Acceptance and Impact of introducing sliding scale payment policy into CBHI in Ethiopia: Randomized controlled trial

Last registered on May 02, 2024

Pre-Trial

Trial Information

General Information

Title
Acceptance and Impact of introducing sliding scale payment policy into CBHI in Ethiopia: Randomized controlled trial
RCT ID
AEARCTR-0011191
Initial registration date
April 05, 2024

Initial registration date is when the trial was registered.

It corresponds to when the registration was submitted to the Registry to be reviewed for publication.

First published
April 16, 2024, 1:01 PM EDT

First published corresponds to when the trial was first made public on the Registry after being reviewed.

Last updated
May 02, 2024, 3:11 PM EDT

Last updated is the most recent time when changes to the trial's registration were published.

Locations

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information

Primary Investigator

Affiliation
Jimma University

Other Primary Investigator(s)

PI Affiliation
Jimma University
PI Affiliation
Jimma University
PI Affiliation
Jimma University
PI Affiliation
Jimma University
PI Affiliation
CEDLAS, UNLP
PI Affiliation
Jimma University
PI Affiliation
University of South Carolina

Additional Trial Information

Status
On going
Start date
2023-11-01
End date
2026-12-31
Secondary IDs
Prior work
This trial does not extend or rely on any prior RCTs.
Abstract
The Ethiopian government initiated a Community-Based Insurance scheme (CBHI) in 2011 to enhance financial protection for the informal sector. As of 2020, 8.4 million households were enrolled in the scheme. However, one of the CBHI's challenges is that households contribute an equal amount of premium (flat-rate contribution in majority of the settings across the country) regardless of their economic status and ability to pay. This adversely affected the lower-income households, sustainability, and equity of the program. To address this challenge, a new policy initiative, a sliding scale (SS-CBHI) contribution system based on households’ economic status is under consideration. This study is intended to evaluate the impact of introducing the new policy initiative using a Randomized controlled trial (RCT).The new policy initiative will be implemented in both rural and urban areas of these regions. Randomization will be conducted at the district level. A total of 5200 households will participate in the study. Households in the control areas will continue paying fixed rates whereas the households in the treatment areas will pay differential rates based on socio-economic status.
External Link(s)

Registration Citation

Citation
Koricha, Zewdie Birhanu et al. 2024. "Acceptance and Impact of introducing sliding scale payment policy into CBHI in Ethiopia: Randomized controlled trial ." AEA RCT Registry. May 02. https://doi.org/10.1257/rct.11191-1.1
Sponsors & Partners

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information
Experimental Details

Interventions

Intervention(s)
Conditions in treatment arm: In the treatment arm of this study, all households within the district, referred to as clusters, have enrolled in a new policy initiative known as the sliding scale CBHI contribution scheme. This initiative bases its contributions on the economic status of households, regardless of age, gender, or family health risks. Under this scheme, households in the treatment district will pay an annual CBHI contribution rate determined by their economic level as stratified by the Ethiopian Health Insurance Service (EHIS). There will be different stratification for urban and rural residents.
Upon successful enrollment, each household will receive a health service coupon, or CBHI identification card, containing pictures of all insured household members. This card grants access to health services at designated health facilities, primarily health centers within their administrative units. Patients can also seek care at district/general hospitals or referral hospitals as per the referral chain within the system. The enrollment and ID card are valid for one year from the issue date and must be renewed annually using similar procedures to initial enrollment. The implementation of these interventions, including enrollment and health services, is overseen by the national EHIS, the Ministry of Health, and regional health bureaus such as the Oromia Regional State Health Bureau and Sidama Regional State Health Bureau.
Conditions in control arm: In contrast, communities in control districts remain under the current fixed-rate CBHI contribution scheme. Enrollment at the household level is independent of family health risks and socioeconomic status, with a fixed annual premium for all households regardless of economic standing. However, there are variations in rates for urban and rural residents. Additionally, 20% premium required for each household member aged 18 years and above. The government fully subsidizes the contribution for the poorest households, ensuring their access to healthcare services.
Intervention Start Date
2023-11-01
Intervention End Date
2026-03-31

Primary Outcomes

Primary Outcomes (end points)
Three primary outcomes with separate indicators and measurements will be assessed for this experiment:
1. Financial protection which includes out-of-pocket health expenditure per capita; catastrophic health expenditure and Impoverishing spending
2. Health service utilization-outpatient and inpatient service utilization
3. Women empowerment for health service utilization explained in terms of gender disparity

Equitable health service utilization and the reduced disparity between household economic statuses
1. Financial risk protections (out-of-pocket health expenditure per capita; catastrophic health expenditure and Impoverishing spending)
2. Indicators of women empowerment and gender equity in healthcare utilization
3. Acceptance (uptake, renewal, dropouts, rates) and beneficiaries’ satisfaction with the new policy initiative

Primary Outcomes (explanation)
1) Financial protection outcomes: Financial protection is one of the primary outcomes in this experiment and it consists of three related outcomes indicators, explained as follows
• Per capital out-of-pocket payment (oop): is defined as the payments made by households at the point, they receive health services (e.g., doctor’s consultation fees, purchases of medication and hospital bills, laboratory, diagnostic etc.) including payments on alternative and/or traditional medicine but excluding expenditure on health-related transportation, food, and any insurance reimbursement. We will calculate the per capita oop payment by aggregating all in 12 months and then dividing by total number of individuals in the households. This out-of-Pocket health payments will be calculated by adding the payments made by households at the point when they received health services, including expenses for registration, consultancy fee if went to private service, drugs, laboratory tests, inpatient beds, and other related items.
• Catastrophic health expenditure (CHE): Impoverishment of expenditure occurs when a non-poor household is impoverished by health payments (become poor after paying for health services). This outcome will be calculated by using different cut-off points (10%, 25%,40%) of households’ annual direct medical costs compared to total health expenditure.
• Impoverishment: Impoverishing health expenditure (impoor) is when household expenditure is equal to or higher than subsistence expenditure but is lower than subsistence expenditure net of out-of-pocket health payments. This outcome will be computed by assessing (using the measures of household financial burden proposed by the WHO) the household’s total expenditure compared to the national poverty line after deducting annual OOP expenditure.

2) Health service utilization is also the other primary outcome, which will be assessed by three indicators.
• Per capita outpatient visits and inpatient admission: This outcome will be measured by the number of outpatient visits and inpatient admissions and will be aggregated at the household level and averaged out across all households to calculate per capita outpatient visits and inpatient admission rates per year.
• Treatment-seeking behavior: This outcome will be calculated by the proportion of individuals seeking modern health care among households with at least one sick member during the last 4 week's recall period for outpatient services and 1 year for admissions.
• Concentration indices: Concentration indices of health service utilization will be computed from total household expenditure to indicate the social disparity in health service utilization. The third primary pot come is women's empowerment and this will measure Women's decision-making power and perception of their health.

4. Women empowerment is another primary outcome, which is intended to measure Women's decision-making power and perception about their health. Impact of Sliding scale -CBHI on women will be examined in three groups of outcomes: women’s decision-making on health care; health care utilization (disparity by gender and wealth status), out-of-pocket expenses, and satisfaction perceptions. The household survey questionnaires will be used to elicit indicators for each dimension. Each category will then be summarized in indexes using standardized methodologies and using gender segregated analysis.

Secondary Outcomes

Secondary Outcomes (end points)
There are two crucial secondary outcomes:
1) Acceptance and satisfaction with the new policy initiative
2) Perceived quality of services and satisfactions with health services
Secondary Outcomes (explanation)
1) Acceptance and satisfaction with the new policy initiative: Beneficiaries’ acceptance of the new policy initiative and user satisfaction and experience will be assessed using an interviewer-administered questionnaire using an acceptance and satisfaction scale. To measure satisfaction and experience with the new scheme, we will use five points likert scale (1=Very satisfied 2=Satisfied, 3=Dissatisfied, 4=Very, dissatisfied and 5=don’t know or can’t Assess) for thirteen aspects of the new policy initiative (sliding based CBHI scheme).
2) Acceptance and uptake of the new scheme: Beneficiaries’ acceptance of the scheme is expressed by participation or enrollment rate into the program, and renewal rate at the follow-up phase. The extent to which the scheme is appealing (acceptance) will be assessed using acceptance scale on five points (1= strongly disagree, 2= disagree, 3= undecided, 4=Agree and 5= strongly agree) using nine acceptance items.
3) Perceived quality of services and satisfactions with health services: Beneficiaries satisfactions with health service provisions and availability will be assessed using thirteen satisfactions and perceived quality items, addressing aspects including time spent with the providers, waiting time, staff, availability of drugs etc.). This satisfaction measure will be gathered separately both for outpatient and inpatient experiences. We will create a satisfaction index as high satisfaction and low satisfaction.

Experimental Design

Experimental Design
This impact evaluation will be based on a two-arm pairwise cluster randomized controlled trial involving one treatment arm and one control arm where assignment to the treatment will be randomized at the district level. Accordingly, 13 clusters (10 rural and 3 urban areas) is randomly assigned to treatment arm ( pilot for new policy initiative) and 13 ( 10 rural and 3 Urban areas) is assigned to the control arm, which means these districts will remain under existing CBHI premium scheme. Within cluster, the study population will be comprised of all households (irrespective of their economic status) in rural and urban clusters (districts) in selected districts of Ethiopia.

Experimental Design Details
Not available
Randomization Method
In this RCT, the district was used as a unit of randomization whereas households under each district will be used as a unit of observation (measurement) and the randomization was done in office by a computer software.
Randomization Unit
The district is used as a unit of randomization whereas households under each district will be used as a unit of observation (measurement). Randomization at the district level is chosen for practical reasons:
• We will randomize districts under each geographic (rural and urban) area by considering implementation and administrative feasibility.
• Because of resource constraints, considering small clusters at the level below that (PHCU and kebele/villages) level yields very low power of the study (<15%), and practically it’s difficult to implement due to methodological challenges (contamination and spillover effects).
• Methodological challenges (contamination and spillover effects). Randomization within geographic areas will help to reduce the imbalance/variation between regions, geographic areas, and district-level covariates.
Was the treatment clustered?
Yes

Experiment Characteristics

Sample size: planned number of clusters
In this RCT, 26 clusters (districts/towns) are included:13 in treatment clusters and 13 in control clusters. Specifically, 18 comparable districts (7 pair-wise rural districts and 2 paire-wise Urban for the Oromia region, and 8 comparable districts- six rural districts and two Urban for the Sidama region was included the experiment.
Sample size: planned number of observations
There will be three survey rounds (baseline, mid-term, and end-line) observations. At each round, a total of 5200 households will be included, including all of their family members.
Sample size (or number of clusters) by treatment arms
The sample is calculated using Stata 16.0 version. Using 1:1 allocation ratio, interclass correlation of (ICC=0.25) and 200 household samples per cluster, the 5200 total sample and 26 total clusters are required for the study. Therefore, we will recruit 200 households per cluster or pair (200 households x26 clusters =5200 households) for treatment and the same size of households in the control areas at each measurement points.
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
For estimation of the sample size, the following assumptions were made: We hypothesized that the program would increase per capita outpatient visits by ten percentage points (minimum Detectable effect=10 standardized mean difference.) to detect the power of 80% with an alpha of 5% with an upper tail. Using a 1:1 allocation ratio, interclass correlation of (ICC=0.6), and 200 household samples per cluster, the 5200 total samples and 26 total clusters are required for the study.
IRB

Institutional Review Boards (IRBs)

IRB Name
Jimma Univeristy Institutional Review Board
IRB Approval Date
2023-05-06
IRB Approval Number
JUIH/IRB/377/23