BRITISH MEDICAL JOURNAL VOLUME 285 23 OCTOBER 1982 Outside Europe Surveillance of patients attending a diarrhoeal disease hospital in Bangladesh BARBARA J STOLL, ROGER I GLASS, M IMDADUL HUQ, M U KHAN, JAMES E HOLT, HASINA BANU Abstract In October 1979 a surveillance system was set up at the International Centre for Diarrhoeal Disease Research, Bangladesh, Hospital at Dacca to study a 4% systematic sample of the 100 000 patients with diarrhoea who come to the hospital for care each year. From December 1979 to November 1980 inclusive, 3550 patients were studied. A recognised pathogenic organism was identified for 66% of patients screened for all pathogens, one-third of whom had a mixed infection with two or more agents. Enterotoxigenic Escherichia coli was the most common enteropathogen detected in all age groups (detection rate 20%), followed by rotavirus (19%), Campylobacter jejuni (14%), and Shigella (12%). Infants and young children (up to 5 years) were most often infected with rotavirus, enterotoxigenic E coli, and C jejuni and older children (5-14 years) had more infections with entero- toxigenic E coli, Shigella, and E histolytica. Surveillance has helped to define the range of disease among patients attending the Dacca Hospital. Sixty-five per cent of patients complained of watery diarrhoea, a presentation that was significantly more common in patients with Vibrio cholerae 0:1 (91%), enterotoxigenic E coli (78%), rotavirus (77%), and C jejuni (71%) than in all patients studied. Dysentery, defined as a history of diarrhoea with blood, was the presenting complaint of 20% of all patients but 55% of those with Shigella. Only patients with V cholerae 0:1 and enterotoxigenic E coli were at increased risk for severe dehydration. In addition surveillance has been used to identify areas where patient care can be improved and to generate new ideas for research. Introduction Since 1962 the International Centre for Diarrhoeal Disease Research, Bangladesh (ICDDR,B), formerly the Cholera Research Laboratory, has operated a diarrhoeal disease hospital International Centre for Diarrhoeal Disease Research, Bangladesh, GPO Box 128, Dacca 2, Bangladesh BARBARA J STOLL, MD, associate scientist ROGER I GLASS, MD, MPH, epidemiologist (on assignment from the Centres for Disease Control, Atlanta, Georgia, USA) M IMDADUL HUQ, MSC, PHD, scientist M U KHAN, MB, DPH, scientist JAMES E HOLT, MA, research associate (Department of Biostatistics, Mount Sinai School of Medicine, New York, USA) HASINA BANU, MA, senior field research officer in Dacca (Dacca Hospital), the capital and largest city of Bangladesh, that provides care free of charge to all patients. Between 1970 and 1980, the number of patients seen at the hospital increased 10-fold from about 10 000 to about 100 000 patients a year, and it became difficult to collect useful informa- tion on all patients. Therefore, in October 1979 a surveillance system was set up to study the epidemiological, clinical, and laboratory characteristics of a sample of the patients who come to the hospital for care each year. We were interested in identify- ing the relative importance of various enteropathogens in urban Dacca and in determining the seasonality and range of disease associated with these agents. We outline the methods used for surveillance and review the data on patients seen from 1 Decem- ber 1979 to 30 November 1980. Materials and methods SURVEILLANCE METHODS A systematic sample of patients was chosen for surveillance based on the number they were assigned at registration. From October 1979 to February 1980 inclusive, every fiftieth patient was selected but in March 1980 the sample size was increased to 4% for statistical reasons. Surveillance patients were seen by the regular hospital staff with emergency cases treated on a priority basis. All patients were first treated in an outpatient area; those requiring further care were admitted to an intravenous treatment centre or a hospital ward. After initial examination and care by a nurse or doctor, the surveillance patient or an adult guardian was interviewed by a special health assistant who collected uniform information on demographic back- ground, recent medical history, presenting symptoms, previous treatment, and treatment prescribed at ICDDR,B. A doctor per- formed a physical examination, including assessment of the state of dehydration. Dehydration was graded as none, mild ( < 5%), moderate (5-10%), and severe (>10%) according to clinical signs.' To assess nutritional state, each child 10 years or younger was weighed and measured at discharge and the weight-for-height was compared with the Harvard standard.2 A rectal swab for culture was obtained from all patients, and a stool sample for microscopical examination was requested. Additional information was collected on patients admitted to the intravenous centre or hospital ward. LABORATORY METHODS Rectal swabs or stool samples were plated directly on Salmonella- Shigella (SS), taurocholate-tellurite-gelatin, and MacConkey's agars. Specimens were also enriched in alkaline peptone water for vibrios and then plated on taurocholate-tellurite-gelatin. The plates were examined for salmonellae, shigellae, and vibrios by standard methods.3 Vibrio-like colonies identified on taurocholate-tellurite-gelatin plates were further characterised and classified as V cholerae 0 group 1 or non-O group 1 (NAG).4 Non-lactose fermenting colonies from MacConkey's and SS agars were screened on Kligler's iron agar slants 1185