The Second National Survey of Infection in Hospitals: methods of data collection and overall impressions

Journal

of Hospital

Infection

(1996)

32, 57-60

SHORT

REPORT

The Second National Survey of Infection in Hospitals: methods of data collection and overall impressions I. M. Thompson Department

and E. T. M. Smyth

of Bacteriology, The Royal Hospitals Trust, Belfast BT12 6BA, UK

Received 31 May 1995; revised manuscript accepted for publication 19 &ly 1995 Summary:

The Hospital Infection Society, in association with the Public Health Laboratory Service and the Infection Control Nurses Association, launched the Second National Prevalence Survev of Infection in Hospitals in 1993. On completion of the survey, 157 hospitals throughout the British Isles had taken part with approximately 37 000 patients surveyed. The survey in the Royal Hospitals Trust, Belfast, included all inpatients who were present in the wards on the survey day. Infection control nurses (ICNs) from a selected number of the other participating hospitals were sent questionnaires and asked about their methods of data collection, and their overall impression of the Second National Survey. All respondents received help in performing the survey, but only 10% provided prior training to ward staff before the survey visit. Sixty-five percent of respondents supplied information to the wards prior to the survey visit. In comparison with the questionnaire respondents, it was found that using a ‘link’ nurse to assist in data collection was much more efficient in relation to the time involved. Many ICNs expressed concern over the amount of time spent on this survey, although in general the value of such a survey was appreciated. Twenty-six percent of respondents indicated that they would not be willing to undertake a future survey. Keywords:

Hospital-acquired

infection;

data

collection;

surveillance

Introduction

The First National Prevalence Survey of Infection in Hospitals was conducted in 1980. Of the 18 163 patients surveyed in 43 hospitals, 19.1 o/ohad infections, about half of which were community acquired.’ The Second Correspondence OlY5-67Ol/Y6/OlOOS7+O~

to:

I. M.

Thompson

$12.00/O

0 1996 The Huspml

57

Infection

Socm~

I. M. Thompson

58 Table Hospital

location

Northern Republic England Scotland Wales

Ireland of Ireland

Totals Mean Median

I.

Time

and E. T. M. Smyth

spent

and number

No. of hospitals involved

of beds

Range of beds surveyed

surveyed Range of times (one person)

: 7 6 1

150-310 270-350 330-650 113-500 281

25-90 60-200 20-123 23-120 40

18 -

6102 339 320

1352 75.1 72

(h)

National Prevalence Survey was undertaken on a much bigger scale with 190 hospitals expressing an interest, and 157 completing the survey. A total of 37 111 patients were surveyed. To obtain the maximum benefit of participation in this survey, close involvement of clinical staff was sought, since they have an in-depth knowledge of the patients in their care and could prepare some data in advance. This was also an ideal opportunity to get clinical staff more involved in infection control matters. This ‘link’ nurse system worked well and allowed us to survey 781 beds, the greatest number of beds surveyed by any of the participating hospitals. To assess how other participants fared, questionnaires were sent out to a small number of infection control nurses (ICNs). Methods

The use of a ‘link’ nurse to facilitate data collection was piloted on two wards. Nurses have 24 h contact and a good communication system, and were therefore suited to this role. Briefing sessions for the ‘link’ nurses were held the week prior to the survey date. This session lasted approximately 24 min and introduced the ‘link’ nurses to the survey, and to the role they would play. To assessthe methods used in other participating hospitals questionnaires were sent out. A small sample of 20 (12.7%) hospitals were identified as having participated in the National Survey. The hospitals were spread geographically throughout the British Isles (Table I). A covering letter assured confidentiality and explained the reason for the survey, and how to fill in the questionnaire. Results

Assistance with survey ICNs were asked what assistance they received in carrying out the survey. Out of the 20 who replied, 18 (90%) ICNs were assisted by a microbiologist.

Data

collection

and

survey

impressions

59

The computer posed problems for many; although not hard to use, many found it slow and cumbersome. Three (15%) of the respondents could not finish the survey as the computer ‘crashed’. Interpreting the Hospital Infection Society definitions did not appear to be a problem for 11 (55%) of the respondents, but nine (45%) felt that the definitions could be improved. The latter group did not indicate in which areas any improvement could be made. Thirteen respondents (65%) received help from ward staff. Of these, 10 (50%) provided verbal and written information, two (10%) provided verbal information and one (5%) written information only. In addition, two (10%) of the 13 respondents held training sessions. Beds surveyed and time invoiced Respondents were asked to estimate the time they spent carrying out the survey and the number of beds they surveyed. Eighteen answered these questions. The times ranged from 20-245 h (mean 75.1 h). The number of beds surveyed ranged from 113-650 ( mean 339). To calculate the rate per bed, the total time in minutes vvas divided by the total number of beds. This was 13.3 min/bed. The authors’ total time was 71 h for a total of 781 beds, giving a rate of 5.5 min/bed. The timings are per person carrying out the surlrey. Willingness to repeat the sun$ey in the future Finally, when asked if they would be willing to undertake a similar survey again, five out of 19 (26%) respondents indicated ‘no’. One respondent answered ‘yes’, but added ‘but not of my own free will’. Another who answered ‘no’ added ‘not given the choice’. Five out of 19 respondents who indicated ‘no’ added ‘too time-consuming’. Discussion

An impression from this small sample is that many found participation in the survey quite stressful and time-consuming. Although the value of carrying out such a survey is understood, it is doubtful that many would freely participate in future surveys. Available time seems to be one of the major factors. As with all surveillance, a balance of existing resources and time available needs to be achieved. Our experience suggests the need to involve staff at ward level, as they can answer questions more readily in relation to the patients in their care. Involvement also encourages them to become more aware of infection control problems. Use of a portable computer to collect data needs to be reassessed. Our analysis using ‘link’ nurses and adequate preparation resulted in a survey time of 55 min/bed. Times in other hospitals were greater, which may be due to increased time seeking information at ward level. \+‘e recommend our method of surveillance preparation to future participants of such surveys.

60

I. M. Thompson

and E. T. M. Smyth

We wish to thank all those infection control questionnaire and to all those ‘link’ nurses who to Mrs Lorna Dickson ICN for her support, Mr Dr J.G. Barr for reviewing the manuscript.

nurses who took the time made our task easier. Our G. McIlvenny for processing

to fill in our thanks also go the data and

Reference 1. Meers PD, in Hospitals.

Ayliffe GAJ, Emmerson J Hasp Infect 1981;

AM et 2 Suppl.:

al. Report 13-17.

on the National

Survey

of Infection