Evaluation of Conditional Cash Transfers (CCTs) for Immunization

Overview

Like many developing countries, Pakistan faces a public health challenge of low and incomplete immunization rates of children, only 54% of children aged 12 to 24 months are fully immunized, which leaves children susceptible to vaccine-preventable diseases. The Expanded Program on Immunization (EPI) is a low-cost and effective health intervention, but the uptake is low, delayed, and completion rates are poor. Door-to-door campaigns can increase coverage, but are extremely expensive. Incentive-based approaches have been rigorously demonstrated to effectively increase take-up and completion rates of immunization, and there is substantial evidence that small incentives can have a large impact on the take up of preventative health behavior in general. There are two major constraints to scaling these findings, however. This study will attempt to find the most effective incentive design that helps increase the coverage of full immunization rates among children between the ages of 0 - 24 months in the city of Karachi, Pakistan. The study proposes to conduct a randomized control trial involving small conditional cash transfers (mCCTs) to determine the optimal CCT amount (high versus low), schedule (flat versus increasing) and design (lottery versus sure payment) that would lead to the highest increase in immunization rates. Interactive Research and Development's digital immunization registry will be used to enrol and randomize the study participants and generate CCTs disbursed through a mobile money transfer platform and mobile top - ups . The three year study will be conducted in Karachi, Pakistan enrolling a sample of 11,200 children, 0-2 years of age. The study aims to provide evidence regarding the most cost-effective way to structure incentives in terms of size, schedule, and design; and address the challenge of delivering small incentives in a way that is inexpensive, logistically simple, and not subject to leakage.

The study is a three - year randomized controlled trial that will experimentally evaluate the impact of CCTs and SMS reminders on immunization rates of children under 2 years of age. The study will investigate the influence of different amounts, schedules and design of CCTs on immunization rates to (1) determine the impact of small incentives on immunization coverage and timeliness in 0-23 months old Pakistani children, (2) measure the relative effectiveness and cost of effectiveness of different types of incentives structures (including the amount, progressivity and certainty of payment) on immunization coverage rates and timeliness in 0-23 months old Pakistani children and (3)determine the impact of SMS reminder (both with and without incentives) on immunization coverage rates and timeliness in 0-23 months old Pakistani children. The study has a cross-cutting design where one experiment (1) will investigate (1a) sharply increasing versus slowly increasing payouts; (1b) high incentive versus low incentive and (1c) simple SMS reminder vs no reminder. A second experiment (2), orthogonal to the first, will also test lottery payouts vs non-lottery payouts and will cross-cut (1a), (1b) and (1c) with (2). Moreover two different incentive disbursement methods will be utilised namely mobile phone credit (air time) and mobile money (hard cash) through a mobile money provider-EasyPaisa. The study will have a total twelve arms. The first ten arms will be the incentive arms with eight using mobile phone credit and two utilising mobile money. Additionally arm eleven will only provide SMS reminders while control arm twelve will neither receive CCT nor SMS. The study will enroll a total of 11,200 participants. Additionally to evaluate the proportion of immunized children, the study will also (4) examine 15% of the enrolled children for sero-survey biomarkers - the gold standard for establishing immunity, among different intervention and control arms Study Site All the study activities will be conducted in EPI centers in Korangi town of Pakistan, a suburban area adjoining Karachi city, located in Korangi district of Sindh province. Korangi town belongs to one of the lower socioeconomic areas in Karachi city. It is ethnically diverse, with low to middle income areas and has the presence of an existing research infrastructure and experience capable of running an mHealth immunization registry. Healthcare services in Korangi town are provided mainly by seven public basic health units, two private hospitals, and three private clinics. Randomization Enrolled participants will be randomly assigned to study arms. The stratified randomization sequence will be created using statistical software through random block sizes. For this study, a block randomization procedure has been put in place with a block size of 48 and six different strata based on enrollment vaccine and gender. These strata are: BCG-Male and BCG-Female, Penta1-Male and Penta1-Female and Penta2-Male and Penta2-Female. Performing block randomization ensures that the effect of gender and enrollment vaccine is removed as confounding variables in determining treatment outcomes. Furthermore, it ensures that the treatment assignments are not over/under-represented in any particular stratum. Study Procedure When a child first comes to any of the participating immunization centers the field worker will approach the caregiver and introduce the study and determine eligibility. From prior experience, it is possible that the caregiver will be unable to recall their mobile number at their first visit to the EPI center. Therefore, if screening criteria is not met and the child is coming for BCG/Penta-1 visit, the field worker will instruct the caregiver to bring the mobile number on the next visit. If the caregiver successfully provides the mobile number up till the time of Penta-2 visit , and meets the rest of the eligibility criteria, the field worker will proceed to taking the informed consent.The field worker will obtain a verbal informed consent by providing a detailed explanation of the information contained in the informed consent form and answering any questions that may arise. Contact cards will be provided to participants containing the contact details of the study team and IRB in case of any queries/complaints and the informed consent will be available in both English, and Urdu. If the caregiver refuses to respond or does not consent, the child will not be enrolled, but will proceed to the vaccinator to get his routine scheduled immunizations. Interviewees will be informed during the informed consent process about the possibility of being approached for enrollment in subsequent antibody assessment or any other study related follow-up information. If the caregiver provides consent, the child will be enrolled in the Zindagi Mehfooz digital immunization registry platform and will receive a unique ID in the form of a QR code pasted on the child's EPI card. The caregiver's CNIC number if provided will also be captured and will be entered into the registry, linking the child's ID to the caregiver's CNIC. Once all information is submitted successfully, the system is pre-programmed to randomly allocate the child to one of several treatment groups within the two interventions. All participants (irrespective of whether they are in the treatment or control arms) will receive the routine EPI vaccinations as per Pakistan's EPI Immunization schedule (one dose of BCG (Bacille Calmette-Guérin) and polio soon after birth, 3 doses of pentavalent (DPT + HepB + Hib) vaccine and oral polio vaccine at 6, 10 and 14 weeks of age,at 6, 10 and 14 weeks, and 2 doses of Measles vaccine and 9 and 15 months of age) and the vaccinator will record the child's immunization data in the EPI immunization card as well as the EPI registry. All the field data and specimen collecting activities will be conducted by a team of one field supervisors and four field workers for every four EPI centers. The biomarker assessment will be done by a professional phlebotomist/nurse trained in finger stick capillary blood collection. Follow-up When the next immunization is due, the caregiver will receive an SMS reminder for the follow-up visit (if in the appropriate treatment arm). At the follow up visit, the field worker will scan the child's QR code and retrieves his/her past immunization data. If the EPI card is not available, child data will be screened through additional identifiers (including child name, father name, CNIC number, mobile number). The immunizations received on the current visit will be recorded in the registry and the caregiver becomes eligible for the cash incentive if they are in the in the appropriate intervention arms. In the control group, no incentive or reminders are received. Antibody Assessment Recruitment and Consent Process After 3 months following the recommended last immunization visit of measles-2 vaccine (children between 18 -24 months), a randomly selected sample of children will be approached for the antibody assessments either when they come for the Measles-2 vaccine at the EPI center or through household visit to the child's house (address of the child and phone number of the caregiver would be recorded at the time of enrolment). Randomization will be carried out in real-time on an android device through an online server. After conducting a short immunization verification survey with caregiver consent, the field worker will proceed with a detailed explanation of the information contained in the written informed consent form for antibody assessment and signatures of the caregiver will be taken as proof of consent. Blood collection will then proceed at the child's house and an interviewer trained in capillary blood collection or a phlebotomist will collect a finger stick capillary blood specimen using standard aseptic techniques into serum separator tubes. Personal protective equipment (gloves, sharps box) will be used by the interviewer/phlebotomist and a maximum of three capillary blood draw attempts will be attempted. To minimize distress and pain, parents may be asked to hold, breastfeed, or provide other distractions for the child. At least one interviewer or field supervisor will be present during the blood collection to assist. The specimen vials will be coded immediately with the child's QR code and a maximum of 0.5 ml of blood will be drawn from each child with a minimum volume of 0.2 ml. If the child is not present at the time of the survey or if the caregiver would like to reschedule the blood collection, the household will be revisited at the time the caregiver prefers. A maximum of three attempts will made to contact the child before moving on to next randomized child. Each finger stick capillary blood sample will be kept in cold box at 4-8°C and transported to Indus Lab Karachi by a designated vehicle within two to four hours of collection. Laboratory Procedures Lab testing will be performed periodically throughout the biomarker specimen collection period when an adequate number of specimens are available for testing. Serum specimen will be tested for measles IgG antibodies (Enzygonost ELISA, Siemens, Germany; reported sensitivity and specificity of 99.6% and 100% respectively) and tetanus toxoid IgG antibodies (Euroimmun anti-TT IgG ELISA; reported sensitivity and specificity of 94.1% and 100% respectively). Serum samples will be categorized as negative, equivocal or positive as recommended by the manufacturers. Samples with equivocal results will be re-tested and, if equivocal again, will be categorized as positive, according to a previous study. Data Management and Analysis Data will be analyzed using STATA 13. The immunization coverage for each vaccine by study arm will be assessed for children under 2 years of age as reported through their digital immunization record. Bivariate and multivariate analysis using a generalized linear model of estimated risk ratios for FIC coverage will be conducted (using α = 0.05 for evaluating statistical significance). For multivariate analysis, covariate selection will be based on a priori knowledge using previous studies and biological plausibility. Primary specification for the analysis The investigators will use intention to treat analysis-that is, all participants will be analysed with the assumption that they remained in the treatment group to which they were initially assigned. The main regression analysis will thus be: Y=a SMS + b Lottery + c High\*FLAT + d High\*SLOPE + e Low\*FLAT+ f LOW\* Slope + e Where: Y is the relevant outcome measure (FIC coverage) The coefficient "a" gives us the impact of the SMS reminder in the control group. The coefficient "b" gives us the impact of lottery versus sure payment, keeping constant the level and structure of the incentive. The coefficients c, d, e and f give us the impact of different incentives (compared to the SMS control group). Antibody Data Analysis There will be three possible outcomes for the antibody analysis namely positive, negative and equivocal. The first two are self explanatory and for the latter the following approach will be used For a well-performed ELISA test, equivocal results are generally caused by four reasons: an imperfect sample condition, such as general protein degradation (low total antibody level) due to imperfect blood sample preparation; a nearly undetectably low level of target antibody in sample; insufficient primary antibody coating; and unspecific binding leading to high background noise. Since the current study is intended to ascertain history of vaccination, it is intended to define positive even when protective antibody level is not induced. Thus, the first three potential causes that lead to equivocal results will be defined positive in this study, while the fourth potential cause will be excluded by observing valid negative control tests.