Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “missing data”

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 1,189 records · Page 66Linked to original sources
PubMed2006

Combined calcium, magnesium and potassium supplementation for the management of primary hypertension in adults.

BACKGROUND: Previous research suggests that increasing dietary intakes of calcium, potassium or magnesium separately may reduce BP to a small degree over the short term. It is unclear whether increasing intakes of a combination of these minerals produces a larger reduction in BP. OBJECTIVES: To evaluate the effects of combined mineral supplementation as a treatment for primary hypertension in adults. SEARCH STRATEGY: We searched the Cochrane Library, MEDLINE, EMBASE, Science Citation Index, ISI Proceedings, ClinicalTrials.gov, Current Controlled Trials, CAB abstracts, and reference lists of systematic reviews, meta-analyses and randomised controlled trials (RCTs) included in the review. The search was unrestricted by language or publication status. SELECTION CRITERIA: Inclusion criteria were: 1) RCTs of a parallel or crossover design comparing oral supplements comprising a combination of potassium, and/or calcium, and/or magnesium with placebo, no treatment, or usual care; 2) treatment and follow-up >=8 weeks; 3) participants over 18 years old, with raised systolic blood pressure (SBP) >=140 mmHg or diastolic blood pressure (DBP) >=85 mmHg with no known primary cause; 4) SBP and DBP reported at end of follow-up. We excluded trials where participants were pregnant, or received antihypertensive medication which changed during the study. DATA COLLECTION AND ANALYSIS: Two reviewers independently extracted data and assessed trial quality. Disagreements were resolved by discussion or a third reviewer. Random effects meta-analyses and sensitivity analyses were conducted. MAIN RESULTS: We included three RCTs (n=277) with between 24 and 28 weeks follow-up. Three combinations of minerals were investigated: potassium & magnesium, calcium & magnesium, and calcium & potassium. One trial investigated combinations of calcium & magnesium and of calcium & potassium, and for each found a statistically non-significant increase in both SBP and DBP. All three trials investigated the combination of potassium & magnesium. None of the trials provided data on mortality or morbidity. The combination of potassium & magnesium compared to control resulted in statistically non-significant reductions in both SBP (mean difference = -4.6 mmHg, 95% CI: -9.9 to 0.7) and DBP (mean difference = -3.8 mmHg, 95% CI: -9.5 to 1.8), although the results were heterogeneous (I(2)=68% and 85% for SBP and DBP respectively).A sensitivity analysis using alternative reported values which accounted for missing data had very little effect on DBP but resulted in a larger, statistically significant reduction in SBP (mean difference = -5.8 mmHg, 95% CI: -10.5 to -1.0). The quality of the trials was not well reported. AUTHORS' CONCLUSIONS: We found no robust evidence that supplements of any combination of potassium, magnesium or calcium reduce mortality, morbidity or BP in adults. More trials are needed to investigate whether the combination of potassium & magnesium is effective.

Adult↗
PubMed1993

Extrapolation of incomplete marker tracks by lower rank approximation.

Motion and deformation of an object such as the heart may be measured by tracking optical or radiopaque markers. In the experimental situation markers may fail to be detected due to occlusion or lack of contrast. As a result a continuous marker track is observed in separated parts, which often cannot be directly identified as corresponding to one marker. This paper presents a method of extrapolating a partly known track by using information provided by the known track part and the available complete tracks of other markers. The extrapolations are obtained by iteratively fitting a lower rank matrix to the set of noisy, incomplete marker tracks. The performance is evaluated with computer-simulated data and data obtained in an animal experiment. In both cases 43% of the available complete tracks were made incomplete by removal of track parts varying in length from 3% up to 44%. For the simulated data comparison of the extrapolations with true signal values results in a root mean square (RMS) error about equal to the noise level. For the animal experiment, when comparing the extrapolations with the measured values, in images of 256 x 256 pixels, the RMS error was found to be +/- 0.5 pixel, which is quite small relative to the total excursion of a marker (20 pixels). Estimation of the missing data by applying BMDPAM (BMDP Statistical Software Inc.) to the same data results in RMS errors which are about twice as high.

Algorithms↗
PubMedDate not supplied

Correspondence analysis of the Spanish National Health Survey.

This report gives a comprehensive explanation of the multivariate technique called correspondence analysis, applied in the context of a large survey of a nation's state of health, in this case the Spanish National Health Survey. It is first shown how correspondence analysis can be used to interpret a simple cross-tabulation by visualizing the table in the form of a map of points representing the rows and columns of the table. Combinations of variables can also be interpreted by coding the data in the appropriate way. The technique can also be used to deduce optimal scale values for the levels of a categorical variable, thus giving quantitative meaning to the categories. Multiple correspondence analysis can analyze several categorical variables simultaneously, and is analogous to factor analysis of continuous variables. Other uses of correspondence analysis are illustrated using different variables of the same Spanish database: for example, exploring patterns of missing data and visualizing trends across surveys from consecutive years.

Adolescent↗
PubMed1999

Understanding Rasch measurement: estimation methods for Rasch measures.

Rasch parameter estimation methods can be classified as non-interative and iterative. Non-iterative methods include the normal approximation algorithm (PROX) for complete dichotomous data. Iterative methods fall into 3 types. Datum-by-datum methods include Gaussian least-squares, minimum chi-square, and the pairwise (PAIR) method. Marginal methods without distributional assumptions include conditional maximum-likelihood estimation (CMLE), joint maximum-likelihood estimation (JMLE) and log-linear approaches. Marginal methods with distributional assumptions include marginal maximum-likelihood estimation (MMLE) and the normal approximation algorithm (PROX) for missing data. Estimates from all methods are characterized by standard errors and quality-control fit statistics. Standard errors can be local (defined relative to the measure of a particular item) or general (defined relative to the abstract origin of the scale). They can also be ideal (as though the data fit the model) or inflated by the misfit to the model present in the data. Five computer programs, implementing different estimation methods, produce statistically equivalent estimates. Nevertheless, comparing estimates from different programs requires care.

Algorithms↗
PubMed1995

[Non-significant in univariate but significant in multivariate analysis: a discussion with examples].

Perhaps as a result of higher research standard and advancement in computer technology, the amount and level of statistical analysis required by medical journals become more and more demanding. It is now realized by researchers that univariate analysis alone may not be sufficient, especially for complex data sets. Additional, and sometimes even contradictory, results may be found using multivariate analysis. During the course of data analysis, a common practice is to include in multivariate analysis only those variables that are statistically significant in univariate analysis. Such a habit is risky as some variables not significant in univariate analysis may become significant in multivariate analysis. In this study, we identify, with examples, four possible scenarios in which the above situation could occur: (1) the effect of unbalanced sample size; (2) the influence of missing data; (3) an extremely large within group variation, relative to between group variation; and (4) the presence of interaction. In addition to detailed analysis steps, raw data sets are also available for readers to verify all the results presented. Although we only used the log-rank test and Cox regression for illustration purposes, the underlying concepts can be applied to other multivariate procedures such as the logistic regression and multiple linear regression.

Analysis of Variance↗
PubMedDate not supplied

Subcutaneous injection survey: psychometric evaluation of a treatment satisfaction instrument associated with a novel HIV medication.

PURPOSE: The 20-item Subcutaneous Injection Survey (SIS) was developed to assess patients' experience with subcutaneous self-injection of an innovative anti-HIV fusion inhibitor (T-20). It measures ease of use and impact of treatment on daily life. This study presents the results of a psychometric evaluation of the SIS. METHOD: The SIS and MOS-HIV data (N = 547) from two multinational, phase III clinical trials of optimized background therapy (OBT) vs. OBT + T-20 were utilized for this evaluation. Item performance and internal consistency reliability were assessed. Exploratory factor analysis was used to identify subscales. Pearson correlations were used to examine the relationship between the SIS and the MOS-HIV scores. RESULTS: Patients were predominately white males. SIS items had low rates of missing data (0.9%-5.5%) and with two exceptions were not redundant, as indicated by generally moderate item-to-item correlations. Three factors were identified: daily functioning, ease of injections, and activities of daily living. The SIS also includes three individually scored items on overall injection interference with daily activities, effect of injections on daily activities, and satisfaction with personal appearance. Cronbach's alphas ranged from 0.71 to 0.87. Correlations between SIS and MOS-HIV subscales were low to moderate. CONCLUSION: The SIS is a reliable instrument for measuring patients' assessment of self-injection in HIV clinical trials. Patients' experiences with self-injection are not strongly associated with health-related quality of life.

Adult↗
PubMed1983

Linked cross-sectional study for evaluating the effect on the microfilarial load of the onchocerciasis control programme.

When analysing incomplete longitudinal data (ie data which includes complete longitudinal results plus cross-sectional results plus partially longitudinal results) one can increase the sensitivity of the data by taking the correlation structure of the data into account. For the described data on onchocerciasis, the adapted method of Rao and Rao, called the Linked Cross-Sectional method, which estimates the mean microfilarial load for each passage and hence the differences between passages, taking into account the correlation between the same individuals measured at different times is such a method that reduces significantly the standard deviations of the mean responses and hence increases their sensitivity. The method takes into account the missing data nature of the results. It can be simply generalized to any number of passages.

Cross-Sectional Studies↗
PubMedDate not supplied

Interactive data collection: benefits of integrating new media into pediatric research.

Despite the prevalence of children's computerized games for recreational and educational purposes, the use of interactive technology to obtain pediatric research data remains underexplored. This article describes the development of laptop interactive data collection (IDC) software for a children's health intervention study. The IDC integrates computer technology, children's developmental needs, and quantitative research methods that are engaging for school-age children as well as reliable and efficient for the pediatric health researcher. Using this methodology, researchers can address common problems such as maintaining a child's attention throughout an assessment session while potentially increasing their response rate and reducing missing data rates. The IDC also promises to produce more reliable data by eliminating the need for manual double entry of data and reducing much of the time and costs associated with data cleaning and management. Development and design considerations and recommendations for further use are discussed.

Child↗
PubMed1996

An evaluation of the TransFER model for sharing clinical decision-support applications.

TransFER is a formal model designed to facilitate the sharing of decision-support applications across institutions with heterogeneous clinical databases. The TransFER model provides a mechanism to automatically customize database queries based on a reference schema of clinical data and an encoded set of database mappings. In this paper, we describe the elements of the TransFER model and we present the results of a formal evaluation we conducted to assess the utility and generality of the model. The results suggest that the TransFER has significant potential for automating query translation and facilitating application sharing, but that further work on the representation of temporal semantics, on the modeling of missing data, and on the optimization of complex queries is required.

Decision Making, Computer-Assisted↗
PubMed1997

Statistical methods for two-sequence three-period cross-over designs with incomplete data.

In clinical trials, and in bioavailability and bioequivalence studies, one often encounters replicate cross-over designs such as a two-sequence three-period cross-over design to assess treatment and carry-over effects of two formulations of a drug product. Because of the potential dropout (or for some administrative reason), however, the observed data set from a replicate cross-over design is incomplete or unbalanced so that standard statistical methods for a cross-over design may not apply directly. For inference on the treatment and carry-over effects, we propose a method based on differences of the observations that eliminates the random subject effects and thus does not require any distributional condition on the random subject effects. When no datum is missing, this method provides the same results as the ordinary least squares method. When there are missing data, the proposed method still provides exact confidence intervals for the treatment and carry-over effects, as long as the dropout is independent of the measurement errors. We provide an example for illustration.

Clinical Trials as Topic↗
PubMed1988

[Reduction of cerebral hemorrhage and respiratory distress syndrome in premature infants by avoiding perinatal asphyxia].

Intra- and periventricular haemorrhage (IVH/PVH) and, under certain conditions, the respiratory distress syndrome (RDS) seem to be typical sequelae of perinatal asphyxia in preterm born infants. Therefore, an association of IVH/PVH and RDS can be expected. We have retrospectively analyzed the data of 118 premature infants born between 1982 and 1986, weighing between 750 and 1499 g. 11 of these had experienced a severe IVH/PVH and a severe RDS at the same time, whereas 75 infants did not develop either of those. (2 of the 118 showed a severe IVH/PVH without evidence of severe RDS whereas 29 developed severe RDS without signs of serious IVH/PVH. 1 could not be evaluated due to missing data). This association of severe intracerebral haemorrhage and severe respiratory distress syndrome was statistically significant (p less than 0.005). The number of severe IVH/PVH has decreased during 1984-1986 in comparison to 1982/83 (4/76 vs. 9/42; p less than 0.05); the incidence of severe RDS has slightly declined. Comparing the perinatal conditions we found that the infants of the years 1984-1986 were more rarely delivered after an interval exceeding 24 h after premature rupture of the membranes (p less than 0.05), were more often delivered by caesarean section (p less than 0.005), and were nearly always primarily cared for by an experienced paediatrician (p less than 0.01). There were no significant differences between these two groups as far as dexamethasone-prophylaxis, mean birth weight, percentage of small-for-gestational-age infants and mean Apgar scores were concerned.(ABSTRACT TRUNCATED AT 250 WORDS)

Asphyxia Neonatorum↗
PubMed2000

Statistical considerations in the intent-to-treat principle.

This paper describes some of the statistical considerations in the intent-to-treat design and analysis of clinical trials. The pivotal property of a clinical trial is the assignment of treatments to patients at random. Randomization alone, however, is not sufficient to provide an unbiased comparison of therapies. An additional requirement is that the set of patients contributing to an analysis provides an unbiased assessment of treatment effects, or that any missing data are ignorable. A sufficient condition to provide an unbiased comparison is to obtain complete data on all randomized subjects. This can be achieved by an intent-to-treat design wherein all patients are followed until death or the end of the trial, or until the outcome event is reached in a time-to-event trial, irrespective of whether the patient is still receiving or complying with the assigned treatment. The properties of this strategy are contrasted with those of an efficacy subset analysis in which patients and observable patient data are excluded from the analysis on the basis of information obtained postrandomization. I describe the potential bias that can be introduced by such postrandomization exclusions and the pursuant effects on type I error probabilities. Especially in a large study, the inflation in type I error probability can be severe, 0.50 or higher, even when the null hypothesis is true. Standard statistical methods for the analysis of censored or incomplete observations all require the assumption of missing at random to some degree, and none of these methods adjust for the potential bias introduced by post hoc subset selection. Nor is such adjustment possible unless one posits a model that relates the missing observations to other observed information for each subject-models that are inherently untestable. Further, the subset selection bias is confounded with the subset-specific treatment effect, and the two components are not identifiable without additional untestable assumptions. Methods for sensitivity analysis to assess the impact of bias in the efficacy subset analysis are described. It is generally believed that the efficacy subset analysis has greater power than the intent-to-treat analysis. However, even when the efficacy subset analysis is assumed to be unbiased, or have a true type I error probability equal to the desired level alpha, situations are described where the intent-to-treat analysis in fact has greater power than the efficacy subset analysis. The intent-to-treat design, wherein all possible patients continue to be followed, is especially powerful when an effective treatment arrests progression of disease during its administration. Thus, a patient benefits long after the patient becomes noncompliant or the treatment is terminated. In such cases, a landmark analysis using the observations from the last patient evaluation is likely to prove more powerful than life-table or longitudinal analyses. Examples are described.

Bias↗
PubMed2005

In a randomized controlled trial, patients preferred electronic data collection of breast cancer risk-factor information in a mammography setting.

BACKGROUND AND OBJECTIVE: We evaluated patient acceptance of an electronic questionnaire to collect breast cancer risk-factor data in a mammography setting. METHODS: We developed an electronic questionnaire on a tablet computer incorporating prefilled answers and skip patterns. Using a randomized controlled study design, we tested the survey in a mammography clinic that administers a paper risk-factor questionnaire to every woman at her screening mammogram. We randomized 160 women to use the electronic survey (experimental group, n = 86) or paper survey (control group, n = 74). We evaluated patient acceptance and data completeness. RESULTS: Overall, 70.4% of the experimental group women thought the survey was very easy to use, compared to 55.6% of women in the control group. Ninety percent of experimental group women preferred using the tablet, compared to the paper questionnaire. Preference for the tablet did not differ by age; however, women > or = 60 years did not find the tablet as easy to use as did women < 60 years. The proportion of missing data was significantly lower on the tablet compared to the paper questionnaire (4.6% vs. 6.2%, P = .04). CONCLUSION: Electronic questionnaires are feasible to use in a mammography setting, can improve data quality, and are preferred by women regardless of age.

Adult↗
PubMed2002

Analysis of randomized controlled trials.

Although the sophistication and flexibility of the statistical technology available to the data analyst have increased, some durable, simple principles remain valid. Hypothesis-driven analyses, which were anticipated and specified in the protocol, must still be kept separate and privileged relative to the important, but risky data mining made possible by modern computers. Analyses that have a firm basis in the randomization are interpreted more easily than those that rely heavily on statistical models. Outcomes--such as quality of life, symptoms, and behaviors--that require the cooperation of subjects to be measured will come to be more and more important as trials move away from mortality as the main outcome. Inevitably, such trials will have to deal with more missing data, especially because of dropout and noncompliance. There are fundamental limits on the ability of statistical methods to compensate for such problems, so they must be considered when studies are designed. Finally, it must be emphasized that the availability of software is not a substitute for experience and statistical expertise.

Data Interpretation, Statistical↗
PubMed1995

The use of the NHS Central Register when estimating patient survival.

National and international comparisons of survival of patients are hampered by the bias introduced when different sources of data are used to confirm a patient's continuing survival. Three methods are proposed for processing survival data which, when used in conjunction, will provide limits for the extent of this bias. Using data on heart valve replacement patients it is shown that the use of the United Kingdom National Health Service Central Register, supplemented by data from routine hospital clinic visits, will lead to a maximum overestimate of survival of 1.7% over 10 years, and in practice the overestimate will be much less. It is proposed that for publications two survival curves should be calculated using the most favorable and unfavorable assumptions about patients "lost to sight". This will provide upper and lower bounds to the survival, the difference between these curves being dependent on the quantity of missing data.

Bias↗
PubMed1998

Quality of life assessment in the International Breast Cancer Study Group: past, present, and future.

UNLABELLED: The past: Since 1986, the IBCSG has been accruing a comprehensive longitudinal health-related quality of life (QL) database in addition to biomedical data of patients with early breast cancer who are receiving or have received adjuvant treatment. Our aim is to establish QL as a complementary outcome in randomized clinical trials and to gain new insight in biopsychosocial interactions. In regard to methodology, the IBCSG has made major contributions to the field through the development of global indicators, cross-cultural validation, impact of timing of assessments and working on practical and statistical issues relating to missing data. The present: In two large-scale clinical trials (IBCSG VI and VII) adjuvant chemotherapy (CMF) had a measurable effect on health-related QL, but contrary to expectations this effect was transient and minor compared with the effect of patients' adjustment and coping after diagnosis and surgery. THE FUTURE: In addition to the assessment of health-related QL, the IBCSG is currently developing and applying a global indicator for a patient-derived adapted utility concept in order to better assess the cost-benefit ratio of adjuvant treatment. However, the real challenge for the immediate future is the question how patients' adjustment can be fostered within primary care.

Breast Neoplasms↗
PubMed2003

Bayesian sperm competition estimates.

We introduce a Bayesian method for estimating parameters for a model of multiple mating and sperm displacement from genotype counts of brood-structured data. The model is initially targeted for Drosophila melanogaster, but is easily adapted to other organisms. The method is appropriate for use with field studies where the number of mates and the genotypes of the mates cannot be controlled, but where unlinked markers have been collected for a set of females and a sample of their offspring. Advantages over previous approaches include full use of multilocus information and the ability to cope appropriately with missing data and ambiguities about which alleles are maternally vs. paternally inherited. The advantages of including X-linked markers are also demonstrated.

Animals↗
PubMedDate not supplied

The use of parent report to assess the quality of care in primary care visits among children with asthma.

OBJECTIVE: To determine the accuracy of parent report and the accuracy of the medical record in documenting physician performance of elements of pediatric asthma care in the primary care setting. METHODS: A convenience sample of 79 English-speaking parents of 4--12-year old children with asthma presenting to medical center--affiliated inner-city primary care pediatric clinics in the Bronx, Dallas, and Chicago was enrolled, and the office visit was audiotaped. Parents were interviewed 1--16 days after the visit by telephone. OUTCOME MEASURES: Accuracy of parent report was the primary outcome. The "reference standard" was an independent evaluation of the audiotaped record of the primary care visit. The National Asthma Education and Prevention Program was used as a guide to select data elements to assess quality of pediatric asthma care during primary care visits. RESULTS: Sufficient documentation was significantly (P <.001) less likely to be present in the medical record than in the follow-up interview for each element of care. When these elements were combined into a cumulative score, 71% of parent interviews but only 37% of medical records scored > or = 5 (out of a possible 6), with 29% of medical records scoring < 3. Parents were able to accurately report (concordance of parent data with audiotape reference standard) whether or not the visit had included performance of 5 of the 6 elements of care. CONCLUSIONS: Our study suggests that parent telephone interview within 2 weeks after the visit is more accurate than the medical record for documentation of the quality of asthma care in pediatric primary care visits. The medical record was not sufficient to assess the quality of primary care related to asthma, primarily because of missing data. Therefore, our data suggest that assessing quality of care using the medical record will not only bias the findings in the direction of more deficient care but will also make improvement in care more difficult. Further validation of our strategy for using parent report to assess the quality of care in primary care visits will require its application in a variety of other primary care settings.

Asthma↗