Missing data in trial‐based cost‐effectiveness analysis: An incomplete journey

Cost‐effectiveness analyses (CEA) conducted alongside randomised trials provide key evidence for informing healthcare decision making, but missing data pose substantive challenges. Recently, there have been a number of developments in methods and guidelines addressing missing data in trials. However, it is unclear whether these developments have permeated CEA practice. This paper critically reviews the extent of and methods used to address missing data in recently published trial‐based CEA. Issues of the Health Technology Assessment journal from 2013 to 2015 were searched. Fifty‐two eligible studies were identified. Missing data were very common; the median proportion of trial participants with complete cost‐effectiveness data was 63% (interquartile range: 47%–81%). The most common approach for the primary analysis was to restrict analysis to those with complete data (43%), followed by multiple imputation (30%). Half of the studies conducted some sort of sensitivity analyses, but only 2 (4%) considered possible departures from the missing‐at‐random assumption. Further improvements are needed to address missing data in cost‐effectiveness analyses conducted alongside randomised trials. These should focus on limiting the extent of missing data, choosing an appropriate method for the primary analysis that is valid under contextually plausible assumptions, and conducting sensitivity analyses to departures from the missing‐at‐random assumption.     

Safety and efficacy of Rift Valley fever Smithburn and Clone 13 vaccines in calves

Two modified live attenuated vaccines against the disease Rift Valley fever (RVF) have been tested for safety and efficacy in young calves. The RVF Smithburn vaccine produced in South Africa and used successfully to prevent and control the disease in endemic sub-Saharan countries was compared to the candidate vaccine RVF Clone 13. Five sero-negative calves per vaccine group were vaccinated with a single dose of each vaccine and tested for antibody response. All vaccinated calves were challenged with a highly virulent RVF virus together with five unvaccinated calves used as control of the challenge. Protection was confirmed in all vaccinated animals as they did not show any clinical signs typical of RVF. A good neutralizing antibody response was induced post-vaccination and no viraemia could be detected post-challenge in calves of both vaccine groups. All non-vaccinated control animals showed clinical symptoms of RVF, high viraemia and were euthanized. This study reported the first case of blindness in cattle resulting from virulent RVF virus infection in unvaccinated calves used as negative controls.     

Methodology for constructing a colour‐difference acceptability scale

Observers were invited to report their degree of satisfaction on a 6‐point semantic scale with respect to the conformity of a test colour with a white reference colour, simultaneously presented on a PDP display. Eight test patches were chosen along each of the +a*, ?a*, +b*, ?b* axes of the CIELAB chromaticity plane, at Y = 80 ± 2 cd.m^?2. Experimental conditions reliably represented the automotive environment (patch size, angular distance between patches) and observers could move their head and eyes freely. We have compared several methods of category scaling, the Torgerson‐DMT method (Torgerson, W. S. (1958). Theory and methods of scaling. Wiley, New York, USA); two versions of the regression method i.e. Bonnet’s (Bonnet, C. (1986). Manuel pratique de psychophysique. Armand Colin, Paris, France) and logistic regression; and the medians method. We describe in detail a case where all methods yield substantial but slightly different results. The solution proposed by the regression method which works with incomplete matrices and yields results directly on a colorimetric scale is probably the most useful in this industrial context. Finally we summarize the implementation of the logistic regression method over four hues and for one experimental condition.     

Effects of drugs affecting noradrenergic neurotransmission in rats with spontaneous petit mal-like seizures

Wistar rats of a strain displaying spontaneous petit mal-like seizures and spike-wave EEG discharged (SWD) were injected i.p. with drugs affecting noradrenergic neurotransmission. The EEG and behavior were recorded. Drugs which decrease alpha-noradrenergic neurotransmission, prazosin (alpha 1-antagonist) and clonidine (alpha 2-agonist), increased SWD and were sedative in a dose-dependent manner. Drugs which increase alpha-noradrenergic neurotransmission, ST 587, cirazoline (alpha 1-agonists) and yohimbine (alpha 2-antagonist), reduced SWD and the latter two caused agitation. Drugs which interact with beta-noradrenergic transmission (salbutamol, isoprenaline and propranolol), monoamine oxidase inhibitors (nialamide and iproniazid), and a noradrenaline reuptake inhibitor (desipramine), did not affect SWD. These findings suggest that noradrenaline participates in the control of petit mal-like seizures in the rat, as in other types of seizures and other animal models.     

Erratum: MRI monitoring of focal cerebral ischemia in peroxisome proliferator‐activated receptor (PPAR)‐deficient mice

NMR In Biomedicine: MRI monitoring of focal cerebral ischemia in peroxisome proliferator‐activated receptor (PPAR)‐deficient mice. Jean‐Baptiste Pialat, Tae‐Hee Chao, Olivier Beuf, Elisabeth Joye, Samir Moucharaffie, Jean‐Baptiste Langlois, Chantal Nemoz, Marc Janier, Yves Berthezene, Norbert Nighoghossian, Béatrice Desvergnre and Marlene Wiart. Published online April 24 2007, Issue 20:3, pp. 335–342; DOI: 10.1002/nbm.1157 . Following publication it has been noted that an author's name was misspelt. As above, the author was spelt as “Samir Moucharaffie”. It should have been spelt as “Samir Moucharrafie”.     

C-reactive protein levels for early detection of early postoperative complications after proximal femoral fracture surgery

Introduction

We have studied the early C-reactive protein (CRP) kinetics after proximal femoral fracture (PFF) surgery and defined a chart of CRP threshold values beyond which early postoperative complications are likely.

Materials and methods

We have prospectively followed up a 42-patient cohort for surgical treatment by osteosynthesis or hip arthroplasty, and CRP was dosed at admission, on day 1, day 3, day 5 and then twice weekly for 30 days at the same time as a platelet count control of low molecular weight heparins. Then, we identified postoperative complications (medical complications like infections other than the surgical site, thromboembolic complications, death and surgical complications like surgical site infections, dislocation, haematoma) and established the evolutive profile of CRP in a ‘complication’ group and a ‘no complication’ group. We developed a threshold value model for complications detection from our results and compared with Chapman’s 500/d formula (d being the postoperative day).

Results

CRP in the ‘complication’ group (20 patients) is on average 19 points higher than the ‘no complication’ group (22 patients) (p = 0.007). Our model, which corresponds to the formula CRP(d) = e^{(4.71 ? 0.06*d)}, is valid from day 1 to day 30 and has a threshold value of 105 mg/L at day 1. From day 5, the correlation between our chart and the 500/d chart is 0.84 (p < 10^?5). Two successive CRP values strictly above our chart at a 48-h interval have a sensitivity of 65% and a specificity of 85% to detect a postoperative complication while only one CRP value above the chart has a sensitivity and a specificity of 48%.

Discussion/conclusion

Early postoperative CRP kinetics of PFF surgical patients can be modelled as a threshold value scale valid between day 1 and day 30. This can be an additional tool in the early detection of postoperative complications, if two CRP values are above the threshold values, without replacing the clinical examination, which remains the ‘gold standard’.

Level of evidence

Prospective observational study.

     

Assessment of individual vaccine status in a vaccinology experts' group

Rationale  Worldwide, experts in vaccinology have promoted the broad annual coverage of health care workers with the influenza vaccine. Furthermore, pertussis vaccination is now recommended for young adults and health care workers working with newborns.
Aim  To analyse the compliance with these guidelines among experts responsible for the development or dissemination of national immunization schedules.
Method  A cross-sectional survey was conducted in a vaccinology workshop group of French experts, using a self-administered questionnaire.
Results  Among 44 experts, the average rate of influenza vaccination was 69.5% (95% confidence interval, 61.6% to 77.3%) between the 2003/04 and 2005/06 flu seasons, whereas the rate of pertussis vaccination during this period was only 30%. The main reasons that the experts gave for not being vaccinated were a lack of time or simply not remembering to do so.
Conclusion  Experts had low coverage rates for influenza and pertussis vaccination. To improve these rates, a multifaceted intervention combining audit and feedback strategy with a vaccine day is planned.