Conceived and designed the experiments: CC PME. Performed the experiments: CC. Analyzed the data: CC JDK JMN. Contributed reagents/materials/analysis tools: EMHE. Wrote the paper: CC EAC AM. Edited paper, guarantor of paper: PME.
The authors have declared that no competing interests exist.
Trachoma prevalence surveys provide the evidence base for district and community-wide implementation of the SAFE strategy, and are used to evaluate the impact of trachoma control interventions. An economic analysis was performed to estimate the cost of trachoma prevalence surveys conducted between 2006 and 2010 from 8 national trachoma control programs in Africa.
Data were collected retrospectively from reports for 165 districts surveyed for trachoma prevalence using a cluster random sampling methodology in Ethiopia, Ghana, Mali, Niger, Nigeria, Sudan, Southern Sudan and The Gambia.
The median cost per district survey was $4,784 (inter-quartile range [IQR] = $3,508–$6,650) while the median cost per cluster was $311 (IQR = $119–$393). Analysis by cost categories (personnel, transportation, supplies and other) and cost activity (training, field work, supervision and data entry) revealed that the main cost drivers were personnel and transportation during field work.
Population-based cluster random surveys are used to provide the evidence base to set objectives and determine when elimination targets have been reached for several neglected tropical diseases, including trachoma. The cost of conducting epidemiologically rigorous prevalence surveys should not be a barrier to program implementation or evaluation.
The costs of conducting population-based prevalence surveys for neglected tropical diseases such as trachoma are often cited as a reason that program managers do not conduct baseline or impact assessments when guidelines suggest they are warranted. The authors conducted a review of actual costs incurred during the implementation of 165 district level surveys in 8 national trachoma control programs to identify the median and mean costs per district and per cluster. In addition, the costs of the principal activities that are the most expensive were measured. The data show that field work is the most expensive activity for a prevalence survey, with personnel (
Trachoma is an eye disease, caused by infection with ocular
Where trachoma is suspected to be a public health problem, the WHO recommends that the prevalence of the clinical signs of the disease are estimated using a cluster random survey methodology at the district level
In brief, the PBPS method employs a multi-stage cluster random survey design to randomly select clusters, and households within the clusters. Once households are selected, all members of the household are examined for clinical signs of trachoma disease using the WHO Simplified Grading System
Trachoma prevalence surveys provide an estimate of the burden of disease at the level of interest, usually the district. These data serve as the evidence base for determining how the SAFE strategy should be employed. For example, where the prevalence of the clinical grade TF (trachomatous inflammation, follicular) exceeds 10% in children aged 1–9 years, the WHO recommends district-wide mass treatment with antibiotics and facial cleanliness and environmental improvements—the “AFE” of SAFE. Prevalence survey data are also used to calculate annual intervention targets and ultimate intervention goals (UIGs), such as the number of people who require trichiasis surgery. These targets are used to plan annual activity budgets, forecast the need for donated pharmaceuticals and other supplies, and monitor progress towards the elimination of blinding trachoma.
Although survey implementation may vary by location, there are currently no data on the cost of trachoma prevalence surveys in the peer-reviewed literature. There are examples in the literature where different survey methods were compared to determine the most cost-effective method to estimate immunization coverage
The analysis of prevalence survey cost data did not involve any research on human subjects. The prevalence surveys reviewed in this paper were conducted in accordance with the Declaration of Helsinki and reviewed by the Emory University Institutional Review Board or the London School of Hygiene and Tropical Medicine (LSHTM) Ethical Committee and each country's respective Ministry of Health. External funding for the prevalence surveys was as follows: LSHTM, The Gambia survey; Helen Keller International, Sikasso Region of Mali; The International Trachoma Initiative and The Carter Center, 18 districts in Ghana; The Carter Center, all other surveys.
A systematic review of trachoma prevalence surveys conducted in Ethiopia, Ghana, The Gambia, Mali, Niger, Nigeria, Sudan, and Southern Sudan was performed February through May 2010. This review of prevalence survey costs included surveys that employed a PBPS methodology to estimate trachoma prevalence at the district level, or the administrative unit equivalent to a district (administrative unit with population of approximately 100–250 thousand people:
A data collection tool was used to collect the actual costs incurred in local currency during survey activities from accounting records in the programs. The tool collected data for four cost activities: training, field work, supervision and data entry. Training included costs such as
The data collected in this study captured the incremental cost of conducting prevalence surveys in the context of an existing national trachoma control program. Ministry of Health and NGO salaries and other associated costs were not included in the analysis. Integrated prevalence surveys (more than one disease measured) were excluded from this analysis. “Headquarters” expenses were not included in the primary analysis of prevalence survey costs. Although beneficial, consultant or other outside technical assistance is not required for a national program to conduct trachoma prevalence surveys. Furthermore, the cost of outside technical assistance is dependent on travel expense policies which are unique to each partner. The cost of Carter Center headquarter support for specific survey activities are reported in this review, but were not included in the district-level cost data, as these costs are organization-specific and cannot be generalized.
Once completed, the cost data forms were verified against the financial reports from the Carter Center, Helen Keller International, LSHTM or the Ministries of Health. In Ghana, Ethiopia and Northern Sudan, exact data on distance traveled were not available; the data reported for these programs' distance traveled are estimates from the national programs.
Data were converted to US dollars using the mean of the weighted average exchange rate from the World Bank (
Based on these observations, the analysis generates the overall costs, the average survey costs per district and average costs per cluster for each observation. Data were first entered into Excel and then analyzed using STATA to generate descriptive statistics for each cost activity. Subsequently, a cost composition analysis was performed. The data were classified into activities as defined in the data collection tool to calculate the proportion of the total cost for each cost activity. Within each of the four activities (training, field work, supervision and data entry), four main cost categories were identified: personnel, transportation, supplies and other. The costs for each category were compared against the total cost for each activity to identify the main cost drivers of survey expenses.
Normally distributed data are presented as the mean and standard deviation (SD). Not-normally distributed data is presented by the median and inter-quartile range (IQR).
A total of 29 observations were collected from eight national trachoma control programs. The cost per district by observation is presented in
National program | Observation | Number of districts | Number of clusters | Number of households per cluster | Number of people examined | Total costs ($) | Cost per district ($) | Cost per cluster ($) | Cost per person screened ($) | Reference |
Ghana | Northern & Upper West | 18 | 720 | 30 | 74,225 | 72,249 | 4,014 | 100 | 0.97 | Yayemain 2009 |
Mali | Kidal | 1 | 20 | 24 | 2,165 | 14,777 | 14,777 | 739 | 6.83 | Bamani 2010 |
Kayes | 7 | 140 | 24 | 13,576 | 13,593 | 1,942 | 97 | 1.00 | Bamani 2010 | |
Koulikoro | 9 | 180 | 24 | 19,342 | 17,505 | 1,945 | 97 | 0.91 | Bamani 2010 | |
Sikasso | 8 | 160 | 24 | 18,795 | 19,046 | 2,381 | 119 | 1.01 | PNLCC | |
Segou | 8 | 160 | 24 | 16,471 | 18,553 | 2,319 | 116 | 1.13 | PNLCC | |
Nigeria | Plateau & Nasarawa | 13 | 260 | 16 | 21,606 | 24,036 | 1,849 | 92 | 1.11 | King 2010 |
Southern Sudan | Jonglei (Ayod County) | 1 | 20 | 20 | 2,335 | 25,409 | 25,409 | 1,270 | 10.88 | King 2008 |
Northern Sudan | Kassala | 10 | 132 | 30 | 10,576 | 35,308 | 3,531 | 267 | 3.34 | FMOH GOS |
Blue Nile | 4 | 45 | 20 | 5,166 | 18,799 | 4,700 | 418 | 3.64 | FMOH GOS | |
Gazeira | 7 | 105 | 20 | 10,466 | 42,049 | 6,007 | 400 | 4.02 | FMOH GOS | |
White Nile | 8 | 120 | 20 | 10,570 | 39,168 | 4,896 | 326 | 3.71 | FMOH GOS | |
Gadarif | 10 | 150 | 20 | 13,682 | 47,839 | 4,784 | 319 | 3.50 | FMOH GOS | |
Sinnar | 7 | 105 | 20 | 9,095 | 34,961 | 4,994 | 333 | 3.84 | FMOH GOS | |
River Nile | 6 | 90 | 20 | 7,528 | 20,632 | 3,439 | 229 | 2.74 | FMOH GOS | |
Red Sea | 10 | 150 | 20 | 9,918 | 40,680 | 4,068 | 271 | 4.10 | FMOH GOS | |
Northern | 5 | 66 | 20 | 11,076 | 36,454 | 7,291 | 552 | 3.29 | FMOH GOS | |
North Kordofan | 9 | 135 | 20 | 10,360 | 37,494 | 4,166 | 278 | 3.62 | FMOH GOS | |
South Kordofan | 9 | 135 | 20 | 10,755 | 41,960 | 4,662 | 311 | 3.90 | FMOH GOS | |
Niger | Magaria | 1 | 20 | 24 | 1,789 | 7,884 | 7,884 | 394 | 4.41 | PNLCC Niger |
Matameye | 1 | 20 | 24 | 1,712 | 7,835 | 7,835 | 392 | 4.58 | PNLCC Niger | |
Nguigmi | 1 | 20 | 24 | 1,659 | 7,866 | 7,866 | 393 | 4.74 | PNLCC Niger | |
Maine Soroa | 1 | 20 | 24 | 1,867 | 7,866 | 7,866 | 393 | 4.21 | PNLCC Niger | |
Maradi Commune | 1 | 20 | 24 | 2,393 | 6,132 | 6,132 | 307 | 2.56 | PNLCC Niger | |
Tessaoua | 1 | 20 | 24 | 1,806 | 6,132 | 6,132 | 307 | 3.40 | PNLCC Niger | |
Gaya | 1 | 20 | 24 | 2,036 | 6,650 | 6,650 | 333 | 3.27 | PNLCC Niger | |
Loga | 1 | 20 | 24 | 1,801 | 6,650 | 6,650 | 333 | 3.69 | PNLCC Niger | |
Ethiopia | Amhara | 5 | 90 | 10 | 5,762 | 7,556 | 1,511 | 84 | 1.31 | Ngondi 2008 |
The Gambia | Lower River & North Bank | 2 | 60 | 25 | 2,990 | 7,815 | 3,908 | 130 | 2.61 | Harding-Esch 2009 |
Total | 165 | 3,203 | 301,552 | 672,897 |
When the costs for each survey activity were compared against the total cost (
Activities | |||||
Training | Field work | Supervision | Data entry | Total | |
Category | |||||
Personnel | 1.9% | 40.4% | 11.3% | 10.9% | 64.6% |
Transportation | 1.6% | 22.4% | 1.7% | 0.0% | 25.7% |
Supplies | 0.9% | 5.3% | 0.0% | 0.0% | 6.3% |
Others | 1.4% | 1.7% | 0.3% | 0.0% | 3.3% |
Total | 5.9% | 69.9% | 13.2% | 10.9% | 100.0% |
Training and data entry activity costs were reported by observation as the cost for each activity. These costs were not always directly related to the number of districts surveyed as some programs did not incur cash costs for these activities. The mean cost of training was $1,342 (SD $659) while the median was $1,791.50 (IQR = $588–$1,816). The mean cost of data entry was $2,548 (SD $3,493) and the median was $1,028 (IQR = $415–$4,431).
Although the cost of outside technical assistance was not factored into the district or cluster level cost analysis, there were 9 observations that were surveyed with at least one representative from The Carter Center Headquarters (Atlanta, Georgia, USA) present, covering a total of 58 districts. The average cost for airfare, hotel, meals and incidentals per person-trip was $1,779 (n = 13, SD = $2,027) from 2006–2010.
It is possible that trachoma control programs do not implement prevalence surveys due to a perception that the costs will be beyond the capacity of the program. However, the results of this analysis show that such surveys are not cost-prohibitive. The range of costs per district varied from $1,151–$25,409, in large part due to differences in accessibility and the number of clusters sampled in each survey. Of the 29 observations, only three surveys reported a cost per cluster exceeding $500: Ayod in Southern Sudan, Kidal in Mali and the Northern Region in Sudan. These surveys were characterized by both high transport and personnel costs. In Ayod County of Southern Sudan, where the average cost per cluster was $1,270 and average cost per person screened was $10.88, vast distances of water-logged and unforgiving terrain made vehicle transport impossible, requiring a chartered airplane to transport staff to airstrips from where they traveled to the clusters on foot over a period of days. These exceptional circumstances therefore required additional staff, working for a longer period of time, and transport by chartered aircraft. In Kidal Region (a desert region of Mali), the second most expensive survey per cluster ($739 per cluster, $6.83 per person screened), the sparse population (80,000) and low population density (less than one person per square kilometer) resulted in the national program treating the region as the domain, with the consequence that the distances between clusters was hundreds of kilometers. To conduct this survey, the program rented vehicles instead of using Ministry of Health and NGO transport due to security concerns in the area. The Northern Region of Sudan ($552 per cluster, $3.29 per person screened) is also on the edge of the Sahara with similar demands on transport and time. Least expensive, at under $100 per cluster, were the surveys conducted in the Amhara region of Ethiopia ($84 per cluster, $1.31 per person screened) and Plateau and Nasarawa States of Nigeria ($92 per cluster, $1.11 per person screened) where
Among the cost categories reported, the
The review of data entry costs also presents new findings for Ministries of Health. Although data entry was not an expense for all surveys reported, data entry accounted for an average of 11% of total survey expenses. In this sample, the incremental cost of data entry ranges from 0% in surveys where existing program staff conducted data entry on existing computers incurring no additional cash cost to 25% of the total cost of the survey where external contractors were hired to complete the work. Survey planners should consider the cost of data entry in their own country context to ensure that costs for double entry, analysis and preparation of printed reports are included in budgets.
By design, we did not capture the cost of each Ministry of Health and NGO employee who contributed time to conduct survey work, the incremental cost effectiveness ratio is likely to be underestimated since these costs were not taken into account. This could be included in the analysis as an opportunity cost. However, since the implementation of prevalence surveys is recommended as the standard monitoring and evaluation framework for trachoma control programs by the WHO, these surveys were within the mandate of the Ministry of Health personnel who were engaged in field work and supervision. Salary costs were excluded as they were considered part of the functional trachoma control program and we sought to establish the incremental cost of conducting surveys in the presence of a program. We also did not include the cost of technical assistance (including travel) for ‘headquarters’ staff. Although the average cost of a person-trip from The Carter Center for technical assistance was $1,779 (SD = $2,027), we considered this to be a non-essential cost for a program, subject to considerable variation between supporting NGOs who have different travel policies, and likely to come from a different operating budget which would not have an incremental effect on the cost of a national program.
The selection of a sample representative of the underlying population presents an opportunity to collect data on multiple conditions and this has been done for trachoma and malaria
Although the data presented show costs from a variety of settings, there are a few limitations. The data in this analysis were reported retrospectively and therefore, it is possible that some costs may not have been captured. For some surveys (Ghana, Ethiopia and Northern Sudan) log book entries for distance travelled were not available and we relied on the local knowledge of the national program to calculate distance travelled. Each of these surveys was conducted in the presence of a functioning trachoma control program; there was no need to purchase new vehicles or make other large capital expenses. Survey work performed in the absence of this infrastructure would be more expensive. New country programs may find it necessary to rent vehicles and seek technical assistance for training survey staff, the costs of which would need to be considered in addition to the incremental costs of conducting a survey presented here.
There are variations in the number of clusters surveyed among the different observations, based on the population of each survey domain, which may affect the comparability of the survey costs among different countries. However, the authors expected variation among national programs due to differences such as
Twenty-six out of the 29 observations were conducted with external funding exclusively from The Carter Center, which may imply the cost estimates are limited to those surveys supported by this NGO. However, there are similarities between the cost per cluster from The Gambia, which was fully funded by LSHTM, districts in Mali supported by Helen Keller International, and districts in Ghana co-sponsored by the International Trachoma Initiative and The Carter Center. This suggests that our findings are not unique to the operating principles of one NGO.
Since transport and
The authors would like to acknowledge the national trachoma control programs in Ethiopia, Ghana, Mali, Niger, Nigeria, Sudan, Southern Sudan and The Gambia for their contribution to the data collection for this paper. The authors also wish to thank Stephanie Palmer for her assistance with formatting the manuscript.