Rosy Tsopra

Burden of cardiovascular disease across 29 countries and GPs’ decision to treat hypertension in oldest-old

Abstract Objectives: We previously found large variations in general practitioner (GP) hypertension treatment probability in oldest-old (>80 years) between countries. We wanted to explore whether differences in country-specific cardiovascular disease (CVD) burden and life expectancy could explain the differences. Design: This is a survey study using case-vignettes of oldest-old patients with different comorbidities and blood pressure levels. An ecological multilevel model analysis was performed. Setting: GP respondents from European General Practice Research Network (EGPRN) countries, Brazil and New Zeeland. Subjects: This study included 2543 GPs from 29 countries. Main outcome measures: GP treatment probability to start or not start antihypertensive treatment based on responses to case-vignettes; either low (<50% started treatment) or high (≥50% started treatment). CVD burden is defined as ratio of disability-adjusted life years (DALYs) lost due to ischemic heart disease and/or stroke and total DALYs lost per country; life expectancy at age 60 and prevalence of oldest-old per country. Results: Of 1947 GPs (76%) responding to all vignettes, 787 (40%) scored high treatment probability and 1160 (60%) scored low. GPs in high CVD burden countries had higher odds of treatment probability (OR 3.70; 95% confidence interval (CI) 3.00–4.57); in countries with low life expectancy at 60, CVD was associated with high treatment probability (OR 2.18, 95% CI 1.12–4.25); but not in countries with high life expectancy (OR 1.06, 95% CI 0.56–1.98). Conclusions: GPs’ choice to treat/not treat hypertension in oldest-old was explained by differences in country-specific health characteristics. GPs in countries with high CVD burden and low life expectancy at age 60 were most likely to treat hypertension in oldest-old. Key Points  • General practitioners (GPs) are in a clinical dilemma when deciding whether (or not) to treat hypertension in the oldest-old (>80 years of age).  • In this study including 1947 GPs from 29 countries, we found that a high country-specific cardiovascular disease (CVD) burden (i.e. myocardial infarction and/or stroke) was associated with a higher GP treatment probability in patients aged >80 years.  • However, the association was modified by country-specific life expectancy at age 60. While there was a positive association for GPs in countries with a low life expectancy at age 60, there was no association in countries with a high life expectancy at age 60.  • These findings help explaining some of the large variation seen in the decision as to whether or not to treat hypertension in the oldest-old.

The impact of three discharge coding methods on the accuracy of diagnostic coding and hospital reimbursement for inpatient medical care

BACKGROUND Coding of diagnoses is important for patient care, hospital management and research. However coding accuracy is often poor and may reflect methods of coding. This study investigates the impact of three alternative coding methods on the inaccuracy of diagnosis codes and hospital reimbursement. METHODS Comparisons of coding inaccuracy were made between a list of coded diagnoses obtained by a coder using (i)the discharge summary alone, (ii)case notes and discharge summary, and (iii)discharge summary with the addition of medical input. For each method, inaccuracy was determined for the primary, secondary diagnoses, Healthcare Resource Group (HRG) and estimated hospital reimbursement. These data were then compared with a gold standard derived by a consultant and coder. RESULTS 107 consecutive patient discharges were analysed. Inaccuracy of diagnosis codes was highest when a coder used the discharge summary alone, and decreased significantly when the coder used the case notes (70% vs 58% respectively, p < 0.0001) or coded from the discharge summary with medical support (70% vs 60% respectively, p < 0.0001). When compared with the gold standard, the percentage of incorrect HRGs was 42% for discharge summary alone, 31% for coding with case notes, and 35% for coding with medical support. The three coding methods resulted in an annual estimated loss of hospital remuneration of between £1.8 M and £16.5 M. CONCLUSION The accuracy of diagnosis codes and percentage of correct HRGs improved when coders used either case notes or medical support in addition to the discharge summary. Further emphasis needs to be placed on improving the standard of information recorded in discharge summaries.

Using preference learning for detecting inconsistencies in clinical practice guidelines: Methods and application to antibiotherapy

Clinical practice guidelines provide evidence-based recommendations. However, many problems are reported, such as contradictions and inconsistencies. For example, guidelines recommend sulfamethoxazole/trimethoprim in child sinusitis, but they also state that there is a high bacteria resistance in this context. In this paper, we propose a method for the semi-automatic detection of inconsistencies in guidelines using preference learning, and we apply this method to antibiotherapy in primary care. The preference model was learned from the recommendations and from a knowledge base describing the domain. We successfully built a generic model suitable for all infectious diseases and patient profiles. This model includes both preferences and necessary features. It allowed the detection of 106 candidate inconsistencies which were analyzed by a medical expert. 55 inconsistencies were validated. We showed that therapeutic strategies of guidelines in antibiotherapy can be formalized by a preference model. In conclusion, we proposed an original approach, based on preferences, for modeling clinical guidelines. This model could be used in future clinical decision support systems for helping physicians to prescribe antibiotics.

Translating Visually the Reasoning of a Perceptron: The Weighted Rainbow Boxes Technique and an Application in Antibiotherapy

In this paper, we propose a technique for translating visually the reasoning of a perceptron. The artificial neuron, or perceptron, is a simplified model of a biological neuron. It can achieve simple reasoning and solve linearly separable problems. Despite its limited reasoning power, it is enough to deal with several real-life problems. The proposed technique is based on rainbow boxes, a technique for overlapping set visualization, which has been applied to the input vectors of the perceptron. We extended this technique, leading to weighted rainbow boxes. It can visualize several input vectors and output values for a single perceptron. We applied this approach to decision support in antibiotherapy, for the determination of the most appropriate antibiotic in urinary infections, by taking into account the properties of each drug (e.g. efficacy, risk of adverse effects, etc). Finally, a user study with 11 physicians showed that most of them found the visualization interesting and easy to read.

Conception d’une interface pour favoriser la prise en compte des recommandations lors de la décision d’une antibiothérapie probabiliste

Background: Increasing the use of available guidelines should improve antibiotics prescription appropriateness. The aim of this work is to develop an original interface to deliver in a concise way access to the guidelines and their decision elements. Methods: We manually analysed all the clinical guidelines available for ambulatory treatment of infections and identified their information content. Information elements were then organised into a conceptual model. According to some ergonomic principles, an interface was defined to support guideline presentation for rapid decision. Results: Analysis of the guidelines shows that decision depends on etiological and therapeutical criteria. Various treatment decisions are observed (no antibiotherapy, prescription of antibiotics, monitoring, microbiological sampling, hospitalization). The interface consists of six fixed parts: a decision table, two information zones, a zone with the reasons for hospitalization, a zone with diseases that are not described in the guidelines, and a zone with access to the original guidelines. Limited space use is achieved by the use of concise phrasing, VCM icons and Mister VCM. Conclusion: All the clinical guidelines are implemented in the interface. This interface will be implemented in the Antibiocarte website.