With subsequent depressions of the button additional items of clinical information could be given such as:
P+ for prior treatment(red)  P- for no prior treatment (green).

VI. Means to instruct and remind the patient

Information which the monitor might communicate to the patient could include:

A) A reminder that medication should be taken

To remind the patient to take medication, the monitor could incorporate a buzzer, an indicator light (such as an LED), or both.  If medication had not been removed from  the monitor by a predetermined time, the buzzer and/or LED could be turned on intermittently (chirps and/or flashes) to alert the patient that it was time to take medication. If the indicator LED was able to display more than one color (e.g. a red/green LED), alternating the color of the flashes could help to call attention  to the indicator. This  flashing and/or chirping would continue until medication was removed or until a second predetermined time had been reached. The times when the annunciation began and ended should be established for each patient based on  his or her daily schedule.

The buzzer has the advantage that it could still be effective if the patient stored the monitor out of sight, but has the disadvantage of possibly annoying other family members who might dispense medication to turn off the chirping. One way to take advantage of both the LED and buzzer would be to have the LED turned on when the medication was scheduled to be taken and have the buzzer only start chirping late in the evening if the patient has not taken medication earlier in the day. This would increase the chance that the patient would be home to take the medication and reduce the chance that the family would be annoyed.

For monitors such as the  strip package monitor dispensing from a single strip or the cap opening monitor containing loose pills, the reminder annunciation could also instruct the patient how many pills he should take. If the annunciator is a buzzer or single LED, it could produce bursts of chirps or flashes with each burst having the same number of chirps or flashes as the number of pills to be taken. If a display which is capable of displaying numbers was used as the annunciator, it could simply display the number of pills to be taken. Such a display could, for example, be a 2-digit display or a 5 x 7 matrix display. The images below indicate how the 2-digit display and the matrix display would indicate that 4 pills were to be taken.

B) A response to a query as to whether medication had already been taken

The monitor should incorporate a query button which the patient could press to obtain an indication of whether he or she needed to take medication or had already taken the current day’s medication. A monitor equipped with a red/green LED could indicate that medication should be taken by a green light and that medication should not be taken by a red light. Similarly, a red/green matrix could use its colors to indicate whether medication should be taken. In the case of the strip package medication monitor for dispensing medication from a single strip or in the case of a cover or cap opening monitor the matrix display or 2-digit display could also indicate the number of pills to be taken when the query button was pushed. To indicate that no pills should be taken, the 2-digit display, lacking different colors, could indicate that no pills were to be taken with the digit 0.

C) An indication of how much medication remains in the monitor

With most monitors, the patient should be able to see the medication remaining in the monitor. However, if the medication is in an opaque container, a display of how much medication remains would be helpful to the patient. With a red/green LED, it would only be practical to display an approximate indication of remaining medication. For example green might indicate that the monitor was over half full, yellow that between ½ and ¼  of the medication remained, and red that ¼ or less remained. If the monitor incorporates a display such as the matrix display or the two digit display which is capable of displaying numbers, then a number could be displayed to indicate the number of days’ medication remaining in the monitor. For all dose removal monitors except the strip package medication monitor for dispensing medication from a single strip, there would be no need to display the number of pills to be taken. So, the numeric display associated with the reminder or initial query would be used to display the remaining medication, if such a display was
needed.

Of course, the amount of remaining medication could not normally be determined by a cover or cap opening monitor since the number of doses removed when the monitor was opened would generally be unknown. So, cap or cover opening monitors could not incorporate a display of remaining medication.


VII. Retrieval of the adherence record

A) Direct Communication with Computers

In clinics the data in medication monitors should be downloaded to a computer where it can be displayed and analyzed in a variety of different ways. Many clinics in both developed and developing countries have computers. When these are available they can be used to view and store the adherence record.

Incorporating a USB interface into the medication monitor would offer the most straightforward means for communicating between the monitor and computer since USB ports have become a ubiquitous computer interface. Incorporating a USB interface would also increase the cost of the monitor about $0.50

On the other hand, a village caregiver would rarely have a computer, especially when visiting patients in their homes. Therefore, it will be extremely desirable for the caregiver to be able to view the patient’s compliance record using an LED display mounted in the monitor itself without relying on additional hardware.

B) Using a Dot Matrix Display to determine the battery status

The red/green matrix display is much better suited to displaying the adherence record than either the single red/green LED or the 2-digit display. So, the following description assumes that the monitor will incorporate a red/green matrix display.

Normally, the battery status and adherence record would not be available to the patient and would only be accessed by the caregiver. The caregiver might access this information by simply pressing a separate “caregiver’s button” which is not accessible to the patient. This could be a button which can only be accessed when the monitor is opened, as it would be when the dispenser is refilled with medication.

When the caregiver pushes the button the first time a line of lights in the matrix display will shown green if the battery is fully charged, yellow if the battery is partially charged, and red if the battery needs replaced.  

The single line of lights was chosen to indicate the battery status since this would not be confused with a display of any other information.

C) Using a Dot Matrix display to retrieve the adherence record

When the caregiver pushes the button again the panel for the most recent refill interval would appear. The left hand upper corner would show flashing green (indicated here by a starburst ) for the day that the device was last filled. Patients who were good compliers and returned on time in four weeks would have 28 green dots, each indicating a day on which medication had been removed. If the patient had a poor adherence record red dots would also be present, with each red dot indicating a day on which medication had not been removed.

If the patient’s monitor had not been refilled for more than 35 days, a second and maybe a third panel would automatically appear after a few seconds in order to show all missed days with red dots. When the last red dot was reached the first panel for the first refill interval would once again automatically appear.

When the caregiver pushes the button again the first panel for the preceding refill interval would appear. Again, if the patient’s monitor had not been refilled for more than 35 days, a second and maybe a third panel would automatically appear after a few seconds in order to show all missed days with red dots. When the last red dot was reached the first panel for the refill interval being reviewed would automatically appear..

Each successive button push would bring up the panel for the previous refill interval.  

When there were no more refill intervals to be displayed and the caregiver pushed the button again the matrix would show a unique symbol which would tell the caregiver that there we no more refill intervals to review. (This symbol could flash to draw attention.)

At this point the caregiver has to decide how to manage the patients based on all adherence records. Should he give a 28-day supply, a two-week supply, should he continue therapy for more than 6 months, and should he plan for additional home visits for educating the patient and family.

Therefore, I think it would be reasonable that the next push of the button brings up the most recent refill interval and subsequent pushes of the button goes through the entire compliance record again.

This matrix display could also distinguish pocket doses from catch up doses. Pocket doses are multiple doses removed on one day before they are to be ingested, usually because the patient wants to have medicine available to take on a trip, Catch up doses are multiple medications removed on one day to make up for doses not taken in an effort to deceive the care giver.

If five pocket doses were removed on one day with four of these to be taken while on a trip, the dot for the day when five doses were removed would first flash green followed by four yellow flashes; and subsequently 4 red dots would appear for days while the patient was on the trip.

By contrast the patient might remove multiple doses on one day in a futile attempt to keep the care giver from knowing he missed taking medication. For instance, if the patient missed taking doses for four days, four red dots would appear followed subsequently by a dot which flashed green once and yellow four times.

With the matrix display, it would normally be easy to distinguish advanced removal from catch up removal, as illustrated by the two calendar displays above. However, more complicated patterns would sometimes be seen; and their interpretation would depend upon the caregiver’s judgement .

While the patterns for pocket doses and catch up doses look similar, the caregiver needs to remember that they represent significantly different degrees of adherence.  If the pattern shows pocket doses were removed the patient may or may not have ingested the medication on the subsequent days when the display shows red dots. If the pattern shows catch up doses were removed the patient clearly did not take medication on the preceding days where the display shows red dots.

D) Using a single red/green LED for summary adherence record and battery status

If reduction of cost is a major consideration, substituting the single red/green LED that costs 12¢ for the red/green matrix LED that costs $1.75 could be considered.

With the push of the button the LED would first flash green multiple times if the battery is fully charged, yellow multiple times if the battery is partially charged, and red multiple times if the battery needs replacement.

With a second push of the button the LED could provide a summary of the percentage of days on which medication had been taken since the last refill as follows.

Subsequent pushes of the button would provide the summaries of the adherence record in previous refill intervals.

When all refill intervals had been displayed the LED would sequentially flash green yellow red multiple times to indicate there were no more refill intervals to review.

This less expensive simple display has several limitations.

First, if the battery needed replacement and multiple red flashes occurred with the first push of the button, the caregiver might incorrectly interpret this as poor adherence.

Second, if the patient was one or more months late in returning for refills of medication, the caregiver would get a record of poor adherence, but might not realize the patient was several months late in returning for a refill, a piece of information useful in planning interventions that would have been picked up if the Red/Green Dot Matrix display had been used

Third, the caregiver could not get a detailed adherence record without entering codes and stepping through the adherence record for each day one by one, a process, which would be time consuming and tedious.

Fourth, the caregiver could not distinguish pocket doses from catch up doses without going through the tedious process of stepping through the adherence record for each day.

Fifth, the caregiver could not get any clinical data stored in the monitor.

E) Use of 2-digit display for reviewing battery status and adherence record

A two-digit display, which costs approximately $0.89, could also be used.

With the first push of the button the display would show the charge of the battery with the letter b and a number from 1 to 9 for the charge remaining in the battery.

With a second push of the button the display would show the percentage of medication, which had been taken since the last refill anywhere from 0 to 99% with 99% displayed when the adherence was perfect.

Subsequent pushes of the button would provide the percentage of medication that had been taken in previous refill intervals.

When all refill intervals had been displayed the display would create a distinctive symbol like repetitively flashing 1,2,3,4,5 multiple times to indicate there were no more refill intervals to review.

This two-digit display while somewhat less expensive than a DOT Matrix Display has several limitations.

First, if the patient was one or more months late in returning for refills of medication, the caregiver would get a record of poor adherence, but might not realize the patient was several months late in returning for a refill, a piece of information useful in planning interventions that would have been picked up if the red/green matrix LED had been used

Second, the caregiver could not get a detailed adherence record without entering codes and stepping through the adherence record for each day one by one, a process, which would be tedious and time consuming.

Third, the caregiver could not distinguish pocket doses from catch up doses without going through the tedious process of stepping through the adherence record for each day.

F) Communicating by Mobile Phones

In order to take prompt and effective corrective action when poor
adherence occurs, the caregiver needs to learn about it as soon as possible. According to a UN report in February 2010 over half of the people in developing countries were mobile phone subscribers. This suggests an opportunity to send the adherence record from a medication monitor by a mobile phone to the caregiver to achieve prompt notification of poor adherence in those areas, which have mobile phone service. The medication monitor could be used to supervise self administered medication and medication given to the patient by a family member, i.e. family DOT.
 
Mobile phone users tend to be relatively well off. TB control programs  must deal with all patients including those who are very poor. If a monitoring system uses mobile phones the health department will have to supply inexpensive mobile phones to a large number of poor patients. I have been told that a used mobile phone in India costs $10.00. (Personal communication William Thies) An organization in the USA called the Charitable Recycling Program collects and recycles old mobile phones.  While only 35% of these phones are useable, this still represents much more than a sufficient number of mobile phones for TB patients, since there are only 9 to 10 million new TB patients in the world each year compared to 4.6 billion mobile phone subscriptions. Furthermore, as more sophisticated smart phones are introduced the relatively inexpensive simple mobile phones are often discarded.
 
While it may be possible to directly introduce the data from a monitor into mobile phones, the following system is proposed as one requiring no modification of the phone itself and therefore should be the least expensive. Furthermore, this system would work with both a cover-opening monitor that determines when the cover of container is opened and with a dose removal monitor that determines when each dose is removed. The two classes of devices are described in a published article. (1)

The data transmission system by mobile phones includes the following:

1) A microcontroller in the medication monitor which records the time when the cover of a cover opening monitor is opened or when a dose of medication is removed from a dose removal monitor,

2) A beeper which is used to produce two different audible tones which are picked up by the microphone of a mobile phone during a data transmission session, and

3) A cradle, which positions the patient’s mobile phone's mouthpiece near the beeper during the data transmission session. The details of the cradle and how it holds the mouthpiece close to the beeper needs to be worked out. It probably will be consist of a relatively simple receptacle that is adjustable for different size mobile phones with a switch to turn on the beeps and tones. It may include Velcro straps to hold the phone in position.
 
The data transmission sessions would normally be held weekly. The patient and the clinic staff would decide on the most convenient time for contacting one another. The one-week time period should be adequate for initiating corrective action if poor adherence occurs. If the patient has consistently good adherence records, the time between data transmission sessions could be extended to every 2 weeks. Either the patient or the clinic staff could initiate the phone call.

Before starting the session, the caregiver would have to make sure he/she was communicating with the correct patient and monitor. This could be accomplished in several ways. If the caregiver called the patient and the patient responded with the right name, the caregiver would be reasonably sure that he was communicating with the right patient unless the caregiver made a mistake and called the phone number of a patient with a similar name. If the patient called the clinic, misidentification because of similar names would be more likely if the patient called the clinic. To avoid such a mistake the monitor could have a sticker that displayed the patient’s clinic number and the patient would be told to report this number to the caregiver before starting the transmission session.
If identifying the patients proves to be a significant problem, it would be possible for the medication monitor or the patients phone to generate the patient’s number with coded audible tones like the tones one hears when dialing a phone number. This would entail a modest additional expense.

Before listening to the adherence record the caregiver would dictate the patient’s name and clinic number into a digital voice recorder (2), which is clipped to the clinic’s mobile phone near the earphone. Subsequently, the caregiver would tell the patient to place his or her mobile phone in the receptacle on his or her monitor. This would activate the switch to turn on a data transmission session consisting of a sequence of distinctive one-second tones and brief beeps. The one-second tone would represent midnight for each day. The beep would represent the opening of the cover of a cover opening monitor or the removal of a dose from a dose removal monitor.

If the patient was using a cover opening monitor and opened the cover every day since the last data transmission session, the caregiver and digital voice recorder would hear a one-second tone for the first midnight since the last transmission session, followed by a beep for opening the cover the next day, followed by a one-second tone for the next midnight, followed by a beep for the next day and so on for all the midnights and days since the last data transmission session. For each day the patient failed to open the cover there would be no beep between the one-second tones representing midnights. The number of times there were no beeps between the one-second tones would represent the number of days that the patient didn’t open the cover. If the one-second tones were separated by 3 seconds, the entire adherence record for one week could be transmitted in 29 seconds.

To help the reader of this description to better understand this method of transmitting tones and beeps I will try to present it in graphic form. The tones and beeps would be represented as follows: ( _ ) for the one-second tone and ( . ) for the beep. If the care giver and digital voice recorder heard

_  .  _  .  _  .  _  .  _  .  _  .  _  . _  The patient opened the cover each  day.
_  .  _     _     _  .  _  .  _     _  . _  The patient failed to open cover  on the 2nd, 3rd,and 6th day.   
_     _     _     _    _    _     _    _    The patient failed to open cover  for the entire week.

After hearing the adherence record, the caregiver would be able talk to the patient giving praise for a good adherence record or expressing concern about poor adherence while stressing the importance of uninterrupted treatment and letting the patient express his concerns. While this process takes staff time and transmission time that must be paid for, the maintenance of human contact with the patient even though it is only voice contact should prove to be helpful in maintaining rapport and improving adherence of most patients. Furthermore, phone reinforcement of instructions for patients with minor defects in the adherence record would take far less staff time than a home visit.

In addition to listening to the adherence record at each transmission session, the clinic would want to keep a permanent adherence record from the beginning of therapy for each patient to decide on the best way to manage patients with good, modestly poor and bad adherence records. Such management could include 1) free cell phone minutes for good adherence records, 2) periodic home visits to reinstruct the less adherent patient and family, 3) ordering strict DOT from someone outside the family for the non-adherent patient, and 4) extending the duration of therapy based on the adherence record.
Creating the permanent record would take two steps.  First, the digital voice recorder that is clipped to the clinic’s mobile phone close to its earphone would create a permanent audible record.  One such digital voice recorder, which also creates a convenient time stamp, is shown on the website (2). The next step would be to create a paper or computer record in written or graphic form.

The simplest low-tech way to achieve a permanent record would be for a staff member to listen to the digital voice recording and enter the adherence record on the appropriate form in the patient’s chart.  If the clinic had a computer the data could be typed in.  This task could be carried out immediately after listening to the transmission record, or at some later point when some staff member has a stretch of uninterrupted time.  I have been told by Mr. Bill Thies, bthies@gmail.com, who works for Microsoft that a program could be written that would convert the tones and beeps into a written or graphic record in a computer.

If the health department has to provide a mobile phone for a patient without a phone, the patient might sell the phone and claim he lost it. Fastening the phone to the monitor with a cable lash would reduce the chance of this occurring.
 
If for any reason mobile phone transmission of the data is not carried out, the monitor’s distinctive tones and beeps can be played when the patient returns to the clinic for a refill of medication to obtain the adherence record. Consequently, no display such as a dot matrix display, discussed in reference (1), would be needed unless the monitor is designed to keep and display clinical information.
This data transmission system could be used for a dose removal monitor with one modification, a longer interval between distinctive tones.  This would allow for additional beeps between distinctive tones when the patient removes more than one dose in a 24 hour period which needs to be noted when patients use dose removal monitors as discussed in reference (1).
There is at least one other possible way the monitor record could be transmitted by a mobile phone. Mobile phones with cameras are almost ubiquitous in the USA.  On Amaazon.com and mobilekarma.com a used camera mobile phone can be found with cost in a $12.00 to $14.00 range. These phones could be used to take a picture of the dot Matrix display of the adherence record shown in the published article (1) or see the description above on this page. Whether a sufficient supply of reliable camera mobile phones can be found in an acceptable price range needs to be determined.

References:

1) Moulding T. A neglected research approach to prevent acquired
drug resistance when treating new tuberculosis patients, INT J TUBERC LUNG DIS 15(7):855–86

2) Website for digital voice recorder that costs $40.00 but is used only in the clinic.  http://www.musiciansfriend.com/pro-audio/rca-vr5320r-1gb-digital-voice-recorder-with-voice-management software/h72408000000000?src=3WFRWXX&ZYXSEM=0&CAWELAID=800589643


VIII. Retrieval of clinical data stored in the monitor

When a patient moves and is treated by a different clinic than the one which initiated his therapy, when a patient begins working with a different caregiver, or when clinical records are lost, it would be useful to have basic clinical records stored in the monitor. This information might  include the patient’s HIV status and sputum status for each clinic visit.

For the caregiver to be able to retrieve this clinical data through the monitor’s display, the monitor would generally need to have either a matrix display or a character display (e.g. 2-digit display).

Rather than being required to step through the complete adherence record to get to the display of clinical data, the caregiver should be able to access this data by entering some special code. In the case of a monitor having a separate caregiver’s button, as described above, this code could be as simple as simultaneously pressing the caregiver’s button and the patient’s button.

The following is an example of how the clinical data might be presented by 5 x 7 matrix display:

Upon entering an appropriate code the display would show the HIV status of the patient. H with a P, if HIV positive, H with a negative sign, if HIV negative, and H with nothing, if the HIV test was not done.

V. Display incorporated into the monitor

Incorporating a visual display into the monitor could enable the monitor to remind the patient when medication should be taken, respond to a simple query by the patient, and offer a means for the caregiver to retrieve adherence records and clinical information directly from the monitor. (See following sections for details.)

There are a many types of displays which could be used to achieve these objectives. The monitor’s display needs to be  relatively inexpensive and have modest power consumption.  We have considered three possible displays:

A) Single red/green LED.

A red/green LED incorporates two light emitting diode elements, one of which produces red light and one which produces green light. When both are illuminated together, the resulting illumination appears yellow. These three colors can be used to inform the patient  whether he should take medication  or not  (e.g. red or green) and to inform the caregiver of the condition of the monitor’s battery, the patient’s overall adherence (good, fair, or poor), and whether the patient removed medication on a particular day.

One inexpensive red/green LED (as of June 24, 2010) is:

JAMECO VALUEPRO Manufacturer’s no. LHG3392         ($0.12 each in thousands)

B) Character display

Character displays use line segments to form characters (numbers or letters).  Both LED and LCD (liquid crystal display) character displays are available. LED character displays are available for displaying a single character or a small number of characters with each display module, while LCD displays are generally designed to show a larger number of characters. Where only a few characters are needed, LED displays are more economical.

Some character displays incorporate enough segments to fully form an alphanumeric character set including all digits and all letters of the alphabet. Others, such as those normally used in calculators, are designed to display only digits, but can also display a limited set of letters. These normally have seven segments for each digit.

A 2-digit seven segment display like that shown at the right would be adequate for use in a medication monitor. Such a display has the advantage, relative to the single red/green LED, of being able to show numeric values. For example, the 2-digit display could show the number of pills a patient should take, the number of doses remaining in the dispenser, or the percentage of days on which the patient removed medication. The 2-digit display can also show a limited number of letters, A, b, c, E, F, H, J, L, n, P, r, and U, which can be useful if clinical data is to be retrieved from the monitor. (See:  Retrieval of clinical data stored in the monitor below.)

The 2-digit display, however, does not share the red/green LED’s ability to use different colors.

One inexpensive 2-digit display (as of June 24, 2010) is:

LITE-ON  Manufacturer’s no.  LTD-482P-YQ          ($0.89 each in thousands from Jameco Electronics)

C) Red/green LED Dot matrix display

A dot matrix display consists of a rectangular array of display elements, commonly LEDs. Dot matrix displays are made in several patterns form 4x4 up to 16x16 or larger. The most common configuration is a 5x7 array which is designed for displaying any single letter or digit as well as numerous other dot patterns. Among these other patterns are simple displays of 2 digits equivalent to those possible with the 2-digit  character display. While many matrix displays use a single color, 5x7 matrix displays are also available with a red/green LED for each dot.

A red/green matrix display combines all of the functionality of the red/green LED with that of the 2-digit display. Furthermore, the 5x7 matrix display seems particularly well suited for displaying a patient’s detailed adherence record since each of its 35 dots can represent one day; and the color of the dot can represent medication removed (green), medication not removed (red), or several doses removed (yellow flashes), as described in the section Using the monitor’s display to retrieve the adherence record below.

One inexpensive red/green matrix display (as of June 24, 2010) is:

LITE-ON  Manufacturer’s no. LTP-2757AA          ($1.75 each  in lots of  500 from Jameco Electronics)

With the next push of the button the results of the first sputa test would appear: S with a P if the patient was sputum positive. S with a negative sign, if the patient was sputa negative, and S with nothing if no sputa test had been done.

With subsequent depressions of the button the sputa status in subsequent months would be given for each of the months since the patient was started on medication. Since sputa tests are usually not done every month, most of these displays would show S with nothing. A patient, who had a positive sputum the first month, no sputa for 3 months and negative sputum on the last month, would have the following sequence of displays.

d+ for known drug resistance (red), d - for known drug susceptible (green), and d 0 for no drug susceptibility test done (yellow).

IX. Further notes on information retrieval

A) Ending the caregiver’s information retrieval session

When the caregiver has finished reviewing information from the monitor, the monitor function needs to return to “patient mode” where pressing the patient’s button elicits an indication of whether the patient should take medication, or not. The caregiver could end her session by pressing and holding the button by which she initiated her display. When the monitor had returned to “patient mode” the caregiver’s display would turn off. If the caregiver failed to return the monitor to “patient mode” after some predetermined time (e.g. 2 minutes), the monitor should automatically return to “patient mode.”

B) Additional display functions

Obviously, additional information could be shown with the monitor’s display.

One example of additional information which could be useful would be an assessment of how consistent a patient was in taking medication near the same time each day. There is some evidence that patients who take medication near the same time each day tend to be more consistent in taking their medication every day rather than missing occasional days. This potentially useful assessment could be represented by different color lights or as a numeric value, or both.

For any additional information shown, decisions would need to be made as to how the information would be presented with the display and how it would be accessed.

TITLE PAGE and ABSTRACT

Background and Optimal Requirements for Medication Monitors for TB Treatment

I.   Cover and Cap Opening Monitors

II.  Dose Removal Monitors

III. Dose Removal Monitors Developed by Other Groups

     A) E-box

      B) U-box

IV. Devices That Prove or Come Close to Proving Medication Ingestion

V. Display incorporated into the monitor

VI. Means to instruct and remind the patient

VII. Retrieval of the adherence record

VIII. Retrieval of clinical data stored in the monitor

IX. Further notes on information retrieval

____________________________________

A Neglected Research Approach to Prevent Acquired Drug Resistance When Treating New Tuberculosis Patient (PDF)

Adapting to New International TB Treatment Standards with Medication Monitors and DOT given selectively