Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “Imputation”

Search indexed PubMed citations on genomics, clinical trials, systematic reviews and public health. Explore titles, authors and supplied subject terms, then open the PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 523 records · Page 29Linked to original sources
PubMed1997

[Hypersensitive urticarial vasculitis after natisedine intake].

BACKGROUND: Various skin and mucosal reactions can be observed after administration of quinidine derivatives. CASE REPORT: A patient who was taking Natisédine (quinidine phenylethyl barbiturate) intermittently and at reintroduction developed a papulopurpuric eruption (without thrombopenia) producing extensive centrifugal annular infiltration and central healing which regressed approximately one week after drug withdrawal. This eruption was associated with moderate 24 h proteinuria. The clinical aspect was that of vasculitis purpura as confirmed histology. Direct immunofluorescence only demonstrated C3 deposits in the vessel walls of the superficial dermis. The quinidine moiety of this drug (currently removed from the formulation) appears to be the responsible agent (imputability score: 13 B3). DISCUSSION: Thrombopenic purpura by synthesis of anti-platelet antibodies induced by quinidine derivatives is well known. Inversely, cases of non-thrombopenic purpura imputable to these same derivatives is uncommon (7 reported cases). The pathophysiological mechanisms involved might be similar (antigenic similarity between the platelet surface and endothelium).

Aged↗
PubMed2005

The global distribution of risk factors by poverty level.

OBJECTIVE: To estimate the individual-level association of income poverty with being underweight, using tobacco, drinking alcohol, having access only to unsafe water and sanitation, being exposed to indoor air pollution and being obese. METHODS: Using survey data for as many countries as possible, we estimated the relative risk association between income or assets and risk factors at the individual level within 11 medium- and low-income subregions of WHO. WHO and The World Bank data on the prevalence of risk factors and income poverty (defined as living on < US$ 1.00 per day, US$ 1-2.00 per day and > US$ 2.00 per day) were analysed to impute the association between poverty and risk factors for each subregion. The possible effect of poverty reduction on the prevalence of risk factors was estimated using population-attributable risk percentages. FINDINGS: There were strong associations between poverty and malnutrition among children, having access only to unsafe water and sanitation, and being exposed to indoor air pollution within each subregion (relative risks were twofold to threefold greater for those living on < US$ 1.00 per day compared with those living on > US$ 2.00 per day). Associations between poverty and obesity, tobacco use and alcohol use varied across subregions. If everyone living on < US$ 2.00 per day had the risk factor profile of those living on > US$ 2.00 per day, 51% of exposures to unimproved water and sanitation could be avoided as could 37% of malnutrition among children and 38% of exposure to indoor air pollution. The more realistic, but still challenging, Millennium Development Goal of halving the number of people living on < US$ 1.00 per day would achieve much smaller reductions. CONCLUSION: To achieve large gains in global health requires both poverty eradication and public health action. The methods used in this study may be useful for monitoring pro-equity progress towards Millennium Development Goals.

Health Services Accessibility↗
PubMedDate not supplied

A search for sex differences in response to analgesia.

BACKGROUND: It is generally accepted that males and females respond differently to painful conditions. With few exceptions, according to the published literature, females demonstrate a lower pain threshold and a lower tolerance of painful stimuli. There is some support in the literature that females experience greater analgesic efficacy than do males after the administration of narcotic analgesics. We compared the analgesic response of females and males to ibuprofen in a post-third-molar extraction dental pain model. METHODS: We performed a meta-analysis of 314 subjects included in the ibuprofen treatment arm of 7 double-blind, post-third-molar extraction dental pain (moderate to severe) studies, which were submitted to the agency electronically. The inclusion and exclusion criteria were practically identical in all studies. Pain relief and pain intensity measurements used the same metrics in all studies and were recorded just before and at least at 0.5, 1.0, 1.5, 2.0, 3.0, 4.0, 5.0, and 6. 0 hours after drug administration. RESULTS: The study included 195 female subjects and 119 male subjects (mean age, 21 years). Other than requiring dental extractions, the subjects were all healthy. Postoperative baseline pain was greater in females than in males to a statistically significant degree (P =.006). Both pain intensity and pain relief scores demonstrated the well-established analgesic effect of ibuprofen in the pooled data set as well as in all the individual studies. Moreover, the mean pain intensity and pain relief scores over time for the female and male treatment groups were not noticeably different at any time point after drug administration, with no imputation for missing values. Analysis of the data using the "baseline observation carried-forward" technique for remedicated subjects (the technique recommended by the Food and Drug Administration for efficacy analysis of acute analgesic medications) produced the same results, which were confirmed by analysis of variance and t tests at each time point of the study. CONCLUSIONS: Our results demonstrated no sex effect on the analgesic response to ibuprofen. These results were obtained under the post-third-molar extraction setting, in which the least possible confounding factors are present. To fully establish the generality of this phenomenon, studies should be carried out in other pain models and using analgesic medications with different mechanisms of action. Arch Intern Med. 2000;160:3424-3428.

Adult↗
PubMedDate not supplied

A randomized trial comparing human e-mail counseling, computer-automated tailored counseling, and no counseling in an Internet weight loss program.

BACKGROUND: Several studies have shown that e-mail counseling improves weight loss achieved in self-directed Internet programs. Computer-tailored feedback offers a population-based alternative to human e-mail counseling. METHODS: One hundred ninety-two adults, aged 49.2 +/- 9.8 years, having a body mass index (calculated as weight in kilograms divided by height in meters squared) of 32.7 +/- 3.5, were randomized to 1 of 3 Internet treatment groups: No counseling, computer-automated feedback, or human e-mail counseling. All participants received 1 weight loss group session, coupons for meal replacements, and access to an interactive Web site. The human e-mail counseling and computer-automated feedback groups also had access to an electronic diary and message board. The human e-mail counseling group received weekly e-mail feedback from a counselor, and the computer-automated feedback group received automated, tailored messages. RESULTS: Retention was 82% at 3 months and 80% at 6 months for all 3 groups. At 3 months, completers in both the computer-automated feedback (-5.3 +/- 4.2 kg) and human e-mail counseling (-6.1 +/- 3.9 kg) groups had significantly greater weight losses compared with the no counseling group (-2.8 +/- 3.5 kg) and these groups did not differ from each other. At 6 months, weight losses were significantly greater in the human e-mail counseling group (-7.3 +/- 6.2 kg) than in the computer-automated feedback (-4.9 +/- 5.9 kg) or no counseling (-2.6 +/- 5.7 kg) groups. Intent-to-treat analyses using single or multiple imputation techniques showed the same pattern of significance. CONCLUSIONS: Providing automated computer-tailored feedback in an Internet weight loss program was as effective as human e-mail counseling at 3 months. Further research is needed to improve the efficacy of automated computer-tailored feedback as a population-based weight loss approach.

Adult↗
PubMed1997

Epidemiologic study of time and space clustering of 4 cases of choroidal malignant melanoma.

OBJECTIVE: To identify common risk factors that might be associated with a cluster of 4 cases of choroidal malignant melanoma that occurred in a manufacturing plant between 1982 and 1985. DESIGN: Survey of choroidal malignant melanoma cases in Caddo and Bossier parishes during the same time frame. METHODS: We identified 4 additional individuals with choroidal malignant melanoma first diagnosed during the study period. Characteristics of the workplace were examined and a questionnaire was administered to all subjects to ascertain exposures to putative carcinogens and to putative risk factors for intraocular malignant melanoma. Observed and expected incidence rates were calculated. RESULTS: The overall incidence of intraocular malignant melanoma for the 2 parishes during the 4-year period was 0.56 cases per 100,000 population per year, similar to the expected rates for this population. The incidence rate for the manufacturing plant employees was 16.5 cases per 100,000 per year. The mean age at diagnosis for workers at the plant was 38.7 years compared with 69.2 years for nonplant employees. CONCLUSIONS: The close occurrence of manufacturing plant cases in time and space and the younger age of the people are consistent with a cluster of intraocular melanoma cases. No specific risk factors were found in the workplace environment to impute a causal association.

Adult↗
PubMed1981

Borderline personality and the Rorschach test.

Rorschach responses of borderline persons, acute and chronic schizophrenics, normals, and neurotics were compared on summary, composite, and fabulized combination scores and on a score reflecting decline in the quality of responses to individual cards. The groups' summary scores were as ego function theory would predict; normals had the highest scores, followed by neurotics, borderline persons, acute schizophrenics, and chronic schizophrenics. In a three-group comparison, discriminant-function analysis correctly classified most of the borderline and acute and chronic schizophrenic subjects. In a two-group comparison, stepwise regression analysis correctly classified most of the borderline and acute schizophrenic subjects. The borderline persons tended to produce more fabulized combination responses and show a greater decline in response quality on each card. The associative drift and sporadic reasoning problems imputed to borderline persons clinically distinguished the borderline sample's Rorschach records.

Adolescent↗
PubMed1998

Prophylactic abdominal drainage after elective colonic resection and suprapromontory anastomosis: a multicenter study controlled by randomization. French Associations for Surgical Research.

BACKGROUND: Only 4 controlled trials have investigated whether prophylactic abdominal drainage was of value after colonic resection. None have been able to find any statistically significant difference, but the number of patients was small and the beta error risk was high. OBJECTIVES: To compare patients who underwent abdominal drainage with those who did not for the rate and severity of complications after elective colonic resection followed immediately by anastomosis of the suprapromontory colon and to compare suction drains with nonsuction drains. PATIENTS: Between September 1990 and June 1995, 319 patients (135 men and 184 women), whose mean age was 67 years (range, 22-95 years), with carcinoma, benign tumors, or colitis, located anywhere between the ascending and sigmoid colons, were included in the study. Patients were comparable for demographic characteristics, except that there were more patients with ascites in the group that did not undergo abdominal drainage (P<.02). INTERVENTIONS: After 2 protocol violations, 156 patients were randomized to the abdominal drainage group and 161 to the no abdominal drainage group. All 317 anastomoses were tested for airtightness intraoperatively and repaired if leakage was found (n=71), and all patients with anastomoses received a routine diatrizoate sodium enema to detect infraclinical leakage. MAIN OUTCOME MEASURES: The postoperative complications possibly influenced by drainage included (1) deep complications for which drainage can lead to early diagnosis, such as generalized or localized peritonitis, intraabdominal hemorrhage, or hematoma; (2) complications believed to be enhanced by drainage, such as an operative wound (an abscess, disruption, or incisional hernia) or pulmonary (microatelectasis) and intestinal obstructions; and (3) complications directly due to the drains, such as ulcerations leading to fistulae, hemorrhages, drainage tract infections, difficulty in removal, intra-abdominal retention, and incisional disruptions. Subsidiary end points were the severity of these complications as assessed by the number of related subsequent operations and deaths. RESULTS: Twenty-six patients overall (8%) had postoperative complications possibly influenced by drainage (9% in the group that underwent abdominal drainage and 8% in the group that did not). This difference was not statistically significant (P<.90). One patient had a fistula directly imputable to drainage. There was no difference between suction and nonsuction drainage (P<.90). CONCLUSIONS: Routine abdominal drainage after colonic resection and immediate anastomosis decreases neither the rate nor the severity of anastomotic leakage. It can, occasionally, be detrimental.

Abdomen↗
PubMed1995

Efficacy and safety of lowering dietary intake of fat and cholesterol in children with elevated low-density lipoprotein cholesterol. The Dietary Intervention Study in Children (DISC). The Writing Group for the DISC Collaborative Research Group.

OBJECTIVE: To assess the efficacy and safety of lowering dietary intake of total fat, saturated fat, and cholesterol to decrease low-density lipoprotein cholesterol (LDL-C) levels in children. DESIGN: Six-center randomized controlled clinical trial. PARTICIPANTS: Prepubertal boys (n = 362) and girls (n = 301) aged 8 to 10 years with LDL-C levels greater than or equal to the 80th and less than the 98th percentiles for age and sex were randomized into an intervention group (n = 334) and a usual care group (n = 329). INTERVENTION: Behavioral intervention to promote adherence to a diet providing 28% of energy from total fat, less than 8% from saturated fat, up to 9% from polyunsaturated fat, and less than 75 mg/4200 kJ (1000 kcal) per day of cholesterol (not to exceed 150 mg/d). MAIN OUTCOME MEASURES: The primary efficacy measure was the mean LDL-C level at 3 years. Primary safety measures were mean height and serum ferritin levels at 3 years. Secondary efficacy outcomes were mean LDL-C levels at 1 year and mean total cholesterol levels at 1 and 3 years. Secondary safety outcomes included red blood cell folate values; serum zinc, retinol, and albumin levels; serum high-density lipoprotein cholesterol (HDL-C) values, LDL-C:HDL-C ratio, and total triglyceride levels; sexual maturation; and psychosocial health. RESULTS: At 3 years, dietary total fat, saturated fat, and cholesterol levels decreased significantly in the intervention group compared with the usual care group (all P < .001). Levels of LDL-C decreased in the intervention and usual care groups by 0.40 mmol/L (15.4 mg/dL) and 0.31 mmol/L (11.9 mg/dL), respectively. Adjusting for baseline level and sex and imputting values for missing data, the mean difference between the groups was -0.08 mmol/L (-3.23 mg/dL) (95% confidence interval [CI], -0.15 to -0.01 mmol/L [-5.6 to -0.5 mg/dL]), which was significant (P = .02). There were no significant differences between the groups in adjusted mean height or serum ferritin levels (P > .05) or other safety outcomes. CONCLUSIONS: The dietary intervention achieved modest lowering of LDL-C levels over 3 years while maintaining adequate growth, iron stores, nutritional adequacy, and psychological well-being during the critical growth period of adolescence.

Analysis of Variance↗
PubMed2002

Effects of hormone replacement therapy and antioxidant vitamin supplements on coronary atherosclerosis in postmenopausal women: a randomized controlled trial.

CONTEXT: Hormone replacement therapy (HRT) and antioxidant vitamins are widely used for secondary prevention in postmenopausal women with coronary disease, but no clinical trials have demonstrated benefit to support their use. OBJECTIVE: To determine whether HRT or antioxidant vitamin supplements, alone or in combination, influence the progression of coronary artery disease in postmenopausal women, as measured by serial quantitative coronary angiography. DESIGN, SETTING, AND PATIENTS: The Women's Angiographic Vitamin and Estrogen (WAVE) Trial, a randomized, double-blind trial of 423 postmenopausal women with at least one 15% to 75% coronary stenosis at baseline coronary angiography. The trial was conducted from July 1997 to January 2002 in 7 clinical centers in the United States and Canada. INTERVENTIONS: Patients were randomly assigned in a 2 x 2 factorial design to receive either 0.625 mg/d of conjugated equine estrogen (plus 2.5 mg/d of medroxyprogesterone acetate for women who had not had a hysterectomy), or matching placebo, and 400 IU of vitamin E twice daily plus 500 mg of vitamin C twice daily, or placebo. MAIN OUTCOME MEASURE: Annualized mean (SD) change in minimum lumen diameter (MLD) from baseline to concluding angiogram of all qualifying coronary lesions averaged for each patient. Patients with intercurrent death or myocardial infarction (MI) were imputed the worst rank of angiographic outcome. RESULTS: The mean (SD) interval between angiograms was 2.8 (0.9) years. Coronary progression, measured in mean (SD) change, worsened with HRT by 0.047 (0.15) mm/y and by 0.024 (0.15) mm/y with HRT placebo (P =.17); and for antioxidant vitamins by 0.044 (0.15) mm/y and with vitamin placebo by 0.028 (0.15) mm/y (P =.32). When patients with intercurrent death or MI were included, the primary outcome showed an increased risk for women in the active HRT group (P =.045), and suggested an increased risk in the active vitamin group (P =.09). Fourteen patients died in the HRT group and 8 in the HRT placebo group (hazard ratio [HR], 1.8; 95% confidence interval [CI], 0.75-4.3), and 16 in the vitamin group and 6 in the vitamin placebo group (HR, 2.8; 95% CI, 1.1-7.2). Death, nonfatal MI, or stroke occurred in 26 HRT patients vs 15 HRT controls (HR, 1.9; 95% CI, 0.97-3.6) and in 26 vitamin patients and 18 vitamin controls (HR, 1.5; 95% CI, 0.80-2.9). There was no interaction between the 2 treatment interventions. CONCLUSION: In postmenopausal women with coronary disease, neither HRT nor antioxidant vitamin supplements provide cardiovascular benefit. Instead, a potential for harm was suggested with each treatment.

Aged↗
PubMed2006

Effect of rimonabant, a cannabinoid-1 receptor blocker, on weight and cardiometabolic risk factors in overweight or obese patients: RIO-North America: a randomized controlled trial.

CONTEXT: Rimonabant, a selective cannabinoid-1 receptor blocker, may reduce body weight and improve cardiometabolic risk factors in patients who are overweight or obese. OBJECTIVE: To compare the efficacy and safety of rimonabant with placebo each in conjunction with diet and exercise for sustained changes in weight and cardiometabolic risk factors over 2 years. DESIGN, SETTING, AND PARTICIPANTS: Randomized, double-blind, placebo-controlled trial of 3045 obese (body mass index > or =30) or overweight (body mass index >27 and treated or untreated hypertension or dyslipidemia) adult patients at 64 US and 8 Canadian clinical research centers from August 2001 to April 2004. INTERVENTION: After a 4-week single-blind placebo plus diet (600 kcal/d deficit) run-in period, patients were randomized to receive placebo, 5 mg/d of rimonabant, or 20 mg/d of rimonabant for 1 year. Rimonabant-treated patients were rerandomized to receive placebo or continued to receive the same rimonabant dose while the placebo group continued to receive placebo during year 2. MAIN OUTCOME MEASURES: Body weight change over year 1 and prevention of weight regain during year 2. Additional efficacy measures included changes in waist circumference, plasma lipid levels, and other cardiometabolic risk factors. RESULTS: At year 1, the completion rate was 309 (51%) patients in the placebo group, 620 (51%) patients in the 5 mg of rimonabant group, and 673 (55%) patients in the 20 mg of rimonabant group. Compared with the placebo group, the 20 mg of rimonabant group produced greater mean (SEM) reductions in weight (-6.3 [0.2] kg vs -1.6 [0.2] kg; P<.001), waist circumference (-6.1 [0.2] cm vs -2.5 [0.3] cm; P<.001), and level of triglycerides (percentage change, -5.3 [1.2] vs 7.9 [2.0]; P<.001) and a greater increase in level of high-density lipoprotein cholesterol (percentage change, 12.6 [0.5] vs 5.4 [0.7]; P<.001). Patients who were switched from the 20 mg of rimonabant group to the placebo group during year 2 experienced weight regain while those who continued to receive 20 mg of rimonabant maintained their weight loss and favorable changes in cardiometabolic risk factors. Use of different imputation methods to account for the high rate of dropouts in all 3 groups yielded similar results. Rimonabant was generally well tolerated; the most common drug-related adverse event was nausea (11.2% for the 20 mg of rimonabant group vs 5.8% for the placebo group). CONCLUSIONS: In this multicenter trial, treatment with 20 mg/d of rimonabant plus diet for 2 years promoted modest but sustained reductions in weight and waist circumference and favorable changes in cardiometabolic risk factors. However, the trial was limited by a high drop-out rate and longer-term effects of the drug require further study. Clinical Trials Registration ClinicalTrials.gov Identifier: NCT00029861.

Adult↗
PubMed2000

Cost analysis of filgrastim for the prevention of neutropenia in pediatric T-cell leukemia and advanced lymphoblastic lymphoma: a case for prospective economic analysis in cooperative group trials.

BACKGROUND: Growth factor use has been shown to ameliorate chemotherapy-induced neutropenia, leading to shorter hospital stays and lower use of parenteral antibiotics, two costly areas of cancer treatment. Prior reports on pediatric patients have shown evidence of cost savings in some studies, but no such evidence in others. In this study a retrospective analysis compared the costs of inpatient supportive care for pediatric patients with T-cell leukemia and advanced lymphoblastic lymphoma enrolled in a Pediatric Oncology Group trial. PROCEDURE: Patients 1-22 years of age were randomized to receive either granulocyte colony-stimulating factor (G-CSF; n = 45) or no G-CSF (n = 43) following induction and two cycles of maintenance therapy. There were no significant differences in neutropenia-related outcomes during the induction phase. During maintenance therapy, G-CSF patients had significantly fewer days to an ANC >500 cells/microl and a trend towards fewer days of hospitalization. Data on resource utilization were tabulated from case report forms. Costs were imputed from national data on hospitalization costs, average wholesale prices of pharmaceuticals, and patient billing information from a single institution. RESULTS: Total median costs of supportive care were $34,190 for patients receiving G-CSF and $28,653 for patients not receiving G-CSF (P > 0. 05 for the cost difference). Sensitivity analyses demonstrated that the total cost difference was not statistically significant, even in scenarios that included reasonable variations in estimates of the range of the length of stay, antibiotic regimen, and dosage and cost of G-CSF. CONCLUSIONS: In the setting of pediatric leukemia, the cost of growth factor may offset potential savings from shorter hospital stays or lower antibiotic use, a finding consistent with that from the Children's Cancer Study Group.

Adolescent↗
PubMed1998

Computer-assisted spine surgery.

The aim of this study was to improve the reliability of pedicle screw insertion. Transpedicle screw insertion may cause neurological, vascular, and mechanical complications. Previous studies of surgical procedures have shown a significant rate of incorrect placement of the screw ranging from 10 to 40%. A new technique that combines preoperative computed tomography (CT) imaging with intraoperative passive navigation was used to perform 64 pedicle screw insertions in the thoracolumbar region. At the same time, 64 pedicle screw insertions were performed manually in the same region and on the same vertebral levels. Surgery was followed in all cases by postoperative radiographs and computed tomography examination, which allowed measurements of screw position relative to pedicle position to be performed. A comparison between the two groups showed that six screws in 64 vertebra (9%) had incorrect placement with the computer-assisted technique whereas 28 screws in 64 vertebra (44%) had incorrect placement with manual insertion. The intraoperative accuracy provided by the computer after registration was better than 1 mm. The good results obtained are similar to those reported in the literature. The cortex penetration observed with the computer-assisted technique was not imputed to computer failure. Errors by the surgeon in acquiring data in the pre- and perioperative steps may explain the six incorrect screw placements. This clinical experience confirms that the accuracy and the reliability of this computer-assisted technique are good.

Bone Screws↗
PubMed1996

The incidence of cystic fibrosis.

We develop a statistical model for estimating cystic fibrosis (CF) incidence among infants born in the U.S.A. that accounts for under-diagnosis due to death prior to diagnosis and we apply it to the Cystic Fibrosis Foundation Patient Registry data for the years 1989 to 1991. The resulting estimate of incidence relative to live births among whites is 1:3419 while that among non-whites is 1:12,163. As a by-product of the modelling approach, estimates of the underlying average diagnosis age given survival to diagnosis are 4.09 years for whites and 4.55 years for non-whites, but this difference was not statistically significant and appears to demonstrate that diagnosis efforts may be approximately the same for whites and non-whites. Also, as another by-product of the modelling approach, CF mortality was estimated as more severe for females than males and marginally more severe for non-whites than whites. A variety of statistical methods underlie achievement of these results, including semi-parametric maximum likelihood, survival analysis, multiple imputation and bootstrap techniques.

Actuarial Analysis↗
PubMed1996

Explained variation in survival analysis.

Several measures of explained variation have been suggested for the Cox proportional hazards regression model. We have categorized these measures into three classes which correspond to three different definitions of multiple R2 of the general linear model. In an empirical study we compared the performance of these measures and classified them by their adherence to a set of criteria which we think should be met by a measure of explained variation for survival data. We suggest that currently there is no uniformly superior measure, particularly as the concepts of either uncensored or censored populations may lead to different choices. For uncensored populations, a measure by Kent and O'Quigley and the squared rank correlation between survival time and the predictor from a Cox regression model appear recommendable choices. For the latter, censored survival times are terminated using a very recent data augmentation algorithm for multiple imputation under proportional hazards. With censored populations, Schemper's measure, V2, could be considered. We give an introductory example, discuss aspects of application and stress the desirability of routinely evaluating explained variation in studies of survival.

Humans↗
PubMedDate not supplied

Models for the incubation of AIDS and variations according to age and period.

The objective of the methods proposed is to provide a parametric model for the incubation of AIDS and to use the chosen parameterization to test for the effect of age at seroconversion, and, after adjusting for markers of immunosuppression, to assess variations in periods corresponding to different levels of use of AIDS therapies at the population level. We compared the fit of Weibull, log-normal and three-parameter logistic models incorporating truncation in prevalent cohort and interval censored data. We showed the advantages by restricting the analysis to follow-up durations of greater than five years to improve estimation of the tail of the distribution for the prediction of long-term survivors. We applied the proposed methods to the combination of 1649 seroprevalent and 476 seroconverters with 1022 and 177 AIDS cases, respectively, who have been followed in the Multicenter AIDS Cohort Study (MACS) up to April 1995. Differences according to age at seroconversion are quantified in terms of relative percentiles and their associated 95 per cent confidence intervals were calculated using methods of multiple imputation. Using the proposed methods, we found that the log-normal model provides a fit as good as the three-parameter logistic; both are close to the non-parametric estimate and are significantly better than the fit of the Weibull model. We found that the older the age at seroconversion, the shorter the time to AIDS (relative percentile = 0.72 for age > or = 40 versus age < 25), and that the incubation of AIDS in calendar periods where treatment has been widely administered has been significantly longer among individuals with low CD4 cell counts.

Acquired Immunodeficiency Syndrome↗
PubMed1997

An examination of methods for sample size recalculation during an experiment.

In designing experiments, investigators frequently can specify an important effect that they wish to detect with high power, without the ability to provide an equally certain assessment of the variance of the response. If the experiment is designed based on a guess of the variance, an under-powered study may result. To remedy this problem, there have been several procedures proposed that obtain estimates of the variance from the data as they accrue and then recalculate the sample size accordingly. One class of procedures is fully sequential in that it assesses after each response whether the current sample size yields the desired power based on the current estimate of the variance. This approach is efficient, but it is not practical or advisable in many situations. Another class of procedures involves only two or three stages of sampling and recalculates the sample size based on the observed variance at designated times, perhaps coinciding with interim efficacy analyses. The two-stage approach can result in substantial oversampling, but it is feasible in many situations, whereas the three-stage approach corrects the problem of oversampling, but is less feasible. We propose a procedure that aims to combine the advantages of both the fully sequential and the two-stage approaches. This quasi-sequential procedure involves only two stages of sampling and it applies to the stopping rule from the fully sequential procedure to data beyond the initial sample which we obtain via multiple imputation. We show through simulations that when the initial sample size is substantially less than the correct sample size, the mean squared error of the final sample size calculated from the quasi-sequential procedure can be considerably less than that from the two-stage procedure. We compare the distributions of these recalculated sample sizes and discuss our findings for alternative procedures, as well.

Anti-HIV Agents↗
PubMedDate not supplied

Incomplete quality of life data in randomized trials: missing forms.

Analysing quality of life (QOL) data may be complicated for several reasons, such as: repeated measures are obtained; data may be collected on ordered categorical responses; the instrument may have multidimensional scales, and complete data may not be available for all patients. In addition, it may be necessary to integrate QOL with length of life. The major undesirable effects of missing data, in QOL research, are the introduction of biases due to inadequate modes of analysis and the loss of efficiency due to reduced sample sizes. Currently, there is no standard method for handling missing data in QOL studies. In fact, there are very few references to methods of handling missing data in this context. The aim of this paper is to provide an overview of methods for analysing incomplete longitudinal QOL data which have either been presented in the QOL literature or in the missing data literature. These methods of analysis include complete case, available case, summary measures, imputation and likelihood-based approaches. We also discuss the issue of bias and the need for sensitivity analyses.

Bias↗
PubMedDate not supplied

Suggestions for the presentation of quality of life data from clinical trials.

Quality of life (QOL) data is complex since it is both multidimensional and longitudinal. This complexity is compounded with its unbalanced nature through missing observations as a consequence of patient non-compliance with assessment schedules, and, for example, in cancer clinical trials data absence due to patient attrition often through death. QOL data poses difficulties for presentation and analysis and hence interpretation. This paper illustrates, using data from a randomized trial of the United Kingdom Medical Research Council Lung Cancer Working Party, a step-by-step approach to presentation of QOL data. This begins with a description of compliance and its relationship with patient attrition caused by death, to a final summary profile to indicate change over time. We recognize that no single summary statistic is likely to be able to encapsulate all the subtleties of QOL data. We stress the importance of examining data graphically before performing detailed analysis and also to facilitate interpretation in the final clinical report. Although a description of analytical methods is not the purpose of this paper, we draw attention to the need for imputing missing values and to the (multi-level) modelling approach to summarizing the data, both essential adjuncts to the less formal methods described here.

Authorship↗