Search

Rationale, aims and objectives: Establishing the relationship between various doses of an exposure and a response variable is integral to many studies in health care. Linear parametric models, widely used for estimating dose–response relationships, have several limitations. This paper employs the optimal discriminant analysis (ODA) machine‐learning algorithm to determine the degree to which exposure dose can be distinguished based on the distribution of the response variable. By framing the dose–response relationship as a classification problem, machine learning can provide the same functionality as conventional models, but can additionally make individual‐level predictions, which may be helpful in practical applications like establishing responsiveness to prescribed drug regimens. Method: Using data from a study measuring the responses of blood flow in the forearm to the intra‐arterial administration of isoproterenol (separately for 9 black and 13 white men, and pooled), we compare the results estimated from a generalized estimating equations (GEE) model with those estimated using ODA. Results: Generalized estimating equations and ODA both identified many statistically significant dose–response relationships, separately by race and for pooled data. Post hoc comparisons between doses indicated ODA (based on exact P values) was consistently more conservative than GEE (based on estimated P values). Compared with ODA, GEE produced twice as many instances of paradoxical confounding (findings from analysis of pooled data that are inconsistent with findings from analyses stratified by race). Conclusions: Given its unique advantages and greater analytic flexibility, maximum‐accuracy machine‐learning methods like ODA should be considered as the primary analytic approach in dose–response applications. [ABSTRACT FROM AUTHOR]

Rationale, aims and objectives Open-label extension (OLE) studies following blinded randomized controlled trials (RCTs) of pharmaceuticals are increasingly being carried out but do not conform to regulatory standards and questions surround the validity of their evidence. OLE studies are usually discussed as a homogenous group, yet substantial differences in study design still meet the definition of an OLE. We describe published papers reporting OLE studies focussing on stated objectives, design, conduct and reporting. Method A search of Embase and Medline databases for 1996 to July 2008 revealed 268 papers reporting OLE studies that met our eligibility criteria. A random sample of 50 was selected for detailed review. Results Over 80% of the studies had efficacy stated as an objective. The most common methods of allocation at the start of the OLE were for all RCT participants to switch to one active treatment or for only participants on the new drug to continue, but in three studies all participants were re-randomized at the start of the OLE. Eligibility criteria and other selection factors resulted in on average of 74% of participants in the preceding RCT(s) enrolling in the OLE and only 57% completed it. Conclusions Published OLE studies do not form a homogenous group with respect to design or retention of participants, and thus the validity of evidence from an OLE should be judged on an individual basis. The term 'open label' suggests bias through lack of blinding, but slippage in relation to the sample randomized in the preceding RCT may be the more important threat to validity. [ABSTRACT FROM AUTHOR]

Background Systematic reviews ( SR) of clinical evidence are rightfully considered the basis for developing recommendations to support decisions in current practice. To avoid bias, SRs are expected to be systematic in their research strategy and are exhaustive. The drawback of the latter criteria relies in the substantial work needed to conduct and keep SRs updated. The objective of this paper is to compare a research strategy based on the review of a selected number of core journals, which we consider a 'pragmatic review' ( PR), with that derived by an SR in estimating the efficacy of treatments. Methods Five clinical areas were considered for the comparison between the two strategies: chronic obstructive pulmonary disease, dermatology, heart failure, renal diseases and stroke. We extracted a systematic sample from all the Cochrane SRs pertaining to each area and were published before April 2010. Two groups of journals were considered in the PR: six general journals that commonly published research for the five clinical areas, and five specialist journals with the highest impact factor in each area. To assess the agreement in the findings of SRs and PRs, we considered both the direction of the estimates and P-values. Results A sample of 27 SRs included 171 overall analyses and 259 subgroup analyses related to primary outcomes. The PR captured one or more clinical trials in 24 of the 27 SRs (89%), and 118 of the 171 overall analyses (69%) were replicated. The PR supported the recommendations to use (or not) the study treatment in 11 of the 13 SRs (85%), which ended with a clinical recommendation. Conclusions We verified in a sample of SRs that the conclusion of a research strategy based on a pre-defined set of general and specialist medical journals is able to replicate almost all the clinical recommendations of a formal SR. [ABSTRACT FROM AUTHOR]

Background Colorectal cancer (CRC) is one of the most common types of cancer in European countries and associated with a high mortality rate. A 16% relative risk reduction (RRR) of mortality was found in a meta-analysis based on four randomized controlled trials (RCT) on CRC screening. The aim of this paper was to scrutinize these trials for potential biases and assess their influence on the screening trials. Methods The four RCTs were reviewed based on the principles of 'Critical Appraisal of the Medical Literature'. Principal investigators of the four RCTs were contacted to clarify uncertainties in their study. Data were collected from The Danish Data Archives. Authors of the Cochrane review were contacted. Results Six biases were identified, of which five favour screening. Three of the biases identified were specific to CRC screening: type of diagnostic method, place of surgery and diagnostic delay. Conclusion The 16% RRR in CRC mortality found in the updated Cochrane review's meta-analysis is overestimated. [ABSTRACT FROM AUTHOR]

The aim of this paper was to compare selected indication parameters for patients scheduled for hip and knee replacement at orthopaedic units in Sweden. Swedish orthopaedic clinics performing joint replacement were invited to enrol in the study. The study time was set to 2 years (from June 2006 to June 2008). The study subjects were patients undergoing hip or knee replacement for osteoarthritis (OA). For data collection, we used a Swedish priority criteria tool based on a translation from a form used in Canada with minor changes. The reliability and validity of the Swedish tool were investigated, with good reproducibility. The questionnaires (one for the doctor and one for the patient) were completed during decision making for surgery. Eleven hospitals enrolled in the study. In total, 2961 patients were included during the study period. Among these, 1662 were hip replacement patients and 1299 were knee replacement patients. The vast majority of patients undergoing hip or knee replacement had findings indicating severe OA, both clinically and radiologically according to the clinical priority tool. Statistically significant self-reported problems with pain at rest, walking and impaired activities of daily living were also observed. There were statistically significant differences in reported indications between the hospitals, both for hip OA patients and for knee OA patients. A clinical priority criteria tool is a useful means of following changes in indications for certain procedures. It could also contribute to explaining differences in case mix when evaluating clinical outcome and patient satisfaction. [ABSTRACT FROM AUTHOR]