NEW UPDATE: Completed trials may now upload and register supplementary documents (e.g. null results reports, populated pre-analysis plans, or post-trial results reports) in the Post Trial section under Reports, Papers, & Other Materials.
The Impact of Packaging and Diagnosis on Adherence to Malaria Treatment in Uganda
Initial registration date
September 03, 2014
September 03, 2014 12:04 PM EDT
Harvard School of Public Health
Other Primary Investigator(s)
Harvard School of Public Health
Additional Trial Information
The failure of patients to adhere to recommended treatment guidelines is a major driver of widespread pathogen resistance, making diseases like malaria, pneumonia and HIV increasingly difficult and expensive to treat. Currently, Artemisinin Combination Therapies (ACTs) are the only fully effective treatment remaining for malaria. Although ACTs have a short 3-day dosing regimen, over 35 percent of patients do not complete the full course of drugs. We conducted a randomized controlled trial in central Uganda with 2,500 households, designed to understand the reasons for poor adherence to the ACT treatment regimen. We also experimented with specially designed packaging and targeted messages to boost adherence. We find that the strongest predictor of adherence is how the patient is feeling (their symptom severity) when they are halfway through the treatment course. We hypothesize that patients who feel better mid-course assume their malaria is cured and discontinue treatment. Consistent with this hypothesis, a sticker affixed to standard ACT packaging that informs people that “malaria is not gone until all tablets are finished” significantly (though modestly) increases adherence, particularly for those patients whose symptoms were resolving early. On the other hand, a message designed to discourage saving pills for future malaria episodes had no significant effect on adherence. We also test a common approach to increasing adherence to ACTs in Africa by using specialized packaging that includes pictorial instructions for illiterate patients, and information designed to raise patients' confidence in the effectiveness of the medication. While this special packaging increases the cost of ACTs by 10 to 50 percent, we find that it has no significant effect on medication-taking behavior or on comprehension of instructions.
Cohen, Jessica and Indrani Saran. 2014. "The Impact of Packaging and Diagnosis on Adherence to Malaria Treatment in Uganda." AEA RCT Registry. September 03.
The objective of the project was to assess the impact of various forms of packaging and messages associated with the ACT drugs on adherence to the treatment regimen. To evaluate this, we randomized the type of packaging/messages individuals received each time they came to the participating shops to purchase ACTs.
The control package in this study was the standard package in which ACTs (specifically Artemether Lumefantrine, brand name Lumartem) were sold in Uganda and elsewhere in Africa. The box had the name, brand and manufacturer of the medication. Inside the box was a blister pack which grouped the pills by dose and day and a paper insert, similar to what is seen inside most medication boxes in the U.S. and elsewhere, which have small print with details about dosing, side effects, etc. The first treatment arm tested the status quo approach to promoting adherence through specialized packaging. It was known as the "CAPSS" package because it was used in Uganda during the Consortium for ACT Private Sector Subsidy pilot program (run by the Uganda MOH, Medicines for Malaria Venture, Population Services International and others). The CAPSS packaging served as a form of branding and quality assurance, and it was also designed to encourage correct use of the product, incorporating features like colorful pictorial instructions on how to take the medicine, principally to assist illiterate patients and caregivers in taking the medication correctly.
The second treatment arm was designed to test, if the CAPSS packaging was successful at boosting adherence rates, whether the improvement was due mostly to the information and pictorial instructions or whether it was also linked to the product quality and differentiation conveyed by the special, glossy packaging. We created a handout to be distributed with the ACTs that was a simple photocopy of the CAPSS package and wrapped it around the control package when distributing the medication at the drug shop Two additional treatment arms were designed to test whether simple messages delivered via stickers attached to the control packaging could also promote adherence. The first sticker was designed to address non-adherence based on the belief that the illness is cured when symptoms have improved or resolved. It said "Malaria is NOT gone until ALL tablets are finished". The second sticker message aimed to discourage the saving of pills for the next malaria episode and internalize the externality associated with non-adherence. It said "Finish ALL tablets. Saving tablets for later can be harmful for malaria control in your community." An additional treatment arm was also included in the study with a sticker that provided the actual (non-subsidized) price of the medicines. However, due to budgetary reasons, this treatment had to be phased out before we obtained a reasonable sample size. The final treatment was a random (cross-cutting) offer to patients of a free rapid diagnosis test to test for malaria.
Intervention Start Date
Intervention End Date
Primary Outcomes (end points)
Our primary outcome is adherence, that is whether the patient completed all the recommended doses of the ACT drugs.
We also look at the number of doses and tablets remaining as additional outcomes of the intervention. Some treatments arms may have an effect on the intensive margin of increasing the number of pills taken without affecting the level of adherence. This increase in the intensive margin is likely to still be beneficial both in treating the disease and in minimizing the likelihood of the development of resistance to the drugs by the pathogen.
Primary Outcomes (explanation)
Secondary Outcomes (end points)
Secondary Outcomes (explanation)
The study took place in the district of Luwero, Uganda, located in Uganda's central region, between November 2010 and September 2011. The study area constitutes catchment areas surrounding nine drug shops that were located in and around three small trading centers in the east of the district (Busiika, Zirobwe and Wabitungu).Surveys were conducted at four points through the study period: at baseline, at the time of ACT purchase, several days after ACT purchase (“follow up”) and at study endline.
In November and December 2010, a team of enumerators traveled to each household in the study area to enroll participants and conduct a baseline survey. Households were then given a Purchase ID card, which enabled any of the household members to purchase ACTs at a 95 percent subsidy at any of the nine selected drug shops. The objective of the project was to assess the impact of various forms of packaging and messages associated with ACTs on adherence. To evaluate this, we randomized the type of packaging/messages households received each time they came to the participating shops to purchase ACTs. The treatment arms were randomly assigned at the shop-day level. That is, an ex-ante schedule was laid out using a random number generator that indicated that Shop 1 got package A on March 1, package B on March 2, and Shop 2 got package C on March 1, control pack on March 2, etc. Surveyors assigned to each shop brought the control or treatment packs for that particular day with them, and both the study team and shop owners were blinded to the treatment assignment until the day of sale.
During the time of the ACT purchase, a short survey was administered at the shop with the patient or with the caretaker if the patient was a young child. The survey primarily concerned the severity of the symptoms that the patient was currently experiencing. In addition, a subset of people were randomly offered a free rapid diagnosis test for malaria. In cases where the patient was not present at the drug shop, the field officer tested the patient at the household. Adherence was assessed through follow-up visits at the household roughly three days after the time of ACT purchase. At the end of the data collection period (in August and September 2011), field officers visited each of the participating households and informed participants the study was ending.
Experimental Design Details
Randomization was done in an office by a computer.
For the packaging and messaging interventions, the unit of randomization was the shop-day, so that each shop was randomly assigned a different type of ACT package each day (either the control pack or one of the five treatment packs).
For the offer of the free rapid diagnosis test, the unit of randomization was the individual (patient). Approximately 23% of people who purchased an ACT were randomly offered a free rapid diagnosis test.
Was the treatment clustered?
Sample size: planned number of clusters
Sample size: planned number of observations
1662 ACT purchases
Sample size (or number of clusters) by treatment arms
Control: 300 ACT Purchases
CAPSS Pack: 300 ACT Purchases
CAPSS Information Only Pack: 254 Purchases
Message 1 ("Malaria is not gone until ALL tablets are finished"): 254 Purchases
Message 2 ("Finish ALL tablets. Saving Tablets for later can be harmful for malaria control in your community"): 254 ACT Purchases
Message 3 (Sticker with actual, unsubsidized price of ACTs affixed to control package): 300 ACT Purchases
RDT Offered (Cross-cutting randomization): 306 Individuals
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
INSTITUTIONAL REVIEW BOARDS (IRBs)
Uganda National Council for Science and Technology
IRB Approval Date
IRB Approval Number
Harvard School of Public Health Office of Human Research Administration
IRB Approval Date
IRB Approval Number
Post Trial Information
Is the intervention completed?
Intervention Completion Date
August 18, 2011, 12:00 AM +00:00
Is data collection complete?
Data Collection Completion Date
September 26, 2011, 12:00 AM +00:00
Final Sample Size: Number of Clusters (Unit of Randomization)
9 drug shops
Was attrition correlated with treatment status?
Final Sample Size: Total Number of Observations
2,517 ACT Purchases
Final Sample Size (or Number of Clusters) by Treatment Arms
Control: 699, CAPSS Pack: 465, CAPSS - Information Only: 392, Message 1 ("Malaria is not gone...etc"): 394, Message 2 ("Finish all your medication...etc"): 434, Message 3 (Actual, unsubsidized price of ACT on Sticker): 133
Is public data available?
This section is unavailable to the public. Use the button below
to request access to this information.