Staircase bioassay: the up-and-down method

Experiments are conducted to estimate the threshold for an all-or-none response. Threshold is defined to be a point above which 50% of the subjects will respond and below which 50% of the subjects will not respond. Examples are death, death in a fixed time period, shock, fibrillation, emesis. Staircase designs, in particular up-and-down trials, produce median (ED50) estimates of given standard error with as few as one-fifth the number of subjects as the traditional designs with preset numbers of tests at each of several levels of stimulus. We discuss these estimates and their efficiency as well as procedures to estimate standard deviation and its use in designing up-and-down trails. The advantages in using several short series in factorial experiments are presented. Suggestions are given for minimizing the complications of sequential designs. Case studies indicate the efficiency of the design for various applications.     

Primary care referrals for lumbar spine radiography: diagnostic yield and clinical guidelines.

BACKGROUND: Primary care requests for radiographs of the lumbar spine have come under increasing scrutiny. Guidelines aiming to reduce unnecessary radiographs by limiting referrals to patients at high risk of serious disease have been widely distributed. Trial evidence suggests that guidelines can reduce radiography referrals. It is not clear whether this reduction has been achieved in routine practice. AIM: This study, using routine data, was conducted to measure trends in pnmary care referrals for lumbar spine radiography at two hospitals between 1994 and 1999. DESIGN OF STUDY: Analysis of primary care requests for lumbar spine radiography from computerised records. SETTING: Addenbrooke's Hospital, Cambridge (1 July 1994 to 30 June 1999), and Ipswich General Hospital (1 July 1995 to 30 June 1999), United Kingdom. METHOD: All primary care requests for lumbar radiography were identified electronically from computerised information systems. A random sample of 2100 radiography reports were classified according to clinical importance. These classifications were used to examine whether the proportion of radiographs demonstrating potentially more serious findings had increased between 1994 and 1999. RESULTS: There was no evidence that primary care referrals for radiography of the lumbar spine had decreased between 1994 and 1999 at either hospital. General practitioners did not progressively refer more high-risk patients for lumbar radiography. Only a small proportion of patients had important radiographic findings that might warrant specialist referral or specific therapy. CONCLUSION: The implementation of diagnostic guidelines offers much to the NHS. However in these two hospitals, the reduction in radiograph utilisation evident in trials was not achieved. Guideline development is a resource intensive process; distribution must be supported by more effective implementation strategies.     

Potential for cost economies in guiding therapy in patients with metastatic breast cancer.

Therapeutic response in patients with advanced breast cancer is conventionally assessed with reference to criteria devised by the International Union Against Cancer. Evidence to date suggests, however, that assessments of equivalent quality may be obtained at lower cost from the use of serum markers. The paper presents estimates of potential cost savings resulting from the use of serum markers in place of conventional assessment and argues that the size of these savings merits the establishment of a randomised controlled trial.     

Computed tomography in blunt abdominal trauma: an analysis of clinical management and radiological findings.

We have retrospectively assessed the computed tomography (CT) findings in 92 patients suffering severe blunt abdominal trauma. Surgical findings and clinical follow-up were correlated with the CT findings. In nine patients CT was first used after emergency surgery and provided baseline data which was useful for further management. In two patients CT did not demonstrate small hepatic lacerations seen during previous surgery. No deaths were recorded. In 16 patients surgery followed CT within 24 h; there was good correlation between the CT and operative findings in 10 patients. However, CT failed to detect significant solid organ injury in five patients and was misleading (false positive) in another patient. There were two deaths amongst these 16 patients. Sixty-seven haemodynamically stable patients were initially managed non-operatively. Fifteen of these 67 patients had normal CT examinations; only one had subsequent laparotomy (for reasons unconnected with the trauma) where no injury was detected; there were no deaths. Of the 52 patients with an abnormal CT examination, 43 were successfully managed non-operatively. There were three deaths, including one where an injury missed at CT contributed to the demise of the patient. After an initial trial of non-operative management, the remaining six patients went to surgery where there was good concordance with the CT findings except for one missed renal injury. Active non-operative management of blunt abdominal trauma is widely accepted in haemodynamically stable patients and this report shows how CT supports this policy of surgical restraint in such cases. However, on review CT missed 13 injuries in nine patients overall; stricter attention to technique and better equipment may lead to improved results in the future.     

Psychoactive substance use disorder in relatives of patients with anorexia nervosa.

We assessed the lifetime prevalence and morbid risk of psychoactive substance use disorder (SUD; alcoholism and drug use disorder) in the first- and second-degree relatives, excluding children, of 34 female patients with anorexia nervosa (AN) and 34 age- and sex-matched controls who had no history of an eating disorder. Diagnoses of relatives were made blind to probands' diagnoses. The prevalence of SUD was 9% in both anorectic and control relatives, and the figures for morbid risk were 14% and 15%, respectively; these differences were nonsignificant. These results suggest that adolescent and adult women with AN do not possess many of the familial factors that predispose to the development of psychoactive SUD.     

Effects of the antimigraine compound zolmitriptan ('Zomig') on psychomotor performance alone and in combination with diazepam in healthy volunteers

Zolmitriptan (Zomig^TM) is a 5HT_1B/1D agonist which has the ability to cross the intact blood-brain barrier to access central as well as peripheral receptors. Because of the potential for central nervous system side effects, this randomized, double-blind, placebo-controlled, 6-period crossover study evaluated the effects of 2.5 and 5 mg doses of zolmitriptan on psychomotor performance and investigated any pharmacodynamic or pharmacokinetic interaction with diazepam. Twelve healthy volunteers received the following "treatments" as single doses: zolmitriptan 2.5 mg, zolmitriptan 5 mg, diazepam 10 mg, zolmitriptan 2.5 mg+diazepam 10 mg, zolmitriptan 5 mg+diazepam 10 mg and placebo. Pre-dose and at 1, 4, 8, and 24 h post-dose, the following validated battery of psychomotor tests was performed: Bond-Lader visual analogue scales (calmness, contentedness, and alertness factors), critical flicker fusion test, choice reaction time (recognition, motor, and total reaction times), finger-tapping test, number cancellation test and digit symbol substitution test. Plasma concentrations of zolmitriptan, its active metabolite, and diazepam and its active metabolites were measured at the same timepoints. Zolmitriptan 2.5 and 5 mg had no effect on psychomotor function when given alone. In contrast, diazepam 10 mg had profound effects, consistent with its sedative properties, but there was no synergism on concomitant administration of either dose of zolmitriptan. Plasma concentrations of zolmitriptan, diazepam, and their respective active metabolites were similar when the two drugs were given alone or in combination.