Developing Texting 4 Relapse Prevention: A Scalable mHealth Program for People with Schizophrenia and Schizoaffective Disorder

Abstract

Symptom relapse in people with schizophrenia and schizoaffective disorder is common and has a negative impact on the course of illness. Here, we describe the development of Texting for Relapse Prevention (T4RP), a scalable, cost-effective text messaging relapse prevention program that queries patients’ early warning signs daily. Results of development focus groups and content advisory teams with 25 patients and 19 providers suggest that patients were enthusiastic about the program and thought that receiving daily messages about their symptoms would help them feel supported. Providers also were positive about the idea, but worried that the program might interfere with patient-provider communication if patients thought that the messages were coming from providers. Patients found the content positive and actionable. The program was improved to address this feedback. This iterative development process that included multiple stakeholders ensured that the program is feasible and acceptable to both patients and providers.

Introduction

Symptomatic exacerbation and subsequent relapse are common for people with schizophrenia and schizoaffective disorder (Robinson et al., 1999). Recurrent relapses have been shown to correlate with poorer treatment response, poorer quality of life, and focal gray matter changes (Briggs et al., 2008; Emsley et al., 2013; Van Haren et al., 2007). Frequent relapses accelerate social impairment and increase the likelihood of having residual symptoms (Hogarty et al., 1991; Shepherd et al., 1989).

Research suggests that early warning signs could be a critical component of relapse prevention: The majority of individuals with schizophrenia and schizoaffective disorder experience early warning signs of relapse 2-3 weeks before relapsing, and approximately 70% of individuals can recognize their symptoms (Birchwood et al., 1989; Eisner et al., 2013; Herz and Melville 1980; Lee et al., 2010; Van Meijel et al., 2006). Moreover, the nature of these symptoms remains consistent over time. To this end, the Program for Relapse Prevention outlines a comprehensive prevention approach that includes monitoring of early warning symptoms; examples include sleep disturbance and the emergence or worsening of auditory hallucinations (Herz et al., 2000). Medication compliance also is emphasized, and there is prompt intervention if symptoms emerge. This model and others like it have been shown to be successful in preventing relapse and hospitalization, but it is offered to only a minority of high-risk patients because of the significant cost associated with its implementation (Eisner et al., 2013; Lee et al., 2010; Van Meijel et al., 2006). High relapse rates may therefore be due in part to the paucity of accessible relapse prevention programs that help patients monitor their early warning signs.

The widespread adoption of text messaging provides novel opportunities to provide cost-effective resources for people who have schizophrenia and schizoaffective disorder and are also at high risk of relapse and symptom exacerbation. Nationally, up to 81.4% of people with serious mental illness report owning a cell phone, and text messaging has been reported as the second most common use of their phones (Ben-Zeev et al., 2013a; Firth et al., 2015). Previous research demonstrates the feasibility and acceptability of text messaging-based programs with this population (Alvarez-Jimenez et al., 2014; Ben-Zeev et al., 2014; Granholm et al., 2012; Palmier-Claus et al., 2013; Spaniel et al., 2008a; Spaniel et al., 2008b). More recently, there has been increasing program development using smartphone apps (Ben-Zeev et al., 2013b; Palmier-Claus et al., 2013). In general, these studies have provided smartphones to participants and trained them on their use, which reduces generalizability to patients who do not have the means to obtain a smart phone and/or know how to use them. Although a study by Ainsworth et al. (2013) reported that patients preferred using an app over text messaging, it still remains true that text messaging is more accessible and does not require a phone data plan. In addition, the need to train people on how to use the program is reduced because most people are already familiar with text messaging. As such, we believe that text messaging continues to represent an exciting opportunity to deliver relapse prevention programming in a low-burden manner to a large proportion of people with schizophrenia and schizoaffective disorder.

The Preliminary Program Design

Texting for Relapse Prevention (T4RP) aims to increase a person’s daily self-monitoring of warning signs while also improving communication with their provider when relapse may be imminent. Before the formative development research with patients or providers, we envisioned the program thusly: Patients and providers would select five personal early warning signs of relapse. T4RP would then query one of these symptoms each day. To reduce redundancy and improve program experience, the symptom queries would rotate throughout the week. If a patient responded that a warning sign was worsening, they would receive a follow-up message that provided a suggestion about how to self-manage the symptom. If the patient indicated the warning sign was not present, they would receive a neutral, positive program response (e.g., “Great – have a lovely day!”). Symptoms that were endorsed the previous day would be asked again in subsequent days until the symptom was no longer present. If the number of symptoms present in a week met a threshold pre-specified by the provider and the patient, the provider would receive an automated alert via email or text message to reach out to the patient to determine their status and collaboratively decide on the most appropriate intervention. So that not all messages were focused on symptoms, patients would also receive a message about why medications were important to their treatment experience in the afternoon.

The initial T4RP program design integrated lessons learned from previous research in several important ways. In contrast to other programs, T4RP allows patients to identify their own set of early warning signs, which are then monitored on a daily basis (Spaniel et al., 2008a; Spaniel et al., 2008b). The program also allows for an individualized rather than automated response when symptom management is required. Like the MATS program, T4RP aims to empower patients by providing coping suggestions to manage their symptoms on their own (Granholm et al., 2012). T4RP additionally provides psychoeducation about the benefits of medication compliance and suggestions for ways to cope with possible adverse effects. As past research has found that patients had difficulties using study-provided phones that required training to use, the preliminary T4RP program design focused on individuals who had cellphones and were using text messaging (Granholm et al., 2012). To address challenges of use among those with lower cognitive functioning, the branching of our proposed intervention was designed to be simple - symptom check-in messages would require a yes/no or numerical response while all other messages would not require a response.

The Current Article

We iteratively developed TR4P, a mobile health program for people who have schizophrenia or schizoaffective disorder and their providers to ensure its salience and acceptability for both patients and providers. The goal of this article is to describe the lessons learned from this user-centered approach to developing the relapse prevention program. Findings will inform future technology-based programs for clinicians working with people with serious mental illness.

Methods

The study protocol was reviewed and approved by both the Chesapeake Institutional Review Board and the Johns Hopkins School of Medicine Institutional Review Board. Development of the T4RP program involved the following: (i) conducting three patient and two provider/administrator focus groups; (ii) content and design development; (iii) review of content by Content Advisory Teams composed of patients and provider/administrators; (iv) further content modification; (v) and alpha testing. Patient and provider/administrator participants were recruited from two community mental health outpatient clinics: Healthcare for the Homeless (HCH) and Johns Hopkins Community Psychiatry Program (JHCPP). Inclusion of multiple sites ensured a diversity of opinions and participants. Informed consent was obtained for all focus group and content advisory team participants, and a Capacity to Consent screen was administered to patient participants to ensure their understanding of the study requirements.

Eligibility for all phases of the development period matched eligibility requirements for the proposed randomized controlled trial to ensure that the feedback would reflect the intended user of the program. Patient eligibility included the following: 1) speaking English, 2) having a chart diagnosis of schizophrenia or schizoaffective disorder, 3) judged to be at their personal baseline by their treating providers, 4) having regular access to a cellphone and using text messaging, and 5) passing a Capacity to Consent screening. Any therapist, psychiatrist, or administrator (hereafter referred collectively as “providers”) working at either site was eligible to participate. Participants could only take part in either the focus group or content advisory teams to promote a diversity of feedback. Because of a smaller sample pool, providers were able to take part in both research activities if they chose to.

Recruitment

Study staff attended clinical staff meetings at both HCH and JHCPP to inform the providers about the study. For both the focus groups and the content advisory teams, providers were invited to participate in the study by self-referral. Providers identified potential patient participants during routine clinic visits and referred interested patients to the study coordinator for eligibility assessment. The study coordinator explained the study, and if the patient remained interested and was deemed eligible, the capacity to consent screen was administered and the informed consent was obtained.

Focus Groups

Three patient focus groups and two provider focus groups were conducted. For the provider groups, a sample size of 6 to 8 was targeted. For the patient groups, a smaller number (i.e., 3-4 people) was identified because we anticipated it would be difficult for people to effectively communicate in larger groups. As shown in Appendices A and B: (Supplemental Digital Content 1 and 2, http://links.lww.com/JNMD/A73, http://links.lww.com/JNMD/A74) both patient and provider participants were asked to describe patient text messaging use, the strategies patients currently used to monitor themselves for a potential relapse, and their feedback about the program idea.

All focus groups were audio recorded with participant permission and subsequently transcribed. A study team member with experience in qualitative interviewing who did not have a prior relationship with the participants led the groups. All patient participants received a $40 gift card. Focus group provider participants received $40 towards continuing education credits.

An inductive thematic analysis was conducted, first considering the breadth of topics mentioned, followed by a consideration of the frequency that they were mentioned and whether specific topics were mentioned in relation to others. We primarily emphasized common themes, but also noted outliers to assess whether there were specific, unusual topics we needed to mention (e.g. if we needed to correct mis-information about medications, we would do that even if only a single person was not well informed).

Content advisory team

Once program content was drafted, content advisory teams were recruited to provide feedback on the messages. First, patient and provider participants individually reviewed about 200 text messages. They were instructed to review the content for tone (“Describe the tone of the text messages. Do they make you feel like you’re being lectured or talked down to, or supported?”), clarity (“Address the structure of the message: is the message clear? Are you able to read it and understand it easily? Are there changes that could be made to make it easier to read?”), appeal (“Are the messages interesting and informative? Boring or dull? Are there changes that could be made to make them more interesting?”), and utility (“Do the messages – especially the coping skills ones – give advice and information that is useful?”). Providers completed their review independently. Patients reviewed the messages in the presence of study staff so that any questions could be quickly addressed.

Following the individual message review, focus groups were conducted to better understand their experience with the messages. Four content advisory team focus groups were conducted: one with patients and one with providers, at each respective treatment site. The patient group discussion script is included as Appendix C (Supplemental Digital Content 3, http://links.lww.com/JNMD/A75). The provider group discussion script is included in Appendix D (Supplemental Digital Content 4, http://links.lww.com/JNMD/A76). Patient participants received a $60 gift card whereas content advisory team provider participants received $60 toward continuing education credits.

Alpha test

To finalize the program content and flow, internal team tests, (alpha tests) were conducted with research team members. The tests included three, six, seven, and nine mock patient participants, respectively.

Results

Focus Groups

A total of 14 patients and 12 providers participated in the focus groups. Patient demographics are shown in Table 1.

Table 1.

Patient participant demographics

	Focus group participants (n=14)		Content Advisory Team participants (n=11)
Characteristic	N	%	N	%
Sex
Male	6	43%	4	36%
Female	8	57%	7	64%
Age
<30 years	1	7%	2	18%
31-40 years	3	21%	3	27%
41-50 years	4	29%	3	27%
51-60 years	3	21%	1	9%
>61 years	3	21%	2	18%
Race
White or Caucasian	3	21%	1	9%
Black or African American	10	71%	10	91%
Other	1	7%	0	0%
Hispanic ethnicity	2	14%	0	0%
Diagnosis
Schizophrenia	5	36%	9	82%
Schizoaffective disorder	9	64%	2	18%
Marital status
Single	7	50%	9	82%
Married	4	29%	1	9%
Widowed	1	7%	0	0%
Unknown	2	14%	1	9%
Employed	1	7%	0	0%
Religion
Christian	7	50%	5	45%
Muslim	0	0%	1	9%
None	2	14%	4	36%
Other	5	36%	1	9%
Medicare	9	64%	5	45%

Access to and use of technology among patients with schizophrenia/schizoaffective disorder

There was a mixture of smart phone and “dumb” phone ownership among patients. Many reported that they had a “government phone,” which was a phone paid for by Medicare and provided them with free texting. All patients agreed that, regardless of the type of phone they had, they could receive and send text messages. The range of messaging frequency varied; many reported sending and receiving a lot of text messages (up to 400 for one participant) whereas some said they might go a day or longer without sending a message.

Patient participants said they liked the emotional connection they felt when they received a text message. As one male participant shared: “I like when somebody texts me because it makes me feel like they’re thinking about me. When they took the time to uh, write out a word or a sentence, you know, it always makes me feel like they thinking about me because they took the time to text me”.

When asked about symptoms that might affect their ability to interact with a text messaging program, most talked about how feeling anxious sometimes elicited auditory hallucinations, making it difficult to text. As one woman shared: “I just, I get impatient with the phone. I get anxiety, when I, you know, trying to spell a word, and then I hear, I hear voices and then, it just, I kind, I gotta calm down, put the phone down for a minute”.

Only a few reported that medication adverse effects impeded their messaging, although blurry vision was reported by multiple people. One person noted that slight tremors sometimes made it difficult to message. Patients also noted that their mood could affect their texting experience, particularly the way that they interpreted the intention of the message.

T4RP program feedback from patients

When the intended text messaging program was described, patients were enthusiastic about it and all agreed that the program sounded supportive. As one man described: “I have very little motivation, happens sometimes. But um, I do, in that case a text would help, in a situation like that. Where I would just, you know respond to it and then get back into my coping skills and deal with it”.

Accordingly, patients suggested that the program have a positive tone and provide concrete suggestions about how to cope with their symptoms. They also recommended including both medication and appointment reminder alerts, as well as inspirational quotes aimed at promoting creativity.

Most patients said that they would be willing to receive 2-4 messages a day from the program and recommended that they be spread throughout the day. Some were concerned about confidentiality, particularly whether messages would disclose information about their illness to others, as well as what was said in their therapy sessions. They suggested using the term “illness” rather than “schizophrenia” to protect their privacy. They also said that they would prefer receiving messages that required simple responses, like a number rating.

Provider feedback about patients’ access to and use of technology

Providers felt that sending clinically oriented messages to patients would be acceptable given that some patients already ask them to text message about upcoming appointments and when their medications are ready at the pharmacy. At the same time, providers reported that they usually call rather than text message their patients as they do not want to message from their personal phone number. Providers said that sometimes patients got a new phone either because they had used up all of their minutes on their current phone and did not want to wait for the next month to get more minutes or their phone was lost or stolen. Providers also noted that when patients were given a new “government” phone, they were given a new phone number as the old phone number was not ported.

T4RP program feedback from providers

Providers agreed that the prevention program would be useful, particularly if they could choose which warning signs the program queried for each of their patients. They were particularly enthusiastic about the ability to monitor symptoms over time.

Providers thought these messages should query whether the symptoms have gotten worse than usual given that, for some patients, their warning signs (e.g., poor sleep) may be persistently present. They suggested that having reports that aggregated symptom information over time would be particularly useful. Providers also talked about the benefit of having a daily monitoring system that would give the providers an idea of how the patient was doing more often than their weekly or bi-weekly appointments, particularly because it was not uncommon for patients to miss their appointments. Some thought that it would be useful to have the symptom report sent to them automatically so they would not have to remember to query it, whereas others preferred it being on demand.

When asked to name coping strategies that they provide to their patients, many providers agreed that distraction-related activities were useful. Ideas included breathing exercises, cleaning, listening to music, walking, reaching out to other people to spend time with them, playing games on one’s phone, taking a relaxing bath or shower, and coloring or painting. For those who are struggling with weight issues, the providers talked about replacing sugary drinks with water, baking instead of frying foods, and choosing healthier proteins and vegetables over red meat.

Providers also talked about the stigma that many patients felt about taking their medications and the need for psycho-education. Providers specifically noted that patients sometimes struggled to remember when to take the medication and to budget to ensure they could pay for them. They also sometimes lost their medication and failed to understand that it needs to be taken consistently.

Providers thought that it may be difficult to keep patients engaged enough to answer the messages, particularly over an extended period. Relatedly, providers were concerned that 2-3 messages per day may feel like too much for the patients. They also raised the possibility that patients might think that the program had a crisis component although it did not and that patients may also send text messages after hours with the expectation that someone would immediately respond to them. There was also concern that some patients might stop attending their clinic appointments because they would feel like they were talking to their provider daily through the program. Providers did not voice concerns that the program might affect clinic flow. Instead, they were excited about the fact that the program could provide daily contact with patients outside of the clinic.

Integrating findings into the program

Using the information gathered from the focus groups, in conjunction with a literature review of what was known about early symptoms of relapse in schizophrenia, the research team created a list of 11 early warning symptoms (See Table 2) and drafted the text message content. Messages queried these identified symptoms, as well as suggested coping skills, and described medication adherence benefits (Eisner et al., 2013; Lee et al., 2010; Van Meijel et al., 2006). By near-universal patient request, inspirational quotes were added into the T4RP program content as patients felt that these would provide hope and help them in their recovery. All messages aside from inspirational quotes were written to be at the 6^th grade reading level or lower. The research team iteratively reviewed and edited messages until they were approved by all team members.

Table 2.

The Early Warning Signs of the Texting for Relapse Prevention Program

• Depressed mood

• Poor concentration

• Appetite disturbance

• Sleep disturbance

• Restlessness

• Feeling tense and nervous

• Isolating

• Feeling irritable

• Relapse to or increase in smoking*

• Relapse to or increase in alcohol and drug use*

• Worsening or re-emergence of auditory hallucinations

• Increased paranoia

^*Was changed from one to two ‘signs’ following the content advisory team feedback

Content Advisory Teams

Eleven patients and seven providers across both sites participated in the content advisory teams. Table 1 shows the demographic breakdown of patients.

Patient feedback about individual messages

In reviewing each of the messages, the majority of the patient feedback was positive. In all cases, people indicated that most of the suggestions were relevant and helpful, that the language was deemed acceptable, and that the reading level was appropriate.

Both messages querying early warning signs and the responses that patients would receive if a symptom was not present were well received. Suggested coping mechanisms were also positively appraised. Demonstrating the acceptability of the messages, some patients indicated that they already used some of these coping skills. Importantly, messages that were not necessarily personally relevant were deemed acceptable. For example, in response to the message: “Reading can be a great way to get your mind off things if you are feeling down. Books that have uplifting messages or quotes can be especially helpful,” a female participant said: “not a reader but someone might be.” Participants also appreciated the light-hearted tone. Messages aimed at promoting the positive aspects of regularly taking one’s medication and suggested coping skills for adverse effects also received encouraging feedback.

At the same time, some messages were noted to be long and a few were confusing. A few patients also thought that some suggestions might be difficult to implement (e.g., exercise).

Provider feedback about individual messages

Providers also had positive feedback about the clarity of the messages and the tone (e.g., “Relatable and supportive”, “Fun and upbeat”). That said, some worried that some messages were “corny” or “glib” and underestimated how hard it may be for patients to cope with symptoms.

Most coping mechanisms were acceptable, although some providers raised concerns that some messages might encourage patients to stop engaging with the world. For example, this message: “Are you having a hard time getting going today? Try watching your favorite TV show or movie. Maybe even grab a snack and have a nice, relaxing time,” prompted a provider to say: “Many of my patients lay around, eat, and watch TV when feeling unmotivated. Not sure if I would promote this behavior in relation to lack of motivation.” Suggestions were made to broaden the inclusivity of some messages. For example, one coping skill read: “Try hanging with friends or going to the day center when you hear voices. When you are around others, the voices can often fade away.” A provider noted that some patients are not part of a day center, so the message could encourage them to join one. Providers were worried that some of the messages were too emphatic or promised positive outcomes for a specific strategy that may not be justified. Providers were also concerned that suggestions involving exercise may be too difficult for some.

Patient feedback in the group discussions

In the focus group discussions, the written feedback was contextualized. Patients agreed that the messages querying their early warning signs made them feel like someone cared about them. At least one female participant said she tried the coping skills after reading them in the individual content review. Patients agreed that the tone connoted either a healthcare provider or a friend. They also agreed that messages to encourage patients to take their medications would not be annoying if they were already taking them. One man explained: “It just shows your provider, doctor is concerned, that’s all.” Patients were asked what their reaction would be to a message that suggested a coping skill that may not be a good fit for them. They said they might try it anyway or consider it interesting but not for them; and not feel bothered by it. We also asked patients whether they thought it would feel repetitive if they received the same message querying an early warning sign each week; they did not deem it problematic.

Some patients felt there were too many exclamation points and some affirmed that they do not have ‘good’ friends and family to reach out to. They suggested that messages also include the idea of joining a group to meet new people. And, consistent with some of the written feedback, not all messages were universally understood.

Specific messages that may have been cognitively demanding were queried for understandability. For example, the message that queries mood asks the patient to rate their mood on a scale of 1 to 10, anchoring 1 as “very depressed” and 10 as one’s “best mood ever.” Everyone agreed that they understood how the question was meant to be answered, and one patient affirmed that it made sense whereas another noted that he journals daily and includes his mood number.

Several questions were asked in response to providers’ written concerns during the message review. We asked patients whether they felt that this message was suggesting unattainable goals: “Exercise is good for your heart. And it also helps you fall asleep at night. Find that old bicycle or hula hoop from when you were younger and give it a try. Or just take a walk, it can’t hurt!” One man explained “I try to find something like more coping - most people don’t have a bike or a hula hoop anymore, I know I don’t, so I try to find something else to replace that that I can do here and now.” To which the moderator followed-up: “So if you got this message, would you be like, “You don’t get me at all”?’ And the patient clarified: “No, I just think I would just, it would be fine, if I know someone who had a bicycle I would ask if I could borrow it and try it or something like that or whatever.”

We also asked patients about the first-person language used in most of the messages: Did it make the content feel more personalized, or did it instead feel like we were lying about the fact that the program is automated? Many said it was acceptable. As one woman explained: “No, I mean, no, I don’t think so because a human, a person actually put this information in the system, just because they’re not the one that’s sending at that very moment, it did originally come from a person.” Others were less enthusiastic. One man said: “Yeah, you can take away that “I”.”

Provider feedback in the group discussions

Much of the provider feedback in the focus groups mirrored what they noted in their written feedback. In addition to concerns that the messages may over-promise, some providers said that some messages did not seem to acknowledge how disturbing symptoms could be. As one female provider said: “You know, the extreme language. I think the word I used was “glib”, like when I felt it was just promising too much, it was like, “Oh, come on, that’s a little too easy.” I mean, because, these are clients that are probably going to be in treatment their whole lives, so their coping skills are not, you know, that great.” Another provider said that some of the coping skills, particularly for depressed mood, may not be realistic.

Some were concerned that the casual message tone would lead to patients feeling that their challenges were under-appreciated: “If it’s too palsy-walsy, they know it’s a text and it’s not a person, um, then, I would just get concerned about that….” This was not a universal concern however. Another provider found the messages “upbeat and almost fun” and said he personally liked that tone.

Some providers thought that there were sufficient messages about emotional and physical coping skills, but that more cognitive skills could be suggested. In addition, providers encouraged the inclusion of a greater diversity of exercises be encouraged, particularly those that could be done by someone who was physically disabled. Similar to previous discussions, some providers voiced concerns about messages that suggested taking a nap if the person is tired or watching television to relax because many of their patients are already relatively inactive and they worried that these messages might promote these behaviors that the providers are trying to extinguish. Some pointed out that these could be helpful coping strategies if time-limited.

There also was discussion as to whether the smoking, alcohol use, and other substance use should be included in one symptom query message. After discussing it, providers agreed that smoking could be asked separately and that alcohol could be included in queries about use of other substances.

Integrating findings into the program

Messages were shortened and new coping skills were added as suggested by patients. Provider concerns about messages that encouraged physical inactivity motivated edits to these messages to include limits (e.g., watch one show) and messages that seemed to indicate that the coping skill would definitely be effective were also re-worded. Furthermore, per provider suggestions, smoking was designated as its own warning sign, making 12 early warning signs, and messages were framed to query a worsening of symptoms. Although providers were concerned that the tone of the messages was sometimes casual, the tone remained unchanged as patients found it acceptable. Example changes to the content are shown in Table 3.

Table 3.

Examples of message improvements as a result of content advisory team feedback

Symptoms	Draft program message	Feedback from content advisory team	Revised message
Sleep disturbance	People who have TVs in their bedroom do not sleep as well. Try moving your TV to another room if you are having trouble sleeping.	A patient pointed out some people need sound to fall asleep.	People who have TVs in their bedroom usually do not sleep as well. If you need sound to fall asleep, try a sound machine or download night sounds to your phone.
Depressed mood	When we are depressed, sometimes we get stuck in our heads. Get out of your thoughts by offering to help someone with something. It will help both them and you!	This message was noted to over-promise by a provider.	When we are depressed, sometimes we get stuck in our heads. Get out of your thoughts by offering to help someone with something. It may help both them and you!
Depressed mood	Singing can help us feel happier, so turn up the radio and sing it out (doing a little dance makes it even better!)	One patient was worried that someone might turn the music up too loud. Another said they were not a singer.	Singing may not be for everyone but it really can help. Maybe turn up the radio (not too loud!) and sing it out (doing a little dance makes it even better!)
Poor concentration	Having trouble concentrating? Maybe it is because you have too much creative energy! Drawing, painting, and coloring are great ways to bring your thoughts back into focus.	One patient thought the beginning was a bit aggressive.	Sometimes, I have trouble concentrating when I have too much creative energy! If this is you, drawing, painting, and coloring are great ways to calm your mind.
Feeling tense and nervous	Slowing your breathing will instantly help you feel less tense and nervous - I promise. Take a deep breathe in through your nose, hold it while you count 5 and then let it out slowly through your mouth.	A patient said he has a hard time calming down.	Having a hard time calming down? Try breathing slowly. Take a deep breathe in through your nose, hold it for 5; then let it out slowly through your mouth.
Feeling tense and nervous	Being creative is a great way to feel less tense or nervous. Find something to draw, paint, or color and see how everything just melts away.	One patient said that art may not be for everyone.	Being creative is a great way to feel less tense. Maybe it will work for you! Find something to draw, paint, or color and see how everything just melts away.
Restless	Get out your restless energy with exercise! Maybe it’s running around the block or walking to the store. Even doing a couple of push-ups can help. Do one thing for 5 minutes and see how you feel.	One patient was worried about getting hurt. Another said she could not do everything listed. A third patient was worried these ideas might not work for everyone.	Get out your restless energy with exercise! Try walking to the store or just around the block. Even stretching helps. Try it for 5 minutes & see how you feel
Restless	Having a hard time sitting still? Listening to music is a great way to calm your body and relax your mind. Try it!	A patient was worried that this might encourage someone to not do anything else for the rest of the day.	Having a hard time sitting still? Listening to music for an hour is a great way to calm your body and relax your mind. Try it!
Using alcohol, drugs, or smoking more	It is harder to say ‘no’ when everyone around you is saying ‘yes’. If you are craving alcohol, drugs, or cigarettes, make an effort to spend time with friends who are sober.	A patient said she has no friends.	It is harder to say ‘no’ when everyone around you is saying ‘yes’. If you are craving alcohol or drugs, make an effort to spend time with people who are sober.
Hearing voices	Wearing headphones can help quiet the voices. Give it a try and see for yourself!	A patient found this message to be confusing.	Sometimes, headphones can help quiet the voices you are hearing in your head. Maybe put some on and see how it works for you!
Hearing voices	Having someone that you can talk to during times like when you are hearing voices can be a great distraction. Call a friend or family member who supports you.	One person worried that this might be too distracting. Another said it might be hard. A third patient pointed out that it was important to find someone who understands what they are going through. A fourth patient said they don’t have friends or family.	When you are hearing voices, distraction can sometimes be helpful. Even if it’s hard, maybe call someone who supports and understands what you’re going through.
Feeling people are talking about me or looking at me	If you feel like people are talking about you or looking at you funny, put on some headphones and tune out to music. Who cares what they say? You have your favorite tune on!	A patient said this might be hard to do.	If you feel like people are talking about you or looking at you strangely, put on some headphones & ignore them. It might be hard, but it’s worth a try!
Adverse effect	Feeling slower because of your medication? Then get things going with a bit of exercise - a walk around the block with a friend is a great start. Stretching helps too.	The word “exercise” seemed to be a trigger for some patients who consistently responded that they could not do it, so it was deleted. We also deleted references to friends and family.	Feeling slower because of your medication? Try a nice walk around the block or some stretching. Anything that gets you moving is a good thing.

Alpha Test

Although the messages had been thoroughly vetted in Word and Excel, receiving them on one’s cell phone as a pseudo-participant was helpful for the research team to improve the user experience as well as the messages. One of the first things noted was that the initial message in the morning seemed out of context and abrupt. Testers also realized that the number of follow-up symptom queries needed to be limited to two previous symptoms; additional symptom queries in the same day became annoying. The program sent reminder text messages to people who did not respond to the initial symptom query. Testers determined that one reminder message for non-response was sufficient. The messages that suggested coping mechanisms also needed to be reworded: Some of these messages assumed a hypothetical (e.g., “If you are feeling anxious today…”). They were rewritten to acknowledge that patients had just replied that they were experiencing a symptom.

Integrating findings into the program

We added introductory messages that preceded morning messages (e.g., “Good morning from T4RP!”) as well as afternoon messages (e.g., “And lastly for today”). Transitions that reflected the participant’s positive response to a symptom query were also added (e.g., “I am sorry to hear you are experiencing that.”). The protocol was updated to ensure that the number of symptom queries and reminders were limited. If, according to the program logic, more than three symptoms were endorsed in the past week, the most recent symptom endorsed was queried.

A summary of lessons learned from the formative research can be found in Table 4.

Table 4.

Lessons learned from the formative research conducted to develop Texting for Relapse Prevention

Focus groups

Text messaging behavior

1. Text messaging was a mode that equalized the ability for all patients with phones to take part in the intervention.

2. There was very little concern that symptoms would get in the way of texting.

3. Providers thought it would be likely that phone numbers might be unstable / change frequently.

Program idea

1. Even though the amount of text messaging that people engaged in varied, there was general agreement that 2-4 program messages a day was an acceptable number.

2. Patients and providers agreed that receiving messages would feel supportive and positive.

3. The proposed content (symptom queries, coping skills, medication psychoeducation) was acceptable to both patients and providers. Inspiration quotes were strongly suggested by patients.

4. Providers were not concerned about how the program would affect clinic flow, but they were concerned that it might impact provider-patient communication if patients thought the program messages were being sent from them.

Content Advisory Teams

1. Although some providers voiced concern that the messages might be too ‘corny’ or sound like a real person, the patients especially appreciated the positive tone of the messages and liked that they felt familiar (e.g., using the first-person).

2. Providers worried that some of the coping skills would encourage sedentary behavior (e.g., watching TV); and that some messages seemed to promise that the coping skill would work for everyone. Both issues needed to be addressed in the messages.

3. Longer messages needed to be shortened for readability.

4. Content needed to acknowledge that not everyone has friends/family available and to provide other ideas (e.g., community centers).

5. Messages also needed to acknowledge that not everyone is mobile and able to do exercise beyond what they can do in a chair.

Discussion

Given the high cost associated with treating people with schizophrenia and schizoaffective disorder, affordable approaches to early detection of relapse and exacerbation of symptoms are urgently needed (Eaton et al., 2008; Lora et al., 2012). Early detection can potentially help patients and clinicians to avert costly and disruptive emergency department visits and hospitalizations by expedient adjustment of medications or crisis intervention. To address this need, we iteratively developed T4RP, a text message-based relapse prevention program for people with schizophrenia or schizoaffective disorder. Input was gathered from both patients and providers to ensure that the program was relevant, user-friendly, empowering, and met the needs of all stakeholders. Findings are encouraging. Patients in both the focus group and content advisory teams welcomed the idea of a text messaging-based program and felt that it would be supportive and helpful in terms of coping with their symptoms. Findings from the content advisory teams suggest that the message tone and reading level were well received by the patient population. Of particular note, patients in the focus group emphasized the importance of including brief, simple messages that only required a short response. This echoes findings from earlier research that examined the feasibility of technology-based interventions with this population and highlights the importance of designing programs that are both easy to use and understand (Granholm et al., 2012). In addition, the use of existing technology may improve engagement in a population that has historically been difficult to keep engaged in treatment (Kreyenbuhl et al., 2009). As transition to life in the community following discharge from a psychiatric admission is a particularly stressful period with increased risk of relapse, the intervention may prove especially useful for recently discharged patients.

Given the potential impact that the intervention could have on clinic flow, feedback from providers also was crucial. Providers were enthusiastic about the opportunities to monitor their patient’s daily progress through the program interface and appreciated that T4RP is envisaged as a supplement to current care and not an alternative. Their feedback highlighted how this program could be envisioned as a program for two populations, patients and providers; and as such, the need to ensure that the program was usable for both groups, although patients were the daily messaging target.

Findings also highlight the importance of obtaining feedback from the intended users of the program. For example, providers but not patients were concerned that the tone of messages sometimes seemed to minimize the challenges posed by some of the symptoms queried (e.g., hearing voices). Eliciting feedback about the content directly from patients provided critical information about what the actual experience with the program content might be. Given that patients were the ones who would be receiving these messages each day, their opinions were given more weight.

The alpha test with research team members provided opportunities to experience the program in vivo, resulting in further refinement of the message language, coping skill suggestions, and the overall timing of messages. The experience of receiving program messages led to rewording many of the messages so that they were more personable and friendlier, and there was a concerted effort to ensure that the messaging “voice” was consistent throughout the program.

T4RP builds on and extends previous text messaging programs in a number of ways: It is more heavily guided by input from individuals with schizophrenia and schizoaffective disorder and their providers. It also provides more flexibility for patients to choose their own signature relapse symptoms and to work with their providers about how to manage their clinical care when symptoms worsen. Another improvement represented by T4RP is the way the program empowers patients to manage their symptoms on a daily basis by helping them monitor their symptoms and also by providing ideas for ways to cope when symptoms are present. T4RP also facilitates communication between the patient and their provider by alerting the provider to reach out to the patient if they meet their pre-specified symptom threshold. In addition, medication compliance is encouraged with psychoeducation, and hope is facilitated through the use of inspirational quotes. Although the development of self-management apps represents a promising mode of intervention delivery, (Ben-Zeev et al., 2013b; Palmier-Claus et al., 2013) we suggest that text messaging is currently more accessible to this specific patient population and, therefore, a program such as T4RP and others using text messaging warrant further investigation.

Limitations

Patients and providers from two different sites were involved in the development of this program. Nonetheless, it is possible that they are not representative of the population of individuals with schizophrenia and schizoaffective disorder. It also should be noted that patients who have cell phones and use text messaging likely differ from those who do not. This program was developed specifically for people who are using this technology; alternative strategies will need to be developed for people who do not use text messaging. In addition, the development of this program only involved English-speaking patients. Programs adapted for other languages and cultures will need to undergo a similarly iterative development process to ensure that the content meaning and tone are appropriate for the population. Finally, it should be noted that the patient sample was majority female. Nonetheless, ten participants were male, likely resulting in a sufficient number of voices from men during the development process.

Conclusion

The process of involving all stakeholders in the development of T4RP, including focus groups and content advisory teams, has led to the creation of a relapse prevention program that has the potential to be used as a cost-effective tool in the symptom management of individuals with schizophrenia and schizoaffective disorder. Providers and patients enthusiastically engaged in the development process and were instrumental in shaping the program content and format. The next step, which is currently underway, is to test the feasibility and acceptability of the program in a randomized controlled trial setting. If found feasible and acceptable, the intervention’s effectiveness in detecting early symptoms of relapse will be tested among a large group of patients in multiple settings.