PMA2013/Ghana Round 1 Indicators

The PMA2020 survey collects annual data at the national (urban and rural) and regional levels to allow the estimation of key indicators to monitor progress in family planning. The resident enumerator model enables replication of the surveys each year, and initially every six months for the first two years to track progress.

For the first round of data collection, referred to as “PMA2013/Ghana,” the survey targeted a sample size of 100 enumeration areas (EAs), which were selected by the Ghana Statistical Service (GSS) to be representative at the national, urban-rural areas, and the 10 administrative regions. The EAs were selected systematically with probability proportional to size with urban/rural stratification in the 10 regions.

Previously, the GSS had selected a sample of 810 EAs for the 2011 Ghana Multiple Indicator Cluster Survey (MICS), and the PMA2013 survey EAs were a subsample of this frame. The rationale for the subsample approach was to have PMA2020 estimates be comparable to the most recent national survey estimate. Four regions (Central, Northern, Upper East and Upper West regions) were oversampled in the Ghana MICS and thus the sample is not self-weighted. GSS provided the revised EA selection probabilities for the PMA2020 sampled clusters for constructing weights.

Ahead of data collection, all households, health service delivery points (SDPs), and key landmarks in each EA were listed and mapped by the REs to create a sampling frame for the second stage of the sampling process. This mapping and listing process took place between August 8th and September 9th of 2013 and preceded data collection in each EA by no more than two weeks. Once listed, households and SDPs were systematically selected by field supervisors using a random start (obtained from a phone-based random number-generating application) and a sampling interval that yielded 42 households. All eligible women in selected households were approached for informed consent to participate in the study.

Up to three private SDPs within each EA boundary were randomly selected from the EA listing. In addition three public health SDPs--at the community (CHPS), primary healthcare (health center), and primary healthcare referral (district hospital--designated to serve the EA population) levels were all selected. Using this multistage sampling procedure and anticipated non-response rates, PMA2013/Ghana had a target sample size of 3,400 females and 150 SDPs. For many EAs, SDP selection included only one (rather than three) public SDP serving that community -- and for many EAs, there were fewer than 3 private SDPs within the EA boundaries. Thus, there were fewer SDPs captured in this survey round than expected.

The survey interviews were conducted from September 4th to November 12th, 2013. Data were not collected in one of the selected enumeration areas due to RE turnover, resulting in a final sample of 3,581 households, 3,758 females, and 149 SDPs across 99 clusters. Weights were calculated based on non-response at the cluster, household, and individual level and applied to all estimates in this report.

PMA2020 uses standardized questionnaires at households and SDPs to gather data that is comparable across program countries and consistent with existing national surveys. Prior to launching the survey in each country, these questionnaires are reviewed and modified by local experts to ensure all questions are appropriate to each setting.

Three questionnaires were used to collect PMA2013/Ghana survey data: the Household Questionnaire, the Female Questionnaire and the Service Delivery Point Questionnaire. These questionnaires were based on model surveys designed by PMA2020/Baltimore staff at the Bill & Melinda Gates Institute for Population and Reproductive Health, KNUST and fieldwork materials of the Ghana Demographic and Health Surveys.

All PMA2020 questionnaires are administered using Open Data Kit software and Android smartphones. The PMA2013/Ghana-R1 questionnaires were in English on the phone and had to be translated into local languages using available translations from similar population surveys and experts in translation. The interviews were conducted in the local language or English in a few cases when the respondent was not comfortable with the local language. Female resident enumerators in each enumeration area administered the household questionnaire and female questionnaire in selected households.

The Household Questionnaire gathers basic information about the household, such as ownership of livestock and durable goods, as well as characteristics of the dwelling unit, including wall, floor and roof materials, water sources, and sanitation facilities. This information is used to construct a wealth quintile index.

The first section of the Household Questionnaire – the household roster – lists basic demographic information about all usual members of the household and visitors who stayed with the household the night before the interview. This roster is used to identify eligible respondents for the Female Questionnaire. In addition to the roster, the household questionnaire also gathers data that are used to measure key WASH indicators, including regular sources and uses of water, sanitation facilities used and prevalence of open defecation by household members.

The Female Questionnaire is used to collect information from all women age 15-49 that were listed on the household roster at selected households. The Female Questionnaire gathers specific information on education; fertility and fertility preferences; family planning access, choice, and use; quality of family planning services; exposure to family planning messaging in the media, and the burden of collecting water on women.

The Service Delivery Point Questionnaire collected information about the provision and quality of reproductive health services and products, integration of health services, as well as WASH practices within the health service post.

Training

The PMA2020/Ghana-R1  fieldwork training started on June 3rd, 2013 with a training of 7 field supervisors, 2 central staff, and 5 rapid response or “Tiger Team” members. This training of trainers was led by PMA2020 staff from the Bill & Melinda Gates Institute for Population and Reproductive Health. These field supervisors then became the trainers for three subsequent Resident Enumerator training sessions that would take place between July 1st and August 24, 2013 in Kumasi, Accra, and Tamale, with a total of 100 Resident Enumerators, with 20 reserve data collectors receiving training.

All training participants were given comprehensive instruction on how to complete the Household, Female and Service Delivery Point Questionnaires. In addition to PMA2020 survey training, all participants received training on contraceptive methods by a Ghanaian obstetrician/gynaecologist.

Throughout the trainings, REs and supervisors were evaluated based on their performance on several written and phone-based assessments, mock field exercises, and class participation. As all questionnaires were completed on project smartphones, the training also familiarized participants with ODK and smartphone use in general. All trainings included three days of field exercises, during which participants entered a mock EA to practice listing and mapping and conduct Household, Female, and Service Delivery Point interviews, recording all responses on their project phones and submitting to a practice cloud server. The trainings were conducted primarily in English, with small group sessions conducted in Twi, Ewe, Ga, Fante, Bono, Dagbani, Frafra, Gonja, Dagarti, Waale and Hausa.

Supervisors received additional training on how to oversee fieldwork and complete Household Re-interviews used to carry out random spot-checks in 10 percent of the households interviewed by REs.

Data Collection & Processing

Data collection was conducted between September and November 2013.

Unlike traditional paper-and-pencil surveys, PMA2020 uses Open Data Kit (ODK) Collect, an open-source software application, to collect data on mobile phones. All the questionnaires were programmed using this software and installed onto all project smartphones. The ODK questionnaire forms are programmed with automatic skip-patterns and built-in response constraints to prevent data entry errors.

The ODK Collect application enabled resident enumerators and supervisors to collect and transfer survey data to a central ODK Aggregate cloud server in real-time. This instantaneous aggregation of data also allowed for real-time monitoring of data collection progress and concurrent data processing while PMA2020 was still active in the field and course corrections could be made. Throughout data collection, central staff at KNUST in Ghana and a data manager in Baltimore routinely monitored the incoming data and notified field staff of any potential errors, missing data, or problems found with form submissions on the central server.

The use of mobile phones combined data collection and data entry into one step and therefore data entry was completed when the last interview form was uploaded at the end of data collection in November.

Once all data were on the server, data analysts cleaned and de-identified the data, applied survey weights, and prepared the final dataset for analysis using Stata® version 12 software. Final data analysis was conducted between January and February of 2014 and the national dissemination workshop was held in March 2014 in Accra, Ghana.

This table shows response rates for household and female respondents by residence (rural/urban) for PMA2014/Ghana-R2. Of the households selected for surveys, a total of 4,111 were found to be occupied at the time of the fieldwork. Of these 4,111, 3,581 consented to the household interview for a response rate of 87.1 %. The response rate was higher in the rural (92.2 %) than urban (82.3 %) area.

In the selected households, 4,160 eligible women age 15-49 were identified and interviews were completed with 3,758 of them (90.3 %response rate). In this case, the participation rate was higher in the urban (91 %) compared to the rural (89.8 %) area.

The principal reason for non-response both at the household and female respondent level was the inability to find respondents at home despite as many as three separate visits to the household. The PMA2013/Ghana response rates are lower than those observed by the GDHS 2008 and may reflect both the growing inaccessibility of respondents as Ghana grows more urbanized, as well as the inexperience of the resident enumerator in reaching respondents. Nonetheless, the PMA2013/Ghana survey response rates are within what is conventionally considered in the literature as acceptable.

To view the sample errors for the PMA2020 indicators described above, download the full indicator report here. For more information about PMA2020 indicators, including estimate type and base population, click here.