Memorandum by the Hospital Infection Society
1. Surveillance should be "bottom-up"
rather than "top-down" as meaningful surveillance is
carried out at the "coal face" and the highest degree
of interest is generated there.
2. Regional/National/International data
is a spin-off from the "bottom-up" approach. National
data is interesting and can serve as a useful benchmark but we
should not lose sight of the singular importance of local surveillance.
3. Core datasets should be agreed nationally.
4. Engagement with professional bodies and
royal colleges is imperative and should be at an early stage to
agree on the core dataset.
5. Once nationally agreed core dataset exists
local teams must have freedom to influence dataset design however
the core dataset must not be negotiable.
6. Collation and processing of data should
be undertaken by any method that works.
7. Paper questionnaires have many advantages:
(b) Always "switched on" as opposed
to finding a computer in which to enter data.
(c) Easily modified to local need.
(d) Totally independent of hospital's IT
system. In other words you can collect data however bad the IT
system is at providing data. Also a lot of the data required,
especially clinical data, will not be on a hospital's IT system.
8. Handheld computers for data collection
sounds "sexy" but have major disadvantages:
(b) Requires software for data collection.
(c) Requires training for staff.
(e) Restricted numbers will cause data collect
to fall on a small number of individuals who will most likely
not be able to cope.
(f) Soon out of date requiring replacement.
9. However handhelds/portable PCs may be
extremely useful as a means of feedback. If you carry surveillance
data on the handheld you may be able to answer questions from
clinical teams "on the fly".
10. Data processing can be done locally
if facilities exist and uploaded electronically to a regional
centre. For those with no local facility for data processing questionnaires
can be sent to the regional centre for scanning.
11. Regional centres must not be too big
so as to prevent them from providing an individual service. Specifically,
the best local involvement is obtained by allowing the local teams
to design their own dataset (containing an unaltered core dataset).
To allow this to happen the regional centre must be capable of
providing a scannable questionnaire (or other suitable means)
to meet local needs. There will be a limit as to how many local
questionnaires that can be facilitated hence there will be a maximum
size of a local centre. As an example I feel that only one centre
will be required in Northern Ireland, Wales and Scotland but numerous
centres will be needed for England.
12. Coordination of the regional datasets
is required to facilitate easy aggregation into supra-regional/national
13. Rapid processing of data and return
of the database to individual trusts is paramount.
14. The persons collecting the data will
need all of the following:
(a) Basic computer skills and access to a
15. Any surveillance process is useless
if we do not attempt to stratify risk among patients.