Objective To provide an overview of drug use in children in three European countries.Design Retrospective cohort study, 2000-5.Setting Primary care research databases in the Netherlands (IPCI), United Kingdom (IMS-DA), and Italy (Pedianet).Participants 675 868 children aged up to 14 (Italy) or 18 (UK and Netherlands).Main outcome measure Prevalence of use per year calculated by drug class (anatomical and therapeutic). Prevalence of “recurrent/chronic” use (three or more prescriptions a year) and “non-recurrent” or “acute” use (less than three prescriptions a year) within each therapeutic class. Descriptions of the top five most commonly used drugs evaluated for off label status within each anatomical class.Results Three levels of drug use could be distinguished in the study population: high (>10/100 children per year), moderate (1-10/100 children per year), and low (<1/100 children per year). For all age categories, anti-infective, dermatological, and respiratory drugs were in the high use group, whereas cardiovascular and antineoplastic drugs were always in the low use group. Emollients, topical steroids, and asthma drugs had the highest prevalence of recurrent use, but relative use of low prevalence drugs was more often recurrent than acute. In the top five highest prevalence drugs topical inhaled and systemic steroids, oral contraceptives, and topical or systemic antifungal drugs were most commonly used off label.Conclusion This overview of outpatient paediatric prescription patterns in a large European population could provide information to prioritise paediatric therapeutic research needs.

The use of either atypical or typical antipsychotic drugs in elderly patients is associated in a dose-dependent manner with risk for community-acquired pneumonia.

OBJECTIVETo compare, in patients with non‐muscle‐invasive low‐grade (pTa/pT1, G1/G2) urothelial cell carcinoma of the urinary bladder, the perceived burden of flexible cystoscopy or surveillance by microsatellite analysis (MA) in voided urine, as such patients are normally recommended to adhere to regular cysto‐urethroscopic surveillance (CUS).PATIENTS AND METHODSIn all, 220 participants of a randomized trial comparing CUS and surveillance by MA were asked to complete questionnaires 1 week after cystoscopy or urine sample collection. We assessed the discomfort and pain reported during CUS, experiences with MA, and physical symptoms, medical consumption and general functioning in the week after CUS/urine sampling.RESULTSWe analysed data from 732 questionnaires (197 patients) completed after CUS and 184 (67 patients) after collecting urine. The introduction of the cystoscope was reported to cause discomfort in 39% and pain in 35% of the responses to the questionnaires; the waiting time for the results of MA was reported as burdensome in 19%. Painful micturition was significantly more frequent in the week after CUS than after MA (30% and 12%, respectively). The frequency of fever (1% and 2%) and haematuria (7% and 6%) was similar in both groups. Older patients reported significantly less pain and discomfort from cystoscopy, and this was not related to having more previous cystoscopies.CONCLUSIONCUS caused pain and discomfort in about a third of patients. The burden of MA appeared fully attributable to the waiting time for the test result. The present results are a further motivation in the search for less invasive surveillance tests.

Asthma drugs are amongst the most frequently used drugs in childhood, but international comparisons on type and indication of use are lacking. The aim of this study was to describe asthma drug use in children with and without asthma in the Netherlands (NL), Italy (IT), and the United Kingdom (UK). We conducted a retrospective analysis of outpatient medical records of children 0–18 years from 1 January 2000 until 31 December 2005. For all children, prescription rates of asthma drugs were studied by country, age, asthma diagnosis, and off-label status. One-year prevalence rates were calculated per 100 children per patient-year (PY). The cohort consisted of 671,831 children of whom 49,442 had been diagnosed with asthma at any time during follow-up. ß2-mimetics and inhaled steroids were the most frequently prescribed asthma drug classes in NL (4.9 and 4.1/100 PY), the UK (8.7 and 5.3/100 PY) and IT (7.2 and 16.2/100 PY), respectively. Xanthines, anticholinergics, leukotriene receptor antagonists, and anti-allergics were prescribed in less than one child per 100 per year. In patients without asthma, ß2-mimetics were used most frequently. Country differences were highest for steroids, (Italy highest), and for ß2-mimetics (the UK highest). Off-label use was low, and most pronounced for ß2-mimetics in children <18 months (IT) and combined ß2-mimetics + anticholinergics in children <6 years (NL). Conclusion: This study shows that among all asthma drugs, ß2-mimetics and inhaled steroids are most often used, also in children without asthma, and with large variability between countries. Linking multi-country databases allows us to study country specific pediatric drug use in a systematic manner without being hampered by methodological differences. This study underlines the potency of healthcare databases in rapidly providing data on pediatric drug use and possibly safety.

BackgroundDistinguishing cases from non-cases in free-text electronic medical records is an important initial step in observational epidemiological studies, but manual record validation is time-consuming and cumbersome. We compared different approaches to develop an automatic case identification system with high sensitivity to assist manual annotators.MethodsWe used four different machine-learning algorithms to build case identification systems for two data sets, one comprising hepatobiliary disease patients, the other acute renal failure patients. To improve the sensitivity of the systems, we varied the imbalance ratio between positive cases and negative cases using under- and over-sampling techniques, and applied cost-sensitive learning with various misclassification costs.ResultsFor the hepatobiliary data set, we obtained a high sensitivity of 0.95 (on a par with manual annotators, as compared to 0.91 for a baseline classifier) with specificity 0.56. For the acute renal failure data set, sensitivity increased from 0.69 to 0.89, with specificity 0.59. Performance differences between the various machine-learning algorithms were not large. Classifiers performed best when trained on data sets with imbalance ratio below 10.ConclusionsWe were able to achieve high sensitivity with moderate specificity for automatic case identification on two data sets of electronic medical records. Such a high-sensitive case identification system can be used as a pre-filter to significantly reduce the burden of manual record validation.