Memorandum by Focus Technologies Inc
Focus Technologies is a company, based in the
United States, with experience in the establishment of electronic
surveillance networks in 10 countries, including the United Kingdom.
A number of barriers to implementation with
direct relevance to the broader surveillance of infectious disease
have been encountered. Some issues have been encountered in other
countriesthe lack of standardised nomenclature and the
diversity of computer connections required while others are specific
to the United Kingdom being a direct result of historical under-investment
in microbiological testing.
1.1 Approaches to surveillance
In broad terms, two models have been widely
used for the surveillance of bacterial populations and their susceptibility
to anti-microbial agents.
1.1.1 Central laboratory testing (ex. TRUST,
SENTRY, Alexander Project)
Clinical microbiology laboratories select clinical
isolates which correspond to the study specifications and send
them by courier to a central laboratory where they are re-tested
using standardised techniques.
1.1.2 De-centralised testing (ex. TSN, EARRS,
Clinical microbiology laboratories send identification
and susceptibility test results to a central repository for collation.
2. THE SURVEILLANCE
2.1 TSN concept and principles
The Surveillance Network (TSN<eptm<rs)
is a de-centralised surveillance system. Data are collected at
regular intervals, usually daily, from hospitals and private laboratories
in 10 countries. TSN participation is voluntary. No financial
relationship exists between participating laboratories and Focus
Technologies. In return for data, Focus Technologies provides
data management services and makes consolidated databases available
to participants thus providing benchmarks against which participants
may compare their local findings. Participants gain access to
TSN data by connecting to an Internet site (www.mrlworld.com).
This site provides a set of tools designed specifically for the
performance of queries upon TSN data.
2.2 Current status of TSN networks
|All TSN Databases||301||861
Table 1 TSN Database Statistics July 2002
Focus Technologies first began recruiting participants in
the United States in 1993. Upon installation, most participants
have been able to provide three years of historical data. Most
TSN databases include reliable data from 1998 onwards.
2.3 Electronic collection of surveillance data
2.3.1 Selection of participants
In order to provide a balanced sample, the geographical location,
size, type (ex. general hospital, teaching hospital, private laboratory,
etc.) and methods used to perform antibiotic susceptibility tests
are taken into account when selecting participants.
To ensure the reliability of the sample, only sites with
acceptable microbiology practices and quality control procedures
are invited to participate. Preference is given to participants
who are able to provide antibiotic susceptibility results in a
quantitative format. Quantitative data allows more sophisticated
quality control procedures to be implemented and can provide a
more detailed view of the evolution of resistance.
2.3.2 Sample size
With TSN Focus Technologies aims to collect in each country
a sample size of between 10 per cent and 15 per cent of all susceptibility
tests performed. This sample size adequate to provide reliable
information to participants and the pharmaceutical industry concerning
the trends in antibiotic resistance. The industry makes use of
the data in the registration of antibiotics, in making strategic
decisions to develop new products and in the marketing of existing
ones. In some countries, the TSN database significantly exceeds
this target sample size. The collection and management of the
quantity of data this presents is made possible only through the
use of electronic data transfer and a high degree of automation.
2.3.3 Data extraction
Data extraction requires a program to interrogate the participant's
Laboratory Information System (LIS). Development of this program
is performed by the manufacturer of the LIS, by the participant
if the necessary competence is available or by Focus Technologies.
2.3.4 Data collected
|Antibiotic tested||The name of the antibiotic
|Qualitative result||Susceptible, Intermediate or Resistant
|Quantitative result||MIC or Zone size where available
|Isolate identification||Genus and species or a code
|Patient identifier||Required to enable removal of duplicate results
|Month and year of birth||Antibiotic resistance is strongly correlated with age
|Gender||Male, female or unknown
|Patient location||In-patient, Out-patient or Intensive care
|Ward or service||For local use only
|Specimen type||Urine, blood-culture
|Specimen identifier||Required to enable removal of duplicate results
|Laboratory code||Code to identify participating laboratory
|Institution||Institution where the patient is being treated
|Region||Institution's geographical region
|Test Method||Antibiotic susceptibility testing method used
18.104.22.168 Encryption of identifiers to protect patient confidentiality
The patient and specimen identifiers are encrypted before
transmission using a key known only to the participating institution.
For additional security, all data fields are further encrypted
and compressed using commercial system, PGP. The participant retains
the PGP public key.
2.3.5 Data transmission
In most cases, participants transmit their data via the Internet
using File Transfer Protocol (FTP). Where local security arrangements
forbid this kind of transmission, the data are sent as an electronic
2.3.6 Data processing
Data processing activities are centralised at Focus Technologies'
head office in Herndon, Virginia in the United States.
Normalisation would not be required if all laboratories adhered
to the same nomenclature for organisms, antibiotics and specimen
descriptions. This is the position of ONERBA, the organisation
that oversees the surveillance of antibiotic susceptibility data
in France. At present there is no rigid standardisation of nomenclature,
apart from that introduced passively through the use of automated
systems, so in order that data from diverse sources may be compared,
Focus Technologies has defined a standard nomenclature for organisms,
antibiotics, specimen types, gender and patient locations. Participants
are not required to use this standard nomenclature when transmitting
data, instead conversion tables are established with each participant
before data transmission begins. The data are automatically converted
to the standardised nomenclature on receipt. If a previously un-encountered
data item is received, it is referred back to the participant
for interpretation and the conversion tables updated.
22.214.171.124 Detection of unusual resistance
Once normalised, data then pass through a rule-based expert
system. This program detects uncommon resistances. The rules for
each country are defined in collaboration with an advisory board
comprising local experts in antibiotic resistance. Many of these
rules will be common between countries (ex. resistance to vancomycin
in Staphylococcus aureus) however where a local frequency of resistance
is typically very low or very high, this will be taken into consideration.
Participants are notified of any result considered to be rare
or unlikely and are asked to confirm by re-testing if they have
not already done so. In the case of extremely rare occurrences,
participants are requested to send the strain to a local reference
laboratory for confirmation. Data, which cannot be verified in
this way, are excluded from the database thereby maintaining database
126.96.36.199 Removal of duplicate results
In order to avoid bias from multiple transmission or repeated
testing, duplicate results are prevented from entering the TSN
database. A duplicate is defined as a second occurrence within
five days of the same organism from the same patient with the
same susceptibility pattern.
2.3.7 Quality control procedures
Data are added to the TSN databases on a monthly basis. Before
being released to the participants, the updated databases are
compared statistically to the previous version.
2.4 Access to surveillance data
2.4.1 Prevention of unauthorised access
The TSN servers are protected by a key-card access system.
Essential computer operations personnel from the Focus Technologies
Information Systems team alone are provided with key-cards.
2.4.2 Participant access
Individuals designated by participating laboratories are
provided with passwords allowing access to TSN data via the Internet.
They may view their own data and the consolidated national and
regional databases. They may not examine data from an individual
institution other than their own.
2.4.3 Third-party access
Advisory Board members are granted access to the national
and regional databases. Subscriptions are commercially available.
2.4.4 Publication of TSN data
TSN data is now frequently quoted in peer-reviewed journals,
conference presentations and posters. Participants are encouraged
to publish their experiences.
3. ADVANTAGES PRESENTED
3.1 Pertinent source of information
Clinical microbiology laboratories provide an excellent source
of surveillance information dedicated as they are to the isolation
and study of disease-causing micro-organisms. TSN attempts to
capture all susceptibility results for all isolates unlike other
studies which tend to focus on a narrow range of organisms and
antibiotics. The reason for this is simple. It is difficult to
predict which organisms and antibiotics will be concerned when
future resistance threats arise. Many susceptibility results for
yeasts and mycobacteria are also received. The lack of standardisation
in these areas precludes the use of this data at present.
3.2 Data volumes
Electronic data transfer and automated handling allows for
a much larger sample than would be possible with a central laboratory
testing or paper report approach.
3.3 Trending capability
TSN networks are based upon a fixed core of participating
laboratories. Automatic data transmission provides a consistent
flow of information. A stable group of participants allows reliable
3.4 Real-time data availability
Data are processed in monthly batches. Data remain in quarantine
for a month while quality control procedures are performed. The
delay between data collection and availability is between one
and two months.
3.5 Interaction with laboratories
When issues arise, they are discussed with the laboratory
concerned. Frequently this leads to improvements in laboratory
practice. Over time this interaction is expected to improve the
reliability of TSN databases.
3.6 Detection of "alert" organisms
TSN's expert rule system provides a means of rapid detection
of rare or previously unencountered resistance. Often these occurrences
occur through misidentification of the organism concerned or other
technical error. The resolution of these incidents may also lead
to improvements in laboratory practice.
4. BARRIERS TO
Despite the enthusiasm clinical microbiologists have shown
for the project, the implementation of TSN in the United Kingdom
has proved more difficult than in other countries.
4.1 Data quality
The primary reason for generating antibiotic susceptibility
results is to treat patients suffering from infectious disease.
The consolidation of this information for epidemiological analysis,
outside of their institution, has not been a consideration for
most clinical microbiology laboratories in the United Kingdom.
4.2 Laboratory practice
4.2.1 Identification of isolates
Identification of micro-organisms to the level of species
is common practice in most developed countries. It is considered
important to the correct interpretation of susceptibility tests.
Manual and automated test systems for the identification of bacterial
and fungal isolates have been available for more than 25 years
and are widely used. In the United Kingdom the precise identification
of bacterial isolates, in particular Gram-negative rod-shaped
bacteria from urine samples, is commonly considered to be unnecessary.
These organisms are frequently grouped together under the pseudo-classification
"coliforms" for which no standard definition exists.
The collection and analysis of this data is of dubious value since
the strains termed "coliforms" may have widely varying
susceptibilities to antibiotics.
4.2.2 Antibiotic susceptibility testing
Automation in susceptibility testing has been adopted more
slowly in the United Kingdom than in other countries through unwillingness
or inability to invest in new technology. The impact on surveillance
|TSN Database||Antibiotics Tested
Table 2 Average number of antibiotics tested per strainTSN
databases July 2002
188.8.131.52 Few antibiotics tested
Instruments such as the Microscan, Vitek or Phoenix systems
that test minimum inhibitory concentrations employ standard test-panels
of between 12 and 20 antibiotics. Automatic inhibition zone readers
are widely used in France where pairs of large square agar plates,
each with a capacity of nine antibiotics are commonly used. The
testing of large numbers of antibiotics permits the recognition
of patterns of resistance. These patterns allow the detection
of a number of the mechanisms employed by bacteria to evade or
counteract the anti-bacterial effects of the antibiotic. Knowledge
of the resistance mechanisms present can provide useful information
to support the choice of therapy. Expert systems are widely used
to automatically perform these analyses. In the United Kingdom,
the tradition has been to test each bacterium isolated against
fewer antibiotics. The source of data is therefore less rich and
the choice of molecules chosen by each laboratory to represent
each class of antibiotics is more varied. This has an adverse
effect upon the ability to consolidate and perform inter-laboratory
184.108.40.206 Qualitative results only
Most automated techniques provide quantitative as well as
qualitative results. Quantitative results allow more detailed
analyses of microbial populations and are therefore of value to
surveillance programmes. Few laboratories in the United Kingdom
are currently able to supply quantitative susceptibility results.
4.3 Testing standards
In the absence of a European Union standard for susceptibility
testing, laboratories in the United Kingdom apply the British
Society for Antimicrobial Chemotherapy (BSAC) or the American
National Committee for Clinical Laboratory Standards (NCCLS) recommendations.
Some laboratories continue to use the method devised by Stokes.
This lack of standardisation complicates surveillance because
results obtained by different methods cannot be combined to provide
an overall statistic.
4.4 Information system incompatibility
The automation of data extraction from laboratories in the
United Kingdom has been greatly assisted when laboratories have
in-house computer expertise. Experience with participants whose
computer support is provided by external consultants has been
less favourable. There is little standardisation between laboratory
information systems. Recent consolidation among suppliers of laboratory
information systems may come greater harmonisation. Where participants
have adopted automated tests producing quantitative results, it
is frequently the case that their laboratory information system
is not designed to accept and store them.
An Institution is a hospital or clinic for which the participating
laboratory performs all microbiological analysis. Back