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 469 records · Page 26Linked to original sources
PubMed1995

A multiple imputation strategy for clinical trials with truncation of patient data.

Clinical trials of drug treatments for psychiatric disorders commonly employ the parallel groups, placebo-controlled, repeated measure randomized comparison. When patients stop adhering to their originally assigned treatment, investigators often abandon data collection. Thus, non-adherence produces a monotone pattern of unit-level missing data, disabling the analysis by intent-to-treat. We propose an approach based on multiple imputation of the missing responses, using the approximate Bayesian bootstrap to draw ignorable repeated imputations from the posterior predictive distribution of the missing data, stratifying by a balancing score for the observed responses prior to withdrawal. We apply the method and some variations to data from a large randomized trial of treatments for panic disorder, and compare the results to those obtained by the original analysis that used the standard (endpoint) method.

Adult↗
PubMedDate not supplied

Ensuring the quality of quality of life data: the Southwest Oncology Group experience.

The Southwest Oncology Group (SWOG) has successfully included quality of life (QOL) questionnaires in selected oncology treatment trials. Extensive quality control procedures have been necessary for obtaining and maintaining good questionnaire submission rates. Since the first QOL study was activated in SWOG in 1990, the Group has found it increasingly necessary to incorporate centralized monitoring of the QOL assessment schedule. Successful quality control strategies are presented. Current submission rates for five SWOG phase III treatment trials (both open and closed) and one chemoprevention trial are presented for those scheduled QOL assessments for which we have follow-up data. Reasons for missing QOL questionnaires and the extent of missing data within submitted QOL questionnaires are described for two different disease contexts: a trial for patients with advanced stage disease, and a trial for patients with earlier stage disease.

Clinical Trials as Topic↗
PubMed2001

Pharmacological treatment for aphasia following stroke.

BACKGROUND: Aphasia describes language impairment associated with a brain lesion. OBJECTIVES: The objective of this review was to assess the effects of drugs on language abilities when given to people with aphasia following stroke. SEARCH STRATEGY: We searched the Cochrane Stroke Group Register (last searched: May 2001), and reference lists of relevant articles to December 1998. We also contacted academic institutions and other researchers to identify further published and unpublished trials. MEDLINE was searched from 1966-1998, and CINAHL from 1982-1998. We searched the International Journal of Disorders of Communication by hand (known by other names in the past), from 1969 to 1998. SELECTION CRITERIA: Randomised controlled trials comparing: ~bullet~Any drug given to improve language, versus no treatment, or versus placebo ~bullet~Any drug given to improve language versus speech and language therapy ~bullet~One drug given to improve language versus another drug given with the same aim DATA COLLECTION AND ANALYSIS: The principal reviewer collected the data, and assessed the quality of the trials with independent data checking and methodological advice. If we could not perform a statistical combination of different studies, we sought missing data. Failing that we provided a description. We sought missing data from authors, or where appropriate, a drug company. MAIN RESULTS: We considered fifty two studies in detail, from which we identified ten trials suitable for the review. In most cases the methodological quality was unassessable, and only one trial reported sufficient detail for us to complete a description and analysis. This study did lose a large number of patients during its course. Drugs used in the trials identified were piracetam, bifemalane, piribedil, bromocriptine, idebenone, and Dextran 40. We found weak evidence that patients were more likely to have improved on any language measure at the end of the trial if they had received treatment with piracetam (odds ratio 0.46, 95% confidence interval 0.3 to 0.7). The evidence is considered weak because of the large numbers of drop outs from the trials identified, who were lost to follow up. Patients who were treated with piracetam were no more likely (considering statistical significance) than those who took a placebo to experience unwanted effects, including death (odds ratio 1.29, 95% confidence interval for difference 0.9 to 1.7). However, the differences in death rates between the two groups, even though not not statistically significant, do give rise to some concerns that there may be an increased risk of death from taking piracetam. We could not determine if drug treatment is more effective than speech and language therapy. We could not determine whether one drug is more effective than another. REVIEWER'S CONCLUSIONS: The main conclusion of this review is that drug treatment with piracetam may be effective in the treatment of aphasia after stroke. Further research is needed to explore the effects of drugs for aphasia, in particular piracetam. If a trial is done, this must be large enough to have adequate statistical power. The safety of the drug should be of primary interest. Researchers should examine the long term effects of this treatment, and whether it is more effective than speech and language therapy.

Aphasia↗
PubMed2000

Home care by outreach nursing for chronic obstructive pulmonary disease.

BACKGROUND: Chronic obstructive pulmonary disease (COPD) is characterised by progressive airflow obstruction, worsening exercise performance and deterioration in health. It is associated with significant morbidity, mortality and costs to health care systems. Care strategies, such as outreach nursing in the community, may reduce this burden. OBJECTIVES: To evaluate the effectiveness of outreach respiratory health care worker programmes for patients with COPD in terms of improving lung function, exercise tolerance and health related quality of life (HRQL) of patient and carer, and reducing mortality and hospital service utilisation. SEARCH STRATEGY: A search was carried out using the Cochrane Airways Group database. Bibliographies of identified RCTs were searched for additional relevant RCTs. Authors of identified RCTS were contacted for other published and unpublished studies. SELECTION CRITERIA: Only randomised control trials of patients with COPD were included. The intervention was an outreach nurse visiting patients in their homes, providing support, education, monitoring patient status and providing liaison with physicians. Interventions that used nurse practitioners who provided therapeutic intervention were also included. Studies in which the therapeutic intervention under test was physical training were not included. DATA COLLECTION AND ANALYSIS: Data extraction and study quality assessment were performed independently by two reviewers. Where further or missing data was required, authors of studies were contacted. MAIN RESULTS: Four studies were found. Three assessed mortality following twelve months of care (n=96, 152 and 301), and one after seven months(n=75). Meta-analysis demonstrated that mortality was not significantly reduced by the intervention, Peto Odds Ratio 0.72; 95 % confidence interval 0.43, 1.21. Post hoc subgroup analysis suggested that mortality was reduced by the outreach nursing programme in patients with less severe disease. Significant improvements in health related quality of life were reported in one study in moderate COPD, but not in a study in patients with severe disease. No changes in lung function or exercise performance were found in the studies where data were available. Hospital admissions were reported in only one study in patients with severe disease and no benefit was observed. REVIEWER'S CONCLUSIONS: Patients with moderate COPD may have mortality and health related quality of life gains from a nursing outreach programme, but there are no data about reductions in hospital utilisation. Patients with severe COPD do not appear to have benefit from such programmes and one large study found no reduction in hospital admissions in such patients.

Community-Institutional Relations↗
PubMed2000

Long-acting beta2-agonists for chronic obstructive pulmonary disease.

BACKGROUND: Chronic obstructive pulmonary disease (COPD) is characterised by airflow limitation which is only partially reversible. Long acting beta2-agonists, effective in the management of asthma,are also recommended for COPD management so it is important to establish whether these drugs are effective in reducing COPD symptoms in view of the potential side effect and cost burden. OBJECTIVES: To determine the effectiveness of long acting beta2-adrenoceptor agonists in improving lung function and quality of life and reducing dyspnoea in patients with COPD. SEARCH STRATEGY: A search was carried out using the Cochrane Airways Group register. Bibliographies of identified RCTs were searched for additional relevant RCTs. Authors of identified RCTs were contacted for other published and unpublished studies. In addition, unpublished studies were also obtained from the pharmaceutical companies that manufacture long acting beta2- adrenoceptor agonists. SELECTION CRITERIA: All randomised controlled trials over four weeks in duration comparing treatment with long-acting beta2-adrenoceptor agonists (salmeterol and formoterol) with placebo in patients with stable non-reversible COPD. Outcome measures included forced expiratory volume in one second (FEV1), peak expiratory flow rate (PEFR), symptom scores, six minute walk distance, quality of life scores, Borg scores for dyspnoea and rescue bronchodilator use. DATA COLLECTION AND ANALYSIS: Data extraction and study quality assessment was performed independently by two reviewers. Where further or missing data were required, authors of studies were contacted. MAIN RESULTS: Thirty three abstracts were identified as potentially relevant. Of these, four randomised controlled trials (RCTs) were included in this review. Two were parallel group studies of 16 week duration and two were cross-over studies with four week treatment arms. All four RCTs assessed the efficacy of salmeterol in COPD. In a 16 week study of salmeterol 50 mcg and 100 mcg twice daily treatment, a significant increase in FEV1 was seen in both treatment groups. The weighted mean difference (WMD) for the increase in FEV1 for the 50 mcg group was 0.10 litre (95% CI: 0.05;0. 15) and in the 100 mcg group the WMD was 0.12 litre (95% CI: 0.06; 0. 17 ). In the two cross-over studies of four weeks treatment, salmeterol 50 mcg twice daily treatment did not show significant increases in FEV1 (WMD = 0.04 litre, 95% CI: -0.06; 0.15). Similarly, morning and night time PEFR was not significantly improved with salmeterol treatment. In a 16 week study, disease-specific quality of life, measured using the St. George's Respiratory Questionnaire (SGRQ), showed a significant improvement after 50 mcg twice daily, but not after 100 mcg twice daily. This improvement exceeded the threshold for a clinically significant change with this questionnaire. General health status, as measured by the Medical Outcomes Short Form 36, did not improve in any of the eight components with either salmeterol dose. No significant difference was demonstrated in the mean change from baseline in the six minute walk distance (WMD = 1.9 metres, 95% CI: -15.4;19.3). Breathlessness was reduced in one study in patients receiving salmeterol 50 mcg twice daily group. Significantly more patients in this group had Borg scores for breathlessness less than three (a score of three indicating moderate dyspnoea) compared to the placebo treated group (Peto Odds Ratio = 0.60, 95% CI: 0.40; 0.88). Neither dose of salmeterol influenced the incidence of COPD exacerbations, (50 mcg: Peto Odds Ratio = 0.74, 95% CI: 0.47, 1.15) and (100 mcg: Peto Odds Ratio = 0.98, 95% CI: 0.64, 1.52). REVIEWER'S CONCLUSIONS: Treatment of patients with COPD with long acting beta2-agonists produces only small increases in FEV1. In one study, a dose of salmeterol 50 mcg twice daily produced a reduction in breathlessness and a clinically significant improvement in quality of life. (ABSTRACT TRUNCATED)

Adrenergic beta-Agonists↗
PubMed2000

Short-acting beta 2 agonists for stable chronic obstructive pulmonary disease.

BACKGROUND: Chronic Obstructive Pulmonary Disease (COPD) is a chronic condition characterised by progressive airflow limitation that is at most partially reversible. Despite the lack of reversibility patients often report symptomatic improvement with short-acting beta 2 bronchodilator medications. They are used on either an "as required" or a "regular plus as required basis" and they may be used in conjunction with other bronchodilator medicines such as ipratropium and methylxanthines. These medicines are used in the management of both stable and acute exacerbations of COPD. This review examined the effect of short-acting beta 2 bronchodilators given by inhalation in stable COPD. OBJECTIVES: To determine the clinical effect and assess the adverse effects of inhaled short-acting beta 2 agonist bronchodilators compared with placebo in stable COPD. SEARCH STRATEGY: Only randomised controlled trials (RCTs) were considered. RCTs were identified using the Cochrane Airways Group database (CENTRAL). In addition, the reference lists of review articles and RCTs retrieved in full were searched for other potentially relevant citations. SELECTION CRITERIA: Only trials with adult patients with stable COPD, as defined by internationally accepted guidelines (ATS, ERS or BTS) were included in this review. All trials had a minimum duration of 7 days of regular treatment with short-acting beta 2 bronchodilators given by inhalation and compared with placebo. Data from trials where beta 2 agonists were used alone or in combination with other medicines (e.g. ipratropium bromide) were used only if there was a direct comparison between beta 2 bronchodilator alone and placebo. DATA COLLECTION AND ANALYSIS: Outcomes were analysed as continuous or dichotomous outcomes, using standard statistical techniques. For continuous outcomes, the weighted mean difference (WMD) and 95% confidence intervals were calculated and for dichotomous outcomes, the odds ratio was calculated with 95% confidence intervals by Peto's methods. Funnel plots were used to test for publication bias. MAIN RESULTS: Thirteen studies were included in this review. Most had small sample sizes and some of the sutides used very short-acting outdated compounds. All the studies used a cross-over design and were of high quality. Spirometry done at the end of study period was measured after administration of treatment (post-bronchodilator) which showed both FEV1 (0.150 L/min, 95%CI: 0. 02-0.28) and FVC (0.310 L, 95%CI: 0.00-0.62) to improve significanly but slightly when compared to placebo. A few studies measured FRC, airway resistance or conductance at the end of the study period. In all cases these measurements were done several hours after treatment, and no significant differences (p>0.05 in all cases) were found between the bronchodilator and placebo groups. Walking test Large increases in 6MW distance was observed in two studies, however one study did not show any positive improvements. There was a large increase in the 12MW distance as shown by one study. Due to the small number of studies reporting this outcome no significant differences were found in the walking distance between the bronchodilator and placebo groups. Peak Flow Rate Both morning (36. 04 L/min; 95%CI: 0.80-71.27) and evening (36.68 L/min; 95%CI: 2. 47-70.89) PEFR were significantly higher during active treatment than during placebo. Symptoms Breatlessness was measured on various scales therefore data that were presented in a suitable form were combined using standardized means for inclusion in the analysis. A significant improvement (-0.33; 95%CI: -0.58 to -0.07 with p=0.01) in the breathlessness score was observed during treatment with beta-2 agonist when compared to placebo. Cough was reported to improve significantly (data not usable) during treatment with beta2 agonist in one study but not in two others. A non-significant decrease in sputum production was reported by Wilson 1980, however four other studies reported no

Administration, Inhalation↗
PubMed2004

Information for pregnant women about caesarean birth.

BACKGROUND: Information is routinely given to pregnant women, but information about caesarean birth may be inadequate. OBJECTIVES: To examine the effectiveness of information about caesarean birth. SEARCH STRATEGY: We searched the Cochrane Pregnancy and Childbirth register, CENTRAL (26 November 2002), MEDLINE [online via PubMed 1966-] and the Web of Science citation database [1995-] (20 September 2002), and reference lists of relevant articles. SELECTION CRITERIA: Randomised controlled trials, non-randomised clinical trials and controlled before-and-after studies of information given to pregnant women about caesarean birth. DATA COLLECTION AND ANALYSIS: Two reviewers independently assessed trial quality and extracted data. Missing and further data were sought from trial authors unsuccessfully. Analyses were based on 'intention to treat'. Relative risk and confidence intervals were calculated and reported. Consumer reviewers commented on adequacy of information reported in each study. MAIN RESULTS: Two randomised controlled trials involving 1451 women met the inclusion criteria. Both studies aimed to reduce caesarean births by encouraging women to attempt vaginal delivery. One used a program of prenatal education and support, and the other cognitive therapy to reduce fear. Results were not combined because of differences in the study populations. Non-clinical outcomes were ascertained in both studies through questionnaires, but were subject to rates of loss to follow-up exceeding 10%.A number of important outcomes cannot be reported: knowledge or understanding; decisional conflict; and women's perceptions: of their ability to discuss care with clinicians or family/friends, of whether information needs were met, and of satisfaction with decision-making. Neither study assessed women's perception of participation in decision-making about caesarean birth, but Fraser 1997, who examined the effect of study participation on decision making, found that women in the intervention group were more likely to consider that attempting vaginal birth was easier (51% compared to 28% in control group), or more difficult (10% compared to 6%). These results could be affected by the attrition rate of 11%, and are possibly subject to bias. Neither intervention used in these trials made any difference to clinical outcomes. About 70% or more women attempted vaginal delivery in both trials, yet caesarean delivery rates exceeded 40%, at least 10% higher than was hoped. There was no significant difference between control and intervention groups for any of the outcomes measured: vaginal birth, elective/scheduled caesarean, and attempted vaginal delivery. Outcome data, although similar for both groups, were not sufficient to compare maternal and neonatal morbidity or neonatal mortality. There was no difference in the psychological outcomes for the intervention and control groups reported by either of the included trials. Consumer reviewers said information for women considering a vaginal birth after caesarean (VBAC) should include: risks of VBAC and elective caesarean; warning signs in labour; philosophy and policies of hospital and staff; strategies to improve chances of success; and information about probability of success with specific care givers. REVIEWER'S CONCLUSIONS: Research has focussed on encouraging women to attempt vaginal delivery. Trials of interventions to encourage women to attempt vaginal birth showed no effect, but shortcomings in study design mean that the evidence is inconclusive. Further research on this topic is urgently needed.

Cesarean Section↗
PubMed2004

Prophylactic platelet transfusion for haemorrhage after chemotherapy and stem cell transplantation.

BACKGROUND: Platelet transfusions are used in modern clinical practice to prevent and treat bleeding in thrombocytopenic patients with bone marrow failure. Although considerable advances have been made in platelet transfusion therapy in the last 30 years, some areas continue to provoke debate, especially the use of prophylactic platelet transfusions for the prevention of thrombocytopenic bleeding. OBJECTIVES: To determine the optimal use of platelet transfusion for the prevention of haemorrhage (prophylactic platelet transfusion) in patients with haematological malignancies undergoing chemotherapy or stem cell transplantation. SEARCH STRATEGY: Randomised controlled trials (RCTs) were searched for in the Cochrane Central Register of Controlled Trials (CENTRAL). Searching was also undertaken on the OVID versions of MEDLINE and EMBASE using an RCT search filter strategy. SELECTION CRITERIA: Randomised controlled trials involving transfusions of platelet concentrates, prepared either from individual units of whole blood or by apheresis, and given prophylactically to prevent bleeding in patients with haematological malignancies and receiving treatment with chemotherapy and/or stem cell transplantation. DATA COLLECTION AND ANALYSIS: All electronically derived citations and abstracts of papers identified by the review search strategy were initially screened for relevancy by one reviewer. Studies clearly irrelevant were excluded at this stage. The full text of all potentially relevant trials was then formally assessed for eligibility by two reviewers independently. Two reviewers completed data extraction independently. Missing data were requested from the original investigators, as appropriate. Disagreements were resolved by discussion with the other reviewers. MAIN RESULTS: Eight completed published trials, with a total of 390 participants in the intervention groups and 362 participants in the control groups, were included in the review for further analysis. The eight studies were classified as: * three trials relevant to prophylactic platelet transfusions versus therapeutic platelet transfusions; * three trials relevant to prophylactic platelet transfusion with one trigger level versus prophylactic platelet transfusion with another trigger level; * two trials relevant to prophylactic platelet transfusion with one dose schedule versus prophylactic platelet transfusion with another dose schedule. The few reports of controlled trials addressing prophylactic versus therapeutic transfusions contained small numbers of patients and were all undertaken over 25 years ago. None of these three studies explicitly clarified whether the lack of a reported difference was a reflection of insufficient power in the trials. The findings of the meta-analyses for this group of three small studies must be interpreted with caution. In contrast, more contemporary trials addressed the question of what platelet count thresholds should apply for prophylactic transfusion; three identified studies broadly compared platelet transfusion thresholds of 10 versus 20 x 109/litre for different clinical groups of patients. There were no statistically significant differences between the groups with regards to mortality, remission rates, number of participants with severe bleeding events or red cell transfusion requirements. However, it was unclear whether the studies had sufficient power to demonstrate in combination non-inferiority in terms of safety of the lower threshold, 10 x 109/litre. Insufficient randomised trials have been undertaken to make clinically relevant conclusions about the effect of different platelet doses. REVIEWERS' CONCLUSIONS: There are no reasons to change current practice but uncertainty about the practice of prophylactic transfusion therapy should be recognised, particularly in the light of concerns about the scenario that blood products, including platelets, could become an increasingly scarce resource in the future and for which adequate alternatives do not exist. Consideration should be given to developing adequately powered trials comparing strategies of prophylaxis versus therapeutic platelet transfusion.

Hemorrhage↗
PubMed1993

The geo-systemic model and mortality in Ubon-Ratchathani.

The Thailand 'Rural Data Base' includes a number of variables in the social, economic, services, housing, agriculture and health fields. This data base comes from a survey at the village level. There are weaknesses in the data base--inaccuracies, missing data, etc ... that make it difficult to discover the structures that theoretically exist in the data using direct analysis methods, like correlation analysis for instance. The 2383 villages of the province of Ubon-Ratchathani, in Thailand, are the units of this study. The objective is to uncover the hypothetical health disparities that exist among the villages and to devise a typology of villages that will show classes of villages that have significant problems and, also, classes of villages that could suggest concrete solutions and be used as models. Some variables of the data base are considered dependent (health indicators) and others are considered independent; they are used to build a model called 'geo-systemic' that is used to expose mortality differentials in Ubon-Ratchathani.

Adolescent↗
PubMed1997

Does organization matter? A multilevel analysis of the demand-control model applied to human services.

The demand-control model (DC model) in occupational epidemiology suggests that health, an individual attribute, is partly determined by work organization, via the interplay of demand and control, job strain. The objective of this study was empirical assessment of the model's tenet of an organizational determination of individual health. An emerging analytic method, multi-level modelling, permits such an assessment. The study encompasses two large Swedish human service organizations. It was based on a nationally representative sample of 291 local organizational units (level 2) with 8296 employees (level 1), a median of 18 employees per unit. 5730 persons (69.1%) completed the questionnaire. Listwise deletion of missing data left a net study base of 4756 individuals in 284 units. Missing data were largely random. Demand and control were measured by standard questions and combined into a job strain index. Two such indices were calculated, one for quantitative demands and one for emotional demands. Individual attributes included age, gender, marital status, having children, social anchorage, and education. There were two dependent variables, self-assessed psychovegetative symptoms (worry, anxiousness, sadness, sleep difficulties, restlessness, and tension) and exhaustion (fatigue, feelings of being used up and overworked), both measured as summative indices. For psychovegetative health, a null model yielded 2.2% level 2 variance, unchanging when individual attributes were included in a random intercepts model. Inclusion of the strain variables rendered level 2 variance non-significant, decreasing level 1 variance by 23% and level 2 variance by 62%. For exhaustion, level 2 variation was 8.3% in the null model and 1.6% in the final model, with strain variables. The strain variables utilized in the DC-model thus draw a substantial part of their variation from the organizational level. It is concluded that the claim of the DC model to rely on organizational factors receives support.

Adult↗
PubMedDate not supplied

Statistical approaches to experimental design and data analysis of in vivo studies.

The objective of any experiment is to obtain an unbiased and precise estimate of a treatment effect in an efficient manner. Statistical aspects of the design, conduct, and analysis of the experiment play a major role in determining whether this goal is met. We highlight some of the more important statistical issues that pertain to in vivo studies. Particular emphasis is placed on the role of randomization, the number of animals, the utilization of repeated measures data, adjustments for missing data, and dealing with multiple causes of death or treatment failure. The discussion is not intended to be a comprehensive guide to all the statistical issues that can occur in animal experiments. Rather, the objective is to acquaint researchers with components of the experiment that will require careful statistical thought.

Animals↗
PubMed2003

Evaluation of a data collection tool (TELEform) for palliative care research.

INTRODUCTION: The Alberta Cancer Board Palliative Care Research Initiative (ACBPCRI) encourages province-wide collaboration on palliative care research projects. Because of geographic differences in information system infrastructure, it is necessary to evaluate and adopt a data collection tool that will span the variability in system hardware and software. We assessed TELEform (Cardiff Sofware Inc., Vista CA), an optical recognition-based technology that scans data collection paper forms and exports data to a computer database. We examined work place suitability, data quality, and effective resource utilization (time and cost) during the data collection tool evaluation. METHODS: Two hospices and two hospitals from the cities of Edmonton and Calgary participated in the revised Edmonton Staging System (rESS) project that used TELEform as its data collection tool. The evaluation was conducted over a period of 7 months. Data source such as e-mail and summary notes collected primarily through meetings and discussions with management, caregivers, researchers, and clerical staff was used to assess work practice and resource utilization. Descriptive statistics was employed to examine data quality and resource utilization. RESULTS: One hundred seventy eight patients were recruited during the 7-month trial. The costs and time involved in staff training, logistic support, and equipment startup were found to be reasonable. Data error and missing data were 0.4% and 0.6%, respectively. We initially encountered several problems with TELEform. The optical recognition tool could not easily pick up handwritten data. Furthermore, it was unforgiving in the sense that an error was not correctable by an eraser on the paper form. Data collectors found TELEform usage to be easy and simple because it retained the familiarity of paper-based recording. CONCLUSION: It is important to develop an information infrastructure to support research project data collection for different health settings across health regions. The TELEform based on optical recognition was able to respond to the need for current information processing. We believe that TELEform is a useful tool in terms of work practice, data quality, and resource utilization.

Data Collection↗
PubMed2005

Measuring quality of life in multiple sclerosis: not as simple as it sounds.

Data from a clinical study presented an opportunity to examine the psychometric properties of the Leeds Multiple Sclerosis Quality of Life scale (LMSQoL), which has undergone limited psychometric evaluation. LMSQoL and Multiple Sclerosis Quality of Life-54 (MSQoL-54) data were collected from 90 people with multiple sclerosis (MS) living in the community. Standard psychometric methods to examine data quality, scaling assumptions, scale to sample targeting, reliability, validity, and responsiveness were employed. The LMSQoL satisfied criteria for data quality (no missing data), scaling assumptions (item-total correlations: 0.24-0.56), reliability (Cronbach's alpha: 0.71), and demonstrated responsiveness (effect size: 0.34). Correlations between the LMSQoL and MSQoL-54 physical (range: -0.02 to -0.50) and emotional subscales (range: -0.38 to -0.65) were similar; the magnitude and pattern was not consistent with predictions based on the construct purported to be measured by the LMSQoL. The LMSQoL satisfied many psychometric criteria in this small study, however, it was difficult to interpret the validity data. From this, two fundamental measurement issues are highlighted. Firstly, current methods of examining rating scales provide only circumstantial evidence of validity; secondly, health-rating scales should be developed on the basis of clear conceptual definitions.

Adult↗
PubMedDate not supplied

[Statistical processing of non-response in transversal epidemiological studies].

In epidemiological surveys, non-response constitutes a great limitation due to the loss of validity and statistical power it represents, whether such a loss occurs due to partial participation (the individual fails to answer certain variables) or due to total lack of participation (the individual does not answer any variable). This paper reviews the scientific literature on the different methods to process statistic data when non-response has occurred in non-longitudinal studies, so as to counteract their effect in such studies. Most statistical methods focus on dealing with partial participation (missing data). These methods, of which there is a great variety, can be classified into two large groups: imputation and complete data. For accurate selection of the study method, it is necessary to study the data matrix beforehand, observing the missing data generation mechanism, as well as the proportion they represent of the total data.

Cross-Sectional Studies↗
PubMed1998

Multiple imputation for early stopping of a complex clinical trial.

It is desirable to have procedures available for stopping a clinical trial early if there appears to be no treatment effect. Conditional power procedures allow for early stopping in favor of the null hypothesis if the probability of rejecting H0 at the planned end of the trial given the current data and a value of the parameter of interest is below some threshold level. Lan, Simon, and Halperin (1982, Communications in Statistics C1, 207-219) proposed a stochastic curtailment procedure that calculates the conditional power under the alternative hypothesis. Alternatively, predictive power procedures incorporate information from the observed data by averaging the conditional power over the posterior distribution of the parameter. For complex problems in which explicit evaluation of conditional power is not possible, we propose treating the problem of projecting the outcome of a trial given the current data as a missing data problem. We then complete the data using multiple imputation and thus eliminate the need for explicit calculation of conditional power. We apply this method to AIDS Clinical Trials Group (ACTG) protocol 118 and to several simulated clinical trials.

Anti-HIV Agents↗
PubMed2005

Inference of missing SNPs and information quantity measurements for haplotype blocks.

MOTIVATION: Missing data in genotyping single nucleotide polymorphism (SNP) spots are common. High-throughput genotyping methods usually have a high rate of missing data. For example, the published human chromosome 21 data by Patil et al. contains about 20% missing SNPs. Inferring missing SNPs using the haplotype block structure is promising but difficult because the haplotype block boundaries are not well defined. Here we propose a global algorithm to overcome this difficulty. RESULTS: First, we propose to use entropy as a measure of haplotype diversity. We show that the entropy measure combined with a dynamic programming algorithm produces better haplotype block partitions than other measures. Second, based on the entropy measure, we propose a two-step iterative partition-inference algorithm for the inference of missing SNPs. At the first step, we apply the dynamic programming algorithm to partition haplotypes into blocks. At the second step, we use an iterative process similar to the expectation-maximization algorithm to infer missing SNPs in each haplotype block so as to minimize the block entropy. The algorithm iterates these two steps until the total block entropy is minimized. We test our algorithm in several experimental data sets. The results show that the global approach significantly improves the accuracy of the inference. AVAILABILITY: Upon request.

Algorithms↗
PubMed2005

Validation of data in the Norwegian Arthroplasty Register and the Norwegian Patient Register: 5,134 primary total hip arthroplasties and revisions operated at a single hospital between 1987 and 2003.

BACKGROUND: The usefulness of a national medical register relies on the completeness and quality of the data reported. The data recorded must therefore be validated to prevent systematic errors, which can cause bias in reports and study conclusions. PATIENTS AND METHODS: We compared the number of hip replacements reported to the Norwegian Arthroplasty Register (NAR), 1987-2003, and to the Norwegian Patient Register (NPR), 1999-2002, with data recorded at a local hospital. The date of operation and the index hip were further validated to find inaccurately recorded data in the NAR. Kaplan-Meier estimated survival curves were compared to evaluate the possible influence of missing data. RESULTS: Of 5,134 operations performed at a local hospital, 19 (0.4%) had not been reported to the NAR. Completeness of registration was poorer for revisions (1.2%) than for primary operations (0.2%). Among 86 Girdlestone revisions (removal of the prosthesis only), 9 (11%) had not been reported to the NAR. Missing data on revisions, however, had only a minor influence on survival analyses. The date of the operation had been recorded incorrectly in 56 cases (1.1%), and the index hip in 12 cases (0.2%). The surgeon was responsible for 85% of these errors. Comparisons with data reported to the NPR, 1999-2002, showed that 3.4% of operations at the local hospital had not been reported to the NPR. INTERPRETATION: Only 0.4% of the data from a local hospital was missing in the NAR, as opposed to the NPR where 3.4% was missing. The information recorded in the NAR appears to have been valid and reliable throughout the entire period, and provides an excellent basis for clinically relevant information regarding total hip arthroplasty.

Arthroplasty, Replacement, Hip↗
PubMed1999

Estimation and comparison of rates of change in longitudinal studies with informative drop-outs.

Many cohort studies and clinical trials have designs which involve repeated measurements of disease markers. One problem in such longitudinal studies, when the primary interest is to estimate and to compare the evolution of a disease marker, is that planned data are not collected because of missing data due to missing visits and/or withdrawal or attrition (for example, death). Several methods to analyse such data are available, provided that the data are missing at random. However, serious biases can occur when missingness is informative. In such cases, one needs to apply methods that simultaneously model the observed data and the missingness process. In this paper we consider the problem of estimation of the rate of change of a disease marker in longitudinal studies, in which some subjects drop out prematurely (informatively) due to attrition, while others experience a non-informative drop-out process (end of study, withdrawal). We propose a method which combines a linear random effects model for the underlying pattern of the marker with a log-normal survival model for the informative drop-out process. Joint estimates are obtained through the restricted iterative generalized least squares method which are equivalent to restricted maximum likelihood estimates. A nested EM algorithm is applied to deal with censored survival data. The advantages of this method are: it provides a unified approach to estimate all the model parameters; it can effectively deal with irregular data (that is, measured at irregular time points), a complicated covariance structure and a complex underlying profile of the response variable; it does not entail such complex computation as would be required to maximize the joint likelihood. The method is illustrated by modelling CD4 count data in a clinical trial in patients with advanced HIV infection while its performance is tested by simulation studies.

Biomarkers↗