Bradykinin and cysteinyl leukotriene concentrations in cell-salvaged blood before and after passage through negatively charged filters during clinical use in cancer patients: a pilot study.

It has been suggested that giving cell-salvaged blood through a leucocyte depletion filter can cause hypotension due to bradykinin released when factor XII and platelets are activated by the negatively charged surface of the filter. We measured the concentration of bradykinin and cysteinyl leukotrienes in cell-salvaged blood sampled before and after passage through a negatively charged leucodepletion filter in 24 consecutive patients with gynaecological or bowel cancer undergoing elective surgery with cell salvage. In no case was an increase in bradykinin concentration observed after passage through the filter; in 23 patients the post-filtration bradykinin concentration was zero (p = 0.007). The change in the concentration of cysteinyl leukotrienes detected during passage across the filter was not statistically significant (p = 0.1). Our findings do not support the suggestion that either bradykinin or cysteinyl leukotrienes are generated in cell-salvaged blood during passage through leucodepletion filters.

Generalized method for adaptive randomization in clinical trials.

A flexible, generalized method of treatment allocation is proposed. The method uses a set of controlling parameters that enables the generic algorithm to produce a family of possible outcomes ranging from simple randomization to deterministic allocation. The method controls balance at stratum level, stratification level and overall without detriment to the predictability of the method. The paper lists the desirable characteristics of allocation methods and shows that the proposed method fulfils the majority and is easy to use in the clinical context, once the coding has been established. An explanation of the method for 2, 3 and 4 treatment group allocations is given. Simulations demonstrate the flexibility of the method.

Refractory depression: mechanisms and evaluation of radically open dialectical behaviour therapy (RO-DBT) [REFRAMED]: protocol for randomised trial.

INTRODUCTION: Only 30-40% of depressed patients treated with medication achieve full remission. Studies that change medication or augment it by psychotherapy achieve only limited benefits, in part because current treatments are not designed for chronic and complex patients. Previous trials have excluded high-risk patients and those with comorbid personality disorder. Radically Open Dialectical Behaviour Therapy (RO-DBT) is a novel, transdiagnostic treatment for disorders of emotional over-control. The REFRAMED trial aims to evaluate the effectiveness and cost-effectiveness of RO-DBT for patients with treatment-resistant depression. METHODS AND ANALYSIS: REFRAMED is a multicentre randomised controlled trial, comparing 7 months of individual and group RO-DBT treatment with treatment as usual (TAU). Our primary outcome measure is depressive symptoms 12 months after randomisation. We shall estimate the cost-effectiveness of RO-DBT by cost per quality-adjusted life year. Causal analyses will explore the mechanisms by which RO-DBT is effective. ETHICS AND DISSEMINATION: The National Research Ethics Service (NRES) Committee South Central - Southampton A first granted ethical approval on 20 June 2011, reference number 11/SC/0146. TRIAL REGISTRATION NUMBER: ISRCTN85784627.

The UK IBDQ-a British version of the inflammatory bowel disease questionnaire. development and validation.

Measurement of health-related quality of life (HRQL) is becoming more important in studies of patients with inflammatory bowel disease. The McMaster IBDQ is the most widely used HRQL instrument for these patients. However, its use with patients in the United Kingdom has not been validated. This study develops and validates a UK version of the McMaster IBDQ (UK IBDQ). The UK IBDQ was tested with two samples of patients for its reliability, validity, reproducibility, and responsiveness. The first sample consisted of 180 patients participating in a randomized clinical trial. The second was recruited from members of the National Association for Colitis and Crohns Disease. Reliability of the subscales and the summary score of the UK IBDQ is demonstrated by Cronbach's alpha and item-total correlations. Their validity is demonstrated by their correlations with SF-36 subscales and an empirical index of disease activity. Good intraclass correlations and responsiveness ratios show their reproducibility and responsiveness. The findings support the reliability, validity, reproducibility, and responsiveness of the UK IBDQ and its acceptability to patients in UK.

Birthweight and perinatal mortality: III. Towards a new method of analysis.

Perinatal mortality is closely related to birthweight. We propose a model that summarizes this relationship and provides a basis for the analysis of perinatal mortality. The components of this model are the frequency distribution of birthweight and the curve of weight-specific mortality. Taken together, these two curves completely describe perinatal mortality for a given population. The perinatal mortality of two populations can be meaningfully compared by plotting each weight-specific mortality curve relative to its own birthweight distribution. By this means, the excess mortality in one population can usually be expressed as the sum of two excess mortalities--one that occurs uniformly over the whole birthweight distribution, the other due to an increased number of small births. To illustrate this method, we analyse differences between white and black infants. We find that the excess mortality of black infants is chiefly due to an excess of small black births, but also to higher mortality over all (adjusted) birthweights. In contrast to other methods of analysis, the proposed method does not assume a priori that a difference in the mean birthweight of two populations is the cause of any difference in perinatal mortality. Furthermore, the proposed method is unbiased; in particular, it is preferable to direct or indirect standardization for birthweight, previously shown to be biased.

Birthweight and perinatal mortality: I. On the frequency distribution of birthweight.

Perinatal mortality is closely related to birthweight. Hence the study of perinatal mortality requires a sound understanding of the influence of birthweight on perinatal mortality. This paper discusses one aspect of this problem--the frequency distribution of birthweight. This distribution is essentially Gaussian but with additional births in the lower tail. It can therefore be divided into two components--a predominant (Gaussian) distribution and a residual distribution. The complete distribution can be summarized by three parameters: the mean and the standard deviation of the predominant distribution, and the proportion of births in the residual distribution. This paper shows that the predominant distribution is composed largely of term births, while the residual distribution is composed almost entirely of small preterm births. It also shows that the three parameters together help to explain the apparent paradox that male infants suffer a higher perinatal mortality than females despite there being fewer light male births.

Birthweight and perinatal mortality: II. On weight-specific mortality.

The study of perinatal mortality requires a sound understanding of the influence of birthweight on perinatal mortality. This paper discusses one aspect of this problem--the pattern of weight-specific mortality. Mortality is very high at the lowest birthweights, falls to a minimum within the range of the most frequent birthweights, but rises again for the heaviest birthweights. Such a curve is best displayed and modelled by plotting the ratio of deaths to survivors on a logarithmic scale. Transformed in this way, perinatal risk may be regarded as the sum of three components--one independent of birthweight, one which decreases linearly with birthweight and one which increases linearly with birthweight. These two lines appear to have slopes of equal magnitude. Each is shown to represent general susceptibility to perinatal problems, rather than the cumulative effect of diseases specific to low birthweight or to high birthweight.

Laparoscopic and minilaparotomy cholecystectomy: a randomized trial comparing postoperative pain and pulmonary function.

BACKGROUND: Upper abdominal surgery is associated with severe postoperative pain and a concomitant reduction in pulmonary function and oxygen saturation. Laparoscopic cholecystectomy is said to result in less postoperative pain compared with open cholecystectomy. METHODS: In a pragmatic, randomized trial, postoperative pain, opiate analgesic consumption, oxygen saturation, and pulmonary function (forced vital capacity, forced expiratory volume in 1 second, and peak expiratory flow rate) were assessed after laparoscopic (n = 67) and minilaparotomy (n = 65) cholecystectomy. RESULTS: Compared with minilaparotomy cholecystectomy, laparoscopic cholecystectomy was associated with lower linear analogue pain scores (median 40 vs 59, p < 0.001), lower patient-controlled morphine consumption (median 22 vs 40 mg, p < 0.001), a smaller reduction in postoperative pulmonary function (mean peak expiratory flow rate 64% of preoperative value vs 49%, p < 0.001), and better oxygen saturation (mean 92.9% vs 91.2%, p = 0.008). CONCLUSIONS: This study confirms that the postoperative pain and pulmonary changes associated with upper abdominal surgery are significantly reduced by the laparoscopic technique. These findings suggest that laparoscopic cholecystectomy may result in a reduced risk of postoperative pulmonary complications.

Can randomised trials rely on existing electronic data? A feasibility study to explore the value of routine data in health technology assessment.

OBJECTIVES: To estimate the feasibility, utility and resource implications of electronically captured routine data for health technology assessment by randomised controlled trials (RCTs), and to recommend how routinely collected data could be made more effective for this purpose. DATA SOURCES: Four health technology assessments that involved patients under care at five district general hospitals in the UK using four conditions from distinct classical specialties: inflammatory bowel disease, obstructive sleep apnoea, female urinary incontinence, and total knee replacement. Patient-identifiable, electronically stored routine data were sought from the administration and clinical database to provide the routine data. REVIEW METHODS: Four RCTs were replicated using routine data in place of the data already collected for the specific purpose of the assessments. This was done by modelling the research process from conception to final writing up and substituting routine for designed data activities at appropriate points. This allowed a direct comparison to be made of the costs and outcomes of the two approaches to health technology assessment. The trial designs were a two-centre randomised trial of outpatient follow-up; a single-centre randomised trial of two investigation techniques; a three-centre randomised trial of two surgical operations; and a single-centre randomised trial of perioperative anaesthetic intervention. RESULTS: Generally two-thirds of the research questions posed by health technology assessment through RCTs could be answered using routinely collected data. Where these questions required analysis of NHS resource use, data could usually be identified. Clinical effectiveness could also be judged, using proxy measures for quality of life, provided clinical symptoms and signs were collected in sufficient detail. Patient and professional preferences could not be identified from routine data but could be collected routinely by adapting existing instruments. Routine data were found potentially to be cheaper to extract and analyse than designed data, and they also facilitate recruitment as well as have the potential to identify patient outcomes captured in remote systems that may be missed in designed data collection. The study confirmed previous evidence that the validity of routinely collected data is suspect, particularly in systems that are not under clinical and professional control. Potential difficulties were also found in identifying, accessing and extracting data, as well as in the lack of uniformity in data structures, coding systems and definitions. CONCLUSIONS: Routine data have the potential to support health technology assessment by RCTs. The cost of data collection and analysis is likely to fall, although further work is required to improve the validity of routine data, particularly in central returns. Better knowledge of the capability of local systems and access to the data held on them is also essential. Routinely captured clinical data have real potential to measure patient outcomes, particularly if the detail and precision of the data could be improved.

Statistical assessment of the learning curves of health technologies.

OBJECTIVES: (1) To describe systematically studies that directly assessed the learning curve effect of health technologies. (2) Systematically to identify 'novel' statistical techniques applied to learning curve data in other fields, such as psychology and manufacturing. (3) To test these statistical techniques in data sets from studies of varying designs to assess health technologies in which learning curve effects are known to exist. METHODS - STUDY SELECTION (HEALTH TECHNOLOGY ASSESSMENT LITERATURE REVIEW): For a study to be included, it had to include a formal analysis of the learning curve of a health technology using a graphical, tabular or statistical technique. METHODS - STUDY SELECTION (NON-HEALTH TECHNOLOGY ASSESSMENT LITERATURE SEARCH): For a study to be included, it had to include a formal assessment of a learning curve using a statistical technique that had not been identified in the previous search. METHODS - DATA SOURCES: Six clinical and 16 non-clinical biomedical databases were searched. A limited amount of handsearching and scanning of reference lists was also undertaken. METHODS - DATA EXTRACTION (HEALTH TECHNOLOGY ASSESSMENT LITERATURE REVIEW): A number of study characteristics were abstracted from the papers such as study design, study size, number of operators and the statistical method used. METHODS - DATA EXTRACTION (NON-HEALTH TECHNOLOGY ASSESSMENT LITERATURE SEARCH): The new statistical techniques identified were categorised into four subgroups of increasing complexity: exploratory data analysis; simple series data analysis; complex data structure analysis, generic techniques. METHODS - TESTING OF STATISTICAL METHODS: Some of the statistical methods identified in the systematic searches for single (simple) operator series data and for multiple (complex) operator series data were illustrated and explored using three data sets. The first was a case series of 190 consecutive laparoscopic fundoplication procedures performed by a single surgeon; the second was a case series of consecutive laparoscopic cholecystectomy procedures performed by ten surgeons; the third was randomised trial data derived from the laparoscopic procedure arm of a multicentre trial of groin hernia repair, supplemented by data from non-randomised operations performed during the trial. RESULTS - HEALTH TECHNOLOGY ASSESSMENT LITERATURE REVIEW: Of 4571 abstracts identified, 272 (6%) were later included in the study after review of the full paper. Some 51% of studies assessed a surgical minimal access technique and 95% were case series. The statistical method used most often (60%) was splitting the data into consecutive parts (such as halves or thirds), with only 14% attempting a more formal statistical analysis. The reporting of the studies was poor, with 31% giving no details of data collection methods. RESULTS - NON-HEALTH TECHNOLOGY ASSESSMENT LITERATURE SEARCH: Of 9431 abstracts assessed, 115 (1%) were deemed appropriate for further investigation and, of these, 18 were included in the study. All of the methods for complex data sets were identified in the non-clinical literature. These were discriminant analysis, two-stage estimation of learning rates, generalised estimating equations, multilevel models, latent curve models, time series models and stochastic parameter models. In addition, eight new shapes of learning curves were identified. RESULTS - TESTING OF STATISTICAL METHODS: No one particular shape of learning curve performed significantly better than another. The performance of 'operation time' as a proxy for learning differed between the three procedures. Multilevel modelling using the laparoscopic cholecystectomy data demonstrated and measured surgeon-specific and confounding effects. The inclusion of non-randomised cases, despite the possible limitations of the method, enhanced the interpretation of learning effects. CONCLUSIONS - HEALTH TECHNOLOGY ASSESSMENT LITERATURE REVIEW: The statistical methods used for assessing learning effects in health technology assessment have been crude and the reporting of studies poor. CONCLUSIONS - NON-HEALTH TECHNOLOGY ASSESSMENT LITERATURE SEARCH: A number of statistical methods for assessing learning effects were identified that had not hitherto been used in health technology assessment. There was a hierarchy of methods for the identification and measurement of learning, and the more sophisticated methods for both have had little if any use in health technology assessment. This demonstrated the value of considering fields outside clinical research when addressing methodological issues in health technology assessment. CONCLUSIONS - TESTING OF STATISTICAL METHODS: It has been demonstrated that the portfolio of techniques identified can enhance investigations of learning curve effects. (ABSTRACT TRUNCATED)

A systematic review of comparisons of effect sizes derived from randomised and non-randomised studies.

BACKGROUND: There is controversy about the value of evidence about the effectiveness of healthcare interventions from non-randomised study designs. Advocates for quasi-experimental and observational (QEO) studies argue that evidence from randomised controlled trials (RCTs) is often difficult or impossible to obtain, or is inadequate to answer the question of interest. Advocates for RCTs point out that QEO studies are more susceptible to bias and refer to published comparisons that suggest QEO estimates tend to find a greater benefit than RCT estimates. However, comparisons from the literature are often cited selectively, may be unsystematic and may have failed to distinguish between different explanations for any discrepancies observed. OBJECTIVES: The aim was to investigate the association between methodological quality and the magnitude of estimates of effectiveness by comparing systematically estimates of effectiveness derived from RCTs and QEO studies. Quantifying any such association should help healthcare decision-makers to judge the strength of evidence from non-randomised studies. Two strategies were used to minimise the influence of differences in external validity between RCTs and QEO studies: a comparison of the RCT and QEO study estimates of effectiveness of any intervention, where both estimates were reported in a single paper a comparison of the RCT and QEO study estimates of effectiveness for specified interventions, where the estimates were reported in different papers. The authors also sought to identify study designs that have been proposed to address one or more of the problems often found with conventional RCTs. DATA SOURCES: Relevant literature was identified from: The Cochrane Library, MEDLINE, EMBASE, DARE, and the Science Citation Index. References of relevant papers already identified experts. Electronic searches were very difficult to design and yielded few papers for the first strategy and when identifying study designs. CHOICE OF INTERVENTIONS TO REVIEW FOR STRATEGIES 1 AND 2: For strategy 1, any intervention was eligible. For strategy 2, interventions for which the population, intervention and outcome investigated were anticipated to be homogeneous across studies were selected for review: Mammographic screening (MSBC) of women to reduce mortality from breast cancer. Folic acid supplementation (FAS) to prevent neural tube defects in women trying to conceive. DATA EXTRACTION AND QUALITY ASSESSMENT: Data were extracted by the first author and checked by the second author. Disagreements were negotiated with reference to the paper concerned. For strategy 1, study quality was scored using a checklist to assess whether the RCT and QEO study estimates were derived from the same populations, whether the assessment of outcomes was 'blinded', and the extent to which the QEO study estimate took account of possible confounding. For strategy 2, a more detailed instrument was used to assess study quality on four dimensions: the quality of reporting, the generalisability of the results, and the extent to which estimates of effectiveness may have been subject to bias or confounding. All quality assessments were carried out by three people. DATA SYNTHESIS AND ANALYSIS: For strategy 1, pairs of comparisons between RCT and QEO study estimates were classified as high or low quality. Seven indices of the size of discrepancies between estimates of effect size and outcome frequency were calculated, where possible, for each comparison. Distributions of the size and direction of discrepancies were compared for high- and low-quality comparisons. FOR STRATEGY 2, THREE ANALYSES WERE CARRIED OUT: Attributes of the instrument were described by k statistics, percentage agreement, and Cronbach's a values. Regression analyses were used to investigate -variations in study quality. (ABSTRACT TRUNCATED)

Estimating the point accuracy of population registers using capture-recapture methods in Scotland.

STUDY OBJECTIVE: To estimate the point accuracy of adult registration on the community health index (CHI) by comparing it with the electoral register (ER) and the community charge register (CCR). DESIGN: Survey of overlapping samples from three registers to ascertain whether respondents were living at the addresses given on the registers, analysed by capture-recapture methods. SETTING: Aberdeen North and South parliamentary constituencies. PARTICIPANTS: Random samples of adult registrants aged at least 18 years from the CHI (n = 1000), ER (n = 998), and CCR (n = 956). MAIN RESULTS: Estimated sensitivities (the proportions of the target population registered at the address where they live) were: CHI--84.6% (95% confidence limits 82.4%, 86.7%); ER--90.0% (87.5%, 92.5%), and CCR--87.7% (85.3%, 90.3%). Positive predictive values (the proportions of registrants who were living at their stated addresses) were: CHI--84.6% (82.2%, 87.0%); ER--94.0% (90.9%, 97.1%), and CCR--93.7% (91.7%, 95.7%). CONCLUSIONS: The CHI assessed in this study was significantly less sensitive and predictive than the corresponding ER and CCR. Capture-recapture methods are effective in assessing the accuracy of population registers.

Randomised comparison of percutaneous angioplasty vs continued medical therapy for hypertensive patients with atheromatous renal artery stenosis. Scottish and Newcastle Renal Artery Stenosis Collaborative Group.

BACKGROUND: Data from randomised studies are lacking on the value of interventional procedures in the management of atheromatous renal artery stenosis. This randomised prospective trial compared the effects on blood pressure (BP) and renal function of percutaneous transluminal angioplasty vs medical therapy in hypertensive patients with both unilateral and bilateral disease. METHODS: A total of 135 eligible patients were identified, of whom 55 (44%) were randomised. Eligible patients had sustained hypertension, with a minimum diastolic BP of 95 mm Hg on at least two anti-hypertensive drugs. Renal artery stenosis was defined by renal angiography as at least 50% stenosis in the affected vessel. All patients were observed during an initial 4-week run-in period on a fixed drug regimen and subsequent changes measured from this 4-week baseline. RESULTS: Blood pressure fell during the run-in period in all groups. In patients with bilateral renal artery stenosis randomised to angioplasty, a statistically significant (P<0.05) fall in BP was observed at latest follow-up (range 3-54 months). The mean fall in BP at latest follow-up in the angioplasty group, corrected for the medical group response, was 26/10 mm Hg. In patients with unilateral renal artery stenosis, no statistically significant or clinically important differences in outcome were observed between the two groups. No significant differences or trends in serum creatinine were observed between or within any group during follow-up. Major outcome events (death, myocardial infarction, heart failure, stroke, dialysis) were similar in the angioplasty and medical groups during follow-up. In the 40/135 patients undergoing angioplasty, serious or potentially serious complications attributable to the procedure were observed in 11 patients, bleeding at the arterial site (8 patients) being the most frequent. CONCLUSIONS: In hypertensive patients with atheromatous renal artery stenosis, percutaneous renal angioplasty results in a modest improvement in systolic BP compared with medical therapy alone. This benefit was confined to patients with bilateral disease. No patient was 'cured', renal function did not improve, and intervention was accompanied by a significant complication rate.

Factors associated with onset of menopause in women aged 45-49.

This paper uses a cross-sectional sample of women aged 45-49 to investigate factors that might be associated with an early menopause. Using logistic regression analysis we found that age, smoking, age of maternal menopause, parity, social class, meat and alcohol consumption were all independently associated with an early natural menopause. Meat, alcohol consumption and maternal menopausal age do not seem to have been previously noted as associated with the timing of the menopause. These associations would merit further study, preferably using prospective data. However, this study in line with much previous work shows that smoking is associated with a reduction in menopausal age.

Respiratory illness in children: what makes parents decide to consult?

Parents of a stratified random sample of 234 children from 21 training practices in north east England were interviewed at home. All these children had been reported in a postal questionnaire as having had a cough between six and 10 weeks before the interview. Interviews covered social characteristics of the family, the severity of the child's cough and the child's previous respiratory history. These data were analysed using the statistical technique of logistic regression. This produced a good model of consulting behaviour. The model showed that a doctor was likely to be consulted if the child had severe symptoms, or if the cough affected the child's behaviour. This suggests that most parents deciding whether to consult the doctor make careful decisions based on what they see as objective criteria. No social characteristic had a significant influence on the decision to consult the doctor over and above the influence of the characteristics and effect of the cough itself.

Identifying asthmatic patients in the community: a new method.

Identifying a random sample of patients in the community has long proved problematic. In 1989 changes were made in the management of adult asthmatic patients referred to specialist clinics in the Grampian Health Board area. In order to estimate the effect of these changes on the management of patients not referred, it was necessary to identify two random samples of adult asthmatic patients treated solely in general practice. As it was felt that existing methods were open to bias and other errors, a method using National Health Service drug prescription forms was devised. Following the computerization of the Pharmacy Practice Division in Aberdeen, a similar method for the identification of a follow-up sample had to be devised. Nearly 400 general practitioners (86% of those eligible) took part in the first sampling in 1989; 96% of those contacted participated in the second sampling in 1991. Both methods were effective in identifying asthmatic patients in the community. Computerization has made the task simpler, less time consuming and, as a consequence, most cost effective.

Developing a valid and reliable measure of health outcome for patients with low back pain.

OBJECTIVE: The authors developed and validated a measure of health outcome for patients with low back pain. RESULTS: A questionnaire was developed incorporating the type of questions asked when taking a clinical history. After testing on a sample of 568 patients, three questions were discarded from the questionnaire. The final questionnaire was found to be reliable and valid in the sense that patients' scores correlated highly with their scores on a general health profile and with GP perceptions of severity. Health status in patients with low back pain was significantly impaired when compared with the general population. CONCLUSION: A clinically based questionnaire, together with a general measure of health, can provide a valid and reliable package for the routine assessment of perceived health in patients with low back pain.

Hormone replacement therapy: compliance and cost after screening for osteoporosis.

OBJECTIVE: To assess HRT compliance and cost 1 year after population screening for low bone density. DESIGN: Postal questionnaire sent out to women 1 year after having a bone density scan. MAIN OUTCOME MEASURES: Compliance with HRT, type of HRT and cost. RESULTS: Compliance with HRT after screening for women with low bone density was 48% for postmenopausal women and 59% for women with a simple hysterectomy. The most often used form of HRT was Estraderm (49%) for women with a hysterectomy, whilst for women with an intact uterus Prempak-C (54%) was the most popular. The average annual cost of treating women who had had a hysterectomy was 67 pounds whilst for women with an intact uterus it was 61 pounds, a non-significant difference. CONCLUSIONS: Approximately 50% of women with low bone density had started HRT within 1 year of having a densitometry examination. Previous reports of the cost advantage of treating women who had had a hysterectomy relative to those who had not has been overstated.

Patient preferences in randomised trials: threat or opportunity?

OBJECTIVES: To assess whether it is feasible to elicit patients' preferences for treatments and then to proceed with randomisation which may allocate those with preferences to their less preferred treatment; and to describe which prognostic variables were associated with such preferences within the context of a randomised trial of an exercise programme for back pain. METHODS: The first 97 patients enrolled in a randomised controlled trial (RCT) for the treatment of back pain were asked about their preferences, health characteristics and other prognostic variables. RESULTS: Fifty-eight (60%) patients preferred to be allocated to the exercise programme whilst 38 (39%) were indifferent; one patient preferred conventional general practitioner (GP) management. No patient refused randomisation. Comparing patients preferring the exercise programme with indifferent patients showed that the former had a higher belief in the effectiveness of the new treatment (P < 0.01), tended to have worse back pain (P = 0.09), had back pain for a shorter duration (P = 0.04), and tended to have had more GP home visits (P = 0.06). CONCLUSIONS: For many randomised trials preference may be an important prognostic variable. In such circumstances, preference should be taken into account in the final analysis. This study demonstrates it is sometimes feasible to randomise patients to their less preferred treatment, thus allowing more robust statistical comparisons between randomised groups. This modification may make RCTs more rigorous and improve their external validity.

Audit: the third clinical science?

In summary, we believe that if clinical audit is accepted and prosecuted as the third clinical science, it has the potential to deliver substantial benefits to patients and health professionals.