Incidence and viral aetiologies of acute respiratory illnesses (ARIs) in the United States: a population-based study.

We conducted prospective, community-wide surveillance for acute respiratory illnesses (ARIs) in Rochester, NY and Marshfield, WI during a 3-month period in winter 2011. We estimated the incidence of ARIs in each community, tested for viruses, and determined the proportion of ARIs associated with healthcare visits. We used a rolling cross-sectional design to sample participants, conducted telephone interviews to assess ARI symptoms (defined as a current illness with feverishness or cough within the past 7 days), collected nasal/throat swabs to identify viruses, and extracted healthcare utilization from outpatient/inpatient records. Of 6492 individuals, 321 reported an ARI within 7 days (4·9% total, 5·7% in Rochester, 4·4% in Marshfield); swabs were collected from 208 subjects. The cumulative ARI incidence for the entire 3-month period was 52% in Rochester [95% confidence interval (CI) 42-63] and 35% in Marshfield (95% CI 28-42). A specific virus was identified in 39% of specimens: human coronavirus (13% of samples), rhinovirus (12%), RSV (7%), influenza virus (4%), human metapneumovirus (4%), and adenovirus (1%). Only 39/200 (20%) had a healthcare visit (2/9 individuals with influenza). ARI incidence was ~5% per week during winter.

Quality assessment of cardiac surgery in Britain.

Data are available for every Cardiac Surgery unit in Britain and in 70 % are identifiable by surgeon. The data are linked to registration of deaths so survival for a range of operations, and associated patient or procedure related factors, can be evaluated. The choice of statistical triggers (outside 99.99 % confidence intervals) and the time frames of reported data (averaged over three years) (See P.285/353 of the report http://www.scts.org/documents/PDF/5thBlueBook2003.pdf) reduces its value as an early warning system but the rigour of data collecting systems and the level of scrutiny required probably ensure that poor performance will be detected.

Monitoring the occurrence of wound infections after cardiac surgery.

The aim of this study was to demonstrate the use of a graphical method for real-time monitoring of the occurrence of surgical wound infection following cardiac surgery. This included developing and incorporating a risk scoring system so that variations in case-mix could be duly accounted for in the monitoring process. We analysed routinely collected data from a London teaching hospital. These data consisted of records for 2146 patients who had undergone cardiac surgery between April 2000 and March 2004 and whose surgical wounds were followed up as part of the local surveillance programme. The risk model was developed using logistic regression analysis with surgical wound infection diagnosed before hospital discharge as the outcome measure. Factors included in the model were the number of surgical wounds, patient age, operations that combined bypass surgery and valve replacement, renal disease and the number of days between hospital admission and surgery. The model was a good predictor of outcomes recorded within an independent data set (Chi-squared=3.81, P=0.58) and we incorporated it into a graphical tool for monitoring outcomes. The risk model and the associated graphical monitoring method could be valuable tools to assist with infection management. If used in real-time, problems with the care process can be quickly identified allowing timely remedial action to be taken.

Estimating the long-term impact of a prophylactic human papillomavirus 16/18 vaccine on the burden of cervical cancer in the UK.

To predict the public health impact on cervical disease by introducing human papillomavirus (HPV) vaccination in the United Kingdom, we developed a mathematical model that can be used to reflect the impact of vaccination in different countries with existing screening programmes. Its use is discussed in the context of the United Kingdom. The model was calibrated with published data. The impact of vaccination on cervical cancer and deaths, precancerous lesions and screening outcomes were estimated for a vaccinated cohort of 12-year-old girls, among which it is estimated that there would be a reduction of 66% in the prevalence of high-grade precancerous lesions and a 76% reduction in cervical cancer deaths. Estimates for various other measures of the population effects of vaccination are also presented. We concluded that it is feasible to forecast the potential effects of HPV vaccination in the context of an existing national screening programme. Results suggest a sizable reduction in the incidence of cervical cancer and related deaths. Areas for future research include investigation of the beneficial effects of HPV vaccination on infection transmission and epidemic dynamics, as well as HPV-related neoplasms in other sites.

Identification of systems failures in successful paediatric cardiac surgery.

Patient safety will benefit from an approach to human error that examines systemic causes, rather than blames individuals. This study describes a direct observation methodology, based on a threat and error model, prospectively to identify types and sources of systems failures in paediatric cardiac surgery. Of substantive interest were the range, frequency and types of failures that could be identified and whether minor failures could accumulate to form more serious events, as has been the case in other industries. Check lists, notes and video recordings were employed to observe 24 successful operations. A total of 366 failures were recorded. Coordination and communication problems, equipment problems, a relaxed safety culture, patient-related problems and perfusion-related problems were most frequent, with a smaller number of skill, knowledge and decision-making failures. Longer and more risky operations were likely to generate a greater number of minor failures than shorter and lower risk operations, and in seven higher-risk cases frequently occurring minor failures accumulated to threaten the safety of the patient. Non-technical errors were more prevalent than technical errors and task threats were the most prevalent systemic source of error. Adverse events in surgery are likely to be associated with a number of recurring and prospectively identifiable errors. These may be co-incident and cumulative human errors predisposed by threats embedded in the system, rather than due to individual incompetence or negligence. Prospectively identifying and reducing these recurrent failures would lead to improved surgical standards and enhanced patient safety.

Is hydrotherapy cost-effective? A randomised controlled trial of combined hydrotherapy programmes compared with physiotherapy land techniques in children with juvenile idiopathic arthritis.

OBJECTIVES: To compare the effects of combined hydrotherapy and land-based physiotherapy (combined) with land-based physiotherapy only (land) on cost, health-related quality of life (HRQoL) and outcome of disease in children with juvenile idiopathic arthritis (JIA). Also to determine the cost-effectiveness of combined hydrotherapy and land-based physiotherapy in JIA. DESIGN: A multicentre randomised controlled, partially blinded trial was designed with 100 patients in a control arm receiving land-based physiotherapy only (land group) and 100 patients in an intervention arm receiving a combination of hydrotherapy and land-based physiotherapy (combined group). SETTING: Three tertiary centres in the UK. PARTICIPANTS: Patients aged 4-19 years diagnosed more than 3 months with idiopathic arthritides, onset before their 16th birthday, stable on medication with at least one active joint. INTERVENTIONS: Patients in the combined and land groups received 16 1-hour treatment sessions over 2 weeks followed by local physiotherapy attendances for 2 months. MAIN OUTCOME MEASURES: Disease improvement defined as a decrease of > or =30% in any three of six core set variables without there being a 30% increase in more than one of the remaining three variables was used as the primary outcome measure and assessed at 2 months following completion of intervention. Health services resource use (in- and outpatient care, GP visits, drugs, interventions, and investigations) and productivity costs (parents' time away from paid work) were collected at 6 months follow-up. HRQoL was measured at baseline and 2 and 6 months following intervention using the EQ-5D, and quality-adjusted life-years (QALYs) were calculated. Secondary outcome measures at 2 and 6 months included cardiovascular fitness, pain, isometric muscle strength and patient satisfaction. RESULTS: Seventy-eight patients were recruited into the trial and received treatment. Two months after intervention 47% patients in the combined group and 61% patients in the land group had improved disease with 11 and 5% with worsened disease, respectively. The analysis showed no significant differences in mean costs and QALYs between the two groups. The combined group had slightly lower mean costs (-6.91 pounds Sterling) and lower mean QALYs (-0.0478, 95% confidence interval -0.11294 to 0.0163 based on 1000 bootstrap replications). All secondary measures demonstrated a mean improvement in both groups, with the combined group showing greater improvements in physical aspects of HRQoL and cardiovascular fitness. CONCLUSIONS: JIA is a disease in which a cure is not available. This research demonstrates a beneficial effect from both combined hydrotherapy and land-based physiotherapy treatment and land-based physiotherapy treatment alone in JIA without any exacerbation of disease, indicating that treatments are safe. The caveat to the results of the cost-effectiveness and clinical efficacy analysis is that the restricted sample size could have prevented a true difference being detected between the groups. Nevertheless, there appears to be no evidence to justify the costs of building pools or initiating new services specifically for use in this disease. However, this conclusion may not apply to patients with unremitting active disease who could not be entered into the trial because of specified exclusion criteria. For this group, hydrotherapy or combined treatment may still be the only physiotherapy option. Further research is suggested into: the investigation and development of appropriate and sensitive outcome measures for use in future hydrotherapy and physiotherapy trials of JIA; preliminary studies of methodologies in complex interventions such as physiotherapy and hydrotherapy to improve recruitment and ensure protocol is acceptable to patients and carers; hydrotherapy in the most common paediatric user group, children with neurological dysfunction, ensuring appropriate outcome measures are available and methodologies previously tried; patient satisfaction and compliance in land-based physiotherapy and hydrotherapy and European studies of hydrotherapy in rare disorders such as JIA.

Computer tools to assist the monitoring of outcomes in surgery.

OBJECTIVE: In recent years, there has been increasing use of analytical and graphical methods to assist the monitoring of outcomes in adult cardiac surgery. In this paper, we present extensions to the basic VLAD methodology that add flexibility and assist in its interpretation. METHODS: Using techniques from probability theory, we have devised graphical tools whereby deviations from expected outcomes can be monitored to see how likely they are to have occurred by chance. The methods are based upon pre-operative assessments of risk and use exact analytical techniques. RESULTS: These tools allow deviations from expected outcomes to be readily assessed and compared with the distribution of chance outcomes. Appropriate colour coding allows interpretation in terms of a temperature gradient. CONCLUSIONS: Exact analysis methods based on the use of pre-operative risk assessment provide a useful means for assisting the interpretation of VLAD charts. Such analysis has the advantage that it is applicable even for relatively short series of operations. Also, it takes specific account of the heterogeneity of case mix when quantifying the variability that is expected. By displaying the overall history of outcomes in a visually intuitive manner, it complements the more formal tools for detecting isolated good and bad runs that are available.

Cardiac ECMO for biventricular hearts after paediatric open heart surgery.

OBJECTIVE: To delineate predictors of hospital survival in a large series of children with biventricular physiology supported with extracorporeal membrane oxygenation (ECMO) after open heart surgery. RESULTS: 81 children were placed on ECMO after open heart surgery. 58% (47 of 81) were transferred directly from cardiopulmonary bypass to ECMO. Hospital survival was 49% (40 of 81) but there were seven late deaths among these survivors (18%). Factors that improved the odds of survival were initiation of ECMO in theatre (64% survival (30 of 47)) rather than the cardiac intensive care unit (29% survival (10 of 34)) and initiation of ECMO for reactive pulmonary hypertension. Important adverse factors for hospital survival were serious mechanical ECMO circuit problems, renal support, residual lesions, and duration of ECMO. CONCLUSIONS: Hospital survival of children with biventricular physiology who require cardiac ECMO is similar to that found in series that include univentricular hearts, suggesting that successful cardiac ECMO is critically dependent on the identification of hearts with reversible ventricular dysfunction. In our experience of postoperative cardiac ECMO, the higher survival of patients cannulated in the operating room than in the cardiac intensive care unit is due to early effective support preventing prolonged hypoperfusion and the avoidance of a catastrophic cardiac arrest.

A health impact assessment model for environmental changes attributable to development projects.

STUDY OBJECTIVE: European Union legislation requires large industrial and civil development projects to undergo environmental impact assessment. The study objective was to identify environmental health risk estimates for these developments from the epidemiological literature and to develop, and apply these within, a mathematical health impact assessment model. DESIGN AND RESULTS: In the UK, good practice guidelines have set out environmental issues to be considered in development projects, but little attention is given to direct health effects. Broad quantifiable risks were identified for four-air, chemicals, noise, and road traffic-of 14 standard environmental effects. A mathematical model was constructed that is based on people moving between different health states over their lifetime. Age related hazard functions are applied to cause specific measures of mortality and morbidity. A hypothetical example for a development creating air and chemical pollutants is given. CONCLUSIONS: A mathematical model applying epidemiological risks to an exposed population can provide quantification of environmental health effects. The approach may in future find application during project development, and by public health regulatory authorities for environmental health impact assessment.

Early identification of divergent performance in congenital cardiac surgery.

OBJECTIVES: Heterogeneous caseload and poorly quantified risk stratification make it difficult to monitor outcomes in congenital cardiac surgery. Reliance on formal statistical hypothesis testing may lead to substantial delays before a pattern of poor outcome can be established. Here, we report alternative methods for alerting surgeons to potential problems at an earlier stage. METHODS: Graphical methods developed for monitoring adult cardiac surgery have been adapted for use in congenital cardiac surgery. To illustrate their potential, we have retrospectively examined mortality data for a single surgeon involving 315 cases. Partial risk adjustment has been carried out according to patient's age and the open/closed categorization of the surgical procedure. Additional information has been derived by ranking procedures in order of their complexity and displaying the proportion of the surgeon's cases in each complexity stratum. RESULTS: The display of a surgeon's mortality data adjusted for age and open/closed category provides an easily understood chart of performance and allows one to identify periods when performance appears divergent, relative to the surgeon's own overall standards. Cases carried out during such periods can then be scrutinized by alternative methods. One such method is to examine caseload complexity during the periods of apparent divergent performance compared with other periods. CONCLUSIONS: These methods, while in no way representing formal statistical tests, provide a means that can alert surgeons to potential problems and help to identify sequences of cases that might benefit from further scrutiny.

Assessment of mortality rates for congenital heart defects and surgeons' performance.

BACKGROUND: In the absence of reliable national data, we have collected results of all operations for congenital heart defects from five departments to assess mortality rates and compare them among surgeons and departments. METHODS: Data relating to all operations (2,718) carried out at the five centers during a period from April 1, 1997 through March 31, 1999. Clearly defined criteria were agreed for the classification of patients into various subgroups. RESULTS: The overall hospital mortality was 4.4% (95% confidence intervals 3.7%-5.3%). Mortality for open-operations was 12.6% in neonates, 5.1% in infants, and 3.5% in children. Mortality rates were 1.1% for coarctation, 0.4% ventricular septal defect, 4.1% atrioventricular septal defect, 2.9% Fallot, 0.9% switch, and 15.6% truncus arteriosus. Although individual surgeons' mortality rates ranged from 1.8% to 7.5%, none of the 12 surgeons' data were above 95% confidence intervals. For individual surgeons, the change in mortality rates between the 2 years ranged between -3.3% and +3.8%. CONCLUSIONS: With 2 years' data available, estimates of mortality rates are more precise as reflected by tighter confidence intervals. There were relatively small data sets for individual hospitals and surgeons, which made statistical evaluation difficult. For setting standards, data from more departments for a longer period will be required. Statistical methods alone cannot be used as a sole arbiter of what is considered acceptable performance.

Computer modelling of a cataract waiting list.

AIM: To compare three different strategies for determining admission dates for patients awaiting cataract extraction after scoring for visual impairment. METHODS: 357 patients attending for assessment for cataract surgery were scored for visual impairment. These scores were used as a basis for ranking patients into three impairment strata. A computer simulation was used to compare 3 years' operation of different admission strategies-a first come first served booking system, a triage booking system, and a waiting list system in which admissions were strictly ordered according to priority stratum. Differences in priority weighted delays before treatment were analysed. RESULTS: Both the triage system and the priority based waiting list system gave considerable reduction in priority weighted delay compared with a first come first served admission policy. The percentage reduction achieved (30%-60%) is strongly influenced by the number of weeks fully booked when the booking systems are introduced. The priority weighted delay of the triage system, where booking decisions were made at the time of the outpatient assessment, was consistently and substantially outperformed by the priority based waiting list system where the decision to allocate an admission date was delayed as long as possible. CONCLUSIONS: There is considerable scope for reducing delays to high priority patients if simple rules are used to determine admission dates. Using these rules, booking patients at the time of the outpatient assessment gives substantially less benefit in terms of reducing delays to high priority patients than if the decision about the admission date is deferred.

Can simulation be used to reduce errors in health care delivery? The hospital drug distribution system.

OBJECTIVES: To construct a simulation model of the hospital drug distribution system, to evaluate the effects of different changes to the system on unavailability-related medication administration errors (U-MAEs), to test the most promising system in a controlled trial and to explore the model's validity. METHODS: A discrete-event simulation model of the drug distribution system was constructed, using data collected on a vascular surgery ward and a renal medicine ward as model inputs. The model's output was the U-MAE rate. The model was used to explore the effects on U-MAEs of different changes to the system. One of the changes predicted to reduce U-MAEs, a patients' own drugs scheme, was introduced on each study ward; U-MAE rates were measured using observations before and after its introduction and compared with those predicted by the model. RESULTS: The model predicted that the introduction of a patients' own drugs system would reduce unavailability-related errors on each ward; in practice, there was a slight decrease on the medical ward but an increase on the surgical ward. Reasons for these findings were explored and four contributing factors identified. Three of these related to failure to follow hospital procedures, the fourth to an issue for which no policy existed. If these factors had been taken into account, the error rates predicted by the model would have been similar to those observed. CONCLUSIONS: Simulation modelling is a potentially useful approach to the study of U-MAEs, although care must be taken to ensure that such models reflect actual practice rather than stated policy.

The planning of cervical cancer screening programmes in eastern Europe: is viral testing a suitable alternative to smear testing?

Cervical cancer screening with human papillomavirus (HPV) DNA testing has potential advantages over conventional, smear testing in that it can predict cases in which invasive cancers are more likely to develop, may be cheaper to implement and improve compliance. In areas of the world where little formalized cervical cancer screening takes place, or where health resources are limited, HPV testing has been suggested as a possible alternative for primary screening. In this paper we demonstrate the use of mathematical modelling to evaluate the effects of setting up screening programmes in Eastern Europe with HPV DNA testing as the primary screening tool and compare it with conventional smear testing. The impact of screening is measured in terms of the life years gained and the resulting resource usage and cost. We investigate several screening options with different screening intervals and age ranges for the target population.

The clinical, cerebrospinal fluid, and histopathologic effects of epidural ketorolac in dogs.

OBJECTIVE: To evaluate the clinical, cerebrospinal fluid (CSF), and histopathologic effects of epidural ketorolac. STUDY DESIGN: Blinded, randomized, placebo controlled study. ANIMALS: Twenty-two adult mixed breed dogs with 16 treatment and 6 control dogs, weighing 14.4 to 29.8 kg. METHODS: Dogs were anesthetized and epidural catheters were placed at the lumbosacral space. Catheter placement was evaluated fluoroscopically. Ketorolac (0.4 mg/kg) or placebo (5% ethanol) was administered epidurally over a 52-hour period, with 5 injections given at 12-hour intervals. At 1, 2, 4, or 8 hours after the first and last injection of ketorolac, dogs were anesthetized and CSF was obtained. Control dogs had CSF sampled 1 hour after the first and last ethanol injection. Neurologic function and pain responses were evaluated before and during the study. Selected dogs were then killed and necropsies performed. RESULTS: None of the dogs exhibited any clinical or neurologic abnormalities during the study. No statistical difference was noted in pain response or CSF analysis between treatment and control dogs. Gross necropsy revealed gastrointestinal ulceration of varying degrees in all treatment dogs. Histopathologic analysis of the spinal cord and meninges revealed minimal focal leptomeningeal phlebitis in 2 of 8 treatment dogs and minor subdural inflammation in 1 control dog. No changes to the neural structures were noted in any dogs. CONCLUSIONS: Epidural administration of ketorolac did not cause clinical signs, alteration in CSF values, or pathologic changes to the spinal cord when used for short duration. Gastrointestinal ulceration was common when ketorolac was administered epidurally at 0.4 mg/kg every 12 hours for 5 treatments. CLINICAL RELEVANCE: This study documented the neurologic safety of epidural ketorolac in dogs before an efficacy trial can be performed. Gastrointestinal ulceration may limit use to short duration or a single injection.

Quantitative detection of oncogenic HPV DNA using hybrid capture to triage borderline and mildly dyskaryotic Papanicolaou smears.

OBJECTIVE: To investigate the use of hybrid capture to triage borderline and mildly abnormal PAP smears to either colposcopy or continued cytological surveillance. METHOD: A prospective cross-sectional study of 207 women with either mild or borderline dyskaryosis referred to the colposcopy clinic at the Whittington Hospital, London was done. Quantitative detection of oncogenic human papillomavirus deoxyribose nucleic acid was performed using the solution hybridization technique Hybrid Capture I (Digene Diagnostics). Positive or negative hybrid capture results were compared to colposcopically directed punch biopsy or loop excision of the transformation zone. RESULTS: Hybrid capture correctly predicted 1/1 micro-invasive cancers, 23/24 CIN 3 and 20/24 CIN 2, giving a sensitivity to detect significant disease of CIN 2 or above of 90%. For CIN 1, 37/66 were HPV positive making the sensitivity to detect all grades of CIN 70%. Of women with a normal cervix, 39/92 were HPV positive and in total 58% of the women were HPV positive and would have been referred to colposcopy. CONCLUSIONS: Hybrid capture was a simple solution hybridization method which provided a sensitive method of detecting CIN 2 and above lesions in women with minor PAP smear abnormalities. However, the sensitivity for CIN 1 and the specificity were low, meaning that further improvements in the technique are needed before it can be used as a triage strategy.

Cervicography to triage women with borderline or mild dyskaryotic cervical PAP smears.

OBJECTIVE: To evaluate six years experience using cervicography to triage women with borderline or mildly dyskaryotic PAP smears to either immediate colposcopy or cytological surveillance. METHOD: Retrospective study of 1,436 women referred to the cervicography clinic of the Whittington Hospital, London. Women with suspicious cervigrams were examined colposcopically and biopsied. Women with normal appearances at cervicography were followed-up with cytology over a period of a least two years. RESULTS: Cervicography detected 281 out of 307 women with CIN 2/3 and 215 out of 264 women with CIN 1. The sensitivity of cervicography to detect CIN 1, 2 or 3 in women with minor PAP smear abnormalities was 92%. The specificity was 39% and 847 (59%) of the women were referred for colposcopy; 215 (15%) were lost to follow-up. CONCLUSIONS: Cervicography was a sensitive method to detect CIN in women with mild or borderline dyskaryosis on their PAP smears. However, the low specificity meant that a high proportion of the women were referred for colposcopy, and a significant proportion of women were lost to follow-up. Thus cervicography is not an efficient strategy for managing women with minor PAP smears abnormalities.

How does functional disability in early rheumatoid arthritis (RA) affect patients and their lives? Results of 5 years of follow-up in 732 patients from the Early RA Study (ERAS)

OBJECTIVES: To assess the impact of rheumatoid arthritis (RA) on function and how this affects major aspects of patients' lives. METHODS: The inception cohort of RA patients was recruited from rheumatology out-patient departments in nine National Health Service (NHS) hospital trusts in England. All consecutive patients with RA of less than 2 yr duration, prior to any second-line (disease-modifying) drug treatment were recruited and followed-up for 5 yr. Standard clinical, laboratory and radiological assessments, and all hospital-based interventions were recorded prospectively at presentation and yearly. The outcome measures were clinical remission and extra-articular features, functional ability [functional grades I-IV and Health Assessment Questionnaire (HAQ)], use of aids, appliances and home adaptations, orthopaedic interventions, and loss of paid work. RESULTS: A total of 732 patients completed 5 yr of follow-up, of whom 84% received second-line drugs. Sixty-nine (9.4%) had marked functional loss at presentation, compared with normal function in 243 (33%), and by 5 yr these numbers had increased in each group, respectively, to 113 (16%) and 296 (40%). Home adaptations and/or wheelchair use by 5 yr were seen in 74 (10%). Work disability was seen in 27% of those in paid employment at onset. One hundred and seventeen (17%) patients underwent orthopaedic surgery for RA, 55 (8%) for major joint replacements. Marked functional loss at 5 yr was more likely in women [odds ratio (OR) 1.63, 95% confidence interval (CI) 1.04-2.5], patients older than 60 yr (OR 1.94, 95% CI 1.3-2.9), and with HAQ > 1.0 at presentation (OR 4.4, 95% CI 2.8-7.0). CONCLUSIONS: Clinical profiles of RA patients treated with conventional drug therapy over 5 yr showed that a small proportion of patients (around 16%) do badly functionally and in terms of life events, whereas around 40% do relatively well. The details and exact figures of cumulative disability are likely to be useful to clinicians, health professionals and patients. The rate of progression and outcome in these patients can be compared against future therapies with any disease-modifying claims.

Likely variations in perioperative mortality associated with cardiac surgery: when does high mortality reflect bad practice?

OBJECTIVE: Several methods exist for estimating the risk of perioperative mortality based on preoperative risk factors; graphical methods such as the variable life adjusted display (VLAD) can be used to examine how an individual surgeon's performance for a series of operations fares against what would be expected, given the case mix. This study aimed to devise a method for assessing the natural variation in outcome in order to assist with making judgements about individual performance, in particular whether seemingly poor performance could have occurred by chance. METHOD: The risk scoring system has been derived and validated locally for cardiac surgery. A method is described for calculating the probability that an observed number of deaths occurs within a sequence of operations if perioperative mortality is regarded as a chance event with an expected value derived from the risk score. To illustrate this method, nested prediction intervals are superimposed onto VLAD plots for a series of 393 isolated coronary artery bypass and isolated valve operations performed by a single surgeon. RESULTS: Using the locally derived risk score, the VLAD plot for the individual surgeon shows a net life gain of about 6 over the predicted number of survivors, which is observed to be within the 90% prediction interval. If the Parsonnet scoring system is used instead of the locally derived risk score, the net life gain is considerably overestimated. CONCLUSIONS: The nested prediction intervals are straightforward to generate and can be integrated into a visually informative display. As an indication of the inherent variability in outcome, they have a valuable role in the monitoring of surgical performance.

Mortality rates after surgery for congenital heart defects in children and surgeons' performance.

BACKGROUND: A public inquiry into surgery for paediatric congenital heart defects in Bristol, UK, underscored the need for reliable data on overall mortality rates, which would allow assessment of individual surgeons' performance. We aimed to gather and report such data for 1 year to provide information for clinicians, researchers, policy makers, and the general public. METHODS: We collected data on all operations (1378) for congenital heart defects done by 11 surgeons in five departments in the UK between April 1, 1997, and March 31, 1998. These operations represented about 36% of all operations done in the UK during that time. Clearly defined criteria were agreed to classify operations into subgroups. FINDINGS: The overall mortality rate for all operations was 4.0% (95% CI 3.0-5.2). No deaths occurred for 67 arterial-switch operations. Mortality rates for coarctation, ventricular septal defect, atrioventricular septal defect, Fallot, and truncus arteriosus operations were 1.1%, 0.6%, 3.6%, 2.3%, and 28.6%, respectively. Although overall mortality rates between surgeons varied (1.6-6.9%), no surgeon's were higher than the 95% CI. The numbers of operations done by individual surgeons were small, which led to wide confidence intervals and made the detection of differences in performance difficult. INTERPRETATION: The participating departments seemed to reach high standards of care for children with congenital heart defects, although more data would be needed to assess performance of individual surgeons. The development of quality standards will be difficult because of the complexity of defects, the different types of operations, and few patients in each subgroup. Collection of larger sets of data for more patients and centres are needed.