According to the CDC Guidelines, “the sensitivity of a surveillance system can be considered on two levels. First, at the level of case reporting, sensitivity refers to the proportion of cases of a disease (or other health-related event) detected by the surveillance system. Second, sensitivity can refer to the ability to detect outbreaks, including the ability to monitor changes in the number of cases over time”86. Mathematically, sensitivity is calculated as:
True Positive + False Negative
At the level of case reporting, according to the trachoma surveillance coordinators and public health staff, the diagnosis of trachoma is not likely to be missed (false negative) or diagnosed as something else (false negative) by practitioners with experience in rural and remote health services. Therefore, assuming rural health practitioners receive appropriate training (the training is not considered complex), the likelihood of true positives is high and the likelihood of false negatives is low, which means the sensitivity of the surveillance system is high. Eye health experts consulted advised that if screening staff are not sufficiently experienced (but they have had training) the staging of trachoma may be noted incorrectly (e.g. TF diagnosed when TI) but it is unlikely that a child who has active trachoma will not to be identified.
It is more difficult to comment on whether the current trachoma surveillance system is able to detect outbreaks. Firstly, the use of the term ‘outbreak’ in the context of trachoma is not appropriate, there are no trachoma ‘outbreaks’. There is no doubt that should trachoma prevalence in screened communities increase, the current trachoma screening program has the potential to detect it, in a relatively timely fashion at the local level. There is concern about the impact on prevalence of the number of cases that might be missed on screening day due to absenteeism. Many stakeholders believe that the prevalence of active trachoma in the children that are screened may not be representative of the prevalence in those that are not screened. Even if the current surveillance system accurately identified an increase in trachoma prevalence, due to the poor timeliness of the current reporting, the increase would not show in the national surveillance system for a considerable number of months post identification at the local level.top of page
Screening coverage is important in determining sensitivity. Current data shows considerable variations in trachoma prevalence year to year in the same communities. For example, in reviewing the National Surveillance Reports for the Katherine region, community KA_10 had a 100% prevalence rate for active trachoma in 2006 based on two children being examined. In the following two years prevalence was 6% and 97% based on 36 and 31 children being examined respectively for 2007 and 2008. There was little difference in the number screened for 2007 and 2008, however there was huge variation in prevalence between the two years (6% and 97%) suggesting that the screened population may not be representative. No information was provided on the screening coverage for KA_10, however the screening coverage was 38% for the Katherine region. If screening coverage was improved (target screening of 80% of children aged 5-9 years), then the variability in prevalence could be confidently attributed to actual changes in the community, thereby improving sensitivity.
As already described only a subset of ‘at risk’ communities are screened as identified by jurisdictional health departments in NT, WA and SA based on knowledge of relevant medical practitioners, PHUs, and/or ACCHSs. The 2008 National Indigenous Eye Health Survey used a different sampling method involving a national, stratified, random cluster sample of Indigenous children (5–15 years) and adults (40 years) in 30 communities across Australia87. This survey produced estimates of active trachoma prevalence that are broadly similar to those derived from the national surveillance system, although some differences are apparent. The survey also identified active trachoma in populations where no active trachoma screening program exists. Thus it can be concluded that the communities currently identified as ‘at risk’ communities do not exhaust the areas in which trachoma may exist. There is no mechanism in the current surveillance system to identify active trachoma in not screened communities.
Findings: With respect to the sensitivity of the national surveillance system the evaluation team finds:
- the current system is highly sensitive at case level (assuming screening staff have received training), accurately diagnosing children with trachoma in screened communities;
- in screened communities, the current system allows for the identification of trachoma at the local level in a reasonably timely fashion;
- for not screened communities, the current system should be supplemented by periodic surveys to ensure that there is a process for detecting any trachoma;
- the adequacy of screening coverage should be assessed by cross-correlating results with the Indigenous Eye Survey and also by further investigating variations in prevalence from year to year.top of page
86CDC (2001) Updated Guidelines for Evaluating Public Health Surveillance Systems. Recommendations and Reports July 27, 2001/50(RR13);1-35
87Fox SS, Arnold A-L, Keeffe JE, Taylor HR (2010). Sampling and recruitment methodology for a National Eye Health Survey of Indigenous Australians. Aust N Z J Public Health 2010. In press