Skip to content

5. Workflow

Jump to bottom
Máximo Ramírez edited this page Nov 8, 2017 · 9 revisions

Context and settings description

General scenario

This page aims to show a general workflow of a household survey supported by the EpiSample tool. In this case, the survey will be carried out in two study areas composed by 4 clusters each. These study areas are remote from the study leader headquarters, so we can assume the worst logistics scenario. That means the field team does not have a local office with the technology infrastructure required to set up a local server where data can be aggregated and backed up.

Logistics

The field team in this case will be composed by 8 field workers (one per enumeration area) and 2 supervisors (one per study area). These 10 professionals will be provided with a mobile device with EpiSample installed and already configure (technical people in headquarters will be in charge of that task).

Furthermore, the mobile devices of the supervisors will be connected to Internet. Only these two devices. We assume that supervisors will be able to find a place inside the study area and with quality-enough mobile signal to send the data through Internet.

During a first phase, field workers will register all households belonging to the targeted clusters. Sampling will be done in each study area by supervisors. And during a second phase, field workers will survey selected households. It is very important to emphasize that we will have two sampling processes (one in each study area) to assure exactly the same number of respondents in each area.

Geographical hierarchy

Hierarchy Name Hierarchy ID Area Name Hierarchy ID Nesting Reference
Study area 1 KENGE
Study area 2 MASIMANINBA
Cluster 1 BANGONGO 1
Cluster 2 MINKWO 1
Cluster 3 GONGO 1
Cluster 4 BWATUNDA 1
Cluster 5 BIBAMBA 2
Cluster 6 PALANGA 2
Cluster 7 KIPEMBE 2
Cluster 8 MUKAKALA 2

Base on this hierarchy, a config.csv file to be loaded in EpiSample was created.

Data collection instruments

This study will have two different data collection instruments:

  1. The census form of the EpiSample App with the following relevant variables (among others not shown here and produced by EpiSample):
    • META_ROW_ID
    • PLACE_NAME
    • HOUSE_NUMBER
    • HEAD_NAME
    • COMMENT
    • LOCATION_LNG
    • LOCATION_LAT
    • LOCATION_ALT
    • LOCATION_ACC
    • EXCLUDED
  2. The example of survey provided with this documentation:
    • HH_ACCEPTANCE
    • PREGNANT_LAST_12_MONTHS
    • RESIDENT_DURING_PREGNANCY
    • CONSENT
    • WHY_NOT_CONSENT
    • OTHER_WHY_NOT_CONSENT
    • VISIT
    • ROOF
    • OTHER_ROOF
    • ASSETS
    • DOB
    • READ
    • MARITAL_STATUS

EpiSample configuration

EpiSample configuration will not be covered in this page due to is not the purpose. For reading how to configure EpiSample refer to the How to install page. However, we will show here some of the configuration parameters that are important to understand this workflow.

ODK Aggregate

The supervisors EpiSample applications will be linked to the ODK Aggregate service hosted in the headquarters servers and exposed to Internet.

Workflow Configuration: ODK Aggregate

Interviewer/Device information

This is the configuration of the supervisor EpiSample application of study area 1 - KENGE:

Workflow Configuration: Interviewer/Device information

Permissions

Among other actions, field workers and supervisors will not be allowed to edit or remove already collected information, both census and survey information. Therefore, Edit Census, Remove Census, Invalidate Census and Deleted Saved Form parameters are disabled.

Workflow Configuration: Permissions

Collect module

The study will invalidate points collected with an accuracy worse than 25 meters during the census phase. Therefore, these points will not be considered during sampling:

Workflow Configuration: Collect

However, we assume that mobile devices have not low quality GPS, so we do not add a delay to increase GPS accuracy.

Send/Receive module

Supervisors EpiSample applications will be configure as Device belongs to team lead and all the applications will enable Turn on and off WiFi automatically to save battery of devices while working in the field.

Workflow Configuration: Send/Receive

Workflow

In this section will be describe the steps to implement the survey described here with EpiSample. This workflow assumes that EpiSample is installed and configured in all the mobile devices and that the server is ready to receive the data weakly from the supervisors devices.

Step 1: Generate a household list

  1. Go outside.
  2. Tap the Collect button on the main menu.
  3. For each household you find:
    1. Tap the Select Place Name drop down at the top and select an enumeration area.
    2. Enter a House Number. If you are using an auto-increment strategy and this is the first collected point, you must introduce manually 1 or other Number for EpiSample to know which is the first numeric value.
    3. Enter the Head Name.
    4. Enter a Comment if you want to. The comment is not necessary, but you can use it to describe the house better to help identify it later. Example, next to school or red door or water well in front.
    5. Wait until the GPS accuracy goes under 25 meters and tap on Save.
  4. Continue until you finish collecting all your cluster houses.

Workflow: Collect

Step 2: Aggregate household list on the supervisor device

In this step we assume that the two devices that are going to communicate are correctly configured: one as the team leader device.

  1. Tap the Send/Receive button on both devices.
  2. Tap connect on both devices.
  3. On the supervisor device, tap the device you wish to connect with (the field worker device).
  4. On the field worker device accept the invitation to connect sent by the supervisor device.
  5. On the supervisor device, tap receive.
  6. On the field worker device, tap send.
  7. All points should be sent successfully, tap OK to confirm.

Workflow: Send/Receive by FW Workflow: Send/Receive by SUP

Step 3: Sample households

Once the household list is completed (that means all the field workers finished with their enumeration areas and all data is aggregated on the supervisor device) the sampling process can be initiated. Sampling should be done from the supervisor device where all the points are aggregated.

  1. Tap Select button on the main menu.
  2. Enter the number of points to select:
    • Main Points: These are the primary points (typically households) to survey.
    • Additional Points: These are the oversampled points (if needed).
    • Alternate Points: These are the extra points which can be used in the event a household from the main points or additional points list is unavailable or unsurveyable.
  3. Tap Select.

This will sample the households. It will also record the date/time of this sampling process in the database and the total number of valid and invalid points (which is the sampling frame). Obviously, it won't select invalid points but these points will be included in the sampling frame.

Workflow: Selection

Step 4: Send sampled households back to the enumerator devices

This is really similar to the step 2 (aggregate household list on the supervisor device). But in this step the communication occurs in the other direction, the supervisor device sends the points to the enumerator devices indicating which of them were selected. Again, we assume that the two devices that are going to communicate are correctly configured: one as the team leader device.

  1. Tap the Send/Receive button on both devices.
  2. Tap connect on both devices.
  3. On the supervisor device, tap the device you wish to connect with (the field worker device).
  4. On the field worker device accept the invitation to connect sent by the supervisor device.
  5. On the field worker device, tap receive.
  6. On the supervisor device, tap send.
  7. All points (those selected and those not selected) should be sent successfully (selected points will be marked as selected), tap OK to confirm.

Workflow: Send/Receive back by SUP Workflow: Send/Receive back by FW

Step 5: Navigate to households and conduct surveys

  1. Go outside.
  2. Tap the Navigate button on the main menu.
  3. Tap the Sort button at the top to sort by distance.
  4. Tap the house you wish to navigate to and use the compass and distance to navigate to the household.
  5. Once in front of the house, long tap the house in the navigate screen and select Start Questionnaire.
  6. Complete the questionnaire and finalize it. This should take you back to the navigate screen and the finalized count should increment.
  7. Go to the next house and repeat as much as you need.

Workflow: Navigate Compass Workflow: Navigate Start Questionnaire Workflow: Navigate Survey Questionnaire

Step 6: Transfer the surveys to the ODK Aggregate server

This is an important limitation of EpiSample. While supervisor devices can aggregate census data, they do not aggregate surveys data. Therefore, each enumerator device must send individually the survey collected data to the ODK Aggregate server. It means, that these devices require a way to be connected to Internet, in this case.

Ideally, EpiSample would provide with the survey aggregation feature and just the supervisor devices would require to be connected to Internet. But this is not implemented.

For transferring both census and survey records to the ODK Aggregate server we assume that the link to the server was established in the settings and follow the next steps:

  1. Tap on the Send Finalized Form button on the main menu.
  2. Choose the data you want to upload (census or survey).
  3. Select those records you want to send and click on the Send Selected button.

Workflow: Collect

Clone this wiki locally