage, international texting costs, and text messaging charges. This technology can serve as a valuable tool to gain a better understanding of health behaviors (eg, smoking, alcohol, illegal drugs, and physical activity) in real life.

Acknowledgments

We acknowledge and thank the nurses and reception staff at the pediatric and family medicine clinics for their assistance to our research assistant during participant recruitment and their support of the study.

Competing interests

None.

Ethics approval

This was a feasibility study that did not meet the definition of research. It was not deemed to be a systematic inquiry to answer a research hypothesis or question. Nevertheless, we did obtain adolescent assent and parent consent from all study participants. The assent and consent forms were written according to the Michigan State University templates and guidelines.

Provenance and peer review

Not commissioned; externally peer reviewed.