Standardized mortality ratios for regionalized acute cardiovascular care.

BACKGROUND: Standardized Mortality Ratios (SMRs) are case-mix adjusted mortality rates per hospital and are used to evaluate quality of care. However, acute care is increasingly organized on a regional level, with more severe patients admitted to specialized hospitals. We hypothesize that the current case-mix adjustment insufficiently captures differences in case-mix between non-specialized and specialized hospitals. We aim to improve the SMR by adding proxies of disease severity to the model and by calculating a regional SMR (RSMR) for acute cerebrovascular disease (CVD) and myocardial infarction (MI). METHODS: We used data from the Dutch National Basic Registration of Hospital Care. We selected all admissions from 2016 to 2018. SMRs and RSMRs were calculated by dividing the observed in-hospital mortality by the expected in-hospital mortality. The expected in-hospital mortality was calculated using logistic regression with adjustment for age, sex, socioeconomic status, severity of main diagnosis, urgency of admission, Charlson comorbidity index, place of residence before admission, month/year of admission, and in-hospital mortality as outcome. RESULTS: The IQR of hospital SMRs of CVD was 0.85-1.10, median 0.94, with higher SMRs for specialized hospitals (median 1.12, IQR 1.00-1.28, 71%-SMR > 1) than for non-specialized hospitals (median 0.92, IQR 0.82-1.07, 32%-SMR > 1). The IQR of RSMRs was 0.92-1.09, median 1.00. The IQR of hospital SMRs of MI was 0.76-1.14, median 0.98, with higher SMRs for specialized hospitals (median 1.00, IQR 0.89-1.25, 50%-SMR > 1 versus median 0.94, IQR 0.74-1.11, 44%-SMR > 1). The IQR of RSMRs was 0.90-1.08, median 1.00. Adjustment for proxies of disease severity mostly led to lower SMRs of specialized hospitals. CONCLUSION: SMRs of acute regionally organized diseases do not only measure differences in quality of care between hospitals, but merely measure differences in case-mix between hospitals. Although the addition of proxies of disease severity improves the model to calculate SMRs, real disease severity scores would be preferred. However, such scores are not available in administrative data. As a consequence, the usefulness of the current SMR as quality indicator is very limited. RSMRs are potentially more useful, since they fit regional organization and might be a more valid representation of quality of care.

Improving the quality of EHR recording in primary care: a data quality feedback tool.

OBJECTIVE: Electronic health record (EHR) data are used to exchange information among health care providers. For this purpose, the quality of the data is essential. We developed a data quality feedback tool that evaluates differences in EHR data quality among practices and software packages as part of a larger intervention. METHODS: The tool was applied in 92 practices in the Netherlands using different software packages. Practices received data quality feedback in 2010 and 2012. RESULTS: We observed large differences in the quality of recording. For example, the percentage of episodes of care that had a meaningful diagnostic code ranged from 30% to 100%. Differences were highly related to the software package. A year after the first measurement, the quality of recording had improved significantly and differences decreased, with 67% of the physicians indicating that they had actively changed their recording habits based on the results of the first measurement. About 80% found the feedback helpful in pinpointing recording problems. One of the software vendors made changes in functionality as a result of the feedback. CONCLUSIONS: Our EHR data quality feedback tool is capable of highlighting differences among practices and software packages. As such, it also stimulates improvements. As substantial variability in recording is related to the software package, our study strengthens the evidence that data quality can be improved substantially by standardizing the functionalities of EHR software packages.

Expected number of asbestos-related lung cancers in the Netherlands in the next two decades: a comparison of methods.

OBJECTIVES: Exposure to asbestos fibres increases the risk of mesothelioma and lung cancer. Although the vast majority of mesothelioma cases are caused by asbestos exposure, the number of asbestos-related lung cancers is less clear. This number cannot be determined directly as lung cancer causes are not clinically distinguishable but may be estimated using varying modelling methods. METHODS: We applied three different modelling methods to the Dutch population supplemented with uncertainty ranges (UR) due to uncertainty in model input values. The first method estimated asbestos-related lung cancer cases directly from observed and predicted mesothelioma cases in an age-period-cohort analysis. The second method used evidence on the fraction of lung cancer cases attributable (population attributable risk (PAR)) to asbestos exposure. The third method incorporated risk estimates and population exposure estimates to perform a life table analysis. RESULTS: The three methods varied substantially in incorporated evidence. Moreover, the estimated number of asbestos-related lung cancer cases in the Netherlands between 2011 and 2030 depended crucially on the actual method applied, as the mesothelioma method predicts 17 500 expected cases (UR 7000-57 000), the PAR method predicts 12 150 cases (UR 6700-19 000), and the life table analysis predicts 6800 cases (UR 6800-33 850). CONCLUSIONS: The three different methods described resulted in absolute estimates varying by a factor of ∼2.5. These results show that accurate estimation of the impact of asbestos exposure on the lung cancer burden remains a challenge.

Legal claims for malignant mesothelioma: dealing with all cases.

BACKGROUND: Apart of medical reasons, a definitive diagnosis of malignant mesothelioma may be required as a basis for a claim of financial compensation although a pathological source of conclusive evidence is missing. Clinical assessment of all available data is then the only option to come to a final conclusion. We evaluated the diagnostic work-up of a large cohort of Dutch patients who applied for financial compensation due to mesothelioma. We determined how often a pathological or clinical diagnosis can be made, and which factors are associated with making the final diagnosis malignant mesothelioma. METHODS: A flow diagram of the diagnostic work-up was constructed for patients that applied to the Dutch institute for asbestos victims between 2005 and 2008 (N=1498). Both pathological and clinical factors that may influence the diagnostic outcome were assessed. RESULTS: In 97 of the 1498 patients (6%) no pathologic diagnosis could be established because of an uncertain diagnosis (N=54), inadequate (N=22) or unavailable tumor samples (N=21). A final pathological diagnosis of malignant mesothelioma could most often be made when biopsy samples were available compared to those in whom only cytological material was available. In patients in who no conclusive diagnosis could be made, clinical assessment was performed. Eighty percent of patients (66/83) who were clinically assessed were considered to have mesothelioma. None of the clinical features analyzed were strongly associated with a confirmed diagnosis of malignant mesothelioma. DISCUSSION: Our study shows that only in a small number of the patients who applied no pathologic diagnosis could be obtained. Based on judgment of clinical experts in the majority of these cases a near to certain diagnosis could be made. Moreover, it is reasonable to obtain biopsy material from patients to increase the chance to obtain a confirmed diagnosis. Therefore, it is important to refer patients early for diagnostic procedures.

Lung cancer risk at low cumulative asbestos exposure: meta-regression of the exposure-response relationship.

PURPOSE: Existing estimated lung cancer risks per unit of asbestos exposure are mainly based on, and applicable to, high exposure levels. To assess the risk at low cumulative asbestos exposure, we provide new evidence by fitting flexible meta-regression models, a notably new and more robust method. METHODS: Studies were selected if lung cancer risk per cumulative asbestos exposure in at least two exposure categories was reported. From these studies (n = 19), we extracted 104 risk estimates over a cumulative exposure range of 0.11-4,710 f-y/ml. We fitted linear and natural spline meta-regression models to these risk estimates. A natural spline allows risks to vary nonlinearly with exposure, such that estimates at low exposure are less affected by estimates in the upper exposure categories. Associated relative risks (RRs) were calculated for several low cumulative asbestos exposures. RESULTS: A natural spline model fitted our data best. With this model, the relative lung cancer risk for cumulative exposure levels of 4 and 40 f-y/ml was estimated between 1.013 and 1.027, and 1.13 and 1.30, respectively. After stratification by fiber type, a non-significant three- to fourfold difference in RRs between chrysotile and amphibole fibers was found for exposures below 40 f-y/ml. Fiber-type-specific risk estimates were strongly influenced by a few studies. CONCLUSIONS: The natural spline regression model indicates that at lower asbestos exposure levels, the increase in RR of lung cancer due to asbestos exposure may be larger than expected from previous meta-analyses. Observed potency differences between different fiber types are lower than the generally held consensus. Low-exposed industrial or population-based cohorts with quantitative estimates of asbestos exposure a required to substantiate the risk estimates at low exposure levels from our new, flexible meta-regression.

Effect of non-persistent use of oral glucose-lowering drugs on HbA1c goal attainment.

OBJECTIVES: The aim of this study was to quantify the effect of non-persistence with oral glucose-lowering drugs (OGLD) on HbA(1c) goal attainment (<7%) in daily practice. METHODS: From the PHARMO Record Linkage System comprising among others linked drug dispensing and clinical laboratory data from approximately 2.5 million individuals in the Netherlands, new users of OGLD in the period 1999-2004 were identified. Patients with a baseline HbA(1c) > or =7% and at least one HbA(1c) measurement in the period of 6-12 months after treatment onset were included in the study cohort. Persistence with OGLD in the first year of treatment was determined using the method of Catalan. In case the first treatment episode overlapped the first HbA(1c) measurement within 6-12 months after treatment onset, a patient was considered persistent at that measurement. Patients with a HbA(1c) <7% were defined as having attained goal. RESULTS: The study cohort included 2023 patients with a mean baseline HbA(1c) of 8.9 +/- 1.8%. Three-quarters (1512 patients) were persistent with any OGLD at the first HbA(1c) measurement within 6-12 months after treatment onset; of these, 861 (57%) were at goal. Of the 511 non-persistent patients, 239 (47%) were at goal. Non-persistent patients were about 20% less likely to attain goal (RRadj 0.82; 95%CI 0.74-0.91), compared to persistent OGLD users. CONCLUSION: Non-persistent use of OGLD leads to a 20% decreased probability of HbA(1c) goal attainment in daily practice. This effect of non-persistence seems modest, but represents around 12 000 new and 10 000 prevalent OGLD users a year in the Netherlands in whom OGLD use could be better controlled.

Bayes' theorem: a negative example of a RCT on grommets in children with glue ear.

Bayesian inference presupposes that practitioners' belief in the effectiveness of medical intervention is the product of prior belief and recent evidence from studies. Although increasingly used, up to now the posterior belief calculated according to the theorem has not been compared with an empirically measured posterior belief. We conducted a RCT, which was preceded by elicitation of prior beliefs among ENT-surgeons, and which was followed by elicitation of posterior beliefs among ENT-surgeons, 1 year after completion of the trial. We compared the posterior beliefs of ENT-surgeons about the effect of grommets in children with glue ears, as predicted by Bayes' theorem with actual measured posterior beliefs. The distribution of the measured posterior beliefs was not in line with the calculated posterior, but almost identical to the distribution of the measured prior beliefs. The results showed that our trial had little or no impact on the beliefs of the ENT-surgeons, i.e. they did not adjust their belief to the extent that was expected according to Bayes' theorem.

Policy relevance of Bayesian statistics overestimated?

OBJECTIVES: The observed posterior probability distributions regarding the benefits of surgery for otitis media with effusion (OME) with expected probability distributions, using Bayes' theorem are compared. METHODS: Postal questionnaires were used to assess prior and posterior probability distributions among ear-nose-throat (ENT) surgeons in the Netherlands. RESULTS: In their prior probability estimates, ENT surgeons were quite optimistic with respect to the effectiveness of tube insertion in the treatment of OME. The trial showed no meaningful benefit of tubes on hearing and language development. Posterior probabilities calculated on the basis of prior probability estimates and trial results differed widely from those, elicited empirically 1 year after completion of the trial and dissemination of the results. CONCLUSIONS: ENT surgeons did not adjust their opinion about the benefits of surgical treatment of glue ears to the extent that they should have done according to Bayes' theorem. Users of the results of Bayesian analyses, notably policy-makers, should realize that Bayes' theorem is prescriptive and not necessarily descriptively correct. Health policy decisions should not be based on the untested assumption that health-care professionals use new evidence to adjust their subjective beliefs in a Bayesian manner.