A) Optimal Medication Monitor(s) and Method of Retrieving Adherence Record.
B) Other Dose Removal Monitors.
Setting the Time
For all the alerting and data retrieving functions, the monitor will have to know the time of day. To make sure the monitor has the proper time of day entered, the clinic worker opens the caregivers cover and enters a time setting mode by pressing the patient's query button while the caregiver’s button is depressed. The LED would flash a number of times corresponding to the present hour as the monitor's time is currently set. For example, if the monitor's time is 5 o'clock, the LED would flash 5 times slowly enough for the clinic worker to count the flashes. The color of the flashes would indicate whether the time was A.M. or P.M., green for A.M. or red for P.M. If the LED flashed 5 times and 5 o'clock was correct (4:30 to 5:30), the clinic worker would press the caregiver’s button once to confirm 5 as the correct time. If 5 o'clock was not correct, the clinic worker would press the patient's query button the number of times corresponding to the correct time, for example 3 times if the time was 3 o'clock (2:30 to 3:30). The LED would then flash 3 times slowly enough for the clinic worker to count the flashes. If the clinic worker had made an error and the number of flashes was incorrect, he or she could press the patient's query button again the correct number of times to get the correct hour. Once the hour was correct, the clinic worker would press the caregiver's button to confirm the time.
Once the time had been confirmed, the LED would then continue to show the color for A.M. or P.M. If the color indicated the correct AM or PM setting, the clinic worker would again press the caregiver’s button to confirm the correct setting. If the AM / PM setting was incorrect, the clinic worker would press the patient's query button once to change an AM to PM setting or a PM to AM setting and, after seeing the correct color the clinic worker would press the caregiver’s button to confirm the AM / PM setting. At this point the monitor would leave the time setting mode.
Retrieval of the Adherence Record with Audible Tones in the Clinic or Transmitted to the Clinic by Mobile Phones.
Note: This means of retrieving the adherence record has the advantage that it can be used in areas with and without mobile phone service, by community workers in the patients' home, and in clinics with and without computers.
Medication monitors can be used without mobile phone transmissions between the patient and the clinic by retrieving the adherence record when the patient returns to the clinic or by having a community worker retrieve the record in the patient’s home. However, corrective action for poor adherence should be more effective if it was known as soon as possible after it occurred. To achieve this objective mobile phone transmission of the adherence record needs to be considered.
Potential of Using Mobile Phones to Send Monitor Record
In developing nations, 79 percent of the population has a mobile phone, according to a 2011 report from the International Telecommunications Union. Wikipedia lists the the usage in each of 62 countries, http://en.wikipedia.org/wiki/List_of_countries_by_number_of_mobile_phones_in_use
This suggests an opportunity for using mobile phones to improve adherence. There are 243 references to experience using mobile phone reminders to HIV patients taking antiretroviral drugs, (usually text messages) but only two randomized control trials (1). There was improved adherence with both trials, and improved viral load suppression in one trial (1),
Attempts to improve medication adherence of tuberculosis patients based on sending a text or voice message to the patient to remind the patient to take medication followed by a text or voice message reply to the caregiver have been described (2) and have been systematically studied in Thailand (3). However, even if a patient consistently responds to reminder messages, the clinic staff has no idea if the patient has taken the medicine or not.
The effectiveness of mobile phones could most likely be improved if used to send the adherence record from a medication monitor to the clinic at short periodic intervals, probably once a week, during which varying degrees of poor adherence may occur. This data will guide the clinic in taking appropriate corrective action. Furthermore, the mobile phone transmissions to the clinic could be used to supervise both-self administered medication and medication given to the patient by a supporter, usually a family member, i.e. family DOT as described in Use of Monitor Record to Improve Adherence. Private physicians could use the transmissions to improve the treatment as discussed in section Improving Private Sector Treatment. For patients with very poor adherence who require Directly Observed Therapy from a neighbor the monitor record transmitted by mobile phones could be used to Supervise DOT as discussed in Use of Monitor Record to Supervise DOT Given by a Neighbor
Possibility of Having a Direct Connection Between Medication Monitors and Mobile Phones.
We recognize that having some form of direct connection between the medication monitor and the mobile phone with a cable would be the most convenient way to use mobile phones. So far we have been unable to create such a system because most mobile phones do not have a port for an external cable.
We have recently learned of one mobile phone with a USB port that may be in a reasonable price range.
Spice M-4262 - Price: Rs.999. or about $18.00.
The development and cost of a data transmission system that uses a phone with a USB port needs to be seriously considered while keeping in mind that a large number of patients in developing nations currently have mobile phones without USB ports that could be adapted for sending the adherence record as described below.
Possibility of Using Existing Phones Without USB Ports
If a medication monitoring system uses mobile phones, the health department will want to keep the costs as low as possible by using the phones owned by the patients. These phones rarely have a USB port. Furthermore, health departments will have to supply inexpensive mobile phones to many poor patients who do not have mobile phones.
We have been told that a used mobile phone without a USB port costs $10.00 in India. (Personal communication William Thies email@example.com ). An organization in the USA called the Charitable Recycling Program collects and recycles old mobile phones (4). While only 35% of these phones are useable, this still represents much more than a sufficient number for TB patients, since there are only 9 to 10 million new TB patients in the world each year compared to 5.9 billion mobile phone subscriptions at the end of 2011. As more and more sophisticated phones are introduced, the relatively inexpensive simple mobile phones that are still usable are often discarded. Therefore, a sufficient supply of second hand usable mobile phones could most likely be obtained with an adequately advertised collection campaign.
Unfortunately, many areas with mobile phone service do not have electric power to recharge the phone. Perhaps recharging the patient’s mobile phone when the patient returns for a refill of medication could solve this problem. However, this would only work if the phone could be modified so that it could only be used for sending the adherence record since extensive use for other purposes would deplete the battery. Perhaps the mobile phone could be recharged with a solar collector. Looking toward another alternative, a Gates foundation website describes a grant for developing an inexpensive recharging system based on naturally occurring soil microbes. (5)
Transmitting the adherence record from medication monitors audibly with mobile phones that do not have a USB port.
If it is not possible to develop a direct connection system using a USB port that is cost effective, the following system based on sending audible tones is proposed as a less expensive way to use mobile phones. This proposed 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 the Sections I and II of this website, the Background section, and in a published article. (6)
Equipment needed for each patient with proposed system.
While the following material refers to sending the adherence record from a Clip Monitor it could be adapted to any medication monitor. The proposed 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 buzzer 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 rack, that holds the medication monitor and positions the patient’s mobile phone's mouthpiece near the buzzer during the data transmission session is shown in Figure 1. The location of the encircled X on the rack should be placed directly under where the buzzer on the monitor will be located. The patient would be instructed to place the mobile phone at the bottom of the rack with its microphone facing upward but located over the encircled X symbol. This positioning of all elements would place the microphone as close to the buzzer as possible to improve the transmission of the tones. The upper level of the rack includes an electrically conductive surface shown in yellow that is interrupted by a switch controlled by a button to activate a data transmission. The space between the bottom and elevated portion of the rack should be chosen to accommodate all reasonable sized mobile phones
Figure 1. Rack for medication monitor and mobile phone for transmission session.
Figure 2. Clip monitor showing support flange and electrical contact strip.
Figure 3. Monitor positioned in rack for mobile phone transmission session..
Figure 2 shows an image of a clip monitor with a flange on one end and a thin strip of yellow electrically conductive material running from the circuit board inside the monitor to the flange. There is a similar flange and electrically conductive material on the other end of the monitor.
Figure 3 shows the monitor turned over with its flanges supporting it. The thin electrically conductive strips make contact with the electrically conductive material on the rack to create a circuit when the switch button is pushed.
Transmission Sessions Weekly or Every Two Weeks.
The adherence 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 the caregiver to take 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 for the transmission session.
Ensuring Data from the Intended Patient’s Monitor is Being Read.
Before starting the session, the clinic worker 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 clinic worker would be reasonably sure that he/she 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. 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 through the mobile phone 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 patient’s 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 clinic worker would dictate the patient’s name, clinic number, date and day of week into an inexpensive digital voice recorder (7) and arrange for the digital voice recorder to record the adherence record while the clinic worker listens to the record. Using a commercially available mobile phone with a speaker that provides a sufficient volume for both the voice recorder and clinic worker to hear the tones and beeps would achieve this. Subsequently, the caregiver would tell the patient to place his or her mobile phone in the rack, and press the button on the rack which completes the circuit between the two bands of electrically conductive material and turns 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. In case there prove to be problems getting the digital voice recorder activated before the monitor starts sending beeps and tones, the medication monitor could be programmed to delay sending the beeps and tones for short periods of time, like five seconds, after the patient depressed the switch button.
Using Red LED and possibly a Buzzer to Alert Patient to Take Medication
The monitor is built with a Red/Green LED for certain functions. However for alerting the patient to take Medication only the Red LED would be used.
Starting in the morning, perhaps 4:AM, the LED could emit two 150 millisecond red flashes every second until the patient removes medication or until midnight.
As soon as the patient removes a dose of medication the LED would go dark
A set schedule to achieve these flashes could be put into the electronic instructions automatically when the time is set.
A limitation in using an LED to remind the patient is that the monitor may be kept in a location where it cannot be seen. To overcome this problem one might think a buzzer could be activated to remind the patient to take medication. Unfortunately, using a buzzer for alerting the patient may turn out to be a two-edge sword. If the buzzer sounds and the patient is not home some family member may be annoyed, remove a dose to turn off the buzzer, and forget to give the medication to the patient when he or she returns home.
One way around this problem would be for the buzzer to chirp at times when the patient is likely to be home. These times could be for 15 minutes in the AM, and again for 3 hours starting at 9:00 PM if no medication had been removed that day. A set schedule of this nature could automatically be put into the electronics when the time is set. However, for patients with irregular work hours, such a pre set schedule would not work. By using special codes when pushing the buttons it may be possible to create the appropriate schedule. However, potential confusion over use of the codes and annoyance of the family caused by the buzzer could create sufficient problems that it would be best not to use the buzzer to alert the patient.
Using Computers to Store and Display Adherence Record
Many clinics in both developed and developing countries have computers. When computers are available and working properly, adherence data could be downloaded to the computer from a USB port on the digital voice recorder and directly from the monitor when it is refilled in the clinic if a 50¢ USB port is included in the monitor. Furthermore, many newer computers have microphones which could directly accept the data from the beeps and tones emitted by the clinic’s mobile phone or the monitor itself. Using a computer to record and display the data would greatly reduce the effort required to interpret the adherence record.
The computer could use the format previously discussed for the Hard Copy of the Adherence Record. The record could be updated and displayed at various times 1) each time there is a mobile phone transmission session, 2) when the patient returns for a refill of medication, and 3) when the staff reviews the record to decide on corrective actions if there is poor adherence.
On the other hand, a village community worker would rarely have a computer, especially when visiting patients in their homes. Therefore, it will be extremely desirable for the community worker to be able to view the patient’s compliance record. This could be accomplished by viewing an LED display mounted in the monitor as described in Retrieval of Adherence record in Patient’s home using red/green LED or by listening to audible tones as described in
Retrieval of the Adherence Record with Audible Tones in the Clinic or Transmitted to the Clinic by Mobile Phones.
Checking the Battery Status
The clinic worker checks the batter status by opening the battery compartment and pushing the caregiver’s button once. This determines the battery status with the LED showing green for fully charged, yellow for partially charged, red for close to empty and in need of replacement. Note: The LED will admit a yellow color if the red and green illuminations occur at the same time. One inexpensive red/green LED for this purpose is JAMECO VALUEPRO Manufacturer’s no. LHG3392 that costs $0.12 each in thousands (as of June 24, 2010).
Retrieval of Adherence record in Patient’s home using red/green LED
For patients who live in an area without mobile phones, for patients who do not have a mobile phone, or for poorly adherent patients who need home visits for additional education and support, a community worker will need to visit the patient at home. At that time she will need to have a way to retrieve the adherence record. This could be done in the same way the clinic worker listens to and records the audible tones and beeps generated by the monitor. However, the red/green LED that is on the monitor, which also creates a yellow flash when both the red and green LEDs are illuminated simultaneously, can also display the adherence record. This display may be easier to for the community worker to understand.
To create this display, the monitor would be placed in the rack with the medication facing down and the red/green LED facing up as shown in Figure 4. This would lead to contact between the conductive strips on the rack and the surfaces of the monitor’s flange that direct the microprocessor to activate the red/green LED. Note that the conductive strips on the monitor’s flanges which contact those on the rack are on the opposite side of the flanges from those which are used to activate the transmission by beeps and tones and complete a different circuit.
Figure 4. Monitor positioned in rack for visually presenting adherence record.
When the community worker pushes the switch button on the rack, the red / green / yellow LED would emit a sequence of flashes that show the adherence record. Yellow would be emitted for each midnight, green for each dose removed between yellow midnight flashes and red for each day the patient fails to remove a dose between the midnight flashes. If the community worker sees multiple red flashes he or she would quickly realize the patient was poorly adherent.
Since the community worker would probably want to spend considerable time with the patient and show the patient and family the adherence record, he or she would probably want to record the removal of 30 doses of medication using a form with red and green boxes shown below.
The record would start with the first day of the portion of the adherence record being transmitted. For each day the patient had taken one dose of medication, the community worker would see a green flash between the yellow midnight flashes and would put a dot or slash mark in a green box starting at the left hand upper corner of the form (the conventional way or writing). If the patient removed two, three, or more doses in one day, the worker would enter the corresponding number of dots or slash marks or write a number in the green box. For each day the patient failed to remove medication the LED would flash red and the community worker would place a dot or slash mark in the next red box instead of the green box. If the patient had removed all the medication from the monitor several days before the community worker visited the patient the community worker would put a number corresponding to the number of days in the red box at the end of the record. If the community worker gave the patient a dose when he or she was in the patient’s home she would put a dot or slash mark in the last green box. The following example record illustrates these points:
1) The patient took out one dose per day for three days
2) Then failed to take out a dose for three days
3) Then took out one dose per day for two days
4) Then took out four doses on one day. The patient probably ingested one of the doses, and took out the other three doses to catch up in and attempt to cover up the fact that he or she had missed taking medicines for three days.
5) Then took out one dose a day for 6 days.
6) Then took out four doses one to ingest and three to take in the next three days while away on a trip. During those three days he took out no doses from the monitor..
7) Then took out one dose for 4 days
8) Then missed taking out doses for 4 days
9) Then took out one dose a day for 8 days after which the monitor was empty
10) Then he failed to return to the clinic for 9 days which created the need for a community worker’s visit.
11) When the community worker made the visit the patient was given one dose.
Showing the record to the patient and the patient’s family should be helpful in impressing on them the importance of regular medication ingestion in addition to returning to the clinic for a refill of medication and consultation with the clinic staff.
If the officials in charge of the program want the community worker to keep a permanent record on the patient, multiple rows of these red and green boxes would be placed on one page for each patient and kept in the community worker’s notebook. Note the notebook would also have a compartment for keeping a rack with an electrically conductive strip for those patients who do not have mobile phone service with the clinic.
In addition it should be noted that the clinic staff may find that this red/green/yellow flash method of recording the adherence record more convenient than the dash and beep method when the patient returns to the clinic.
Creating the Permanent Hard Copy Record
While it is possible to consolidate a complete hard copy record from short hard copy records made at each transmission session (a few days to 2 weeks), this can get complicated. A simpler low tech (non computer) way of creating a permanent record would be to have the monitor upload the entire adherence record since the last time the monitor was filled each time the patient returns for a refill of medication. To achieve this, the monitor needs a simple, unique signal to instruct it to transmit this entire record rather than just the record since the previous time the record was sent. The circuit which normally instructs the monitor to send a record by LED flashes along with the circuit which normally instructs the monitor to send a record by beeps and tones can be used together to create this simple, unique signal if both circuits are completed at nearly the same time. The monitor rack shown in Figure 5 could be used to complete both circuits nearly simultaneously. In this rack, each of the flanges on the monitor slides into a slot which incorporates 2 spring contacts which contact the conductive strips on the 2 sides of the flange. These spring contacts would be wired to the switch incorporated into the rack so that pressing the switch button would complete both circuits. Nearly simultaneous completion of both circuits would be a signal to transmit the entire record since the last refill.
Possibility of using Camera Phones
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 DigiCircle.com and mobilekarma.com a used camera phone can be found with cost in a $12.00 to $20.00 range. These phones could be used to take a picture of the dot Matrix display of the adherence record shown in the published article (4) or the section of the website Using a Dot Matrix display to retrieve the adherence record. Whether a sufficient supply of reliable camera phones can be found in an acceptable price range needs to be determined.
Another obstacle in using such a camera phone is whether the mobile phone networks in developing countries are capable of transmitting the image data, and what additional costs may be involved in transmitting the pictures.
Addendum: Alternate Probably Less Optimal Means of Activating the Transmission of the Adherence Record.
With the system previously described which uses a button on the rack that holds the cell phone to initiate the transmission of the adherence record, there might be problems in making electrical contact between one or both of the thin electrical strips on the monitor and the electrically conductive surface of the rack. If this occurs, one might consider placing an additional button on the monitor itself for the patient to initiate adherence transmission sessions. Normally, when such a button was pushed the adherence data would start the first day since the last data transmission session.
However, a patient or a child might push the button when he/she was not in mobile phone contact with the caregiver. If this occurred any subsequent pushing of the button would not reveal the compliance record starting with the first day since the last time the data was transmitted to the caregiver.
This problem could be dealt with by telling the patient to quickly push the button two times to transmit the adherence record for the last two weeks, three times for three weeks, or four times for 4 weeks.
Because of these potential problems we believe the system of having the data transmission button on the rack as described in Figures 1,2, and 3 is preferable. To avoid problems in making electrical contact The contact strips on the monitor and the areas of the conductive strips on the rack which they contact should be gold plated in the same way that contacts on highly reliable low current electrical connectors are plated to minimize corrosion and improve electrical contact.
Figure 5. Monitor positioned in rack which makes contact with both sides of the monitor’s flanges.
Since it may be easier for the clinic worker to interpret and record when the adherence information is transmitted by LED flashes, the monitor would transmit these flashes when the clinic worker presses the switch button on the rack. The clinic worker could then transcribe the record onto a form like the following.
While the monitor holds 15 doses of the four-drug combination and 30 doses of the two-drug combination, the form provides 45 spaces for each recording. The additional spaces accommodate extra marks in the red rows for those occasional times when a patient is late in returning for a refill of medication after the last dose was taken.
The patient may or may not have removed a dose of medication from the monitor on the day he or she returned to the clinic. If the patient had not removed a dose the clinic worker should give the patient a dose to ingest, either by removing a dose from the monitor if any doses remain or by taking a dose from the clinic supply. When this dose is taken, the clinic worker would add a dot to the record’s green row for the present day, as described more fully above in the section Retrieval of Adherence record in Patient’s home using red/green LED.
Some clinics may have computers which have been set up to receive adherence records transmitted by beeps and tones. For those clinics, it would be desirable to to have the monitor transmit the adherence record since the last refill by beeps and tones for direct communication to the computer. Therefore, the monitor should transmit the record since the last refill both by LED flashes and by beeps and tones.
Having this record transmitted by beeps and tones as well as by LED flashes also provides the clinic worker with the option of recording the beeps and tones on a digital voice recorder to create a recording from which the hard copy record could be transcribed at some later time, if needed.
Patterns Heard when the Adherence Record Is Transmitted
There are several ways of transmitting the tones and beeps. They could start at the first day of the interval being transmitted and work forward towards the last day of the interval, or start at the last day of the interval and work backward to the beginning of the interval. The following description is based on the beeps and tone starting the first day of the transmission interval.
If the patient was using either a cover-opening monitor or a dose-removal monitors the first sound would be the first midnight tone for the interval being transmitted. Subsequently, the clinic worker would hear a series of beeps and midnight tones. If the cover had been opened or a dose of medication had been removed every day during the interval the record could be graphically shown in the following manner.
__. __ . __ . __ . __ . __ . __ . __ . __ . __ . __ .
If the patient failed to open the cover or failed to remove a dose from a dose removal monitor for one or more days, [non removal day(s)] there would be no beeps between the midnight tones on the days when the patient failed to open the cover or remove medication as shown below.
__ . __ __ __ . __ . __ __ __ __ . __ . __ . __
If the patient was three days late in returning for a refill and had run out of medication the sequence of tones and beeps would occur as follows.
__ . __ . __ . __ . __ . __ . __ . __ __ __ __
If the patient or a member of the family opened the cover of a cover-opening monitor two or more times or removed two or more doses from a dose removal monitor on one day, [an excess removal day], there would be two or more beeps between each midnight tone as shown below.
__ . __ . __ . __ ..__ . __ .… __ . __ . __ . __ .
In the case of a cover-opening monitor an excess removal day may or may not indicate that multiple doses were removed on that day.
In the case of a dose removal monitor an excess removal day clearly shows that more than one dose was removed on that day and may or may not be associated with poor adherence as shown in the following two examples.
Catch up Removals in Attempt to Deceive Caregiver
When a patient fails to remove medication for one or more days, he might attempt to deceive the caregiver and remove the doses he had forgotten to take at a later point in time to “catch up”. If so the adherence record would appear as follows. The three days with no dots were non-removal days. The four dots on the subsequent excess removal day suggests the patient took out one dose to ingest and three doses to catch up. Since the non-removal days precede the excess removal catch up day, the excess removals most likely do not represent doses that were ingested.
__ . __ . __ . __ __ __ __ . __ …. __ . __ .
Advanced Removal of Pocket Doses to be Taken Later
On the other hand a patient may remove multiple doses on one day to have medication to take while away on a trip (pocket doses) without taking the medication monitor on the trip. In this case the excess removal days are called advanced removal days and precede the non-removal days as shown below. This type of record can also be seen if a patient accidentally removes a strip of medication and reinserts the medication-containing strip back into a clip from which medication had been removed. See: Clip Monitor for WHO’s Packaged Medications.
__ . __ . __ …. __ __ __ __ . __ . __ . __ .
When this record is found the patient probably ingested only one of the four doses on the excess removal day and saved three doses to take while away on a trip. When excess removal day(s) precede and are close to the non-removal day(s) the probability that these doses were intentionally removed to be taken later is increased. While there is no perfect way to identify advanced removal days we suggest using the following definition. When excess removal days precede non-removal days by one, two, or three days they are probably advanced removal days,
Immediate Recording of Record if Necessary
The clinic worker should have no difficulty interpreting the adherence record if there are no days on which multiple beeps are heard. When multiple beeps occur on one day the clinic worker may need to listen to the record again by playing the record on the digital voice recorder and entering the record on the following form which has the midnight tones displayed as vertical lines with appropriate spaces between the lines for the worker to enter a beep with dot(s) or slash marks(s),
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 or monitor rack with a cable lash would reduce the chance of this occurring.
Slot to receive monitor flange
1) Horvath T, Azman H, Kennedy GE, Rutherford GW. Mobile phone text messaging for promoting adherence to antiretroviraltherapy in patients with HIV infection. Cochrane Database of Systematic Reviews 2012, Issue 3.
2) Barclay E Lancet 2009 373 2009 35-36 Text messages could hasten tuberculosis drug compliance
3) Kunawararak P, et al Tuberculosis Treatment with Mobile-Phone Medication Reminders in Northern Thailand, Southeast Asian J Trop Med Public Health Vol 42 No. 6 November 2011 1444-1452.
5) Gates foundation website describing research in obtaining electrical power from soil microbes. http://www.seas.harvard.edu/news-events/press-releases/gates-grant
6) Moulding T. A neglected research approach to prevent acquired
drug resistance when treating new tuberculosis patients, INT J TUBERC LUNG DIS 2011;15(7):855–86