Silicon Valley Musings

Notes from my email conversation with Dr. David Green.

October 13th Q&A

> SW: I read you are using SMS messages to remind TB patients to take their
> medicines. I was wondering how this relates to the WHO’s DOTS program.
>

> I was wondering why the DOTS approach was not feasible or appropriate
> for the patients for which you are using the SMS messages. Do your
> patients also have periodic visits from health workers? In other
> words, is your SMS solution an enhancement to the DOTS method or is it
> an alternative where DOTS is not practical or feasible?

DG: Patients are given the option of opting out of the DOTS system in order to go onto our programme (I have the support of WHO in this - if your interested I can send on the open letter they have given me to deal with criticisms that I have received that I am undermining DOTS). The opt out option is given under very specific circumstances - i.e. that the patient has successfully completed the first month of DOTS and that they have a cell phone. I think that DOTS suffers from the general deficiencies of the “one size fits all” approach which I criticise in my paper on adherence (see http://www.on-cue.co.za , click on “Adeherence” - the hypothesis and thinking has come some way since I posted that on the site, but the core ideas are contained therein). …

Patients on our system are seen once per month at the clinic when they are presented with the next month’s supply of medication.

>
> SW: I am curious about the logistics of your program. How does the patient
> get their medications? Do they get the full six months of medication
> at the beginning of the program or in regular intervals over the six
> months? If it’s the later, is their an SMS message to remind them to
> refill their prescription?

DG: The patient presents monthly at the clinic for medication. They are reminded by SMS to come in for these visits.

> SW: Does your program have any feedback mechanism? That is to say, do you
> know that the SMS has been received or read? Does the patient send an
> SMS-based acknowledgement when they take their medicine? Does your
> system raise exceptions to the health care provider if the SMS
> messages are not being received or (if possible) acknowledged?

DG: At the moment all the technology is in place for feedback (either return SMS messages from the patient, or an SMS only device built into the base of the pill box which automatically SMS’s back when the pill box is opened). This has not yet been implemented. Also, this can be escalated by SMS or other means (e-mail) to care providers. We are toying with the idea of only sending out reminders if we do not get an incoming message from the pill box by a certain time, if this state persists then having the patient visited by a community based care provider to sort out potential problems (e.g. battery flat!) or provide support, assistance and guidance (e.g. discussing side effects or other problems associated with compliance)

We do get input on when the SMS is delivered to the handset, but cannot tell if it has been read.

> SW: I read that the original implementation used a single simple message
> every day and this quickly became boring. You mention that you now
> vary the message and use humor and other information in the reminder.
> This is a particular area of interest to me. I would like to talk with
> whoever devised the current database of reminders and strategy for
> varying the messages.

DG: I co-ordinated the compilation of the current database of messages. We had no clear cut strategy - we took input from patients who suggested messages, dreamt up our own and asked all involved to forward ideas for messages.

> SW: I was wondering if reading proficiency or literacy is an issue. Do you
> tailor the messages in response to the patient reading level? Are
> Multimedia Messaging Service (MMS) enabled phone becoming pervasive in
> your patient population? Have you explored using MMS based reminders?

DG: We assume literacy, and in about 10% of cases this turns out to be an incorrect assumption. MMS is not really practical at the moment … our patient population uniformly has older phones, as (if?) MMS capable phones enter the second hand market it may become an option. Interestingly, literacy on cell phones is a highly functional thing … if you see what I mean. Let me give you an example - one patient says that she cannot read the messages, but she can see that it is not a “Please call me” message from her daughter or a prompt to pick up voice mail, and based on the time of day that it comes, it prompts her to take her medication!

October 15th Q&A

>
> SW: I would be interested in reading the open letter from the WHO. Is the
> WHO interested in providing SMS based reminders for TB treatment in
> other parts of the world?

DG: Attached [who_support_letter.pdf]. The WHO would support such initiatives but would not get involved in funding or providing this themselves.

>
> SW: You state, ?We are toying with the idea of only sending out reminders if we do not get an incoming message from the pill box by a certain time ??
> Do you have an active development team for this effort? Do you have a
> list of enhancements you are currently working on or a “wish list” of
> features you would like to build?

DG: We have a developer, who leads an ad hoc development team which we pull together as and when required. We have a small list of enhancements which we are currently working on and an extensive wish list!!

> SW: You mentioned a pill box that can send SMS messages. Does this
> actually exist? If so, did you build it or is it a commercial product?

DG: We developed and made five prototypes of the SMS enabled pill box. They work really well. The ability to manufacture them is in place and, depending on quantities (whether they are to be individually made, or machines tooled for a more automated production etc.) we can roll them out relatively quickly. I would only do this on order, as I don’t have the capital to make them on spec.

>> SW You state ?I co-ordinated the compilation of the current database of messages?? The message database is interesting to me. Do you feel the database is
> as effective as it could be, or is there room for improvement?

DG: The database could, without doubt, benefit from improvement. There has been little to no thought on what the best content, sequence etc etc of the messages should be.

>
> SW: Did you consider using pre recorded voice message as apposed to or in
> addition to the SMS messages?

DG: It is possible, using the GSM standard on one of our networks, to send “text” messages to landlines. The text is “read” as the phone is answered. This was not really satisfactory as the landline is not with the patient all the time. As to “reading” messages over the cell phone, this would require a unified messaging system (different protocols for texting a number and dialing a number) - a unified messaging system is on our wishlist - and, of course, voice calls increase the cost substantially.