Comparing measures of variation.
Explore the source record for details and available documents.
Humans↗
Biomedical subjects
M Goldacre
Publications and source records attributed to M Goldacre.
Interpreting hospital death rates. Can be difficult.
Explore the source record for details and available documents.
Data Collection↗
Measures of early postoperative mortality: beyond hospital fatality rates.
OBJECTIVE: To quantify the short term risk of postoperative mortality in ways which take account of deaths after discharge and the background risks of death in patients who come to operation. DESIGN: Analysis of linked abstracts of hospital admission records and death certificates for common operations. SETTING: Six health districts in the Oxford region. SUBJECTS: Records of 223,529 operations performed in 1980-6. MAIN OUTCOME MEASURES: In hospital fatality rates, case fatality rates, and standardised mortality ratios at selected time periods during the year after operation and the ratio of early (< 30 days) to late (90-364 days after operation) fatality rates. RESULTS: Fatality rates throughout the year after operations performed after emergency admissions were generally higher than those for similar operations performed after elective admissions and higher than expected from population rates. Examples were prostatectomy, hip arthroplasty, inguinal herniorrhaphy, and cholecystectomy. Common elective operations such as inguinal herniorrhaphy and cataract operations showed no early peak in mortality, but others did. These included transurethral prostatectomy (ratio of early to late mortality 2.0; 95% confidence interval 1.3 to 2.6), hysterectomy (3.2; 1.5 to 6.6), hip arthroplasty (3.8; 2.5 to 5.4), and cholecystectomy (6.9; 4.3 to 11.1). CONCLUSIONS: Temporal profiles of death rates in the year after operation show which operations have early peaks in mortality and which do not. Emergency and elective operations have very different profiles and should be analysed separately. For elective operations for conditions which pose no immediate threat to life the ratio of early to later fatality rates provides a measure of increase in mortality after operation while allowing for the background risk of death in the patient groups.
Chi-Square Distribution↗
How many patients are admitted in districts other than their own, and why?
Patients may be admitted across administrative boundaries for a number of reasons. We compared cross-boundary flows into a teaching district for individual operations and diagnoses with flows into a group of non-teaching districts. We have attempted to distinguish cross-boundary care that is the result of natural geographic factors from that which reflects specialized supra-district hospital practice. Hospital activity data were analysed for patients treated in six districts in the Oxford Regional Health Authority area in the calendar years 1979-1986 and the financial year 1990-1991. In 1979-1986, 21.1 per cent of patients admitted in the teaching district, and 9.2 per cent in the non-teaching districts, were from other districts. Cross-boundary flows varied greatly between specialties. The non-resident workload in the teaching district increased a little over time from 21.1 per cent (1979-1986) to 24.5 per cent (1990-1991); no appreciable change was seen in the non-teaching districts. For many common conditions, the proportion of non-residents admitted in the teaching district was similar to that in the non-teaching districts, at about 10 per cent or less. For other conditions, it was higher than this but similar in both the teaching and non-teaching districts. Conditions were also identified for which the proportion of non-residents treated was clearly much higher in the teaching district than elsewhere. Geographical convenience accounted for about half the non-resident workload in the teaching district. Specialized supra-district care constituted a small but important part of the total workload.(ABSTRACT TRUNCATED AT 250 WORDS)
Catchment Area, Health↗
Estimating incidence and prevalence of treated psychiatric disorders from routine statistics: the example of schizophrenia in Oxfordshire.
STUDY OBJECTIVE: To use routine statistical records to estimate the incidence and prevalence of treated schizophrenia. DESIGN AND SETTING: Analysis of linked records in Oxfordshire (population 540,000) for all people in contact with specialist psychiatric services from 1975-86. SUBJECTS: Records of 685 people with a diagnosis of schizophrenia as an inpatient and a further 294 people who received specialist psychiatric care for schizophrenia outside hospital without any record of inpatient care. MEASUREMENTS AND MAIN RESULTS: The measures most commonly recorded in psychiatric statistics, first admission rates for people in whom schizophrenia was recorded at their first psychiatric admission, were 8.7 per 100,000 males and 5.6 per 100,000 females. First contact rates for people in whom schizophrenia was recorded at any time in the study period and in any setting were 15.1 per 100,000 males and 11.4 per 100,000 females. Whichever patient population was analysed, the broad profile of schizophrenia by age, sex, and calendar time was similar. CONCLUSIONS: First admission rates for schizophrenia, as identifiable in current routine information systems, are useful indicators of the general pattern of disease but are inadequate absolute indicators of treated incidence. These data are limited to the first ever contact. Reliable information about the treated incidence of disease requires information systems which incorporate information about when and where each diagnosis was first made. Reliable information about treated prevalence requires systems which also incorporate data about death, recovery, and migration into and out of the study population.
Adolescent↗
Geographical variation in hospital admission rates: an analysis of workload in the Oxford region, England.
OBJECTIVE: To measure variation in hospital admission rates between health districts in part of the English NHS, comparing a wide range of medical and surgical conditions. DESIGN: Retrospective analysis of interdistrict variation using linked routine hospital admission data. Comparisons were also made with levels of variation reported from the USA. SETTING: Oxford Regional Health Authority, 1979-86. SUBJECTS: Six district health authorities--total study population 2.1 million people, 1.6 million hospital admissions. MAIN MEASURES: Age and sex standardised hospital admission rates for resident populations for individual operations and diagnoses; systematic components of variation (SCV). RESULTS: Of 118 standard operation groups, 38 (26% of surgical workload) showed high variation (SCV 16 or more) and 40 (36% of surgical workload) showed low variation (SCV < 4). Operations (SCV) with very low levels of variation included prostatectomy (0.1), inguinal herniorraphy (0.9), and cholecystectomy (1.3). Rates were more variable for myringotomy (3.7), hysterectomy (4.3), dilatation and curettage (5.6), and tonsillectomy (6.2). The SCV was high for only four of the 40 commonest medical causes of admission, and was low for 18 of them. CONCLUSIONS: Most admissions in the Oxford region were for conditions that did not show a great deal of variation in admission rates. The level of variation for many surgical procedures was less than that reported in studies from the USA. Variation was no greater for medical causes of admission than for surgical conditions. Large scale variation may not be an inevitable consequence of autonomous clinical practice.
Age Factors↗
Motor neuron disease in a defined English population: estimates of incidence and mortality.
Linked statistics from hospital records and death certificates were used to study the incidence of and mortality from motor neuron disease in a defined English population. The incidence of motor neuron disease, measured as first-admission rates for the disease, was studied from 1963 to 1985 and death certificates for the patients admitted to hospital were obtained to the end of 1990. The average annual first-admission rate for motor neuron disease was 2.1/100,000 men (95% confidence interval (CI) 1.9 to 2.4) and 1.7/100,000 women (95% CI 1.5 to 1.9). First-admission rates increased with age and peaked in the age range 65-84 years. Motor neuron disease was recorded on the death certificate for 86% of patients who died while they had the disease and there was no appreciable change over time in the recording of motor neuron disease as the underlying cause of death. The admission and mortality data derive from different sources, hospital statistical abstracts and death certificates respectively, but trends over time in the two data sets were similar. There was an increase in mortality during the period covered by the study, as there has been in mortality from motor neuron disease nationally, and the increase in mortality in the Oxford region was accompanied by an increase in first-admission rates. It is concluded that the increase in mortality from motor neuron disease probably occurred as a result of an increase in the diagnosed incidence of the disease rather than changes in death certification practice.
Adult↗
Future of small and specialist medical services.
Explore the source record for details and available documents.
Health Services Needs and Demand↗
Suicide after discharge from psychiatric inpatient care.
People with a history of psychiatric disorder are at higher risk of suicide than people without such a history. The policy of reducing inpatient care in psychiatry has probably meant that some of the risk of suicide has shifted from the hospital to the community setting. We have quantified the risk of suicide within a year of psychiatric discharge in a population-based study in the Oxford health region, UK. We calculated suicide rates per 1000 person-years at risk (time from discharge to death, subsequent readmission, or the end of the study) and the standardised mortality ratio (SMR) for suicide, taking the value among the general population as 1. Among male patients the SMR for suicide (defined by coroner's verdict of suicide) in the first 28 days after discharge from inpatient care was 213 (95% CI 137-317); the equivalent SMR for female patients was 134 (67-240). The result was similar when we defined suicide more broadly as a suicide, open, or misadventure verdict. The suicide rate in the first 28 days after discharge was 7.1 (4.1-12) times higher for male patients and 3.0 (1.5-6.0) times higher for female patients than the rate during the remaining 48 weeks of the first year after discharge. Most of the patients studied (both those who committed suicide and those who did not) had been psychiatric inpatients for only a short time. The findings confirm that there is significant clustering of suicide soon after discharge from psychiatric care. Skilled support after discharge for high-risk patients in the community is essential. Audit of suicides that occur soon after discharge may help identify the patients at highest risk and thereby reduce the number of avoidable deaths.
Adolescent↗
Risk of aseptic meningitis after measles, mumps, and rubella vaccine in UK children.
Cases of aseptic meningitis associated with measles/mumps/rubella vaccine were sought in thirteen UK health districts following a reported cluster in Nottingham which suggested a risk of 1 in 4000 doses, substantially higher than previous estimates based on cases reported by paediatricians (4 per million). Cases were ascertained by obtaining vaccination records of children with aseptic meningitis diagnosed from cerebrospinal fluid samples submitted to Public Health Laboratories or discharged from hospital with a diagnosis of viral meningitis. Both methods identified vaccination 15-35 days before onset as a significant risk factor and therefore indicative of a causal association. With both, half the aseptic meningitis cases identified in children aged 12-24 months were vaccine-associated with onset 15-35 days after vaccine. The study confirmed that the true risk was substantially higher than suggested by case reports from paediatricians, probably about 1 in 11,000 doses. However, the possibility that the aseptic meningitis induced by vaccination was largely asymptomatic and a chance laboratory finding in children investigated for other clinical conditions, particularly febrile convulsions, could not be excluded. Comparison of national reports of virus-positive mumps meningitis cases before and after the introduction of this vaccine indicated that the risk from wild mumps was about 4-fold higher than from vaccine. Altogether, 28 vaccine-associated cases were identified, all in recipients of vaccines containing the Urabe mumps strain. The absence of cases in recipients of vaccine containing the Jeryl Lynn strain, despite its 14% market share, suggested a higher risk from Urabe vaccine. A prospective adverse event surveillance system using the study methods is currently being established to assess the risk, if any, from the Jeryl Lynn strain which is now the only mumps vaccine used in the UK.
Child, Preschool↗
Multiple hospital admissions in a calendar year.
Hospital in-patient workload is routinely measured as episodes of care. We report on the extent to which counts of episodes of care differ from counts of patients treated in different specialties and in different age groups. Linked records of hospital care in a population of 1.9 million people, collected over an 11-year period (1976-1986), were analysed. The all-ages multiple admission ratio (the number of admissions per 100 people admitted in the same specialty and year) varied between specialties from 102 to 171. Medical specialties tended to have higher ratios than surgical ones. The influence of age on multiple admission ratios varied between specialties, although in general the ratios increased with increasing age. There were progressive but small increases in multiple admission ratios over the period studied in a number of specialties but, by and large, stability over time was more striking than any change. The information presented could be used to estimate person-based admission rates from available episode-based data where the former are not available. This should be helpful both in managing hospital resources and in purchasing care on behalf of resident populations. Purchasers in particular should be aware of numbers of people being treated as well as the numbers of episodes of care provided.
Adolescent↗
Population-based trends in treatment rates in psychiatry in Oxfordshire, 1975-1986.
Routinely collected abstracts of medical records for patients in Oxfordshire were used to identify each individual's first contact with specialist psychiatric in-patient, out-patient or community care services over a 12-year period (1975-1986). During this period first-contact rates for people under 65 years of age declined by 3.3 per cent. The decline was observed for contacts with specialist care outside hospital as well as for in-patient care. It is therefore not attributable simply to a shift away from in-patient care. The decline was seen in most major diagnostic groups including schizophrenia, affective psychosis and the neuroses, and it is therefore not attributable to changing diagnostic practices between these groups. Increases in first-contact rates in this age group were found for alcohol-related disorders in women and self-harm in men. First-contact rates for people aged 65 years and over increased by 3 per cent per year. The condition that mainly accounted for the increase over time was dementia, which was by far the commonest recorded psychiatric diagnosis in this age group. Although the increase was seen in all modes of care, it was the domiciliary visiting service that contributed most to the upward trend. The implications of these trends for service planning and their limitations as measures of changes in morbidity are discussed.
Adolescent↗
Morbidity following pelvic inflammatory disease.
OBJECTIVE: To examine patterns of morbidity following hospitalisation for pelvic inflammatory disease (PID). DESIGN: Cohort study using Oxford Record Linkage Study data. SETTING: Oxfordshire and West Berkshire. SUBJECTS: One thousand three hundred fifty-five women discharged from hospital for the first time with a diagnosis of pelvic inflammatory disease during the interval 1970-1985, together with 10,507 control women discharged with various other diagnoses. MAIN OUTCOME MEASURES: Hospital admission for abdominal pain, gynaecological pain, endometriosis, hysterectomy, and ectopic pregnancy. Data were not available in this analysis on pregnancies other than ectopic pregnancy. RESULTS: In comparison with the controls, women with a diagnosis of PID were ten times more likely to be admitted for abdominal pain, four times more likely to be admitted for gynaecological pain, six times more likely to be admitted for endometriosis, eight times more likely to be admitted for hysterectomy and ten times more likely to be admitted for ectopic pregnancy. CONCLUSIONS: Serious sequelae of PID are common, even in a study unable to examine the effects of the disease on fertility. The data on hysterectomy are particularly interesting; the operation in women with pelvic inflammatory disease is most often a consequence of the inflammatory process itself.
Adolescent↗
Use of hospital inpatient care in adolescence.
Epidemiological information about detailed patterns of physical morbidity within the adolescent age group is not generally available. To illustrate the distinctive patterns of morbidity indicated by the use of hospital inpatient care, hospital admission rates in the Oxford region (1979-86) were analysed at each single year of age from 10 to 19 years. At the age of 10 years 22% of general hospital admissions were to paediatrics, 24% to general surgery, 23% to ear, nose, and throat surgery, and 20% to trauma and orthopaedics. By 14 years of age only 6% of general hospital admissions were to paediatrics. By 16 years of age 24% of general hospital admissions of young women were to gynaecology and 40% of admissions of young men were to trauma and orthopaedics. The most common reason for hospital admission in young men was head injury and the second most common was appendicectomy. Termination of pregnancy was the single most common reason for admission for girls aged 15 and 16 years; childbirth and terminations were the most common reasons for admission in girls aged 17-19 years and over. Self poisoning was also common in older teenage girls. Younger girls were admitted most commonly for tonsillectomy. Most admissions of adolescents are thus for surgical rather than medical reasons and some of the most common individual reasons for admission are attributable to behavioural factors rather than disease processes.
Abortion, Legal↗
Ovarian cancer and ABO blood groups.
OBJECTIVE: To determine whether the distribution of ABO blood groups in women with ovarian cancer differs from that in the general population in a large, defined English region. DESIGN: Analysis of record abstracts of hospital care held in the Oxford record linkage study supplemented with data from the Oxford cancer registry. SETTING: Oxford Regional Health Authority area. SUBJECTS: A total of 1261 women who had ovarian cancer between 1968 and 1986 with ABO blood groups recorded on the Oxford Record Linkage Study and cross checked against the cancer registry comprised the study group. MEASUREMENTS AND MAIN RESULTS: The relative incidence of A:O and B:O blood groups in women with ovarian cancer were compared with the general population in the same region. Ovarian cancer was more common in women of blood group A than in others, with a relative incidence of 1.17. In particular, adenocarcinomas were the most common type of tumour and were associated with blood group A. The association was more striking in married women than in single women probably reflecting differences associated with parity. CONCLUSION: The association between ABO blood groups and ovarian cancer found in this English population is similar in size to that reported from several other populations. Childbearing is known to reduce the risk of ovarian cancer and our findings suggest that the blood group association may be most apparent in married, parous (that is, relatively low risk) women.
ABO Blood-Group System↗
Computerised linking of medical records: methodological guidelines.
OBJECTIVES: To report on the development of computer assisted methods for linking medical records and record abstracts. DESIGN: The methods include file blocking, to put records in an order which makes searching efficient; matching, which is the process of comparing records to determine whether they do or do not relate to the same person; linkage, which is the process of assembling correctly matched records into a time sequenced composite record for the individual; and validation checks and corrections, in which any inconsistencies between different records for the same person are identified and corrected. SETTING: The dataset comprising the Oxford record linkage study which includes hospital inpatient records and vital records. RESULTS AND CONCLUSIONS: Probability matching, using an array of identifiers, achieves much higher levels of correct matching than is generally achievable by exact character by character comparisons. The increasing use of information technology to store data about health and health care means that there is increasing scope to link records for research and for patient care. Sophisticated methods to achieve this on a large scale are now available.
England↗
ABO blood group and gall stone disease.
Explore the source record for details and available documents.
ABO Blood-Group System↗
Incidence of disease after vasectomy: a record linkage retrospective cohort study.
OBJECTIVE: To determine whether vasectomy is associated with an increased risk of several diseases, and in particular testicular cancer, after operation. DESIGN: Retrospective cohort study using linked medical record abstracts. SETTING: Six health districts in Oxford region. SUBJECTS: 13,246 men aged 25-49 years who had undergone vasectomy between 1970 and 1986, and 22,196 comparison subjects who had been admitted during the same period for one of three specified elective operations, appendicitis, or injuries. MAIN OUTCOME MEASURES: Hospital admission and death after vasectomy or comparison event. RESULTS: The mean durations of follow up were 6.6 years for men with a vasectomy and 7.5 years for men with a comparison condition. The relative risk of cancer of the testis in the vasectomy cohort (4 cases) compared with that in the other cohorts (17 cases) was 0.46 (95% confidence interval 0.1 to 1.4), that of cancer of the prostate (1 v 5 cases) 0.44 (0.1 to 4.0), and that of myocardial infarction (97 v 226 cases) 1.00 (0.8 to 1.3). There was no evidence of an increase associated with vasectomy in the incidence of a range of other diseases. CONCLUSIONS: Vasectomy was not associated with an increased risk of testicular cancer or the other diseases studied. With respect to prostatic cancer, while we found no cause for concern, longer periods of observation on large numbers of men are required.
Adult↗