Available literature on best practice in public health surveillance is limited particularly when searching for literature that refers to both Indigenous populations and trachoma. The available and relevant literature is summarised in this section, firstly by looking at best practice in public health surveillance systems broadly, then looking more specifically at literature that discusses trachoma surveillance.
2.3.1 Best practice in public health surveillanceA relevant review of surveillance systems was carried out by Wilkins et al. (2008) . Although the review is of public health surveillance systems in general, it is of systems in developing countries, which has some parallels in terms of the economic and environmental conditions of the population for whom trachoma is most prevalent in Australia. Firstly, the authors reiterate the paucity of published information on lessons to be learned from evaluations of public health surveillance systems. The situation does not seem to have changed in 2010.
This study assessed public health surveillance systems in five countries in Latin America, Africa, the Middle East and Asia between 1992 and 2000. The assessments were undertaken by staff of the US Centers for Disease Control and Prevention, who visited each of the countries in teams of two to six members and interviewed high-level ministry officials and other ministry staff, and personnel from all levels of the health care systems. They also directly observed data collection and analysis tasks. The assessment was of information sources, scope, and dissemination of information. A retrospective review of these assessments was then undertaken, from which a standardised assessment tool was built that identified six attributes of good information systems for public health surveillance activities. The attributes were based on those from CDC Updated guidelines for evaluating public health surveillance systems , and included timeliness, accuracy, simplicity, flexibility, acceptability and usefulness. The researchers then field-tested this tool reviewing a second system within one of the five countries originally assessed. top of page
The results were that of the six systems assessed, each was deficient in four of the six attributes, with two systems having five deficiencies and three systems having six deficiencies. The most common problems identified amongst the six systems were in relation to the design of the system (i.e. complexity, forms, and responsiveness to needs of decision makers), people (lack of motivation of personnel, and inadequate training and/or supervision), and dissemination of information from the systems to allow action. Specific results relating to each of the attributes were as follows:
- Timeliness was considered a problem in five of the six systems reviewed. The most common issues were delays in receiving information due to lack of motivation or training of personnel, communication issues (e.g. expense of transmitting information), and poor or non-existent case definitions. Inefficient or insufficient use of computers was also identified as problematic in two of the five systems reporting timeliness as an issue; computers were available in both instances but data entry was slow or behind schedule.
- Accuracy was problematic in all six systems reviewed to some degree. Data flow was the major issue, that is, how and to whom cases and conditions were to be reported was unclear, and data and information were not used at the point of collection and therefore opportunities for the correction of errors did not present. In some cases forms were difficult to complete and contributed to poor quality data. Poor training and supervision of personnel were also factors relating to this attribute.
- Simplicity was a problem in five of the six systems. Multiple reporting forms and lack of clarity of responsibilities in responding were the major contributing factors.
- Flexibility was a problem in four of the six systems. The main issue was lack of routine mechanisms for review and adaptation of systems to reflect changing needs of users.
- Acceptability was a problem in all six systems. All national decision makers rated their systems as unacceptable; and local decision makers and data providers reported that the systems were unresponsive to local needs.
- Usefulness was a problem for all six systems. The main issue was the lack of dissemination of data from these systems.
2.3.2 Best practice in trachoma surveillanceThe single article returned as a result of the search terms ‘public health surveillance’ and ‘best practice’ and ‘trachoma’ was by Usma et al. (2009)6, which was a population-based, cross-sectional study of the prevalence of refractive errors and eye diseases in urban and rural school children. The objective of the study was to assess the prevalence of refractive error and common ocular diseases in school-aged children in urban and rural populations in and around Hyderabad, India. A total of 3,314 school children were reviewed: 1,789 from urban areas and 1,525 from rural areas. Results were reported generally in relation to visual impairment, and also specifically in relation to trachoma. With trachoma, the study found that it was the leading cause of ocular morbidity in the rural group compared with the urban group. It recommended that regular and periodic eye check-ups for school children be undertaken as a necessity, especially in economically backward classes of the community. No other advice on best practice surveillance activities for trachoma or other eye issues were provided.top of page
There is some literature on estimating the prevalence of trachoma, which may be relevant to the current Australian arrangements. Ngondi et al. (2009)7 reviewed the literature in relation to survey methods for trachoma to obtain accurate prevalence figures. The reviewers highlight the importance of quantifying the disease burden associated with trachoma for the purposes of planning, implementing, monitoring and evaluating trachoma control programs. They reiterate the difficulties associated with capturing accurate trachoma prevalence data amongst the populations most affected due to geographic remoteness, political marginalisation, lack of up-to-date population census data, high rates of mobility/migration, insecurity and seasonal inaccessibility due to weather and poor road infrastructure.
The reviewers compare three survey methods: cluster random sampling, trachoma rapid assessment and acceptance sampling trachoma rapid assessment. Cluster random sampling is a population based study of prevalence involving the selection of non-overlapping sub-populations or clusters based on geographic or political boundaries. The approach usually uses a two-stage process, which selects villages (clusters) to be surveyed as the first step, and households within those villages as the second. Trachoma rapid assessment uses a convenience sample drawn from high risk communities. Because it is not based on probability sampling, rapid assessment is not meant as a means of estimating prevalence, although it is often used to do so. Finally, acceptance sampling trachoma rapid assessment takes a sample of individuals and continues to test for active trachoma until a predetermined number of cases are identified or a maximum sample size has been met.
The latter approach, that is, reaching a maximum sample size, is the approach that lends itself to estimates of prevalence. The reviewers conclude that cluster random sampling has the most benefits in relation to measuring trachoma prevalence, and recommend that it remain the ‘gold’ standard for trachoma surveys. However, they also identify methodological deficiencies in cluster random sampling that have been reported in the literature, such as the way in which sample size is calculated and lack of standardisation amongst trachoma graders.
The current arrangements for reporting of trachoma in Australia use the rapid assessment approach to estimating prevalence. This would be problematic if it were not for alternative targeted studies of trachoma using the gold standard approach. Fortunately one was undertaken during 2008 (reported in 2010)8, the first national study of trachoma prevalence since 1980 . The researchers point out that the findings using this approach are similar to those obtained from the national surveillance system10,11. However, a quick review of the survey results reveals apparently large differences in prevalence between the national surveillance data and the National Indigenous Eye Health Survey results12.
Myatt et al. (2005)13 also examine prevalence using lot quality assurance sampling to classify individual communes in Vietnam by the prevalence of active trachoma among children in primary schools. Apart from discussing the value of lot quality assurance sampling as means to efficiently target resources for intervention against active trachoma, the researchers also describe the school-based approach to screening for trachoma. They point out the benefits of school-based screening, which are that it ensures that the group with the highest prevalence of the active disease is targeted, it reduces sampling overheads by having the children in one place rather than having to travel to different places, it increases the efficiency with which the examination can be carried out due to the higher co-operation of children of school age (i.e. aged 6-11 years in this case) compared with younger children, and allows the ability to cover larger populations than would be feasible with community-based sampling due to children coming into the school from several villages.
The researchers also point out the disadvantages of school-based screening, which are that the sampling is limited to days when school is in session (which excludes weekends, public holidays and examination periods), that there is a selection bias in that although school is compulsory for all children in this age group, there is lower attendance for some sub-populations, and in that some children with active trachoma may not attend school due to other related aspects such as diarrhoea from living in unhygienic conditions.top of page
6Uzma N, Kumar BS, Khaja Mohinuddin Salar BM, Zafar MA, Reddy VD (2009). A comparative clinical survey of the prevalence of refractive errors and eye diseases in urban and rural school children. Canadian Journal of Ophthalmology, 44(3): 328-33.
7Ngondi J, Reacher M, Matthews F, Brayne C, Emerson P (2009) Trachoma survey methods: a literature review. Bulletin of the World Health Organisation, 87(2):143-51.
8Taylor HR, Fox SS, Xie J, Dunn RA, Arnold AL, Keeffe JE (2010). The prevalence of trachoma in Australia: the National Indigenous Eye Health Survey. Med J Aust, 92(5):248-53.
9Royal Australian College of Ophthalmologists (1980). Report of the National Trachoma and Eye Health Program, Sydney.
10Tellis B, Fotis K, Dunn R, Keeffe J and Taylor H (2007). Trachoma Surveillance Report 2006/7, National Trachoma Surveillance and Reporting Unit, Centre for Eye Research Australia.
11Taylor HR, Fox SS, Xie J, Dunn RA, Arnold AL, Keeffe JE (2010). The prevalence of trachoma in Australia: the National Indigenous Eye Health Survey. Med J Aust, 92(5):248-53.
12Tellis B, Fotis K, Dunn R, Keeffe J and Taylor H (2009). Trachoma surveillance report 2008, National Trachoma Surveillance and Reporting Unit, Centre for Eye Research Australia.
13Myatt M, Mai NP, Quynh NQ, Nga NH, Tai HH, Long NH, Minh TH, Limburg H (2005). Using lot quality-assurance sampling and area sampling to identify priority areas for trachoma control: Viet Nam. Bulletin of the World Health Organisation, 83(10):756-63.