Th J M Verheij

Confounding by indication in non-experimental evaluation of vaccine effectiveness: the example of prevention of influenza complications

Randomised allocation of vaccine or placebo is the preferred method to assess the effects of the vaccine on clinical outcomes relevant to the individual patient. In the absence of phase 3 trials using clinical end points, notably post-influenza complications, alternative non-experimental designs to evaluate vaccine effects or safety are often used. The application of these designs may, however, lead to invalid estimates of vaccine effectiveness or safety. As patients with poor prognosis are more likely to be immunised, selection for vaccination is confounded by patient factors that are also related to clinical end points. This paper describes several design and analytical methods aimed at limiting or preventing this confounding by indication in non-experimental studies. In short, comparison of study groups with similar prognosis, restriction of the study population, and statistical adjustment for dissimilarities in prognosis are important tools and should be considered. Only if the investigator is able to show that confounding by indication is sufficiently controlled for, results of a non-experimental study may be of use to direct an evidence based vaccine policy.

The Wheezing Illnesses Study Leidsche Rijn (WHISTLER): Rationale and design

The Wheezing Illnesses Study Leidsche Rijn (WHISTLER) was initiated in December 2001 as a single-centre prospective birth cohort study and will include a population-based sample of at least 2000 healthy newborns. The aims of WHISTLER are to investigate determinants for wheezing illnesses (including neonatal lung function, viral infections, asthma-susceptibility genes and endotoxin exposure) and to derive a comprehensive risk score, that is appropriate for use in primary health care and allows for efficient planning of early preventive strategies. Baseline examination includes a questionnaire evaluating known risk factors for wheezing illnesses; anthropometric measurements; measurements of infant and parental lung function; and sampling of infant and parental DNA. Participants will be followed for respiratory events using data from a daily respiratory symptom questionnaire; visits to the general practitioner (primary health care visits, drugs prescriptions and hospital referral); viral sampling during wheezing episodes; and house dust sampling. Based on actual neonatal care practice and embedded in a larger epidemiological study, the Utrecht Health Project, WHISTLER will provide an unique framework to address issues in childhood respiratory disease that are currently insufficiently understood. In particular, WHISTLER will provide a well-balanced view on the prognostic power of neonatal lung function and genetic and environmental factors (including viral infections and endotoxin exposure) to predict wheezing illnesses from birth to young adulthood and beyond. In the scope of prevention, WHISTLER is expected to support the design of solid based prevention measures to reduce respiratory morbidity, mortality and associated costs, and to improve quality of life.

Which determinants should be targeted to increase influenza vaccination uptake among health care workers in nursing homes

Although health care workers (HCWs) have been recommended to be immunized against influenza, vaccine uptake remains low. So far, research on determinants of influenza vaccination among HCWs has been limited by design, population or theoretical framework. Therefore we conducted a questionnaire study in Dutch nursing homes to assess which demographical, behavioural and organisational determinants were associated with influenza vaccine uptake among HCWs. We were able to accurately predict vaccine uptake based on a 13-item prediction model including two demographical, nine behavioural and two organisational determinants developed with data from 1,125 respondents (response rate 60%). To further increase influenza vaccine uptake, implementation programs should target these determinants.

Erectile dysfunction in primary care: prevalence and patient characteristics. The ENIGMA study

The availability of adequate treatment for erectile dysfunction (ED) triggers studies into the prevalence of ED in the general population. Yet, previous studies showed different prevalence estimates partly due to differences in patient selection, in (unclear) definitions of ED and in assessment. ENIGMA has been designed to study the prevalence of ED in the general population of The Netherlands, using the WHO definition with a description of the way of assessment. In all, 5721 mail surveys were sent to all men, aged 18 y and older in 12 general practices in The Netherlands. A total of 5601 were included in the study and 2117 (38%) were completed. A total of 38% of the men reported to have ever had some kind of erectile problem. The prevalence of ED was 17% (6% mild, 4% moderate and 7% complete). Age, diabetes, cardiovascular diseases, penile disorders, irradiation in the pelvic region, relational problems, fear for failure, surmenage, medication use and regular consumption of alcohol were independently related to ED. Men with ED were less content with their (sexual) life and had less confidence in sexual performance. Presence of ED was negatively related to affected happiness in life. ED is commonly found in men and is related to age, medication, comorbidity and lifestyle factors. Men with ED perceive a lower quality of (sex)life. Doctors should be aware of the presence of ED and its consequences in patients.

Clinical effectiveness of conventional influenza vaccination in asthmatic children.

Influenza immunization rates among young asthmatics remain unsatisfactory due to persistent concern about the impact of influenza and the benefits of the vaccine. We assessed the effectiveness of the conventional inactivated trivalent sub-unit influenza vaccine in reducing acute respiratory disease in asthmatic children. We conducted a two-season retrospective cohort study covering the 1995-6 and 1996-7 influenza outbreaks in 22 computerized primary care practices in The Netherlands. In total, 349 patients aged between 0 and 12 years meeting clinical asthma-criteria were included; 14 children were lost to follow-up in the second season. The occurrence of physician-diagnosed acute respiratory disease episodes including influenza-like illness, pneumonia. bronchitis, bronchiolitis, asthma exacerbation and acute otitis media in vaccinated and unvaccinated children were compared after adjustments for age, prior health care and medication use. The occurrence of acute respiratory disease in unvaccinated children was 28% and 24% in the 1995-6 and 1996-7 season, respectively, and was highest in children under 6 years of age (43%). The overall pooled clinical vaccine effectiveness was 27% (95% confidence interval -7 to 51%, P = 0.11) after adjustments. A statistically higher vaccine protectiveness of 55% (95% CI 20-75%, P = 0.01) was observed among asthmatics under 6 years of age compared with -5% in older children (95% CI -81 to 39%). The occurrence of acute respiratory disease among asthmatic children during influenza epidemics is very high, notably in the youngest. Influenza vaccination may reduce morbidity in asthmatic infants and pre-school children. However, larger, preferably experimental, studies are needed to establish the benefits of vaccination, notably in older asthmatic children.

Evaluation of a web-based intervention to reduce antibiotic prescribing for LRTI in six European countries: quantitative process analysis of the GRACE/INTRO randomised controlled trial

BackgroundTo reduce the spread of antibiotic resistance, there is a pressing need for worldwide implementation of effective interventions to promote more prudent prescribing of antibiotics for acute LRTI. This study is a process analysis of the GRACE/INTRO trial of a multifactorial intervention that reduced antibiotic prescribing for acute LRTI in six European countries. The aim was to understand how the interventions were implemented and to examine effects of the interventions on general practitioners’ (GPs’) and patients’ attitudes.MethodsGPs were cluster randomised to one of three intervention groups or a control group. The intervention groups received web-based training in either use of the C-reactive protein (CRP) test, communication skills and use of a patient booklet, or training in both. GP attitudes were measured before and after the intervention using constructs from the Theory of Planned Behaviour and a Website Satisfaction Questionnaire. Effects of the interventions on patients were assessed by a post-intervention questionnaire assessing patient enablement, satisfaction with the consultation, and beliefs about the risks and need for antibiotics.ResultsGPs in all countries and intervention groups had very positive perceptions of the intervention and the web-based training, and felt that taking part had helped them to reduce prescribing. All GPs perceived reducing prescribing as more important and less risky following the intervention, and GPs in the communication groups reported increased confidence to reduce prescribing. Patients in the communication groups who received the booklet reported the highest levels of enablement and satisfaction and had greater awareness that antibiotics could be unnecessary and harmful.ConclusionsOur findings suggest that the interventions should be broadly acceptable to both GPs and patients, as well as feasible to roll out more widely across Europe. There are also some indications that they could help to engender changes in GP and patient attitudes that will be helpful in the longer-term, such as increased awareness of the potential disadvantages of antibiotics and increased confidence to manage LRTI without them. Given the positive effects of the booklet on patient beliefs and attitudes, it seems logical to extend the use of the patient booklet to all patients.

Impact of various questionnaires on the prevalence of erectile dysfunction. The ENIGMA-study

The prevalence estimates of erectile dysfunction (ED) vary considerably across studies. These differences may be attributed to used definitions of ED. Quantitative data on the effect of different definitions of ED on the prevalence are lacking, because precise information on the used definition and questionnaire is often absent. Aim of this study was to quantify the effect of using different questionnaires for ED on the prevalence estimates. In all, 5721 mail surveys on sexual problems and ED were sent to all men (aged>18 y) in 12 general practices in the middle of the Netherlands of which 2117 were completed. The questionnaire contained Enigma (WHO), International Index of Erectile Function (IIEF), Cologne Erectile Inventory (KEED) and one question (Boxmeer, Krimpen). The prevalence of ED based on the various questionnaires and the effect of these questionnaires on risk factor relationships was compared. IIEF gave the highest age specific and overall ED prevalence, KEED the lowest. The difference in prevalence was 16.8%. The agreement (kappa coefficient) between the various ED definitions varied from 0.52 (IIEF & KEED) to 0.95 (Enigma & Boxmeer). The number of risk factor relations were similar for the Dutch studies, reduced for the IIEF and KEED. This study provides evidence that differences in questionnaires to assess ED have a considerable effect on the (age specific) prevalence estimates and little on the risk factor relations. The number of questions of the survey appears not to be responsible for differences in the prevalence of ED and risk factor relations, however they affect the response rate. Uniform use is strongly recommended, since a ‘golden standard’ for ED assessment (by questionnaire) is lacking. A short questionnaire with one or two questions is recommended for example the one from the Boxmeer-study. These data may be used to adjust (age-specific) prevalence rates comparing ED prevalence in the open population across studies.

Analytical performance, agreement and user-friendliness of five C-reactive protein point-of-care tests

Abstract Background. Point-of-care (POC) C-reactive protein (CRP) testing is increasingly used in primary care to assist general practitioners (GPs) in the diagnostic workup for various complaints. The present study compares analytical performance, agreement and user-friendliness of five of these POC CRP tests. Methods. The following five POC CRP tests were evaluated: Afinion and NycoCard Reader II (both Alere), Eurolyser Smart 700/340 (Eurolyser), QuikRead go and QuikRead 101 (both Orion Diagnostica). Results were compared with those of a standard immunoturbidimetric method performed on a routine analyzer (Olympus AU 2700, Beckman Coulter). Analytical performance and agreement with the laboratory standard for the five different POC tests were analyzed. Subsequently, user-friendliness of the POC tests was assessed. Results. Within-day CVs varied from 2.6% (QuikRead go) to 19.4% (Eurolyser Smart 700/340) for low CRP values (< 20 mg/L), and 1.1% (QuikRead go) to 17.5% (Eurolyser Smart 700/340) for high values (> 100 mg/L). Between-day CVs varied from 4.6% (Afinion) to 30.5% (Eurolyser Smart 700/340) for low values and 4.0% (QuikRead go) to 18.0% (Eurolyser Smart 700/340) for high values. With high CRP values (> 100 mg/L) agreement with the laboratory standard systematically decreased for all POC tests. Regarding user-friendliness Afinion and Eurolyser Smart 700/340 were judged easiest to operate. Conclusions. Analytical performance, agreement, and user-friendliness of the POC CRP tests varied considerably, yet overall four devices showed adequate analytical performance and agreement.

Design of the Dutch prevention of influenza, surveillance and management (PRISMA) study

Rationale and design of a study on the cost-effectiveness of the Dutch influenza vaccination campaign are described. During two influenza epidemics, about 75,000 primary care patients recommended for influenza vaccination are included. Cases have fatal or non-fatal influenza, pneumonia, otitis media, acute respiratory disease (ARD), heart failure, myocardial infarction, depression or diabetes dysregulation. Per case four controls are sampled, frequency matched on age and high-risk co-morbidity (<18 years, 18-64, >/=65 healthy, >/=65 with co-morbidity). Baseline and outcome data are retrieved from patient records. During the 1999-2000 influenza A epidemic 5891 (7.9%) high-risk children, 24,848 (33.2%) high-risk adults aged 18-64 years, 18,484 (24.7%) elderly with co-morbidity and 25,527 (34.1%) healthy elderly had been included. The mortality rate was 5.2 per 1000 and 2035 non-fatal outcome events were recorded (incidence rate 27.2/1000).

Huisartsen volgen de herziene NHG-Standaard OMA

BackgroundDutch primary care has over 100 practical evidence-based treatment guidelines. The second revised guideline for Acute Otitis Media (AOM) was published in 2006.4 Antibiotics are indicated for children with AOM younger than 6 months, with severe or exacerbating illness, or with increased risk of complications. New in this guideline are groups of children for whom antibiotics can be considered: children younger than 2 years with bilateral otitis media, children with a discharging ear at first presentation (meta-analysis showed that these subgroups benefit more than others from antibiotics)9, and children with symptoms lasting longer than 3 days.10AimTo investigate whether general practitioners (GPs) comply with the revised guideline AOM.MethodsInsight in GPs’ prescribing behaviour was obtained by mirroring 198 detailed described consultations involving children with AOM to the revised guideline.ResultsAntibiotics were prescribed to 55% of patients with AOM, amoxicillin in 83.5% of prescriptions. Over-prescription (prescribed when not indicated) and underprescription (not prescribed when indicated) were very rare; the guideline was followed in 96% of the consultations. However, most of the children (60%) fell in the category ‘antibiotics can be considered’, which gives GPs the choice of whether to prescribe or not. There was consensus regarding the GPs’ treatment decisions in this category: fever and bilateral otitis (irrespective of age) in particular led to the prescription of antibiotics.ConclusionThe second revised guideline AOM is adhered to in Dutch primary care. However, due to inclusion of the category ‘consider antibiotics’ the guideline does not provide a clear treatment advice for the majority of patients presenting with AOM in Dutch primary care. As resistance problems are globally increasing, prudent use of antibiotics lists high on the international agenda. To decrease prescription of antibiotics for children, defining a more specific evidence-based prescribing advice for AOM remains necessary, and will provide more clarity for physicians and parents.SamenvattingAchtergrondIn 2006 verscheen de tweede herziene standaard Otitis Media Acuta (OMA), met onder meer richtlijnen voor behandeling van OMA met antibiotica. Naast de kinderen voor wie antibiotica zijn geïndiceerd, kan de huisarts nu bij 3 groepen een antibioticum overwegen: kinderen jonger dan 2 jaar met dubbelzijdige otitis, kinderen die bij eerste presentatie een loopoor hebben, en kinderen bij wie de klachten langer dan 3 dagen duren.VraagstellingIn welke mate handelen huisartsen naar deze OMA-richtlijn?MethodeWe spiegelden 198 gedetailleerd beschreven OMA-consulten aan de richtlijn om inzicht te krijgen in het voorschrijfgedrag van de huisarts.ResultatenOver- en onderprescriptie kwamen zelden voor. In 96% van de consulten handelde de huisarts in overeenstemming met de richtlijn. Van de kinderen viel 60% in de categorie ‘overwegen’; voor hen definieert de richtlijn geen duidelijk beleid. Binnen deze categorie blijken huisartsen wel eenduidig te handelen: vooral koorts en (leeftijdsonafhankelijke) bilaterale otitis leiden tot prescriptie.ConclusieDe tweede herziene standaard OMA wordt goed nageleefd. Wel rijst de vraag of huisartsen werkelijk anders zijn gaan voorschrijven door de herziene richtlijn, of dat prescriptie alleen gerichter lijkt door de categorie ‘antibiotica overwegen’.