Colon Cancer Systems

thank you so much for participating in ourthird webinar in egrp's transforming epidemiology through advanced methods, in other words theteam webinar series. today's webinar will focus on mhealth tools to give you ideas forways of facilitating epidemiologic research. as a reminder, all team webinars are recordedand will be available on the egrp team webinar webpage. and the website is there as you cansee at the "grants.cancer.gov/event/team." today's presenter is dr. miller. he is anassociate professor of internal medicine, social sciences and health policy and epidemiologyand prevention at wake forest school of medicine. dr. miller provides primary care servicesat the downtown health plaza clinic that provides a patient center medical home to underservedresidents of winston, salem, north carolina.

he is an expert in cancer control and preventionand is currently the pi of an nci study examining the effects of the multi level clinic basedmhealth intervention to improve colorectal cancer screening rates particularly amonglow literacy patients. today in the webinar you will briefly review the main options availablefor web based surveys. explain why they develop their own program and discuss key featuresthat should be included in web based surveys for use by patients. or in the case of epidemiologicresearch participants. chelsea, can you go to the next slide? at the end of dr. miller'spresentation, we'll open the lines for questions. right now, though, all lines will remain mutedto ensure that there are no disturbances. and there is a "go to webinar" control panelthat will open in the upper right hand corner

of your computer screen once you've enteredthe meeting. if you click on the orange arrow, you can expand the control panel and if youhave technical difficulties or questions during the webinar, just type your questions, typeyour concern into the questions box and they'll go straight to the staff so other people willnot be seeing that. if you need to view the live closed captioning, please click on thelink that appears in the chat box. again, if you're having trouble with any of that,just type your question in the chat box and send it to the staff. chelsea, can you goto the next slide? for discussion if you would like to be unmuted, and after question byphone, please right click your, the function of the raise hand button. it's below the orangearrow to be unmuted. and if you would like

to type in a question and have a staff memberread it to the group, please type it into the questions box. i think we can go aheadand get started and we can we're going to turn over control to dr. miller now. and thankyou so much. great. thank you very much. i appreciate thisopportunity to share with you some of the work we've been doing at wake forest and lessonswe've learned along the way. if my screen isn't visible please just give me a shoutout. but we'll go ahead and get started. so as an overview for the session today, whati'm going to do is start by discussing the rationale for incorporating mhealth toolsin health care. and then i'll talk about some design considerations for tablet surveys andthe mhealth system that we ended up developing

to support colorectal cancer screening. thenlastly, i'll close by sharing some suggestions and the lessons we've learned for applyingtext messages to future studies or other needs. so to start with what is mhealth? simply putit is supporting medical and public health practice with mobile devices. and there afew definitions out there. this definition is the one that's supported by the world healthorganization. but the key feature in all of them is that these devices are wireless. andby definition, being wireless is what makes them mobile. now in health care we've beentalking about mhealth as if it's this new great thing, but the truth is that the corporateworld has latched on to this years before we have. and just as ways of example, i wantto ask you to consider which is easier for

you to do? to schedule a primary care visitwith your physician or your doctor, or to book a trip to paris while requesting a windowseat, and by the way you'd like a vegan meal. and as you think about this also ask yourselfwhich of these two things could you accomplish at 1 o'clock in the morning? now you mightbe thinking "well my health care system has recently implemented a patient portal andi can now request an appointment online ". "and that is true, some practices have done that,but then ask yourself when did i gain that ability? was it within the last 12 monthsor was within the last 12 years? or as another example, would it be easier for you to getyour lab results from blood you had drawn at 9 o'clock this morning, or to confirm thatyour paycheck that was issued at 9 a.m. was

directly deposited to your account as youexpected? a lot of the reason why health care's been a little slower to adopt some of thesetechnologies is we're still stuck in this traditional model. where care is very doctorcentric and based on in person communication. as a result the care we deliver is often disjointedand siloed. we don't share information well and, frankly, can be inefficient. and i'mallowed to say all these things because i'm a primary care doctor myself. and so i'm partof the problem and part of the system here. now on the other hand where we are tryingto move is to this more team based model that incorporates new communication tools and doesa better job of sharing information in real time across inpatient, outpatient spheresand even to other practices. and with some

of the these tools hopefully increasing ourreach into patient populations that have been more difficult to meet their needs. the onedisease that our group was really interested in exploring first was colorectal cancer.and that's because although it's the fourth most common cancer in the united states, it'sactually the second most common cause of cancer death-- second only to lung cancer in termsof mortality. currently one third of americans are unscreened and that's totally unacceptablefor two reasons. first, that's just a very large proportion of people to have unscreened,but second of all colorectal cancer is the disease that we can actually prevent withour screening options-- not just detect early, but keep it from coming on in the first place.and there are a variety of screening options

which we know are effective, ranging fromtesting stool yearly for blood on down to getting a colonoscopy every 10 years. howeverpart of the challenge is that with all of these options, what this means is that inour current doctor focused model of care, we're relying on a health care provider whodoesn't have any spare time to ideally explain to the patient, "you need colorectal cancerscreening, here's why, here's all the options that are available, here's the risks and benefitsof each option," and ideally helping patients make a choice. obviously there are a lot ofbarriers to that happening in the real world. on the patient side, patients have concernsabout some of these tests. they're worried that a colonoscopy could be painful or thatdealing with my stool sample is messy or embarrassing.

some patients aren't even aware of screeningas being an option or why they need it. some patients feel that this is just too difficultfor them to do. they don't know how i'm suppose to collect this stool sample at home. andthen low health literacy can compound all of these barriers. on the health care providerand the system side, there's time pressures. as a physician i never look down and have5 extra minutes in a visit that i don't know what i'm going to do with, and i'm not uniquein that. also we have multiple hand offs in our process. once a test is ordered it goesthrough multiple steps where information can get lost or dropped. and once the patientleaves the office, we real don't have any system in place to help support them afterthe visit is over. so to summarize the current

system of colorectal cancer screening is reallybased on this clinician patient discussion to form a screening recommendation, ideallyhelp that patient to choose screening. but of course doctors don't have the time thatrequires. once a test is ordered, it's handed off to somebody else who has to then ensurethat it gets scheduled or a kit gets handed out. and once the patient goes home, we reallydon't have any system in place to help the patients if they have questions or problemsthereafter. so we asked could we apply some of these mobile technologies to help all ofthese different stages of the screening process. and we developed this system that we calledmcrc. it's the name of a study we're doing-- the multi level crc intervention. and it has3 main components. first there's this m path

program that we display on an ipad. that wegive to patients immediately before their doctor's visit. and the program explains tothem about their need for screening, encourages screening and helps them make a screeningdecision. then on the provider and system side we've put in place some standing ordersso that if patients say "yes, i do want to be screened and i know what tests i want"they can go ahead and essentially do self service ordering and order their screeningtest themselves. once they leave we then offer follow up support through text messages orperhaps emails if the patients want to support them along every step of the way until thescreening is completed. looking at the demographics of internet use in the united states as awhole, it would look like the time is right

for some type of technology based interventionlike this. back here in 2003 this blue bar, less than 2/3 of americans were regular internetusers. and in the most recent years of data collection--2013 we're up to close to 80%of americans using the internet. now of course this isn't true for everybody, there is thisdigital divide in terms of computer literacy and computer skills. and we know that internetuse is less common among those with less education, older age, rural residency and those withlower numeracy skills. unfortunately those are often exactly the people we want to reachwhen it comes to doing colorectal cancer screening or any other disease state in which thereare health disparities. for example in colorectal cancer screening, on the whole about 2/3 ofamericans are screened. however, the screening

rates are lowest among those with the lowestlevels of education, income, and insurance status. down, if you look in the graph younotice that among those with less than high school education only 50% are screened. samething with low income. same thing with not having health insurance. and unfortunately,our current main stream survey programs that are out there, most of these are web based,assume that individuals have some basic computer and literacy skills. if you just look at avery typical web based survey that is out there now, or even a tablet based survey,there are certain features that we all take for granted and know how to use, such as thefact that if i click on a calendar icon, a calendar will appear letting me choose mydate of birth. or that a little down arrow

in a blank box means there's a drop down list.and i know what a drop down list is. i know that i can select the description that bestdescribes me. now the current web based interfaces out there may not be that intuitive for peoplewho have never touched a computer. but here again the corporate world that's use to tryingto reach out to all comers, not just those who use the internet -- and there are someplaces in america where just about everybody goes. i would venture to guess that just aboutevery american has set foot into supercenter or a grocery store at left once in the lastyear. and if we look in these places, we can see that there are some examples of very easyto use touch screen interfaces. the fact that all of you are on the webinar with me todaytells me that you're comfortable with technology.

and so you probably use these devices withoutpausing to think about the different features that make them easier to use. but i wantedto go ahead and point out some of the features that we take for granted. for example thefact that they have large buttons that are easy to see if you press here this screenwill advance with clear instructions. a lot of these will talk to you. in my grocery storeas soon as i put my basket down next to this machine, it automatically tells me welcomeand it instructs me what to do to get the process started. a lot of these incorporatephotographs. so if you have trouble reading, you don't need to read. you can find the pictureof the bananas and choose what you want. this is the same for when it comes to selectingyour payment type. there are pictures of cash,

pictures of credit cards to guide you. andthen once you select your payment method, a lot of them now are incorporating theseanimations and videos showing you an actual person swiping your card, signing a keypad,so you know how to use them. we wanted to try to borrow as many of these features aswe could to then make an ipad survey program that anybody could use, even those who havenever touched a computer or never touched an ipad before. and the program we made wecall "mpath" it stands for mobile patient technology for health. we are currently testingthis in a randomized control trial in a variety of out patient practices. and for our studywhat we're doing is contacting patients who are scheduled for routine primary care visitwho are in the age category that they need

colorectal cancer screening. we confirm onchart review and by a initial outreach phone call that they're due for colon cancer screening.and then we invite them to participate in our study by coming in ahead of their primarycare doctor visit to enroll in this study. once they enroll, we then start the mpathprogram. it starts with a splash screen that gives our study logo and a start button. andthe only training we provide the participants is to say, "when you're ready to start pressthe start button." and at that the research assistant leaves the room, says "if you getstuck, if you need help open the door, i'll be right outside" but we leave the participantalone to walk through the program. we did, with the first 20, we stayed in the room tomake sure there weren't major issues, but

once we were convinced that it was workingwell, we wanted to really leave them alone so we would know that this is a feasible technologypeople can use without any help. once the participant presses the start button, theyget an initial screen that thanks them for being in our study. and then they get thissimple instruction screen explaining that they're going to see one question on eachscreen and to choose an answer they just touch the answer they want. and then when they'resatisfied with their answer to press the next arrow to go to the next screen. we wantedto make it very simple so we limited it to one question per screen. we wanted big buttons,big fonts, so that people with perhaps poor visibility could read it. and then also becauseof literacy barriers, we not only paid attention

to using very simple language at a low educationlevel, but we also have a narrator that reads every question. so when the screen comes onthe narrator reads, "what's your favorite ice cream?" and then as the patient choosesthe choice, if they click "chocolate" the narrator will read back "chocolate" if theythink �oh that's not what i meant,� they can click other choice, the narrator willadd that option. and then when they're happy with their choice they click the "next" buttonto move on. after that one practice question, the program then runs through a variety ofour live survey questions that we're interested in for our study. and following that seriesof questions, which includes collecting some of their basic demographic information, theyare then randomized to see either our colorectal

cancer screening video, which is based onthe choice "choice decision" out of unc chapel hill. or a control program about diet andexercise, the colorectal cancer screening video, though, includes a variety of pictures.it includes some animations, some video testimonials about patients talking about their experienceswith different tests. and then we end with a table summarizing the different featuresof the stool test for blood vs. the colonoscopy. following that the patients then answer acouple post video questions it ends then for those who are in the colon cancer video arm,by giving them the option to have a test ordered for them. and it says "hey, if you want, ican order a test for you. i'll first ask you which test you want and then i'll order it.do you want me to order a test, yes or no?"

and if they do say yes, it then asks themwhich test they would like. now i do want to pause here and point out that the narratoris reading everything, but if somebody reads at a high grade level and they find that annoying,we have a sound switch at the bottom that they can switch to "off" to mute that. interestinglywe are casually observing that very few people are muting it, even those who have higherlevels of education and reading skills don't mind the narrator's voice. if they do choosea screening test to be ordered they get a nice supportive feedback message encouragingthem in the choice that they've made. and then they are told that we would like to helpthem out by sending them some messages to support them as they prepare for the testand asking them if we can send those by text

or by email, or whether they did don't wantany at all. when they do choose their method for getting messages delivered, they thenhave to do the most complicated part of the program from the user standpoint which isto actually enter in their own mobile phone number. so this is a screenshot showing thementering the mobile phone number. if they had chosen an email then the default ipadkeyboard automatically appears so they can enter their email. and then to ensure thatwe get accurate information, once this program ends, once the data collector comes back inthe room, the data collector does confirm that their phone number and e-mail was correctso we can make sure we're getting accurate data and testing that aspect of it. in termsof the text messaging program, we created

two pathways-- one for the fecal blood testingand one for colonoscopy. and each pathway asks the participant to respond at least once.this lets us confirm at least some level of engagement by the participant and also someidea of how many people react to this. so as examples on the left you see one of ourtext messages asking them to text 1 for yes or 2 for no if they're getting ready to dotheir test kit in the next few days. and then on the right is a simple message in the colonoscopyarm, that just says "hey, if you don't mind, please text back once so i know you got thismessage." what happens behind the scenes of the text messaging program can get very complicated.this is our flow chart for the fobt pathway. and the way you read this is each box representsa different text message that gets sent by

the system. and the numbers represent thenumber of days delay between messages. so the f2 message gets sent 3 days after thef1 message. now if you do offer the participants the opportunity to respond back with sometype of information, you obviously have to be prepared to react to that information.so in the f2 message which is one that asks for a yes or no response, there's one pathwaythat will be progressed along. if they answer yes, one if they answer no. but you also haveto keep in mind that, well it's possible they won't say anything and then you have to askyourself what should the system do, what message should come back? or that they might sendyou something other than a yes or a no that you can interpret, at least a machine caninterpret and how will you respond to that?

all of the text messages are sent via a thirdparty application program interface piece of code that links it to a web based nationalservice that sends out the text message. we chose this option because it's really cheap.it costs us less than a penny to send each individual message. and the accounts are verycheap to purchase, which eases dissemination. once the studies over if other practices wantto adopt our text messaging program, they just need to sign up for their own accountfor a nominal few dollars and pay the minimal cost of a few pennies a month for the messages.a nice feature is that the service also maintains comprehensive logs of every message goingin and out. so from a data collection standpoint, and even as a clinical practice standpoint,you can see exactly what patients and participate

are communicating back to you. now for theemail arm, we did not want to ask people to respond to an email with free text becausethat can be very difficult for a machine to interpret. so instead we created these htmlbuttons that get embedded in the emails, since everyone now is using html email and not plaintext email. so we have one button for a yes, one button for a no. and if they're on theirsmartphone they can just touch that button. if they're on their desk top pc, a laptop,they can just click it and it will automatically hit a webpage on our server that logs theresponses being a yes or a no. so in a lot of ways interpreting the responses from theemails is easier than from the text messaging. as of last week we had enrolled 145 patientsin this study. you will see that a lot of

the patients who are enrolling are membersof these more difficult to reach communities with significant barriers to care. 60% ofour samples so far are from the black community. 1/3 have not graduated from high school. only1/3 have gone beyond high school to any level. 80% have household incomes less than $20,000dollars. about 1/2 are not regular internet users. but then you'll see that just abouteverybody has a cell phone. and over half of the sample are texting at least weekly.the good news is that just about everybody has an unlimited texting plan which givesus some reassurance that doing a text message intervention won't be costing our patientsor our participants money. i want to point out that while 78% said, "i have an unlimitedtexting plan," 10% said, "i don't know," which

means they very well may have a unlimitedplan as well. and then pretty much all the others had a monthly allowance. it's veryrare that somebody's paying per message. looking at the usability data we split this out alongliteracy lines because we did know that those with low literacy, which correlates with lownumeracy typically don't use technology as much and we wanted to make sure this was accessibleto everyone. and we were pleased to see that over 90% of people in both groups say theuser friendliness is good or awful. nobody said it was less than good in the adequateliteracy group. we only had one person say "awful "so i guess if you offer a choice beprepared to get that response given back to you. but we have thick skin and we were ableto handle that. over 95% though, said they

did like our program more than a brochure.and looking at a real hard measure of usability, could they get through this without ever havingto open that door and get the research assistant. over 90% were able to do that. in terms ofthe acceptance of the text and email messaging, we were really pleased to see that 3 /4 ofthose who were eligible to get text-- who were offered text, actually signed up forthem because we did not require cell phone ownership, or texting ability to be in thisstudy. we really wanted to get all comers so we could determine the reach of this. andthe most common reason for not accepting texts or emails was that they just don't normallyuse the technology. in terms of getting responses back, 3 /4 of the people who get text messagesdo respond to us compared to less than half

of the email, but the email group is very,very few people at this point. for the accuracy of information entered, all 7 users who requestedemails did enter their email address correctly. and in the texting group, 27 of the 29 peopleentered their cell phone number correctly. one of the two people that didn't, the onlyone mistake they made was not entering their area code. looking at the percentage of peoplewho are signing up for these text or email messages by demographic, it really does showthat we are reaching in to these groups that typically are marked by health disparities--the target groups we are wanting to reach. among those who don't use the internet, 2/3's of them are signing up for the text messages. when you look at those with low income, lessthan high school education, low literacy,

3 /4 or more are signing up for electronicmessaging. what this graph also show though is that just about everybody in these lowincome, low education groups are choosing the text option not the email option. choosingemail was very rare. so i want to switch gears and talk a little bit about applying mhealthand these types of things to other research studies. there's a lot of reasons to thinkabout mhealth in terms of your own studies or even some kind of public health outreachprogram. among the many reasons, one of the main is that they're low cost. it does notcost a lot to buy an ipad for a clinic. is doesn't cost a lot to send a text or a textmessage. and a lot of patients that we want to reach already own these technologies andhave cell phones even if you want to look

at going into developing countries, wherecell phone ownership is becoming very common. from a project management standpoint you reallyare able to administer these programs from afar. you can decide to run your survey programon a secure webpage from your own server. or if you wanted to install locally on theipad, which is what we have done, we installed the program via a secure web link. so ourdata collectors out in the field if we issue an update to the program, they just navigateto our study server, click on a secure link, and the program will automatically updateand install. the department of health and human services published a very nice reviewof the evidence for using health text messages just last summer and in the report they citethe literature which points out that now 90%

of u s adults have cell phones. and this highprevalence of cell phone ownership persists even for those households with low incomeor low education. you also can see that a lot of adults are texting, particular if youwant to reach 18 to 50 year olds, over 95% of those demographics are texting regularlyso you are really safe to you use this type of modality. but even when you look at 50to 64, 75% of adults are now texting and you only see this drop off when you get to 65and older. the report also cites 7 systematic reviews which cover 60 different studies.and in these studies positive changes are seen-- smoking cessation, lifestyle changes,disease self management around diabetes and medication adherence. and also just attendingappointments. to help us are determine what

makes a good text message vs. a bad text message,and even to determine what our target population thought about their doctor's office textingthem in the first place, we conducted 4 focus groups as we were developing our technology.and in the focus groups we asked them about the concept of having doctors send you textmessages. and then we also had them react to a variety of sample messages we had made.this was one message that they particularly liked, it says "hang in there, jim, i knowthe prep and clear diet can be challenging but it is well worth it. remember you willneed to have someone to take you to the test tomorrow." in the main themes from our groups,what we learned is that good text messages are ones that are affirming. saying thingssuch as "glad you decided to be tested." "it

was nice to meet you." things like that. orthings that were supportive. just that little "hang in there" they really liked also, though,that they're informative. they give them some piece of new information or a reminder, suchas "hey, don't forget you need a driver tomorrow." and then also that they're personalized. fortwo reasons-- one, it just makes it feel more personal and like someone is more interestedin them. but also they pointed out that it lets me know this message is really for me.and this wasn't a mistaken text sent for somebody else. now a word of caution. shorthand didnot go over well. we, you are limited to less than 160 characters with a text, so we wantedto know what they thought of these abbreviations. and at least in this 50 to 74 year age group,they did not like it at all. emoticons though

are currently untested, so if you are lookingfor your next research project, that might be fertile ground for you to look at. in termsof concerns about text messages. there were mixed reactions about whether or not textswere impersonal. some people felt that "no, this is nice, this lets me know my doctorcares about me" and others felt like this is just another thing that gets in the wayof person to person communication. confidentiality was mentioned as a concern. there are definitelylimits to what types of information gets sent by text. in particular if someone else wereto read their text they wouldn't want that other person to know everything that was goingon in their health care at that time. and there is also a limit to the amount of textwe can send in terms of the number of texts

before it gets to be nagging, or the amountof encouragement before it turns into nagging. also there was this feeling that most of thetexts should be uni-directional. we really shouldn't request complex information backfrom them. for example, we asked them about two different ways to communicate back. onewe said, "could we ask you to text the date that you're planning to collect your stoolsamples as mmddy." and they really did not favor that. on the other hand, simple thingslike text 1 for yes, 2 for no, people had no problems with. we have learned some lessonsfrom looking at the messages that come back into us. first users don't always text backwhat you asked, so as an example, when we asked people, "please text 1 for yes or 2for no, some things we'll get back are "not

yet." one person texted back the word "one"o-n-e instead of the number 1. and one person literally texted 2. we said text 2 so theytext back "text 2." the point is as you design your systems, when you think about what themachine should react to, as a yes or no response, you probably need to be a little broader inwhat you accept as an affirmative or negative response. and as another example to this simplerequest, "text back 1 so i know you got this." one person gave us exactly what we wanted"one." and then other people just would want to give you a little more personal information.so again just keeping in mind how are you going to program your machine to respond tothese different pieces of information? i want to conclude by acknowledging my collaboratorson this project. really a fantastic group

of individuals that contributed a lot of expertise.and of course the funding from the national cancer institute that made all this work possible.so with that, i will turn it back to the nci staff and i'll be happy to take any questionsor comments people might have. okay, thank you dr. miller, that was a greatand very informative presentation. as a reminder to everybody if you would like to be unmutedand ask your question by phone, please click on the "raise hand" button. it's below theorange arrow and the operator will unmute you. also, if you're a little shy and don'twant to ask your question in person, you can take your question into the question box andthe operator will go ahead and read the question to the group. chelsea, do we have any questionsq'd up yet?

i see that donna rivera has her hand raised.i'll unmute donna and you can go ahead and ask your question. donna, you're unmuted.donna, are you on the phone? [silence] all right, we can move on to any of our next questions.okay, i'll go ahead and so i was wondering how difficult-- i know you were alluding tothe programming that needs to take place if you're going to, depending on what kind ofquestion-- answers, to this text question, that you're going to be asking, but how difficultis it to really consolidate answers to the text? i'm just thinking in terms of applyingthis for research studies where there might be hundreds of replies to text messages andwe need to pull those together for analysis. how-- can you comment on any of that?

sure. so are you talking about respondingto texts like interpreting what this text means so i know what to send out next? ormore in terms of doing qualitative analysis on the replies you're getting?i was thinking the qualitative analysis. sure. so the nice part is that no matter whichsystem you use, it's probably going to give you a log of every single text that comesin, links to somebody's mobile phone number which you could then link to a pid or whicheverstudy identifier you're doing. and because you have actually have the the verbatim responseof what they're sending back, i would think you could code it qualitatively just likeyou would a focus group transcript and in depth personal interview or anything likethat. in terms of technology aspect for knowing

how to interpret it, we kind of do rapid revisioncycle. so you start off by asking 1 for yes, 2 for no. you monitor what comes back. whenyou get surprised and realize somebody types in o-n-e, you then can quickly go back andchange the code so that it will accept o-n-e as a valid response as well. and you can doa little bit of learning as you go. okay, great. thank you.we have a couple questions that have come in from online. sheri schully writes, howwould other investigators go about getting access to the mpath survey and the api fortext messages? can they contact you directly? they definitely may. i would be happy to helpout in any way that i can. in terms of the api, the service we're using is a companycalled twilio, t-w-i-l-i-o. there are a number

of companies out there that will offer textmessaging services, but twilio actually provides this piece of code, called the api. and ifyou work with a programmer who is building the text messaging program for you, they literallytake that api, plug it in their program and it serves as the pipe that connects any messageyou want to go out to this texting service, so then it automatically gets sent by theirservice. thank you. our next question comes from jasminejacobs with the cdc. was the voice of the narrator used to read questions to patientsrobotic? if so how did participants respond to this voice?no. that's a good question. we actually used one of our local public radio announcers andshe does some voice-overs on the side and

so we used her to do the narration. and ithink having, this is just my personal feeling, but i think having a real voice makes a bigdifference. dr. miller. i had another question about couldusing this kind of technology for reminding people for following up for research studies,such as coming in for biospecimen collection aspects, or doing at home specimen collections,are there specific things that you might want to draw to people's attention, specificallyfor biospecimen aspect? sure. i think that's a great example to thinkabout and it would be a really easy way to use this technology. if you know when theyneed to have come in to drop off the samples or to do a test or have a study visit, thenit becomes very easy to send these out because

you know the date, you know what the messageshould say. i do think that the focus groups we've done and what some other studies havedone that looked at text messages, they really do let us know that a personalized text messageis better received than a generic one. so as much as you can personalize that text messageto the participant, the better. the coding to send out something like that would be reallysimple, so it shouldn't be too hard to do. you do need to pay attention to potentiallyhipaa issues around text messaging and your irb can help guide you with that in termsof what information is included in these texts, making sure that patients are consenting tohave this information sent over the airways. so would that, that kind of draws off intomind consenting for a study. would you be

able to do that kind of thing using thesemhealth technologies or is that something that you still would like to have people comein and do it in person? in terms of consenting to do a study, likecould you-- right.somebody. so personally i think that the way we consent people now is not very effective.there's so much we have, this is just my personal opinion, but there are so many things thatwe have to make sure we explain to patients. i think we've lost, we've really kind of lostthe main thing we're trying to accomplish which is do patients really understand thepoint of this study and the main risks and the main benefits. and if you hand somebodya 12 page form and they read at a third grade

level, we know that, one, they're not goingto comprehend it. and, two, if you have trouble understanding it, you tend to ask fewer questions,paradoxically. so i think using a mobile health program like something on an ipad that canwalk them through the consent very simply, maybe even showing some images to reinforceit, would be a fantastic way to improve the consent process. you would have to work withyour local irb to see how they would want that documented in terms of proving somebodysaw it and then what validates a signature would they have to then sign something afterwardsthat says, "yes, i saw this program." but i think it would have the potential to reallyprove the consent process and really help participants by making sure they're fullyinformed.

perfect. thank you. chelsea. i think we havea couple questions that came in from the chat box.we do. a couple questions that came in from on line. nicky decker writes thank you forthe nice webinar. how do you feel about using a program like whatsapp? do you know whetherthere is a secure private option in this program? so the program is called, "what's up?"whatsapp. whatsapp-- so i do not know whatsapp but iwill definitely be interested to look into it afterwards so i'm afraid i can't commenton whatsapp to know more about it. all right. our next question comes from davidishmael. what about risk factors for specific health education mainly for cardiovasculardiseases, risk factors, messages. how can

we handle that? like filtering out messagesto a group that has a specific risk factor-- like diabetes.sure, you know there have been studies. a lot of the text messaging studies that havebeen done to date are around things like diet and exercise, or smoking cessation, whichdo involve a lot of branching logic in there, meaning that people can text back "i'm cravinga cigarette today and they will get a different message based on that. or they might get aspecific lifestyle modification message. so you can take those same principles and applyit to a specific risk factor. if, say you did a survey to collect different risk factorsand then you identified somebody with certain risk factors and you wanted to provide advicefor how they might be able to mitigate those,

you could create customized pathways to sendout any of a number of messages. thank you. our next question is from lisamoy. pew�s internet research study showed differences in cell phone and technology useby ethnic groups. how do you think this type of user interface could adapt for these differences?so when you look at the difference in cell phone use by ethnic groups, you really don'tsee it breaking down along racial lines. it tends to break down-- nor as much about incomelines. it tends to be primarily along age lines where we know that older adults areusing cell phones a little less and particularly text messaging a little less. but when youlook at things like cell phone ownership and even smartphone ownership, in some of yourminority populations and in lower income populations,

you don't see that divide there as well. soi think the real challenge when it comes to, if you're doing text messages there aren'tas many challenges. if you are doing a survey, something that runs on a tablet, on the web,on a smart phone, then i think you do need to pay attention to usability issues makingit very easy to use. and also probably in how you frame it to people. for example, alot of older adults just don't like the idea of using a computer. however they very wellmight be going down to their grocery store and checking themselves out at the kiosk andthey don't think of that as a computer. so perhaps even rebranding this and not evencalling it a computer could help. i can tell you anecdotally, over half of the people wehave enrolled have never touched an ipad.

and they've all come out of it very happy,proud, that, "hey look at this, i can actually use an ipad now." so some of it's also gettingover the fear of the technology. great. nicky decker wrote in with a littlebit of follow up information for you just saying "whatsapp" messenger is an instantmessaging app for smart phones. it can also be used internationally. in addition to textmessaging what's can be used to send images, video, and audio media messages. so thankyou for that, nicky. thank you very much.our next question is from jim rovan. does the text service maintain records of all textsthat are sent? if so, do you have any control over how long that information is retained?so the service maintains a log of every outgoing

and incoming text. in terms of how long thoserecords are maintained, i do not know and i'm not sure if what's behind the questionis, you know, how can have confident are you that you have back up? i know that we do regularlydownload and backup those logs so in case they disappear one day we have several backupcopies. of course the flip side that we will have to look into is also with irb approvaland things like that. you know, there's a limit to having data longer than we want itto be so we'll have to get that the data wiped off of the third party server eventually.okay. our next question is from sally acerolo anthony. do you have any experience with useof mobile phone applications in research? if yes, what challenges did you encounterand how did you overcome those challenges?

so we had not tried to use any smartphoneapps in research ourselves. the, you know, in one way you could say the ipad programcan be displayed on an iphone. although the survey gets a little small. although we havenot done any studies with focusing purely on smartphone apps.and while you're waiting for another question i'll mention with the whatsapp, program outthere that sends text mails, i know with google voice, you can send text messages as well.when we were just trying to examine this concept we used some google voice to send texts andthat's totally free. there again, though, before using google voice you would definitelywant to talk with your irb about hipaa issues and other things like that.we have a question from david ishmael. he

is asking about sms. he asks "do you haveany experience of frontline sms for sending an sms message?frontline. can he clarify what he means by frontlinesms.this is why i love doing these talks. i should mention because there is so much great stuffthat's out there that i always learn of something else out there that i didn't realize eitherand it's a great chance just to share our experiences.we can wait -- oh, he says that it's an open source bulk sending sms software.got it. no. so i have not used that at all. when it comes to developing these text messagingprograms there are a few, some of them are start ups, some of them are affiliated withuniversities that will create your own text

messaging program for you. they will be responsiblefor sending and receiving the text. they tend to not be cheap and so when we priced it outit was much cheaper for us to work with our programmer and just develop our program ourself for the limited scope of what we were doing. but if you were doing a very complicatedstudy or you are now, when you want to think about doing text messages there are companiesout there that will offer to run the whole thing for you and they have security complianceand hipaa compliance and everything else you would need.well, great, i don't think there are any more questions coming in that i see. chelsea, doyou see anything else? we just had a new question come in.okay, great. perfect.

from alice king. she asks, do you have anyexperience with using text messages with the adolescent population? have parents been receptive?no. i have not. but is a, well there's are some publications out there about that. iwould reference you to have that hhs report that came out last summer. in there you'llfind a great bibliography and that might help you find those studies quickly. and i do knowthat there have been studies reminding parents of their children's need for vaccinationsand they will were a little varied in the parents responsiveness to getting texted aboutvaccines and reminded that they needed vaccines. but those were focused on messages sent toparents not to the children or adolescents themselves.lisa moy also many some follow up for you

from her previous question. she wanted torefer you to pew's "african american and technology use" which breaks down mobile broadband differencesby race and ethnicity. one item that may be of interest is internet usage vs. cell phoneboth basic and smart phone which could have implications in a text or web outreach situation.yes, i will definitely look at that. there is a difference when you look at broadbandand internet. what i was talking about earlier was really specific to cell phone, which doesn'tinclude cell phones globally, not necessarily smart phones specifically. but i appreciatethat reference. okay. if there's nothing else that anybodywants to add we can go ahead and conclude this webinar-- we can conclude it a littlebit early. i just wanted to remind everybody

that this webinar as well as our previoustwo webinars-- the one by dr. dana rollinson, which was our first one that we had, and oursecond webinar which was by dr. jim lacey, they're all available, and this one will beavailable as well, free to view and you don't have to register for viewing it. you can passthis information along to anybody else as well that might not have registered for thiswebinar. but it's available on our website and you can see it here on this final slide.thank you very much, dr. miller, and pleased i'm sure that you're willing to take otherpeople's questions. after this webinar if they contact you directly.absolutely, i'll be happy to help. great, thank you so much.thank you.