English Health Region Research Articles

Bed-sharing in the absence of hazardous circumstances: is there a risk of sudden infant death syndrome? An analysis from two case-control studies conducted in the UK.

ObjectiveThe risk of sudden infant death syndrome (SIDS) among infants who co-sleep in the absence of hazardous circumstances is unclear and needs to be quantified.DesignCombined individual-analysis of two population-based case-control studies of SIDS infants and controls comparable for age and time of last sleep.SettingParents of 400 SIDS infants and 1386 controls provided information from five English health regions between 1993–6 (population: 17.7 million) and one of these regions between 2003–6 (population:4.9 million).ResultsOver a third of SIDS infants (36%) were found co-sleeping with an adult at the time of death compared to 15% of control infants after the reference sleep (multivariate OR = 3.9 [95% CI: 2.7–5.6]). The multivariable risk associated with co-sleeping on a sofa (OR = 18.3 [95% CI: 7.1–47.4]) or next to a parent who drank more than two units of alcohol (OR = 18.3 [95% CI: 7.7–43.5]) was very high and significant for infants of all ages. The risk associated with co-sleeping next to someone who smoked was significant for infants under 3 months old (OR = 8.9 [95% CI: 5.3–15.1]) but not for older infants (OR = 1.4 [95% CI: 0.7–2.8]). The multivariable risk associated with bed-sharing in the absence of these hazards was not significant overall (OR = 1.1 [95% CI: 0.6–2.0]), for infants less than 3 months old (OR = 1.6 [95% CI: 0.96–2.7]), and was in the direction of protection for older infants (OR = 0.1 [95% CI: 0.01–0.5]). Dummy use was associated with a lower risk of SIDS only among co-sleepers and prone sleeping was a higher risk only among infants sleeping alone.ConclusionThese findings support a public health strategy that underlines specific hazardous co-sleeping environments parents should avoid. Sofa-sharing is not a safe alternative to bed-sharing and bed-sharing should be avoided if parents consume alcohol, smoke or take drugs or if the infant is pre-term.

Soft methods, hard targets: regional alcohol managers as a policy network

Regional Alcohol Managers (RAMs) were employed in the nine English health regions over 2008–2011. Their mission was to impact on the ‘hard target’ of Alcohol-Related Hospital Admissions (AHRAs) through the ‘soft methods’ of persuasion and influence: working with local partners on evidence-based interventions. Drawing on a qualitative evaluation, this article shows how a central government policy imperative (ARHAs) led to ‘government at a distance’ responses, including the introduction of RAMs. The processes involved in shaping and delivering this function bore the hallmarks of a complex, interactive policy network model, involving individuals whose bearings and roles were flexible and sometimes ambiguous. While there were overlaps and blurring of boundaries, there were three levels of policy network: central government, regional and local. As the ‘network in the middle’, the RAMs were pulled in both directions by conflicting agendas but were also able to have an impact on central and local policy.

Volume of print media coverage and diagnostic testing for influenza A(H1N1)pdm09 virus during the early phase of the 2009 pandemic

BackgroundImpact of an infectious disease on public health diagnostic health services may be affected by the volume of media coverage which can amplify risk perception and increase demand for services. ObjectivesTo examine the association between volume of newspaper reports and laboratory testing for influenza A(H1N1)pdm09 in one English health region during the early phase of the pandemic. Study designCross-sectional retrospective review identifying newspaper articles on A(H1N1)pdm09 in major regional (sub-national) newspapers from 27 April 2009 through 5 July 2009, and comparing the weekly frequency of articles with the weekly number, and positivity rate, of laboratory-confirmed cases of A(H1N1)pdm09 during the same time period. ResultsA positive correlation (r=0.67; p=0.02) was seen between the volume of school-related articles and the number of laboratory-confirmed cases. Increased testing during the most intense period of the pandemic was mainly seen in school-aged children (5–15years) and adults (≥16years). Adults accounted for the highest number of tests, but had the lowest positivity rates, which were highest among school-aged children. As the volume of media coverage decreased this was followed one week later by a fall in the number of tests and positivity rates in each age-group. ConclusionThe results presented suggest a temporal association between volume of media reporting and number of laboratory tests. The increased volume of media reporting, in particular the intense school-related coverage, may have raised population concern leading to an increased demand for diagnostic testing. These results have potential implications for future pandemic preparedness planning.

Communication of carrier status information following universal newborn screening for sickle cell disorders and cystic fibrosis: qualitative study of experience and practice

To describe and explore current practice, methods and experience of communicating carrier status information following newborn screening for cystic fibrosis (CF) and sickle cell (SC) disorders, to inform practice and further research. Three linked qualitative studies. All nine health regions in England. Child health screening coordinators in all English health regions, health professionals communicating results to parents and parents of newborn carriers. A preliminary phase of semi-structured telephone interviews with child health screening coordinators in all nine English health regions, and thematic analysis of data; semi-structured face-to-face interviews with purposeful samples of 67 family members of 51 infants identified by universal newborn screening as carriers of CF or SC with data analysis by constant comparison; and semi-structured telephone interviews, and focus groups, with a key informant sample of 16 differing health professionals currently tasked with communicating results to parents in a range of ways, with thematic analysis of data. Methods for and respondents' experiences of communication of carrier results varied considerably within and between regions, and within and between SC and CF contexts. Approaches ranged from letter or telephone call alone, to in-person communication in the clinic or at home, with health professionals from haemoglobinopathy, CF, screening and genetics backgrounds, or from community and primary care, such as health visitors with SC carrier results. Health professionals identified pros and cons of different methods, preferring opportunity for face-to-face communication with parents where possible, particularly for CF carrier results. They were concerned by regional variations in protocols, the lack of availability of translated information on SC carrier results, and the feasibility of sustaining more 'specialist' involvement at current levels, particularly for SC carriers. Parents were often poorly prepared for the possibility of a newborn carrier result. Some had felt overloaded by screening information received during pregnancy or prior to newborn screening, or found this information failed to meet their needs. Opportunity for face-to-face communication of results was valued by parents of SC carriers and appeared particularly necessary for those without prior knowledge of SC carrier status or where English was not their first language. Indirect communication of results by letter appeared effective and feasible for parents more aware of SC carrier status from antenatal or earlier experience, and where this communication contained an unambiguous opening statement emphasising 'your child is not ill'. Face-to-face communication of CF carrier results by professionals with screening, CF or genetics backgrounds worked well for parents, but communication and information was crucially lacking at the earlier stage of repeat blood spot testing, creating considerable distress among half of respondents. Respondents had no particular preference for the type of health professional who communicated results to them, as long as they were well informed and could answer their queries. Parents regarded carrier results as valuable information gained fortuitously. Methods of communication of newborn carrier results vary considerably across England. Parents' needs for timely and appropriate information may not be met consistently or adequately. Respondents' experiences suggest a need for greater recognition of communication with individuals occurring across a screening pathway, rather than as a discrete event.

Twenty-year trends in diagnosis of life-threatening neonatal cardiovascular malformations

Background:Infants with cardiovascular malformations are usually asymptomatic at birth. Earlier diagnosis is likely to improve outcome.Objective:To examine trends in the diagnosis of potentially life-threatening cardiovascular malformations.Methods:Ascertainment of all cardiovascular malformations...

Sudden Infant Death Syndrome and the time of death: factors associated with night-time and day-time deaths

To investigate the diurnal occurrence of Sudden Infant Death Syndrome (SIDS) and interaction with established risk factors in the infant sleeping environment. A 3 year population-based case-control study, in five English Health Regions. Parentally defined day-time or night-time deaths of 325 SIDS infants and reference sleep of 1300 age-matched controls. The majority of SIDS deaths (83%) occurred during night-time sleep, although this was often after midnight and at least four SIDS deaths occurred during every hour of the day. The length of time from last observed alive until the discovery of death ranged from <l to 14 h but was not significantly different from the corresponding sleep period amongst the controls. Amongst the day-time deaths, 38% of the infants were observed alive 30 min prior to discovery and 9% within 10 min. The risk of placing infants asleep on their side was more marked for day-time deaths (interaction: P = 0.0001) nearly half of whom were found prone, while the risk associated with paternal smoking [OR = 3.25 (95%CI: 1.88-5.62)] was more marked for night-time deaths (interaction: P = 0.02). The adverse effect of unsupervised sleep recognized for night-time practice [OR = 5.38 (95%CI: 2.67-10.85)] was also significant for day-time sleep [OR = 10.57 (95%CI: 1.47-75.96)]. Significantly more (P = 0.002) unobserved SIDS infants (24.8%) were found with bedclothes over the head compared with those SIDS infants where a parent was present in the room (11.3%). SIDS can happen at any time of the day and relatively quickly. Parents need to be made aware that placing infants supine and keeping them under supervision is equally important for day-time sleeps.

Abstracts for the British Congenital Cardiac Association Annual Meeting: The Barbican, London, 24–25 November 2005: Oral Presentations: Twenty year trends in recognition of life-threatening neonatal cardiac malformations

Background: Babies with cardiovascular malformations are usually asymptomatic at birth. Earlier diagnosis is likely to lead to a better outcome. This study examines trends in the timing of diagnosis of potentially life-threatening cardiovascular malformations. Methods: Ascertainment of all cardiovascular malformations diagnosed in infancy in the resident population of one English health region in 1985–2004. Babies considered to have a life-threatening cardiovascular malformation included all those with hypoplastic left heart, pulmonary atresia with intact ventricular septum, transposition of the great arteries, or interruption of the aorta; and babies dying or undergoing operation in the first 28 days with coarctation of the aorta, aortic stenosis, pulmonary stenosis, tetralogy of Fallot, pulmonary atresia with ventricular septal defect, or total anomalous pulmonary venous connection. Results: Of 688,167 live born babies in the 20 years of the study, 4444 had cardiovascular malformations diagnosed in infancy (6.5 per 1000). Cardiovascular malformations were potentially life-threatening in 685 (15%). Over the time of the study, 8% were recognised prenatally, 58% postnatally before discharge from hospital, 31% in life after discharge and 3% after death. Malformations most likely to remain undiagnosed at discharge were coarctation of the aorta (63%), interruption of the aortic arch (50%), and total anomalous pulmonary venous connection (47%). Over the 20 years of the study, the proportion of babies diagnosed antenatally increased from around 1% to 20% and no case was first diagnosed after death in the last 10 years. However, the recent proportion going home without a diagnosis remains around 25%. Conclusions: Overall one third of babies with potentially life-threatening cardiovascular malformations leave hospital undiagnosed. In recent years better antenatal diagnosis has reduced this proportion to one quarter. Better early recognition of such babies is unlikely to be achieved by clinical examination and is more likely to come from further improvements in antenatal diagnosis and more widespread adoption of routine pulse oximetry.

Implementing a national treatment service for dependant smokers: initial challenges and solutions

Before 1999, few treatment services for nicotine-addicted smokers existed in England. When national treatment services were introduced, those responsible for setting them up liaised closely with primary care health services. Setting up an entirely new national service, treating a new category of patient (smokers motivated to stop) was an ambitious aim and this paper documents the problems encountered in the early stages of this process. To describe the principal challenges encountered and solutions employed by those setting up the services during the initial period of smoking cessation service implementation. Qualitative, semistructured interviews with 50 smoking cessation staff in two former English health regions conducted in autumn 2001. Two principal factors which slowed the initial development of smoking cessation services were: (i) the lack of a work-force with experience in smoking cessation methods and (ii) the fact that services were set up outside existing primary and secondary care health services in England. As few training courses in smoking cessation were available, many services provided their own in-house training for staff appointed as smoking cessation advisers. Consequently, senior service staff devoted a lot of effort to training new staff which meant that they had less time to spend on other important tasks which were necessary for service implementation. Smoking cessation services needed to develop relationships with primary care health services in order to generate referrals and find venues for the delivery of smoking cessation interventions. Liaising with primary care physicians was time-consuming, however, and some primary care physicians were opposed to the ideas that service staff had for the interface between primary care and smoking cessation services. As new smoking cessation services were not set up within existing primary or secondary health care services, service staff had to spend large amounts of time on this process of negotiation and overcoming scepticism from some primary health care physicians. If smoking cessation services are set up in other countries, rapid implementation would be facilitated by ensuring that adequate numbers of health professionals trained in smoking cessation methods are available to staff services. Additionally, locating new smoking cessation services within existing health providers' services may speed up service implementation, but this option may not suit all health systems.

Targeting smokers in priority groups: the influence of government targets and policy statements

Cessation services were instructed to make special attempts to attract smokers from three 'priority groups': those who were young, pregnant and economically disadvantaged. Progress with attracting priority groups was not assessed formally, but services were set monitored targets to encourage throughput of smokers. Initial research suggests that services have been successful in attracting smokers who live in disadvantaged areas. This paper investigates how smoking cessation services responded to targets and instructions to attract priority groups and discusses the relative impact these on service development. To describe how monitored throughput targets influenced the development of smoking cessation services, including attracting priority groups. To describe the range of priority groups that smoking cessation services targeted, methods used and reported progress with this. Postal surveys of English smoking cessation coordinators conducted in April 2001 and April 2002. Seventy-eight qualitative, semistructured interviews with cessation service staff in two former English health regions conducted in autumn 2001 and 1 year later. A total of 69.3% of coordinators responded to the first survey (79% to the second survey). In the first survey 91% reported targeting priority groups (100% in the second survey). The proportions (second survey in brackets) who reported targeting the different priority groups were: pregnant women 86% (99%), economically disadvantaged 79% (100%) and young smokers 20% (75%). Interviews showed that coordinators gave the greatest priority to reaching monitored targets as they came under pressure to achieve these. Service staff were generally unclear about how to attract priority groups and developing strategies for this was hindered by the need to meet throughput targets. Locating services in poor areas was thought to attract economically disadvantaged smokers and specialist staff were being appointed to work with pregnant smokers, but otherwise there was little evidence of active strategies for attracting priority groups being applied in practice. Monitored targets for smoker throughput ensured that services quickly began to treat smokers, but this rapid implementation diverted service staff from devising methods for attracting priority group smokers. Coordinators found reaching priority groups challenging and, particularly in the case of young smokers, would have appreciated clear instructions for this aspect of service implementation. Those implementing services in other countries should consider whether similar targets would be helpful to stimulate service development within their health systems.

The prevalence and characteristics associated with parent–infant bed-sharing in England

Aims: To investigate the characteristics of parent–infant bed-sharing prevalence in England. Methods: Data on night-time sleeping practices from a two year, local, longitudinal study and a three-year, national, cross-sectional study...

Sudden infant death syndrome and social deprivation: assessing epidemiological factors after post-matching for deprivation.

As part of the confidential enquiry into stillbirths and deaths in infancy (CESDI), a 3-year population-based case-control study was specifically designed to look at risk factors associated with sudden infant death syndrome (SIDS) after the dramatic fall in incidence. The study was conducted between 1993 and 1996 in five English Health Regions (population 17 million) with parental interviews for each death and four age-matched controls. The aim of this analysis was to investigate the extent to which epidemiological characteristics associated with SIDS were particular to the syndrome or more general markers for socio-economic deprivation. One control was reassigned to each case post-matched for infant age, time of sleep and socio-economic status using components of the Townsend Deprivation Score. The post-matched analysis involved 323 SIDS infants and 323 controls with a similar socio-economic profile. Notable factors significant in the original univariable analysis that became non-significant after post-matching included young maternal age (median: 23 years 4 months SIDS vs. 23 years 11 months post-matched controls), being an unsupported mother (13.6% SIDS vs. 11.1% post-matched controls) and being bottle-fed (56.7% SIDS vs. 55.4% post-matched controls). Other factors, although clearly related to deprivation, such as parental smoking, remained significant in both the univariable and multivariable post-matched analyses.

Public Participation in the New NHS: No Closer to Citizen Control?

Over the last decade support for increasing public participation in decisions regarding the planning and delivery of health services has become a familiar feature of the policy agenda for the UK National Health Service. This paper reviews current Labour policy towards public participation and reports on the response of primary care groups (PCGs) to recent Labour directives to make patient and public involvement an integral part of the way they work, presenting the findings of a survey conducted in one English health region. The experience of these PCGs suggests that, despite the diverse backgrounds of board members, there is marked consensus between local and central decision makers as to their understanding of public participation. Whilst academic debates have tended to conceptualize participation in dualist terms as a form of consumerism or of citizenship, the survey data suggest that in the context of local implementation public participation is framed within a new public management perspective which values it as an aid to organizational learning. The findings of this study highlight obstacles to securing effective public participation, including a lack of substantive guidance regarding policy implementation that produces uncertainty amongst local decision makers as to how best to proceed. The inherent limitations of public participation within the new public management paradigm suggest that democratic renewal, one of the goals of the government’s modernization agenda, is unlikely to be achieved.

Invasive Haemophilus influenzae disease: an ecological study of sociodemographic risk factors before and after the introduction of Hib conjugate vaccine.

This study examines the impact of H. influenzae type b (Hib) conjugate vaccine on sociodemographic risk factors for invasive H. influenzae disease in the 2 years before and immediately after the introduction of Hib conjugate vaccine. An ecological study design was used and cases were identified using active surveillance employing several surveillance systems. The study population comprised all children aged < 5 years resident in the West Midlands, an English health region, with laboratory confirmed invasive disease 2 years before (1990-1992) and 2 years after (1992-1994) the introduction of Hib conjugate vaccine. Selected sociodemographic variables derived from the UK census were obtained for all census enumeration districts in the region. Each variable was then ranked and divided into six categories. Linear associations between disease rates and sociodemographic variables were examined. Overall, there was a significant reduction in the incidence of invasive H. influenzae disease. In the pre-conjugate vaccine era there were trends of decreasing disease incidence with increasing child population density (p = 0.012) and total population density (p = 0.0023). In the post-conjugate vaccine period, total population density (p = 0.0275) remained significant and a trend of increasing disease incidence with increasing population mobility (p = 0.0012) was seen. Although Hib conjugate vaccine has resulted in a dramatic reduction in disease incidence changes in sociodemographic risk factors were identified in the post-conjugate vaccine period, particularly population mobility. Our results may have implications for current and future vaccine strategies.

Health needs assessment and needs-led health service change: a survey of projects involving public health doctors.

To describe the criteria that define an effective health needs assessment and to explore which factors are important for the delivery of effective health needs assessment in the English National Health Service (NHS). A postal questionnaire to all public health doctors in an English health region to describe health needs assessment activity, including initiating factors, methods and outcomes. This was followed by semi-structured interviews with public health professionals and others involved in 10 purposively selected needs assessments. A response rate of 62% identified a total of 102 health needs assessments undertaken between 1993 and 1998. A number of themes emerged as being important in the impact of health needs assessments on policy and planning. These included careful design, methodological rigour, decisive leadership, good communication, involvement and ownership of the work from relevant stakeholders, support from senior decision-makers, appreciation of the political dynamics and engagement with local priorities, availability of resources and, finally, an element of chance. These themes can be categorised broadly into contextual factors and quality or robustness of the work. Our findings suggest that, although methodological and analytical quality are necessary characteristics of effective health needs assessment, they are not sufficient without a favourable political environment. There appear to be conditions under which needs assessment is more likely to be effective in terms of its influence on policy and planning. However, it is clear that needs assessment does not occupy a central position in health service decision-making, remaining vulnerable to a range of factors over which those responsible for its conduct have little or no control.

Does waiting for total hip replacement matter? Prospective cohort study.

To assess the impact on the outcome of total hip replacement of the length of timing spent waiting for surgery. One hundred and forty-three orthopaedic and general hospitals provided information about aspects of surgical practice for each total hip replacement conducted between September 1996 and October 1997 for publicly and privately funded operations in five English health regions. These data were linked to patient information about hip-related pain and disability status (measured using the Oxford Hip Score) before operation and at 3 and 12 months after. Data were analysed using multiple regression analysis. Questionnaires were completed by surgeons for 10,410 (78%) patients treated during the recruitment period and by 7151 (54%) patients. Twelve months after total hip replacement, the majority of patients experienced substantial improvements in hip-related pain and disability (as measured by the Oxford Hip Score). Those patients who started with a worse Oxford Hip Score before the operation tended to remain worse after the operation. Worse pre-operative score was associated with an increased length of either outpatient or inpatient wait, and this trend remained after the operation. The relationship between waiting time and outcome remained after adjustment for possible confounding variables. A consistently worse score was observed in public compared with private patients at all three time-points. In addition, in both sectors, those patients who were socially disadvantaged had a worse score than more socially advantaged patients both before and after the operation. Waiting for surgery is associated with worse outcomes 12 months later. Longer-term outcome needs to be considered to see if this association persists.

Neonatal screening for haemoglobinopathies: the results of a 10-year programme in an English Health Region.

Neonatal identification of sickle cell disease can significantly reduce mortality and morbidity during the first 5 years of life. During a 10-year period, 414,801 neonates were screened by isoelectric focusing. The most common variants detected were haemoglobins S, C, D and E. Two hundred and fifty of the samples tested were homozygotes or compound heterozygotes, and 6554 samples were heterozygotes for the common variants. The gene frequencies in the population tested were calculated from this data for the most common variants. They were: S, 0.0057; C, 0.0014; D(Punjab(Los Angeles)), 0.0007; E, 0.0005. Additionally, 16 babies had beta thalassaemia major and 405 had rarer variants, of which six had never previously been described. Knowledge of the distribution of these inherited diseases is useful in healthcare planning and appropriate allocation of resources, while counselling targeted at appropriate couples enables informed parental choice and may prevent disease.

Inequalities in health: approaches by health authorities in an English health region.

In 1995 the Department of Health published Variations in health: what can the Department of Health do? This recommended that health authorities should have a comprehensive plan for identifying and tackling variations in health. We investigated how health authorities in the South and West Region were taking forward this work. Semi-structured interviews and reviews of documentation were conducted in all health authorities in the South and West Region of England. All health authorities viewed tackling inequalities in health as important; however, explicit strategies did not exist and Health of the Nation targets were a vehicle for determining priorities of inequalities. Explicit corporate commitment was often weak. Analyses were being conducted to determine the magnitude of local health inequalities and to assist in designing appropriate interventions. The importance of alliance working was highlighted; much work was being done although success was variable. Efforts are being made throughout the South and West region to tackle inequalities in health. Although strategic vision at the corporate level was often lacking, there was evidence of commitment to taking the inequalities agenda forward within public health directorates. Strengthening of primary care and alliance working roles is essential. Recent national strategy documents, forthcoming legislation, and a review of health inequalities recognize the health effects of inequalities and require health authorities to collaborate with local partners to tackle these, and will offer opportunities to improve corporate commitment and alliance working. Uptake and success of these opportunities will have a major influence on progress in tackling health inequalities.

A Regional Collaborative Audit of the Practice and Outcome of Carotid Endarterectomy in the United Kingdom

Objectives to evaluate performance and outcome of carotid endarterectomy (CEA) against agreed audit standards within one English health region. Design a prospective collaborative audit over twelve months (November 1994 to October 1995) involving all surgeons undertaking CEA within one English health region. Methods audit standards were agreed by all participating surgeons at the outset based on existing national guidelines. Data were abstracted from clinical notes. Outcomes were reviewed by clinicians 30 days post-surgery. A confidential individualised report of the results was provided to each surgeon. A survey of participating surgeons sought to evaluate the audit process. Results ten surgeons performed 139 CEAs on 134 individuals (64% men). Median per surgeon was 12 (range 1–44). Audit standards were generally achieved: 114 (82%) patients had symptomatic carotid stenosis of 70–99%, 14 (10%) were asymptomatic. The median time from first referral to hospital to operation was 4.8 months (interquartile range 3.0–7.3). The rate of disabling stroke or death at 30 days was 2.2% (95% confidence interval (CI) 0.4–6.4%). Surgeons valued the audit.Conclusions the study showed that in the study area CEA was performed predominantly on high-risk patients with low subsequent surgical mortality.

Accuracy of hepatic adverse drug reaction reporting in one English health region

The lack of specific markers for drug induced liver injury suggests that inaccurate reports of hepatic adverse drug reactions are likely to be common, with the diagnosis based on circumstantial...

Are Variations in the Use of Carotid Endarterectomy Explained by Population Need? A Study of Health Service Utilisation in Two English Health Regions

Objectives:to describe variation in utilisation of carotid endarterectomy (CEA) within two English health regions and explore relationships between use, need and proximity to services.Design:consecutive case series of operations. Comparison at a population level with district stroke mortality, hospital admissions and material deprivation.Main outcome measures:standardised utilisation rates for CEA and measures of inter-district variability. Spearman's rank correlation coefficients for associations between variables.Results:variation in utilisation rates was considerable (14-fold difference across district populations). More individuals had bilateral surgery in the Yorkshire region than in the Northern (11.7% vs. 5.5%, p=0.002). There was no association between utilisation rates for CEA and district stroke mortality (r=−0.06, 95% CI −0.41 to 0.30) or admission rates for stroke (r=0.17, 95% CI −0.2 to 0.49). There was a strong relationship between residence in districts where services were located and higher utilisation. Rates of CEA were lowest in the regions» most affluent wards.Conclusion:use of CEA varies widely, depending on area of residence. Variation is not a consequence of differences in need, but reflects clinical practice and supply of services. There is evidence to suggest unmet need for CEA.