Search

Objective: To investigate the effects of laboratory testing policies, particularly selective testing, rule-based reporting and isolate identification, on estimates of community antimicrobial resistance., Materials and Methods: Antibiotic resistance estimates were analysed from an all-Wales dataset for approximately 300 000 community isolates of common pathogens., Results: Selective testing policies were often associated with markedly increased resistance, particularly for second-line testing. Site-specific testing tended to yield variant resistance estimates for eye and ear isolates. Estimates from rule-based reporting deviated markedly from test-result-based reporting. Urinary isolates reported as Escherichia coli showed greater susceptibility than those reported as undifferentiated urinary 'coliforms'. The proportion of isolates tested for an antibiotic by a laboratory was a useful indicator of selective testing in this dataset. Selective testing policies had invariably been applied where the proportion of isolates of a species tested against an antibiotic was <90%. As this proportion fell with increasingly selective policies, divergence from pooled-all-Wales non-selective estimates tended to increase, with a bias to increased resistance., Conclusions: Selective testing, rule-based reporting and urinary coliform identification policies all had significant effects upon resistance estimates. Triage based upon the proportion of isolates tested seemed a useful tool in assigning analysis resources. Where <20% of isolates were tested, selective policies with inherent bias to increased resistance were common, the low number of isolates gave high potential sampling errors, and little confidence could be placed in the resistance estimate. Where 20-90% of isolates were tested, detailed analysis sometimes revealed resistance estimates that might be usefully retrieved. Where >/=90% of isolates were tested, there was no evidence of selective testing, and inter-laboratory variation in estimates appeared to be safely ascribable to other effects, e.g. methodology or real variation in resistance levels.

Objective: To determine if headgear use by rugby players was associated with a reduced risk of head or facial laceration, abrasion, or fracture., Methods: An emergency department based case-control study in South Wales, UK, with cases being rugby players treated for superficial head and facial injuries and controls being their matched opponents during the game. A review of videos of the 41 games in the 1999 Rugby World Cup was also carried out to compare with the case-control study. Odds ratios (OR) and 95% confidence intervals (95% CI) were used to measure association between exposure (headgear wearing) and outcome (head and facial injuries)., Results: In the case-control study, 164 pairs were analysed, with headgear worn by 12.8% of cases and 21.3% of controls. Headgear use was associated with substantial but non-significant reductions in superficial head (OR = 0.43, 95% CI 0.13 to 1.19) and facial (OR = 0.57, 95% CI 0.21 to 1.46) injuries. The video study followed 547 players over 41 games, during which there were 47 bleeding injuries to the head. Headgear use significantly reduced the risk of bleeding head injury in forwards (OR = 0.14, 95% CI 0.01 to 0.99, p = 0.02), but not in backs. There was also a higher risk of facial injury among forwards, but this was not significant., Conclusions: The combined results suggest that headgear can prevent certain types of superficial head injuries in players at all levels of the game, but the evidence is strongest for superficial head injury in elite forwards. A randomised controlled trial would be the best way to study this further.

It is believed that food hygiene precautions in domestic kitchens are an important strategy in efforts to reduce the incidence of sporadic food poisoning, but recent research has shown that people who have suffered food poisoning handle the same types of foods and adopt similar food hygiene precautions in their kitchens to the rest of the population. This suggests the need to examine other factors. A case-control study of sporadic Salmonella food poisoning was conducted to investigate several domestic kitchen risk factors. Measures of perception of risk, knowledge, and control associated with food poisoning in case and control respondents are reported here. It was found that perceived personal risk from food poisoning in the home was less than perceived risk to other people. In contrast, ratings of personal knowledge about food poisoning and personal control over food poisoning in the home were seen to be greater than other people's knowledge and control. There were no differences between the cases and the controls in their ratings of knowledge about food poisoning or their control over food poisoning. However, cases perceived their personal risk from food poisoning to be higher than controls. Both case and control samples exhibited optimistic bias but this was reduced in the case sample, suggesting that experience with food poisoning may reduce optimistic bias.

Domestic kitchen food handling risk factors for sporadic salmonella food poisoning are largely unknown. We compared food consumption and food handling practices, opportunities for cross contamination and refrigerator temperature control, in 99 households in South East Wales in 1997/8 with a case of salmonella food poisoning, and control households matched for electoral ward. On univariate analyses, cases were significantly more likely than control respondents to have purchased free-range eggs in the preceding week, and more likely than control households to have handled frozen whole chicken in the previous week, and to handle raw chicken portions at least weekly. In multivariate analysis, only consumption of raw eggs and handling free-range eggs were significant risk factors, independent of the age structure of the family and of the season.

The prevalence of antibiotic resistance in community-acquired infections is rising but in contrast to popular perception there has been little published work on its epidemiology. This systematic review evaluates the published evidence on the relationship between antibiotic prescribing and antibiotic resistance of organisms causing community-acquired urinary tract infection. Fourteen papers met the inclusion criteria and these reported on five ecological studies and ten studies of individuals. Only one ecological study provided good evidence of a link with prescribing rates. The remaining studies had no control for population differences in demographics and/or no comparison population. Studies at the individual level lacked clear case definitions and statistical power. Until the epidemiology of antibiotic resistance is better defined the design of effective interventions will not be possible.

In all temperate countries campylobacter infection in humans follows a striking seasonal pattern, but little attention has been given to exploring the epidemiological explanations. In order to better characterize the seasonal patterns, data from nine European countries and New Zealand have been examined. Several European countries with weekly data available showed remarkably consistent seasonal patterns from year to year, with peaks in week 22 in Wales, week 26 in Scotland, week 32 in Denmark, week 30 in Finland and week 33 in Sweden. In Europe, the seasonal peak was most prominent in Finland and least prominent in Scotland and Austria. In New Zealand the seasonality was less consistent since the peak was more prolonged. Possible explanations for the seasonal peaks are discussed. Research into the causes of campylobacter seasonality should help considerably in elucidating the sources of human infection.

Background: Sports injuries sustained by children are worrying because they prevent and deter participation in physical activity. Before we can address such injuries we need to understand the size of the problem and whether there have been changes in occurrence. A study of sports injuries to children, carried out in a Cardiff Accident and Emergency department in 1983, provided the data against which to compare data gathered in 1998., Methods: Data on all sports injuries to children aged 16 and under treated between September and December 1998 were compared with those reported for the same hospital, age group, injury and period in 1983., Results: A total of 953 injuries were treated in 1998, representing an increase of 54 per cent [95 per cent confidence interval (CI) 44-64 per cent]. The male:female distribution remained constant and the majority of injuries were due to rugby and soccer. The number of females injured playing rugby and soccer increased and a wider range of sports led to injuries for both males and females. Amongst 10-15-year-olds injury risk increased from 1 in 78 for boys in 1983 to 1 in 22 in 1998 (p < 0.0001). For girls, the increase was from 1 in 117 to 1 in 55 (p < 0.0001). The number of soccer- and rugby-related fractures increased by 52 per cent (95 per cent CI 22-87 per cent)., Conclusions: Sports injury rates have increased considerably over 15 years. With minimal population change, little variation in minor injuries and only small improvements in data capture, the main reason for change appears to be increased participation.

Background: Increasing numbers of outbreaks of Group C meningococcal disease in teenagers and young adults led to a new policy in the UK in 1999 of vaccinating all new college students. The largest of these outbreaks involved seven students in one university, six of whom were from one hall of residence, and two of whom died., Methods: Control of the outbreak involved close medical surveillance of resident students, mass chemoprophylaxis and vaccination, and wide dissemination of daily information bulletins. Investigation of the epidemiology of the outbreak involved searching for the network of close contacts between cases, a prevalence survey of carriage of meningogocci and a case control study of risk factors for carriage., Results: Clinical cases could be linked by a discrete network of social contacts within the halls of residence, but the Group C epidemic strain (2a P1.5) was not detected in 454 students (upper 95% confidence interval 0.7%). Carriage of any meningococcal strain (19%) was associated with patronage of the campus bar (OR = 3.0, 0.99-9.1)., Conclusion: Important factors in the control of the outbreak were rapid institution of mass chemopropylaxis and immunisation of residents, and involvement of student organizations in the dissemination of information about the disease and its control. The role of campus bars in dissemination of the carriage of meningogocci deserves further investigation.

Residents near the Trecatti landfill site located in South Wales, United Kingdom, expressed concern about odors and health effects they attributed to site emissions. The authors compared routinely collected, population-based, health data from potentially exposed electoral wards (i.e., United Kingdom electoral tracts) with data from both wards nearby, matched for socioeconomic deprivation scores, and with wards where residents were likely to attend the same hospital. Mortality rates were higher for all causes and neoplastic diseases (but not respiratory disease) in the exposed wards, but there was no change in rates after the site opened. Hospital data revealed a transient increase in admissions for asthma during the 3 yr that preceded the peak in odor complaints. The birth prevalence of congenital malformations was raised in the exposed wards, but the authors could not exclude a possible artifact resulting from differences in reporting practices between hospitals. The absence of environmental monitoring in the community during the period of public concern was a significant weakness of this study.

A database of the frequency of human inherited disorders is being established for use in a clinical context, in medical research, for epidemiological studies, and in the planning of genetic services. Each entry includes the disease name categorized by organ system, an Online Mendelian Inheritance in Man (OMIM) number, the mode of inheritance, the population origin, a prevalence and/or incidence rate and a literature reference. The Frequency of Inherited Disorders Database (FIDD) currently contains 1,580 entries relating to 280 different Mendelian disorders. FIDD will be prospectively maintained and can be accessed at http://www.uwcm.ac.uk/uwcm/mg/fidd/. A more refined and systematic literature search that will serve to expand the size, scope and scale of the database is currently in progress. The coverage of neurological and neuromuscular disorders is however considered to be nearly complete. In this first description of FIDD, Huntington disease was used to illustrate the structure and scope of the database as well as its potential scientific utility. A total of 100 published articles on the prevalence of Huntington disease were appraised. Prevalence and incidence rates varied between different ethnic groups and between different countries. Possible reasons for this variation are discussed.

Objective: To describe the frequency, nature and location of acute chemical incidents in Wales, and the morbidity in employees, emergency responders and the general public who were exposed., Design: Active multi-agency community-based surveillance system., Setting: Wales, 1993-5., Main Outcome Measures: Frequency, nature and location of incidents, populations potentially exposed and with symptoms., Results: Most of the 402 incidents identified were not associated with sites governed by the Control of Industrial Major Accident Hazard Regulations but with smaller industrial sites and commercial premises. About two in every thousand of the estimated 236 000 members of the public considered to be at risk from exposure reported symptoms, which were mainly nausea, headaches, and irritation of the eye, skin and respiratory tract. The most commonly reported chemicals that members of the public were exposed to were smoke toxins, miscellaneous organics, toxic gases and flammable gases. A health authority was reported to be involved in only 34 (8%) of the incidents and in only 3 of the 29 incidents where more than 100 members of the public were exposed., Conclusion: A geographically defined, multi-agency surveillance system can identify high risk locations and types of incidents, together with the chemicals most likely to be involved. Such ongoing surveillance information is essential for appropriate policy making, emergency planning, operational management and training.

Increasing awareness and concern about possible long-term health effects of exposure to chemicals after major chemical incidents has opened up a new area of public health activity. However, progress is hampered by a lack of low level exposures in potentially highly susceptible subgroups of the populations, and by a lack of expertise in public health. The challenge is to extend the toxicological and epidemiological experience developed in the occupational setting to the community level. To better define the problem population surveillance of acute chemical hazards, first set up in Wales, has been extended throughout the UK, and is also being piloted internationally. Further work is needed to address specific problems of the epidemiological follow-up of populations exposed, the assessment of the psychological consequences at the community level, and risk perception and communication. Organisational arrangements for the public health management of incidents and their sequelae must move beyond short-term funding and market competition and must incorporate the expertise of the occupational disciplines.

Objectives: To compare indices of health in a population living near a landfill site with a population matched for socioeconomic status and to review environmental monitoring data., Design: Ecological study with small area statistics and environmental reports., Setting: Electoral wards in valleys of South Wales., Subjects: Populations in the five wards near the landfill site who had formally complained of odours (exposed population), and comparison populations in 22 wards in the same unitary authority within the same fifth of Townsend score., Outcome Measures: Mortality, rates of hospital admission, measures of reproductive health (proportion of all births and stillbirths of infants weighing <2500 g; rates of admissions for spontaneous abortion; rates of all reported congenital malformations). Environmental data on site emissions., Results: There were no consistent differences in mortality, rates of hospital admissions, or proportion of low birthweight infants between the two populations. There was an increased maternal risk of having a baby with a congenital abnormality in residents near the site, both before its opening (relative risk 1.9; 95% confidence interval 1.3 to 2.85; P<0.001) and after (1.9; 1.23 to 2.95; P=0.003). Environmental monitoring showed that hydrogen sulphide from the site was probably responsible for odours., Conclusions: The area surrounding the landfill site has an increased rate of reported congenital malformations, which predated the opening of the landfill, although the cluster of cases of gastroschisis post-dated its opening. Several chemicals emitted from the site, including hydrogen sulphide and benzene, were found in air samples in the nearby community. Further studies of the reproductive risk in such communities are needed to examine the safety of waste disposal sites.

In a school outbreak of meningococcal disease in Wales, we compared risk factors for the carriage of Neisseria meningitidis B15 P1.16 with carriage of any meningococci. Students had throat swabs and completed a questionnaire. Sixty (7.9%) carried meningococci; risk for carriage was higher in those >14 years of age.

Surveillance of human VTEC O157 has been reported in several countries, based on laboratory evidence. The incidence is generally less than 10 per 100000 (the highest incidence is in children), with regional variations and a marked seasonality. Laboratory selection criteria and reporting have contributed to, but cannot entirely explain, variations within and between countries. Surveillance data and outbreaks in definable cohorts indicate that the spectrum of illness ranges from diarrhoea through acute bloody diarrhoea, with about 5% of cases developing haemolytic uraemic syndrome; less than 50% of patients report frank blood in their stools. Studies of sporadic cases have associated illness with beef products (particularly if undercooked and eaten outside the home), cooked sliced meat meals and contact with a household member with diarrhoea. Outbreaks have been attributed to contaminated foods (including beefburgers) and water, animal contact and person-to-person spread. Secondary transmission by the primary case in a household is of particular concern, and household transmission has been estimated at 4%, with patients excreting for around 10 days following onset. Recommendations for control have highlighted measures on farms, in slaughterhouses, retail and catering food premises, and in the home.

We investigated an outbreak of Salmonella enteritidis involving at least 19 British tourists returning from one hotel in another European country. A retrospective cohort study of 47 hotel guests identified lasagne as the most likely vehicle of transmission (RR 11.5; 95% CI 3.0-44.1; P < 0.0001). However, difficulties in information exchange and lack of formal mechanisms to agree on the aims of the cross-national investigation hampered efficient management of the outbreak. The factors leading to contamination of the food vehicle were not identified and therefore specific action to prevent reoccurrence could not be taken. There is need to develop protocols for cross-national investigations of outbreaks in Europe which should include specifying objectives, roles and responsibilities of investigators and control agencies, with formal reporting of the outcome of the investigation.

Study Objective: To investigate whether residents in the vicinity of the Sea Empress tanker spill suffered an increase in self reported physical and psychological symptoms, which might be attributable to exposure to crude oil., Design: Retrospective cohort study; postal questionnaire including demographic details, a symptom checklist, beliefs about health effects of oil and the Hospital Anxiety and Depression and SF-36 mental health scales., Setting: Populations living in four coastal towns on the exposed south Pembrokeshire coast and two control towns on the unexposed north coast., Patients: 539 exposed and 550 unexposed people sampled at random from the family health services authority age-sex register who completed questionnaires., Main Results: Adjusted odds ratios for self reported physical symptoms; scores on the Hospital Anxiety and Depression and SF-36 mental health scales, in 1089 people who responded out of a possible 1585 (69%)., Conclusions: Living in areas exposed to the crude oil spillage was significantly associated with higher anxiety and depression scores, worse mental health; and self reported headache (odds ratio = 2.35, 95% CI 1.56, 3.55), sore eyes (odds ratio = 1.96, 95% CI 1.06, 3.62), and sore throat (odds ratio = 1.70, 95% CI 1.12, 2.60) after adjusting for age, sex, smoking status, anxiety, and the belief that oil had affected health. People living in exposed areas reported higher rates of physical and psychological symptoms than control areas. Symptoms significantly associated with exposure after adjustment for anxiety and health beliefs were those expected from the known toxicological effect of oil, suggesting a direct health effect on the exposed population.

The prevalence of salmonella contamination in hens' eggs remains high. A case control study of sporadic cases of Salmonella enteritidis infection in Wales in 1997 confirmed the association of infection with consumption of raw or lightly cooked eggs.

Objectives: To evaluate the opportunities and limitations of using laboratory data to enhance sentinel general practice surveillance of influenza., Design: Descriptive study of active sentinel surveillance of clinically diagnosed influenza in general practice and passive total population surveillance of laboratory reports of influenza A, influenza B, Mycoplasma pneumoniae, and respiratory syncitial virus infections., Setting: Wales., Subjects: Total sentinel practices population (currently 228,130); population of Wales (2,913,000, 1994 mid-year estimate)., Main Outcome Measures: Simplicity, flexibility, acceptability, sensitivity, predictive value positive, representativeness, and timeliness of a surveillance system. Rate of influenza and other respiratory infections., Results: Sentinel general practice surveillance of influenza in Wales is simple, flexible, acceptable, timely, representative, and relatively sensitive. Current laboratory surveillance is complex and less timely than sentinel practice surveillance but is complete and has a relatively high positive predictive value. For the period January 1993 to September 1996, peaks in rates of influenza reported by sentinel practices during winters 1993/94 and 1995/96 were temporally associated with increased rates of laboratory confirmed influenza A and respiratory syncitial virus, whereas the peak in 1994/95 was associated with increased rates of laboratory confirmed influenza B, M pneumoniae, and respiratory syncitial virus., Conclusions: Timely laboratory data can add value to influenza data already obtained from sentinel general practice surveillance. However continuous audit is essential to resolve the possible limitations of either surveillance system.

Study Objective: To establish the source of a community outbreak of Salmonella typhimurium definitive type 124., Design: Two stage case-control study., Setting: Three districts in south east Wales., Subjects: Cases of salmonella food poisoning and community controls., Main Results: An initial case-control study identified an association between illness and eating ham (odds ratio 4.50, 95% confidence intervals 1.10, 21.8) and also found a possible association between illness and food bought from delicatessen stores (odds ratio 5.03, 95% confidence intervals 1.01, 32.3). However, only after a second stage case-control study was a single common ham producer identified as the source (odds ratio 25.0, 95% confidence intervals 2.33, 1155)., Conclusion: Sequential case-control studies are an important and underused tool in the investigation of community outbreaks.

Objectives: To evaluate the success of the south Powys hydatid control programme by analysis of trends in cystic disease in humans and sheep and dog infestation., Design: A review of hospital admissions for human hydatid disease in 1984-90, abattoir prevalence surveys of hydatid cysts in adult sheep, arecoline acetarsol and coproantigen surveys of prevalence of Echinococcus infestation in dogs., Setting: All hospitals in England and Wales, three abattoirs, and dog populations in mid ands south east Wales., Subjects: Residents of England and Wales admitted to hospital between 1984 and 1990 with a new diagnosis of human hydatid disease (International Classification of Diseases (ICD), ninth revision, code 122) acquired in the United Kingdom., Results: The average annual incidence of human hydatid disease in Powys, mid-Wales, fell from 3.9x10(-5) in 1974-83 to 2.3x10(-5) in 1984-90. Age specific incidence rates in Wales declined over this period only in children, and no cases occurred in children (<15 years) in Powys. Two Welsh children who lived in Gwent and mid-Glamorgan were infected. Prevalence of hydatid cysts in old sheep from south Wales declined during the control period, but in 1993 prevalence of cysts was 13%. Prevalence of E granulosus infestation was zero in the control area in 1993, but it was 2.4% in Powys dogs outside the control area in 1989 and 9.2% in dogs in Gwent in 1991., Conclusions: Human hydatid disease has been successfully controlled in south Powys but cystic echinococcosis is still endemic in sheep in mid-Wales, and there is a focus of infection in humans, sheep, and dogs in the bordering areas of Gwent and mid-Glamorgan. There is considerable potential for an upsurge in human cases if control measures are relaxed.

An outbreak of Salmonella enteritidis PT 6B food poisoning, the first reported of this recently described phage type, resulted in illness among 46 of 49 members of a camping group in North Wales, 33 of whom were hospitalized. Epidemiological evidence (P < 0.0001) indicated that a lemon meringue pie was the vehicle of infection. Fresh shell eggs, stored after purchase at ambient temperature, appear to be the most likely source of infection, with multiplication during preparation and subsequent storage of the pie a significant contributory factor. Campers may be at greater risk than others and should consider the use of cold boxes for the transport and storage of eggs, and avoid the preparation of lightly cooked egg products under these basic conditions.

Epidemiology investigations of outbreaks of infectious disease have to be carried out rapidly but must be methodologically sound. In 14 of 25 consecutive Salmonella outbreaks in Wales from 1986 to 1990 case-control or cohort studies were undertaken. Food vehicles of infection were identified by statistically significant associations with cases in 13 of the 14, including three outbreaks where there were less than 10 cases. The particular problems of epidemiological field investigations are discussed.

A common source outbreak of small round structure virus (SRSV) gastroenteritis affected 81 patients and 114 staff in four hospitals served by one central hospital kitchen. Eating salad items was found to be significantly associated with illness. In a cohort study of a staff buffet function eating turkey salad sandwiches was associated with illness (relative risk = 2.4; 95% CI = 1.4-4.1; P = 0.003), and a case control study of patients in one hospital showed an odds ratio of 6.6 (95% CI = 1.0-71.6; P = 0.04) for eating tuna salad and becoming ill. One of two food handlers who prepared the salads became ill the day following food preparation; she also had a young child at home who had been ill with a gastrointestinal illness during the previous two days. Contamination of food by mechanical transmission of the virus from the child via clothes and hands of the mother, or pre-symptomatic faecal excretion in the mother are possible explanations of contamination of food.

The routine use of ELISA and complement fixation tests in the diagnosis of suspected clinical cases of hydatid disease was evaluated. In the ELISA test, dialysed and filtered sheep cyst fluid was used as antigen and two positive cut-off points--+3SD and +2SD of the mean absorbance values of the control sera--were evaluated. The predictive values of ELISA tests were 82% and 90% for positive tests, and 86% and 82% for negative tests, respectively with the two cut-off points. In a population survey of blood donors and veterinary workers in Powys, 4% and 8%, respectively, had ELISA values above the lower cut-off point. However, it would not be appropriate to use the same test for diagnostic population screening in Wales since the predictive value of the test is likely to be very low in this setting. Serological surveys with the ELISA may be of use in monitoring the progress of the South Powys Hydatid Control Programme. The use of cumulative percentages was found to be a useful method of comparing whole distributions of results in different populations.

The associations between infecting dose, incubation period, and the severity of disease were examined in a large outbreak of Salmonella typhimurium which occurred at a medical conference in Wales in 1986. Persons who had eaten two or more pieces of the chicken vehicle had, on average, shorter geometric mean incubation periods than those who had only eaten one piece: 16.6 hours (95% confidence interval (CI) 13.5-20.5) compared with 20.7 hours (95% CI 19.0-22.6) (t = 1.97, p < 0.05). Incubation period was negatively correlated with the maximum frequency of diarrheal stools (r = -0.46, 95% CI -0.56 to -0.33), the maximum temperature reached (r = -0.34, 95% CI -0.50 to -0.16), the duration of symptoms (r = -0.41, 95% CI -0.53 to -0.26), and the amount of time taken off from work (r = -0.54, 95% CI -0.65 to -0.41). Those with shorter incubation periods were more likely to have been hospitalized. There was no association between chicken consumption and any of the measures of severity. The authors discuss the evidence that incubation period is inversely related to dose, the use of incubation period as a marker for dose, and the role that individual differences in susceptibility play in determining both the incubation period and the outcome.

In Wales, in 1988, 119 patients with meningococcal disease were identified, so giving a crude annual incidence of 4.2 patients per 100,000 population. The combined classical clinical features of fever, vomiting, neck stiffness, headache and purpuric rash were identified in only 9% of patients. Fever and vomiting were the commonest symptoms, both being present in 60% of patients. A rash was noted in 77% of patients but neck stiffness in only 39%. Rash was more common in children, headache and photophobia in adults. A total of 13 patients died, the fatality rare increasing with age from 3% in infants to 20% in older teenagers and adults. Only 15% of 75 patients admitted to hospital by general practitioners were known to have received intravenous or intramuscular penicillin before admission as recommended by the Chief Medical Officers of the Health Departments in the U.K. Only 24% of patients received rifampicin to clear nasopharyngeal carriage before or at discharge from hospital. Altogether, 375 household contacts of patients were identified. At least 84% of them received chemoprophylaxis.

An outbreak of hepatitis A associated with a Middle school involved 23 cases; 17 were pupils attending the Middle school, one was a teacher, one was a relative of a case, and four were from the associated First school, of whom three had siblings in the Middle school. The probable source case was a male pupil infected by a sibling who had contracted hepatitis A while abroad on holiday. A questionnaire survey and salivary IgG and IgM anti-HAV testing of the pupils demonstrated a statistically significant association between infection and the use of a changing room toilet for defecation. An inspection of the school showed that toilets lacked toilet paper, soap and hand towels. Advice was given to pupils, parents and staff on hygiene. Human normal immunoglobulin was administered to susceptible family contacts, pupils and staff at the school. The school outbreak might have been prevented if the source case for the school had been given immunoglobulin when his sibling developed hepatitis A.

In 1985 an outbreak of ornithosis affected 13 of 80 (16%) workers in a duck-processing plant. New employees were three times more likely to become cases than established employees. The highest attack rate was in those on the production line. Following the outbreak, an occupational health scheme was set up to monitor the health of new recruits to the company. Serological evidence of recent infection was demonstrated in 18 of 37 (49%) new employees tested in the first 3 months of employment. Five (14%) also had clinical evidence of ornithosis. Veterinary investigation of the ducks demonstrated a high proportion with asymptomatic chlamydial infection. It is suggested that ornithosis may be more common in duck processors than is currently supposed. Strategies to reduce occupational risks are discussed.

Three consecutive outbreaks of Salmonella enteritidis PT4 occurred in Wales in 1989 in which epidemiological and microbiological investigation established eggs as the likely source although kitchen inspection and food preparation histories suggested other vehicles of infection. This paper examines the contribution of analytical epidemiology in attributing causation, with particular reference to those limitations which are regarded as inherent in epidemiological evidence. Such evidence, implicating eggs in the three outbreaks, fulfilled 6/7 widely accepted criteria for causation; data to assess the seventh were lacking. Collaboration between different agencies and professionals in investigating outbreaks is very important.

There are several anecdotal accounts of the association between outbreaks of influenza and meningococcal disease. The exceptional increase in the number of cases of meningococcal infection 2 weeks after an influenza A outbreak in England and Wales during November and December, 1989, provided an opportunity to investigate the relation between the two events. Patients with meningococcal disease in December, 1989, were more likely than age-matched controls to show serological evidence of recent influenza A infection (odds ratio 3.9, 95% Cl 1.2-13.9). The most likely explanation for the association is immune suppression induced by influenza A, though a lowering of mucosal resistance to meningococcal invasion may also be a factor. Public health authorities should be aware of the association and should be prepared to alert medical practitioners and the public to the increased risk of meningococcal disease when influenza A outbreaks occur.

Objective: To investigate the epidemiology of invasive disease due to Haemophilus influenzae type b, the clones responsible, and the antibiotic resistance of the isolates., Design: Prospective population based analysis of clinical and epidemiological data collected for Gwynedd during 1980-90 and in the whole of Wales during 1988-90., Setting: 19 hospitals in Wales; all medical microbiology laboratories in Wales participated., Patients: 82 patients with confirmed invasive infections caused by H influenzae type b in Gwynedd during 1980-90 and 207 in Wales during 1988-90., Main Outcome Measures: Clinical and epidemiological measures; analysis of the clonal types of the isolates based on the electrophoretic mobilities of 17 metabolic enzymes; and antibiotic resistance., Results: The annual incidence of H influenzae type b infections in Gwynedd was 3.2 cases/100,000 and in Wales was 2.5 cases/100,000. Most cases occurred in children aged under 5 years, the highest annual incidence being in those aged under 1 (84.6/100,000 and 56.9/100,000 in Wales). The cumulative risk of acquiring H influenzae type b disease by the fifth birthday was one in 456 in Gwynedd and one in 578 in Wales. Fifteen per cent of cases in Gwynedd and 7% of those in Wales occurred in adults. Predominant clinical conditions were meningitis in children and pneumonia in adults. In Gwynedd 2/70 (3%) children and 5/12 (42%) adults died. Long term neurological sequelae occurred in 8% (4/48) of children who survived haemophilus meningitis. Children presenting with infection were usually the youngest members of their family. No secondary household cases were identified. 100 of 128 (78%) strains were of a single clone, electrophoretic type 12.5, and 4/207 (1.9%) isolates from Wales were resistant to both ampicillin and chloramphenicol., Conclusions: The annual rate of infection in children aged under 5 in four Welsh counties was 12-44% higher than that previously published for the United Kingdom. The study emphasises the potential value of a vaccine effective in early infancy and provides baseline data to assess its efficacy after its introduction. Alternatives to ampicillin and chloramphenicol should be used as first line, empirical treatment for severe infections that might be caused by H influenzae type b in Wales.

An outbreak of pertussis in primary school-children in the St David's area of Pembrokeshire provided the opportunity to estimate pertussis vaccine efficacy. The estimate of efficacy was 88% when notified cases were used, but this fell to 68% when all children with bouts of coughing for two or more weeks were included. Notified cases were significantly less likely to have been vaccinated than other cases with similar symptoms. Therefore vaccine efficacy estimates based upon notified cases are likely to be biased. However, even the lower estimates suggest that pertussis immunisation is highly desirable and efforts to improve coverage should be increased.

A case-control study was carried out to test the hypothesis that the rise in the rate of Campylobacter jejuni infection in the Brigend area of South Wales during May was due to the consumption or handling of milk from bottles that had been attacked by birds. 32 of 36 cases meeting the case definition were interviewed, along with 2 controls per case, matched for age, sex, and area of residence. There were strong associations between campylobacter infection and doorstep delivery of milk bottles, a history of milk bottle attack by birds, milk bottle attack by birds during the week before illness, and consumption of milk from attacked bottles during the week before illness. There was a very strong dose-response relation between frequency of bird attack and illness. Controls with a history of milk bottle attack by birds were more likely than cases to have taken preventive measures against bird attack and consumption of contaminated milk. Although few people witnessed the attacks, the likely culprits are magpies (Pica pica) and jackdaws (Corvus monedula).

A technique for simultaneously measuring pulsation, translation and diffusion of a radioactive tracer in the spinal cerebrospinal fluid (CSF) has been described. To evaluate the accuracy and precision of this method for measuring CSF oscillation in the spine, an apparatus was designed that moved a radioactive source sinusoidally with known amplitude along a line. With list mode acquisition of radioactive emissions, the amplitude of oscillation was measured for a series of peak-to-peak amplitudes ranging from 0.01 mm to 3.8 mm. For oscillations larger than 0.35 mm peak-to-peak, measured values were within 6.1 percent of the expected values. For the spinal CSF, oscillation larger than 0.4 mm should be measured accurately with this technique.

The first recognized outbreak of haemorrhagic colitis due to Escherichia coli O 157.H7 in the United Kingdom affected at least 24 persons living in East Anglia over a 2-week period. The illnesses were characterized by severe abdominal pain and bloody diarrhoea of short duration. Eleven patients were admitted to hospital and there was one death. Patients were mainly adult women who had not eaten out of the home in the 2 weeks before onset. Unlike previously reported outbreaks hamburgers were not the vehicle of infection, and a case-control study suggested that handling vegetables, and particularly potatoes, was the important risk factor.

In a two year prospective survey of outbreaks of salmonella infection in hospitals in England and Wales 55 outbreaks were identified. Reports of investigation of these outbreaks were reviewed for evidence of food borne infection and cross infection. Food borne infection probably accounted for only six outbreaks, but these made up 40% of the 15 outbreaks in which there were more than five patients and staff with symptoms. Person to person transmission was the probable mode of spread in most outbreaks. It is recommended that in addition to bacteriological investigations clinical and epidemiological data should be collected to implicate food or other possible common vehicles of infection. Bacteriological screening of patients and staff who do not have symptoms may be unrewarding in the absence of epidemiological data to define groups at risk of infection.

In one general practice over a six-week period three patients with atypical pneumonia were shown serologically to have psittacosis. All three patients had been exposed to the same flock of ducks but veterinary investigation three months after the first case had been reported did not show the ducks to be infected. However, in a retrospective review of laboratory-diagnosed psittacosis, a likely source of infection was found in 11 out of 24 cases and in most of these full medicoveterinary investigation would have been warranted. Early diagnosis, reporting quickly to the medical officer for environmental health and early liaison with the divisional veterinary officer are essential prerequisites if psittacosis is to be controlled. If psittacosis were made a statutory notifiable disease this would encourage a more energetic approach to its control. The proposed cutbacks in the veterinary investigation service will make full investigations more difficult.

Over three months Cryptosporidium oocysts were identified in faecal samples from 43 (5%) of 867 patients presenting to their general practitioners with gastrointestinal symptoms. Cryptosporidium was the second most common enteric pathogen identified. Of the 867 patients, 329 were children aged under 5, of whom 24 (7%) excreted Cryptosporidium. A characteristic clinical presentation of infection with Cryptosporidium was recognised--namely, mild gastroenteritis with four to six watery, mucoid, and offensive motions a day, which lasted for one to two weeks. The source of infection was not identified, but direct contact with farm animals was not a feature and no association with a common water supply could be established.

Presumed or confirmed ornithosis developed in 15 of 46 veterinary surgeons within 3 weeks of their attending a training course on the supervision of poultry processing plants. A duck processing plant was the probable source of infection. The attack rate in those visiting this plant was 44% (15/34) and the rate in those who had close contact with duck feathers was 1.7 times the rate in the remainder of the group who visited the plant. The incubation period ranged from 6 to 20 days. This is the second outbreak of the duck-associated ornithosis to be recognised in the U.K.

In a multicentre case-control study of necrotising enterocolitis risk factors were found to vary with birthweight of cases. In very low birthweight cases the risk factors identified were those associated with prolonged or recurrent hypoxia (recurrent apnoea, respiratory distress, assisted ventilation, and umbilical artery catheterisation). In heavier birthweight infants the risk factors were, in contrast, related to hypoxia at birth (low 1 minute Apgar score and endotracheal intubation at birth) and umbilical vessel catheterisation used in exchange transfusions. Contradictory findings in published case-control studies carried out in the USA may be due to differences in patient populations and management policies. Hypoxia and umbilical vessel catheterisation should still be considered as risk factors for necrotising enterocolitis.