Family and Childhood Development: Kizazi Kijacho ('The Next Generation') - a Cluster Randomised Controlled Trial

i) Parenting only:
Existing CHWs will be trained to use an innovative digital application for the delivery of integrated ECD services to mothers who are at least 20 weeks pregnant and less than 32 weeks pregnant for a period of 15 months. CHWs will provide tailored ECD services (e.g., prompting messages tailored to child age and triggering follow-up visits conditional on changing conditions), covering all aspects of the Nurturing Care Framework (Health, Nutrition, Responsive Caregiving, Early Learning, Safety and Security). Real time data will be recorded by the CHWs in each visit using the application. Data will include information on visit attendance, activities conducted, home environment, caregiver practices, and CHW observations. From when the target child is 6 months old, group sessions will be organised by CHWs, focusing on caregiver-child interaction and stimulation activities.

ii) Parenting+Unconditional Cash Transfer:
In addition to the Parenting Intervention, pregnant women in the study sample will receive a bi-monthly unconditional mobile money transfer of 77,000 TZS (33 USD) from 5-7 months pregnancy over a period of 15 months (7 transfers in total). The amount 77,000 TZS equals the average of maximum and minimum amount transferred on a bi-monthly basis to similar pregnant women under Tanzania Social Action Fund (TASAF), i.e., Tanzania's National Cash Transfer program.

iii) Unconditional Cash Transfer: Fixed amount
Households, in addition to the CHWs delivering health and nutrition services as usual, will receive a fixed bi-monthly unconditional mobile money transfer each of 109,000 TZS (47USD) from 5-7 months pregnancy over a period of 15 months (7 transfers in total). The bi-monthly cash transfer equals the sum of (i) the average of maximum and minimum bi-monthly cash transfer amount disbursed per family under Tanzania Social Action Fund (TASAF), i.e. Tanzania’s flagship nationwide social protection scheme (77,000 TZS, 33 USD) and (ii) the bi-monthly cost per additional family of delivering a parenting intervention (32,000 TZS, 14 USD). The transfer will be randomly assigned between mothers and fathers/spouses within each community, where in half of the eligible households, mothers will receive the transfer and in the other half, fathers/spouses (or household head where the father/spouse is not available) will receive the transfer.

In addition to the previous intervention groups, there is another UCT only treatment study group where the level of the cash amounts vary across communities. In addition to the CHWs delivering health and nutrition services as usual, communities will be randomised to receive one of the bi-monthly unconditional mobile money transfer amounts: 32,000 TZS (14 USD), 77,000 TZS (33 USD), 109,000 TZS (47 USD) from 5-7 months pregnancy over a period of 15 months (7 transfers in total). The transfers will be randomly assigned between mothers and fathers/spouses within
each community, where in half of the eligible households, mothers will receive the transfer and in the other half, fathers/spouses (or household head where the father/spouse is not available) will receive the transfer. The rationale for the levels of the varying UCT amounts is to keep them comparable with i) the bimonthly cost of the parenting program per family,
ii) the bi-monthly cash transfers disbursed under TASAF, and iii) the sum of the bi-monthly cost of the parenting program and the bi-monthly cash transfer disbursed under TASAF.

2022-10-26

2024-04-30

Our primary outcome of interest include Household Decision making and Child Development.

Please see the attached Pre-analysis Plan for details.

The aim is to sample and survey for the study a total of 3525 pregnant women aged 18 years or above, who are living in the select study communities and who are at least 20 weeks pregnant and less than 32 weeks pregnant at the time of the baseline data collection survey visit to the study community region, Tanzania.

The study will randomly sample 258 public Health Dispensaries (with at least one officially registered Community Health Worker (CHW) working at the facility) across all 7 districts (and all 8 district councils) in the Dodoma region, Tanzania, to participate in a clustered multi-arm Randomised Controlled Trial (cRCT). The 258 Health Dispensaries (HDs) will be randomised to a (i) Control group (81 HDs) where CHWs deliver Early Childhood Development (ECD) services as per existing government guidelines, (ii) Parenting group (88 HDs) where existing CHWs will be trained to use an innovative digital application for the delivery of integrated ECD services for a period of 15 months, from 5-7 months pregnancy onwards and, (iii) an Unconditional Cash Transfer (UCT) only group (89 HDs) where CHWs deliver ECD services as per existing government guidelines but where the study sample of families will receive a bi-monthly UCT fixed amount of 109,000 TZS (equivalent to 47USD) for 15 months (7 transfers in total). The randomisation will be stratified by district council and by whether there is more than one community in the HD catchment area.

Within each of the HD catchment areas in the Control group, one village (in
rural areas) or one ‘mtaa’ (in urban areas) served by the HD and where at least one officially registered CHW is available to work will be randomly sampled. For the 88 Parenting HDs and the 89 UCT only HDs, all villages/mtaas (with at least one available officially registered CHW) will be included in their catchment area to become part of the study. In total, that will give 390 study villages/mtaas in the study sample.

Within each of the selected study villages/mtaas, one CHW will be selected
whose catchment area will become the geographic area of interest, i.e., the
study community. The study community can be the entire village, a hamlet (subvillage) or an mtaa, depending on the size of the CHW's catchment area. This gives a total of 82 Control communities, 155 Parenting communities, and 155 UCT communities in the study. Within the Parenting and UCT only study groups, then second layer of randomisation will be done. In the Parenting group (154 communities across 88 HDs), communities will be randomly assigned, stratified by HD, to either
one of the following two treatment arms: (i) Parenting only (77 communities)
and (ii) Parenting+UCT (77 communities) where the Parenting Intervention will be delivered along with a bi-monthly unconditional mobile money transfer of 77,000 TZS (33 USD) from 5-7 months pregnancy over a period of 15 months (7 transfers in total). In the UCT only group (155 communities across 89 HDs), study communities will be randomly assigned, stratified by HDs, to either one of two treatment arms: (i) UCT only fixed amount (80 communities) where families will receive a fixed bi-monthly cash transfers each of 109,000 TZS (47 USD) over a period of 15 months (7 transfers in total) and (ii) UCT only vary amount where 77 communities will be randomly allocated to one of the following bimonthly UCT amounts: 32,000 TZS (14 USD), 77,000 TZS (33 USD), 109,000 TZS (47 USD) over a period of 15 months (7 transfers in total). In each of these two study arms, further randomisation will be done whether the mobile money transfer is given to the father/spouse or the mother. 10 eligible women per community will be randomly sampled to participate in the study, except the bi-monthly UCT vary amount group, where only 5 eligible women per community will be randomly sampled.

Such a design allows to assess the relative cost-effectiveness of the Parenting and/or UCT only fixed amount interventions, and indeed provide insights into the value of adding a parenting component to a social protection program such as the Tanzania Social Action Fund (TASAF).

Additionally, the study will also explore CHW performance, quality of
care delivered and other fidelity indicators to analyse impacts based on
implementation effectiveness.

Randomization will be done using Stata using a reproducible seed.

The randomisation will be done in the following steps:
• All public HDs will be identified in the Dodoma region with at least one
officially registered CHW, 258 HDs in total.
• These study HDs will be randomised to either (i) the Control group (81
HDs), (ii) the UCT only group (89 HDs), or (iii) the Parenting group
(88 HDs). The randomisation will be stratified by district council and
by whether there is more than one village or ’mtaa’ (city block in urban
areas) served by the HD (i.e. in the HD catchment area).
• In the 81 Control HDs, one village/mtaa with at least one officially registered CHW will randomly sampled to participate in the study. In the 88
Parenting HDs and the 89 UCT only HDs all villages/mtaas (with at least
one officially registered CHW) in their catchment areas will be included to
participate in the study. In total, that will give 390 study villages/mtaas
in the study sample.
• Within each of these study villages/mtaas, one officially registered CHW
will be randomly selected and include the catchment area of that CHW as a study ’community’ in the study sample, where study community is
defined as CHW catchment area. This could be the entire village/mtaa
or a sub-village/mtaa, depending on the size of the village/mtaa and the
number of CHWs working in the village/mtaa.
• Within the UCT only group (155 study communities across 89 HDs),
study communities will be randomly assigned, stratified by HD, to: (i)
UCT only fixed amount treatment arm (80 communities) and (ii) UCT
only vary amount treatment arm (75 communities). In the Parenting
group (154 communities across 88 HDs), study communities will be randomly assigned, stratified by HD, to: (i) Parenting only treatment arm (77
communities) and (ii) Parenting+UCT treatment arm (77 communities).

Yes

The total number of clusters will include 258 health dispensaries in 390 CHW community catchment areas.

A total baseline study sample will include 3525 women.

The final sample in each experimental group will be:

1. Parenting (88 health dispensaries, 154 communities, 1540 households)
a. Parenting only (88 health dispensaries, 77 communities, 770 households)
b. Parenting + UCT (88 health dispensaries, 77 communities, 770 households)

2. UCT (89 health dispensaries, 155 communities, 1175 households)

a. Fixed bi-monthly UCT (89 health dispensaries, 80 communities, 800 households)
i. Mother as target beneficiary (89 health dispensaries, 80 communities, 800 households)
ii. Father/spouse as target beneficiary (89 health dispensaries, 75 communities, 375 households)

b. Varying bi-monthly UCT (89 health dispensaries, 75 communities, 375 households)
i. Mother as target beneficiary (89 health dispensaries, 75 communities, 187 households)
ii. Father/spouse as target beneficiary (89 health dispensaries, 75 communities, 188 households)

3. Control (81 health dispensaries and communities, 810 households)