Search

Study Objective: The aim was to identify and compare rates of admission to hospital of preschool children for medicinal and non-medicinal poisoning in a defined population., Design: The study was an analysis of computerised abstracts of hospital inpatient records for poisoning., Setting: Six districts in southern England covered by the Oxford record linkage study., Subjects: The subjects were children under six years of age residing in the six districts from 1975 to 1986., Measurements and Main Results: 6152 children, drawn from an average annual resident population of 164,000 children in 1975-1986, experienced 6562 hospital admissions for poisoning before six years; 3702 (56.4%) were attributed to medicinal and 2860 (43.6%) to non-medicinal agents. Of the latter, 646 (9.8% of the total) were recorded under the International classification of diseases code, described as "noxious food" (almost exclusively plant material). Average annual admission rates in children under six were 1.88 per 1000 for medicinal agents and 1.45 for non-medicinal substances. Analgesics accounted for 28.1% of the admissions for medicinal poisoning; berries and mushrooms for 97.4% of the plant materials; and corrosive aromatics, acids and alkalis for 22.0% of the other non-medicinals. Admission rates were higher in males than females in each category of poisoning. In children aged 1-4 years there was a significant decrease in admission rates between 1975 and 1986, averaging per annum 5.8% for medicinal poisoning, 6.9% for non-medicinal poisoning (excluding plant material), and 12.8% for plant material poisoning. Significant seasonal variation was found for each class of poisoning., Conclusions: Admission rates for medicinal and non-medicinal poisoning in preschool children declined between 1975 and 1986. The decline could reflect a change in thresholds for admission although, if so, this would be against the general trend in paediatric medical admissions, which is upward. An alternative explanation is a decline in the incidence of poisoning.

Study Objective: The aim was to report on the extent to which death certificates which specify that death occurred in hospital can be matched and linked with routine hospital inpatient information systems., Design: The study involved linkage of hospital records which specified that death occurred in hospital to corresponding death certificates; and linkage of death certificates which specified that death occurred in hospital to corresponding hospital records., Setting: Six health districts in southern England covered by medical record linkage., Subjects: Records were examined of patients aged 65 years and over, which specified that death occurred in hospital between 1979 and 1985., Main Results: 98.2% of hospital record abstracts which specified that death occurred in hospital were linked by our standard computer-based techniques to death certificates. Conversely, however, only 94.4% of death certificates which specified that death occurred in hospital could be linked to the abstracts of corresponding hospital inpatient records. A major factor contributing to the latter failures may be a difference of definition of what constitutes a death "following hospital admission" in patients who die shortly after arrival at hospital., Conclusions: Linkage of hospital records to death certificates is both feasible and desirable. Error rates are generally small; but hospital inpatient record abstracts corresponding to death certificates for deaths in hospital may not invariably exist when death occurs shortly after the arrival of the patient at hospital.

Data from the Oxford Region were used to study trends in hospital admission rates and demographic profiles of hospital care in the National Health Service in oral and maxillofacial surgery. In a defined population of 1.9 million people, admission rates for inpatient care, day case care, and outpatient attendance rates all almost doubled between 1975-1985. Population-based age-specific admission rates were much higher in teenagers and young adults than in other age-groups. In these age-groups admission rates for females were about double those for males. Whilst perhaps not surprising to the oral and maxillofacial surgeon, these demographic profiles of workload are strikingly different from those found in most other hospital specialties.

In the Oxford region 8,903 residents with oesophageal or gastric cancers were registered by the Oxford Cancer Registry between 1974 and 1988. For most of this time the Registry's practice was to code all adenocarcinomas of the oesophagus as adenocarcinomas of the cardia of stomach. There was a significant increase over time in the registered incidence of adenocarcinoma of the cardia, no appreciable change in the incidence of oesophageal cancers registered as such, and a decline in registered incidence of cancers of the stomach other than cardia.

It has been widely recommended that children should be admitted to hospital only if treatment cannot be provided at home, and that durations of hospital stay should be minimized. We have used record linkage to calculate a statistic which is not yet commonly available, the total time spent in hospital by children per year, and have compared this between six districts in the Oxford Region. Hospitalization rates for children who stayed a total of two days or less per year in hospital increased over time; rates for children who stayed more than two days declined. Mean and median total days spent in hospital varied between districts but, in absolute terms, the differences were small at less than one day's difference between districts per year per child admitted. Inter-district variation decreased over time, although the variation in use of day case care which remained suggests that some districts could still do more work in this fashion.

We have studied the amount of time spent in hospital by elderly people in their final year of life in 1976, 1981 and 1986 in six districts covered by the Oxford record linkage study to determine the extent of differences between districts. There was consistent but insubstantial variation between the districts. Variation declined over time, suggesting that the use of hospital care for the elderly may, in this respect, be nearing consensus. We conclude that the measure of total time spent in hospital by the elderly in their last year of life was not a powerful discriminator between districts in this region. When record linkage is implemented elsewhere in England, it will be possible to make wider geographical comparisons to determine whether important variations exist elsewhere.

To identify the clinical conditions associated with substantial time spent in hospital by children aged 1-14 years, records of children admitted to hospital in 1975, 1979, and 1984 were studied. Analysis was by linkage of abstracts of routine records of hospital inpatient care in six districts in southern England covered by the Oxford record linkage study. The total time spend in hospital in the acute specialties each year was calculated by summing the lengths of stay of all episodes of care for each child in each year. First, admissions with long median times in hospital per child admitted were identified. These included, notably, fracture of femur and, in the later years, leukaemia, other malignant neoplasms, and congenital disorders of metabolism. Second conditions were identified which accounted for large numbers of children with lengths of stay of five days or more. These included, in particular, congenital anomalies, asthma, and appendicitis. Third, conditions were identified which accounted for the largest numbers of bed days used. These included congenital anomalies, hypertrophy of tonsils and adenoids, asthma, otitis media, appendicitis, and head injury. Median time spent in hospital per child admitted declined for most conditions but increased for leukaemia, other malignant neoplasms, and congenital disorders of metabolism. Admission rates for children who spent five days or more in hospital each year declined for all common conditions except asthma which increased. Total numbers of beds used increased for asthma and otitis media but declined for all other common conditions.

Objective: To report on trends in workload patterns in gynaecology using linked statistical data., Design: Retrospective analysis of linked abstracts of hospital inpatient and day case records for patients treated in the National Health Service in gynaecology., Setting: Six health districts in the south of England covered by the Oxford record linkage study., Subjects: Records for hospital admissions to gynaecology (excluding obstetric admissions) from 1975 to 1985., Measurement and Main Results: Inpatient episodes increased by 23.5% and day case episodes increased by 13.1%. More people treated contributed about 90% and increased readmissions contributed about 10% to the increase in workload. The workload was decreased by strike action in 1975 and 1981-2. Average length of stay decreased substantially and consistently over the 11 years. Emergency readmissions increased annually by an average of 2.7%. Admission rates in 11 groups of surgical procedures accounting for 85% of all gynaecological inpatients are reported, and increases occurred in 10 of the 11 groups. For example, average increases in annual admission rates were 1.0% for sterilization, 1.9% for legal abortion and 8.2% for biopsy of the cervix, the rate for dilatation and curettage decreased by 1.4%., Conclusions: The increase in admission rates in gynaecology was almost entirely due to increases in numbers of people treated. The rise would have been even greater if the increase in private patients had been considered. The increase may reflect increased expectations on the part of patients and their doctors, advances in technology and increased bed availability due to declining lengths of stay.

Trends in admission rates, lengths of stay, and clinical case mix (adjusted for multiple admissions per person) for Otolaryngology from 1975 to 1985 are described using the Oxford Record Linkage Study. Person-based admission rates increased generally, most strikingly amongst children and adolescents. Both length of stay per episode and total number of days an individual stayed in hospital each year decreased. Operations on tonsils and adenoids decreased averaging 2.2% per annum; admissions for diagnoses for otitis media with effusion increased averaging 8.2% per annum; for the operations of myringotomy and tympanostomy tube insertion increased averaging 9.2% per annum; for other diagnoses related to the ear increased averaging 5.7% per annum; and admissions for malignant neoplasms decreased averaging 2.0% per annum. Clearly more people are being treated for a changing case mix. The largest increases are occurring with children and adolescents.

Objectives: To examine trends over time in the rates of admission to hospital for accidents of preschool children and to study patterns of repeated admissions for accidents in these children., Design: Analysis of linked, routine abstracts of hospital inpatient records for accidents., Setting: Six districts in the Oxford Regional Health Authority covered by the Oxford record linkage study., Subjects: Records for 19,427 children aged 5 years and under at the time of first recorded admission to hospital., Main Outcome Measure: Number of admissions to hospital., Results: Records were analysed in three groups: person based annual admission rates were calculated for each calendar year; each child's first recorded admission in 1976-85 was identified, and the child's record was followed up by linkage for one year from that admission; each child's first recorded admission in 1976-81 was identified and followed up for five years. Overall, 19,427 children from an average annual resident population of 163,000 children in 1976-86 had 20,657 admissions for accidents before they were 6 years of age. Of these admissions 13,983 were for injuries, 5717 for poisonings, and 957 for burns. Admission rates declined after 1976 for poisoning, but no substantial changes over time were found in admission rates for injuries or burns. A total of 17,724 children were followed up for one year and 10,889 for five years; 470 (2.6%) of the children who were followed up for one year and 926 (8.5%) of those followed up for five years had at least one further admission for an accident. Of those followed up for one year the 4 and 5 year old children were least likely and those under 1 and 1 year old were most likely to have a further admission for an accident. The number of children who had more than one accident was greater than would be expected if accidents were random occurrences. Those who had a poisoning at first admission were more likely to have another poisoning than an injury or burn; and those who had a burn at first admission were more likely to have another burn., Conclusions: Hospital admissions for accidents in children are common: on average 1 child in 88 in this population was admitted each year. Multiple admissions are uncommon but none the less occur more often than would be expected by chance.

Objectives: to report on trends in ophthalmology workload using linked statistical data;, Design: analysis of linked abstracts of hospital inpatient and day case records for ophthalmology;, Setting: six districts in Southern England covered by the Oxford record linkage study;, Subjects: records for hospital admissions to ophthalmology from 1975 to 1985;, Measurements and Main Results: Over the period of study, the number of episodes of inpatient and day case care increased by 16.3%. Notable increases in age-specific admission rates were seen among the elderly. An increase in the number of individuals treated contributed about 67% and an increase in multiple admissions per individual contributed about 33% to the increase in admission rates. Both average length of stay per episode and total time in hospital per individual decreased consistently during the 11 years; and there was no increase in emergency readmissions over time. No significant changes over time were found in admission rates for retinal detachments and defects or for glaucoma. There was a statistically significant increase averaging 4.8% per annum in admission rates for cataract, and a significant decrease averaging 5.2% per annum in admission rates for strabismus and other disorders of binocular eye movement between 1975 and 1985., Conclusions: Age-specific admission rates in ophthalmology are much higher in the very young and old than in other age groups. Patterns of work in the specialty are therefore particularly affected by variation in the age distribution of the population. The increase in cataract surgery reflected both the increase in numbers of old people in the population and an increase in age-specific operation rates for cataract. Attempts to alter attitudes and behaviour of the elderly regarding eye diseases, disability, and facilities for treatment may have had a positive impact on the use of services. The decrease in admissions for children probably reflects the impact made by child health surveillance programmes. As the child screening programme has expanded, the admission rates for strabismus have decreased.

Study Objective: The aim was to assess the extent to which routinely collected data on poisoning in adolescents reflected deliberate self poisoning and, in doing, so to assess the accuracy of the diagnostic information on poisoning in the routine hospital abstracts which form the joint data base of Hospital Activity Analysis and the Oxford Record Linkage Study (ORLS)., Design: A comparison was made (a) of all eligible ORLS records during the study period with an independent source of records; and (b) of a random sample of records from an independent source with ORLS., Setting: Records of patients admitted to the John Radcliffe Hospital in Oxford were used., Subjects: These were (a) patients aged 10-20 years between 1980 and 1985 with a diagnosis of poisoning by drugs and medicaments in ORLS; (b) a random sample of 500 patients selected from the self harm monitoring files at the hospital (12 patients were not eligible for inclusion in ORLS and were therefore excluded from the rest of the study)., Measurements and Main Results: The recorded diagnosis was compared on the records selected from the two files. Of the 1123 events of poisoning identified in ORLS, 1081 (96.3%) were correctly coded as poisoning and 1065 (95%) of these were deliberate self poisoning. Of the 488 cases from the monitoring files, 467 (95.7%) of all cases had a correct diagnosis of injury or poisoning on the ORLS file. Of the 453 poisoning cases 436 (96.2%) were correctly recorded in ORLS., Conclusions: Deliberate self poisoning in adolescents can be identified through routinely collected hospital statistics. A very high percentage of the diagnostic information on poisoning in ORLS files is correctly recorded.

Objectives: To determine whether among people aged 65 and over those who died at advanced old age spent more of their last year of life in hospital than those who died younger, and whether the increase in longevity in the elderly between 1976 and 1985 was accompanied by increased time spent in hospital in the last year of life., Design: Linkage of death records to abstracts of records of hospital inpatient care in the preceding year of patients' lives., Setting: Six health districts in England covered by the Oxford record linkage study., Results: People who died at advanced ages (85 and over) were less likely than people who died at younger ages (65-84) to have been admitted to hospital in the last year of life. Once admitted the very old tended to spend longer in hospital than others. The mean total time spent in hospital by the elderly in the year before death (based on all deaths including those among people not admitted at all) showed no appreciable change over time. The median time in hospital based on all deaths increased by about three days between 1976 and 1985. During that time there was a gain in life expectancy in the population of about one year from the age of 65., Conclusion: The gain in life expectancy in this population was not at the expense of any substantial increase in time spent in hospital in the final year of life.

Linked hospital and death records for 10-20-year-olds admitted with a diagnosis of poisoning were analysed. Between 1980 and 1985 there was a significant decline in admission rates, which was mainly attributable to a decline among 16-20-year-old females. There was no evidence of a decline in the admission rates among 12-15-year-olds. Ten per cent of the study population had at least one further hospital admission for poisoning during the mean follow-up period of 3.6 years. Female admission rates were substantially higher than those in males but readmission rates, given a first admission, were similar. The death rate in the study cohort was significantly higher than would be expected in the general population of this age and nearly all the deaths were from violent or unnatural causes.

We report on factors related to the prevalence of bacteria in the vagina of 1498 women who were attending a family planning clinic. Anaerobes were significantly commoner in women with an abnormal vaginal discharge on clinical examination, in women with a history of a troublesome vaginal discharge and in women who used an intrauterine contraceptive device (IUCD) than in others. The association between anaerobes and abnormalities of vaginal discharge was found both in women who used the IUCD and in women who did not. We found no significant associations between symptoms and other bacterial species or combinations of species. Coliforms were significantly commoner in women who did not use tampons, in women who used the diaphragm and in women who had received recent antibacterial or antifungal treatment than in others. Lactobacilli were significantly commoner in women who used the contraceptive pill, and were significantly less common in women who harboured Candida albicans, anaerobes, coliforms or enterococci and in women who had been treated with antibacterial drugs, than in others. We found no important associations between the presence of bacteria and the patients' age, parity or social class.

The authors report the findings from an on-going of 1764 vasectomized men followed up by means of the Scottish medical record linkage system for a total of 4500 man-years after operation. The authors view the results so far as reassuring regarding the safety of vasectomy, although the mean length of follow-up (2.6 years) is still short. The study will continue. Some features of this method of follow-up are described and discussed.

Data from the Oxford Record Linkage Study were analysed to determine the amount of work undertaken in day case surgery for 12 surgical conditions in five districts in the Oxford Region during the years 1976 to 1985. Record linkage was used to study readmission rates, comparing day surgery with inpatient care. The use of day surgery gradually increased in some conditions (eg, termination of pregnancy, female sterilisation) but did not increase from a fairly low base for others (eg, inguinal hernia repair, operations on varicose veins and haemorrhoids). There were striking differences between the districts in the use of day case care. For example, the use of day case care as a percentage of all hospital admissions for termination of pregnancy varied from 1% in one district to 24% in another; that for dilatation and curettage varied from 1% to 43%; and that for female sterilisation varied from less than 1% to 35%. Emergency readmission rates after day surgery were similar to those following inpatient treatment. We conclude that the use of day surgery for some conditions judged suitable for day care is still low and, even within one region, variation in the use of day surgery is considerable. The reasons for continued reluctance in some places to undertake more day surgery merit investigation.

Women attending a family planning clinic were studied to determine the relation between cervical erosion and clinical and social characteristics. The appearance of the cervix was recorded without knowledge of the women's symptoms. The prevalence of erosion increased with parity but, when the effects of other factors were controlled, decreased in women aged 35 and over. Erosion was significantly more common in women taking the "pill" and less common in women using barrier methods of contraception than in others. There was considerable variation between doctors in the reporting of erosion. No association was found between erosion and postcoital bleeding, dyspareunia, backache, or dysuria. There was a significant but modest association between erosion and vaginal discharge and a suggestion that erosion may sometimes be associated with nocturia and frequency of micturition. Vaginal flora was similar in women with and without erosion. Cervical erosion should not be regarded as pathological in asymptomatic women, nor should it be assumed necessarily to be the cause of symptoms in women with genitourinary complaints.

The case-fatality rate for acute bacterial meningitis in childhood is still disappointingly high. A review of 687 cases of acute bacterial meningitis including 72 deaths, identified in a defined population of children under ten years of age, was done to determine where patients were treated and where they died. Ten per cent of all hospital admissions were to infectious-disease hospitals, 16 per cent to teaching hospitals, and 74 per cent to general, non-teaching ("regional-board") hospitals. The case-fatality rates associated with these hospitals were three per cent, three per cent and 11 per cent respectively. The difference between regional-board and other hospitals was smaller, although still present, when children who died within a few hours of hospital admission were excluded from the analysis. Twenty-two per cent of all deaths (10 children who died outside hospital and six who were certified dead on admission) occurred before specialist care was reached. It may be possible to reduce the mortality from this disease by improving facilities for treatment in some hospitals. There is, however, an important residue of deaths--those which occur before hospital care is reached and, no doubt, some of those which occur soon after hospital admission--which would be unaffected by such improvements.

We performed two epidemiologic studies using routine abstracts of medical records to test the hypothesis that vasectomy may predispose men to cardiovascular disease. In a case-control study 1512 men who were under 55 years of age and had a history of myocardial infarction, stroke, or hypertension were matched with 3024 controls with other conditions; 2.4 per cent of the cases and 2.7 per cent of the controls were identified as having undergone vasectomy (risk ratio, 0.9; 95 per cent confidence limits, 0.6 to 1.3). In a cohort study data covering a mean period of 6 1/2 years after surgery were available on 1764 men who had had a vasectomy and on three comparison cohorts of men who had had other minor surgical procedures. There was no evidence of an increased risk of cardiovascular disease associated with vasectomy. There is no consistent evidence from our studies to support the hypothesis that in the short term vasectomy predisposes young men to cardiovascular disease.

Children with squints and old people with cataracts account for most of the work of ophthalmologists. Recent demographic changes have led to fewer children and more elderly patients needing treatment. So far the increased workload has been met by shortening the time patients spend in hospital, but with striking variations between districts.

Yeasts and fungi other than Candida albicans were isolated from the vagina of 86 out of 1498 women (5.7 per cent) who attended a family planning clinic and they comprised 27.7 per cent of all fungal species obtained. These fungi were significantly more common in women aged 35 years and over and in those who used the diaphragm than in others. They were no more common in women with an abnormal vaginal discharge on examination than in others; and they were no more common in women with genitourinary symptoms than in women without.

The discrepancy between the length of the waiting list and eventual admissions from the waiting list was investigated by comparing data from two different sources of routine statistics in the Oxford region. It was estimated that about 28% of the waiting list comprised patients who were not eventually admitted to hospital within the region.

Some basic features of the prehospital management of children with acute bacterial meningitis were ascertained by a retrospective review of case notes. A diagnosis of intracranial infection was considered in 63% of general practitioners' referral letters and was more often considered for older than younger children. 47% of children had received antibacterial chemotherapy before hospital admission. Such treatment did not appear to have jeopardized the outcome. The possibility of delay in the admission of children to hospital at weekends is discussed.

The report of the Resource Allocation Working Party recommended that revenue allocations to health authorities should be based, in part, on national patterns of bed usage and local standardised mortality ratios for conditions aggregated according to the chapters of the International Classification of Diseases (ICD). Similar criteria are now being considered for planning purposes by regions. The extent to which diseases which commonly result in the use of hospital care are also common causes of deaths within their ICD chapter was studied. National utilisation figures show that most beds in ophthalmology, ear, nose, and throat surgery, gynaecology, and consultant dentistry, and an estimated one-third or more of the beds used in general surgery, neurosurgery, and plastic surgery, are used for the treatment of conditions which are uncommon causes of death, both in absolute terms and relative to their ICD chapters. It seems unlikely that the requirements for care of patients with these diseases can be measured simply, either by all-causes mortality statistics, or by the use of mortality statistics ascribed to the ICD chapter which such diseases share with other, more common, causes of death. Consideration needs to be given to the diseases treated by each specialty in deciding whether and how to apply mortality statistics in planning for and funding the specialty.

Some of the main features of neonatal meningitis, studied in a defined population of infants, are reviewed. The incidence rate was twenty-six cases per 100 000 live-births. The case fatality rate was 43%. Gram-negative intestinal bacilli were the commonest infecting organisms. Streptococci accounted for an important minority of cases. It is stressed that neonatal meningitis usually presents as a non-specific illness, without the clinical signs of meningitis which are familiar in older patients. Twenty-three of the seventy-six infants in this study (30%) weighed 2500 g or less at birth. Another eight infants were born to mothers whose membranes had been ruptured for 24 hr or more.

The length of time that patients spend on waiting lists is a topic of current concern. Calculating the proportion of patients who have been on a waiting list for a long time by taking a census of patients on the list at a single point in time will tend to yield a higher estimate than that obtained by calculating waiting times of patients admitted to hospital during a period of time. To illustrate this point the waiting times of patients in the Oxford region as measured by SBH 203 returns ("census" data) were compared with those as measured by the Hospital Activity Analysis ("event" data). As expected, the SBH 203 census returns showed a higher proportion of patients who had waited over a year compared with the "event" measure of all admissions. This difference, which is analagous to the difference between prevalence and incidence in epidemiology, should be considered when using data from these sources to calculate waiting times.

The relation between changes in inpatient workload, measured as increases or decreases in the number of inpatients admitted from the waiting list, and the overall length of the waiting list was studied. Overall trends in admissions from the waiting list, the influence of seasonal patterns, and the impact of industrial action on admissions were also studied. The hypothesis was that when admissions from the waiting list increased the length of the waiting list would decrease and vice versa. No such simple relation was found. In fact, if anything, as the number of admissions from the waiting list increased so did the length of the waiting list. This result could be due to inconsistencies in compiling waiting list data or to the use of waiting lists to improve organisational efficiency. It is also possible, and perhaps likely, that the ability to meet need in admitting patients to hospital influences patients and their doctors to translate previously unmet need into demand for hospital services.

Vaginal swabs were taken from 1498 women attending a family planning clinic. The flora was assessed in the absence of any information about the women to whom the swabs related. Yeasts and fungi were present in 311 women (21%) and were no more prevalent among "pill" users than others. Candida albicans was significantly associated with vulval itching and with a vaginal discharge described as heavier than normal or curdy on clinical examination, though these abnormalities were present in only a minority of women with the organism. Trichomonas vaginalis was found in 14 women (1%) and was associated with abnormalities of vaginal discharge in all but one. Gram-negative anaerobic bacilli were significantly more common in women with a troublesome vaginal discharge and those who used an intrauterine device than others. No associations were found between fungi other than C albicans or the other bacteria sought and either symptoms or clinical abnormalities of vaginal discharge.

Significant space-time clustering was found for cases of meningococcal disease. This result was entirely accounted for by the occurrence of a small number of sibling pairs, and clustering was no longer found when these were omitted from the analysis. Meningococcal disease should still be regarded as potentially communicable between siblings. However, in this region of England the current level of infectivity of the disease outside the family is low. Space-time clustering was not convincingly demonstrated for haemophilus meningitis. This emphasizes that, even with diseases of known microbial aetiology, evidence for such clustering may be difficult to obtain. Only a small number of cases of haemophilus meningitis occurred in single-child families. Cases of single-child families tended to occur in older children than the remainder. Although children under three years of age are most susceptible to haemophilus meningitis, it is likely that the organism is usually introduced into the family by an older sibling.

Several different sources of data, including statutory notifications, were used to identify cases of acute bacterial meningitis in childhood in a defined population. Only half the cases of meningococcal meningitis and less than one quarter of ofther types of bacterial meningitis had been notified. Notified cases differed from the remainder with respect to causative organism, age, outcome, and type of hospital of admission. Thus notifications are an unreliable measure of incidence for this disease, and studies of infectious disease based on notified cases alone should be interpreted with considerable reserve. It is suggested that infectious-disease surveillance should be both clinical and microbiological and that data from different sources need to co-ordinated. The reason for reporting each disease (or organism) should be specified and and reviewed periodically.

A census was taken of outpatient bookings at all hospitals and health centres in Oxfordshire for the main medical and surgical specialities. Nine per cent of all bookings were to peripheral clinics (that is, those outsie the two main medical centres at Oxford and Banbury). About half of all bookings in the area were made to clinics within three of four miles of the patients' homes. The work load at peripheral clinics was mainly local: 75% of all patients at these clinics lived in the town or parishes adjacent to the town where the clinic was held. Peripheral clinics were associated with a lower work load at central clinics, but a higher overall work load from the town in which they were sited.

Readmission rates after inpatient care were studied by using routinely collected data from the Oxford record linkage study for 1968-85. Discharges from hospital and subsequent admissions were identified for people who were both resident and treated in the area covered by the linkage study. Rates were calculated for readmissions within 28 days after discharge from the first, index event. Readmission rates for elective readmissions after elective index admissions rose from 3.5% in 1968 to 7.1% in 1985. Those for elective readmissions after immediate (emergency or accident) index admissions rose from 2.4% to 3.5% during the same period. Emergency readmissions after an immediate index admission rose from 4.0% to 7.0%, and emergency readmissions after an elective index admission rose from 1.3% to 2.5%. All these increases were significant. The rise in elective readmissions may in part reflect a trend towards planned discharge with the expectation of readmission. The rise in emergency readmissions, which has been fairly gradual over many years, may, in some cases, be due to pressure on resources and inappropriately short lengths of stay. Further evidence is required to confirm or refute this. Readmission rates are one of the few potential measures available from routine statistics for assessing outcome, but due consideration must be given to issues of method and interpretation.