Introduction
69. This chapter looks at HAI surveillance systems in England and Wales and defines some basic principles.
HAI Surveillance Systems in England and Wales
England
70. The HAI surveillance system for English hospitals, known as the Nosocomial Infection National Surveillance Scheme (NINSS), is located in the Nosocomial Infection Surveillance Unit at the Public Health Laboratory Service Headquarters, Colindale.
71. The overall aims of NINSS are to develop standard protocols and definitions for the surveillance of HAI, to estimate the incidence of HAI, and to provide data on HAI that are comparable within the hospital, between hospitals, and over time.
72. To date, two protocols have been developed in collaboration with hospital clinicians, viz: a protocol for the surveillance of hospital-acquired bacteraemia and a protocol for the surveillance of surgical site infection. In the longer term, it is intended that protocols will also be developed for urinary tract infection and lower respiratory tract infection.
73. The choice of hospital-acquired bacteraemia as the first module of surveillance to be developed for NINSS was based on:
74. Surgical site infection was chosen as the second module since, in many hospitals, it is the second most common cause of HAI. The protocol targets surgical procedures which are common and/or associated with a high risk of infection and which are likely to require more than three days of post-operative hospital stay. The surgical procedures are grouped into categories with each category containing a set of clinically similar procedures. Twelve categories have been targeted to date. Examples include abdominal hysterectomy, coronary artery bypass graft, hip prosthesis, large bowel surgery and vascular surgem The system is designed to be as flexible as possible to enable participants to choose which categories they wish to target in accordance with the resources at their disposal.
75. In each case, the surveillance system is based on active, prospective, hospital-wide surveillance and, in order to obtain sufficient data, each period of surveillance is carried out for at least three months. In the case of the surgical site infection protocol. all patients are followed up from the day of surgery until discharge, or for a maximum of 30 post-operative hospital days, whichever is the least. Extended terms of follow-up apply where surgery involves the insertion of an implant. The Scheme is selective insofar as it is targeted at specific infections, and intermittent (i.e. undertaken every three months to obtain comparative data between hospitals and aggregate data over time).
76. Participation in the Scheme is voluntary. Approximately 50 hospitals have piloted the protocol for the surveillance of hospital-acquired bacteraemia and another 50 have piloted the protocol for the surveillance of surgical-site infection. A total of 8 hospitals have piloted both protocols.
Wales
77. In Wales, an HAI Surveillance System is being developed with funding from the Welsh Office. This builds on the surveillance systems already in place in Wales, using existing IT as far as possible. The four components of the system comprise Methycillin-Resistant Staphylococcus Aureus (MRSA) surveillance, hospital outbreak surveillance and surveillance of alert organisms, alert conditions and clinical surveillance.
78. The project is managed by a small project team, consisting of two consultant microbiologists, one consultant epidemiologist and, working exclusively on the project, one clinical scientist and one part-time infection control nurse. The project team is itself overseen by the Welsh Hospital Infection Strategy Group, with multidisciplinary membership, including representatives from the Welsh Office. Consultants in Communicable Diseases Control (CCDC), microbiologists, infection control nurses and chief executives.
79. Surveillance of MRSA is the longest established component of the system, with continuous, Principality-wide surveillance in place since 1996. This uses the existing system of communication, COSURV, between laboratories and the Communicable Diseases Surveillance Centre, (CDSC), (Wales), for the transmission of standardised data on every new isolation of MRSA, leading to the production of monthly reports to participants. MRSA data collection and analysis is based on laboratory of isolation, thereby securing the anonymity of individual hospitals.
80. This was supplemented in 1997 by a more detailed survey of control policies, screening methods, prevalence and incidence data and outcomes. There are plans to repeat these additional studies on a regular basis. A comprehensive picture of the impact of MRSA in Wales has been developed, together with a review of the changes over time.
81. A scheme for the reporting of hospital outbreaks of infection started in 1996 and was applied across the whole of Wales in January 1997. This is a paper-based system with monthly reports from infection control teams. Quarterly reports, (including descriptions of outbreaks of interest), are issued to participating centres and CsCDC. This has been successful in defining the database of outbreak control experience and has highlighted variation in reporting practices.
82. Specific software (InControl) has been developed for the surveillance of alert organisms, alert conditions and clinical surveillance. This has been piloted in three acute trusts since September 1997 and is now being made available to all infection control teams throughout Wales. The system provides for local surveillance based on local needs, together with facilities for export of the data to CDSC Wales in a standardised format to facilitate comparisons.
83. The system places particular emphasis on minimising data entry and the appropriate and timely availability of results. The range of alert organisms, alert conditions and clinical surveillance can be determined at a local level dependent on local interests and availability of resources, although there is a minimum dataset of alert organisms and alert conditions. Ease of data entry is achieved by automatically downloading patient and organism information from the Telepath Laboratory System (an interface with the ACT CILMS system has been written but is not yet in use). Datasets and definitions are based on those of the NINSS project in England.
Basic Principles of Hospital Acquired Infection Surveillance
Basic Principles
84. Wherever possible, surveillance systems should be simple in design and use routinely available data. Accurate data should be collected in a standardised manner from a representative sample of the population under observation, using agreed definitions. Data collection from the population or sample should be complete. Data analysis should take account of variations in internal and external risk factors for HAI, and standardised rates should be reported. For surveillance to be effective, it is vitally important that data are reported promptly and that they are perceived to be of value to those who receive them.
Approaches to Surveillance
85. To be effective, surveillance should be carried out at both local and national level. At local level, (e.g. ward/area, hospital, health centre, nursing home), the data collected is collated and analysed locally and reported back to local clinicians by the Infection Control Team. Preventable factors are identified and mechanisms for their reduction or elimination are implemented locally. At national level, the collection, collation and interpretation of a representative set of HAI data., based on common definitions, provides comparative information about the national burden of HAI. It also allows the identification of emerging problems which may have common factors and a national distribution, and which may not be apparent at local level.
Data Collection., Databases, Data Analysis, Data Reporting.
86. Data collection should be carried out using agreed, standardised definitions. Data may be derived from routine systems, such as routine laboratory reports of isolates, or sought by the Infection Control Team. For optimal efficiency, maximum use should be made of routinely available data supplemented by specific surveys or other investigations, where necessary. Methods of obtaining data must be simple and minimum datasets clearly specified. Appropriate denominator data should be obtained.
87. Databases should be computerised, using programmes which are user-friendly and easy to interrogate.
88. Data analysis should include the calculation of standardised rates, using the denominators collected for this purpose. This allows for variations in risk factor distribution and ensures comparability of data from different sources.
89. Data reporting should be prompt and sent to all relevant parties, particularly those who collected the baseline data and their managers. Wherever possible, data should be transmitted electronically.
Data Compatibility
90. Rigorous standardisation of protocols will be a prerequisite if meaningful comparisons are to be made within the UK. This means that data collection protocols and definitions must be agreed and supported through regular exchanges and communications between all participants. True comparisons will only be possible if the studies use the same definitions and a common methodology.