Surveillance system design must take into account not only the objectives and outputs that the system has to deliver, but also the need to be able to operate the system in a consistent and affordable way over a prolonged period of time.

Surveillance systems must be capable of capturing data in a consistent and sustainable way, since fluctuations in reporting may mask or mimic genuine changes in the incidence of the disease, exposure or hazard that is the subject of surveillance. While the attainment of a measure of true disease occurrence within a population through surveillance is probably unrealistic for most diseases and surveillance systems, achievement and maintenance of high levels of ascertainment is still desirable. It is particularly important for the early detection of outbreaks and for surveillance of uncommon diseases of public health importance and/or diseases that are the subject of elimination programmes. It also provides a level of reassurance that the surveillance system is providing a relatively representative picture of the distribution of disease within the target population.

Completeness and consistency of reporting to surveillance systems can be affected by a range of system design and operational factors. Statutory or mandatory reporting has been used as a mechanism for achieving high reporting rates, although there is evidence to suggest that this does not guarantee complete, or even high levels of, reporting. Surveys in the UK have demonstrated very low levels of reporting of some statutorily notifiable infections [1], with little evidence that this can be improved by increasing the payment made to physicians for reporting cases [2]. It is also the case that concerns have been raised that making some diseases notifiable may result in patients being reluctant to seek medical attention, because of fears about being notified, which was a significant factor in the decision not to make HIV or AIDS notifiable in the UK. Another strategy that may be used to achieve high reporting rates is to adopt a sample-based or sentinel approach to surveillance, so that resources can be invested in achieving higher reporting rates from a smaller number of reporters e.g. by providing training and support, such as in the form of specialised software for reporting. For rare diseases it is also possible to consider surveillance based on active reporting, where reporters are actively prompted or reminded to report cases on a regular basis. A good example of this latter approach is the British Paediatric Surveillance Unit, which sends out reporting cards every month to all paediatricians in the UK, requesting that they report cases of a list of 10-12 conditions that are the subject of active surveillance [3]. 

Although these approaches can achieve higher rates of reporting, they are not applicable to all diseases or circumstances, and they do not always achieve the desired result. As a general rule, surveillance systems based on capture of data from health services are most likely to be sustainable and achieve acceptable levels of coverage of the target population where the design of the system is coherent with the infrastructure of the healthcare systems within which patients are seen. Where specialist clinics provide the majority of care for a particular disease or group of diseases, such as is the case with sexually transmitted infections in the UK, surveillance based on data reporting from those clinics can often achieve higher quality (with regards to diagnostic validity of reports and compliance with reporting) information for a given cost than would be possible through systems based on universal reporting. National publicly funded health services often have an established culture of central reporting, and as such surveillance based on voluntary (or mandatory) reporting to a national surveillance centre may be more acceptable than in countries with devolved or largely privately funded healthcare. To some extent the increasing use of information technology within healthcare services, and the development of electronic patient records, may overcome some of the barriers to reporting, in that the effort required of clinical staff should become smaller, although this will require the development of standards for electronic data exchange, and clarification of any data protection issues.

The quality and completeness of reporting to surveillance systems is also likely to be better where reporting makes use of data collected for clinical or other operational purposes is captured as a by-product of routine clinical or administrative processes. Once again, the development of electronic patient information systems is likely to make this easier. Other opportunities for capturing data without requiring new or additional effort by clinical staff include making use of laboratory requesting or result reporting data, use of forms completed for the purposes of claiming for payment from health insurance companies, and use of pharmacy records (e.g. for dispensing of vaccines or disease-specific medications).

 References

1. Cartwright KA. Meningococcal meningitis. British Journal of Hospital Medicine 1987;38(6):516, 521-4

2. McCormick A. The notification of infectious diseases in England and Wales. Communicable Disease Report. CDR Review 1993; 3(2):R19-25

3. British Paediatric Surveillance Unit. 19th Annual Report 2004-05. Royal College of Paediatrics and Child Health 2005. London. (http://www.bpsu.inopsu.com/publications/annual_reports/annual-report_2005.pdf)