SEARCH · PubMed Health

Results for “missing data”

Explore indexed PubMed citations for clinical trials, systematic reviews and public health research. Read source abstracts and follow each citation to its original PubMed record.

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

At least 1,369 records · Page 76Linked to original sources

PubMedDate not supplied

Statistics versus statistical science in the regulatory process.

This paper reviews the established practice of providing evidence to regulatory authorities about the claimed properties (such as efficacy and safety) of new pharmaceutical products. The established conventions and procedures are contrasted with scientific concepts and principles. The following issues are discussed: (a) recruitment of subjects and its connection to treatment heterogeneity; (b) the measurement process and the handling of missing data; (c) data transformation and the use of generalized linear models; (d) model selection and model checking; (e) the 'cult of the single trial' and the use of prior information; and (f) hypothesis testing and the P-value culture.

Area Under Curve↗

PubMed2000

Utilization of multiple imperfect assessments of the dependent variable in a logistic regression analysis.

Often, in biomedical research, there are multiple sources of imperfect information regarding a dichotomous variable of interest. For example, in a study we are conducting on the relationship between cocaine use and stroke risk, information on the cocaine use of each study patient is available from three fallible sources: patient interviews; urine toxicology testing, and medical record review. Regression analyses based on a rule for classifying patients from this information can result in biased estimation of associations and variances due to the misclassification of some subjects and to the assumption of certainty. We describe a likelihood-based method that directly incorporates multiple sources of information regarding an outcome variable into a regression analysis and takes into account the uncertainty in the classification. The method can be applied when some sources of information are missing for some subjects. We show how the availability of multiple sources can be exploited to generate estimates of the quality (for example, sensitivity and specificity) of each source and to model the degree to which missing data are informative. A fitting algorithm and issues of identifiability are discussed. We illustrate the method using data from our study.

Adult↗

PubMed1997

Analysis of gene-smoking interaction in lung cancer.

The Louisiana Lung Cancer Dataset, consisting of 337 extended pedigrees, is analyzed to determine whether a major Mendelian gene interacts with cumulative tobacco smoking (pack-years). The proportional hazards model is utilized, as it is a natural framework for estimating relative risks while adjusting for variability in age of disease onset. Segregation analyses show evidence that a Mendelian gene is segregating in these families, with the most parsimonious model, including sex, pack-years, pack-years squared, and a dominant major gene. The estimated frequency of the high-risk allele is 2% and carriers are estimated to have relative risk of 17.3 for developing lung cancer, compared to noncarriers. The addition of a gene x pack-years interaction does not significantly improve the fit of the model, indicating that on a multiplicative scale, these two factors independently influence lung cancer risk. Smoking history is missing for 23% of the study subjects and degree of "missingness" depends on disease status, age, and birth-year. To account for the nonrandomness of the missing data, a Markov chain Monte Carlo method for covariate imputation is proposed and implemented. Results from this analysis also support a nonsignificant gene-smoking interaction and an allele frequency of 2%, but a lower genetic relative risk (9.0) compared to the "complete case" analysis.

Age of Onset↗

PubMed2000

Comparison of pramipexole, fluoxetine, and placebo in patients with major depression.

Pramipexole, a dopamine D2 receptor agonist, was tested in 174 patients with major depression, with or without melancholia and without psychotic features. Three daily dose levels (0.375 mg, 1.0 mg, and 5.0 mg) were compared to fluoxetine (Prozac) at 20 mg and placebo in a randomized, double-blind, parallel-group study. After a 1 week placebo run-in period, patients were treated for 8 weeks, had a post-study follow-up (week 9), and were evaluated primarily with the Hamilton Psychiatric Rating Scale for Depression (HAM-D), the Montgomery-Asberg Depression Rating Scale (MADRS), and the Clinician's Global Impressions-Severity of Illness scale (CGI-SI). All patients who received one dose of study medication were included in the observed-case analysis (no missing data were replaced). Results indicated that by endpoint (week 8), patients receiving pramipexole at the 1.0 mg per day dose had significant improvement over baseline compared to the placebo group by measure of the HAM-D, MADRS, and CGI-SI. Significant improvement in this dose group was seen at other timepoints as well. The most obvious improvement was seen in the pramipexole 5.0 mg group, although a substantial dropout rate for this group precluded statistical tests vs. placebo late in the study. Patients taking fluoxetine also showed significant improvements at endpoint on the MADRS and earlier in the study on the HAM-D. No new or unusual safety concerns were generated during this study. Pramipexole helped safely alleviate the symptoms of depression at 1.0 mg per day and especially in those patients who could tolerate the escalation to 5 mg per day.

Adult↗

PubMed2000

Amacrine cells of the anuran retina: morphology, chemical neuroanatomy, and physiology.

Amacrine cells are third-order retinal interneurons, projecting their processes into the inner plexiform layer. Historically, they were not considered as neurons first. By the middle of the 20th century, their neuronal nature was confirmed, and their enormous diversity established. Amacrine cells have been most successfully subdivided into morphological categories based on two parameters: diameter of the dendritic field and ramification pattern in the inner plexiform layer. Works combining anatomy, physiology, and neurochemistry are scarce and in the case of the anuran retina, the situation is even worse. Correlation between morphology, neurochemistry, and physiology is little studied. Here we try to build up a database and pinpoint some of the missing data. Obtaining those could help to better understand retinal function. Sporadic attempts did not make it possible to develop a comprehensive catalog of morphologically distinct amacrine cell types in the anuran retina. The number of morphologically identified amacrine cells currently stands at 16. The list of neurochemically identified distinct cell types can be given as follows: five types GABA-containing cell types with secondary markers and at least one without; two glycinergic cell types and one interplexiform cell where glycine colocalizes with somatostatin; one dopaminergic amacrine cell and also a variant of this with interplexiform morphology; two types of serotoninergic cells; three NADPHdiaphorase-positive cells, one substance P-positive cell type without identified second marker; one CCK-positive cell type without identified second marker and the calbindin positive cells (at least one but potentially more types). This adds up to 19 cell types, out of which two are interplexiform in character. This is more than that could be identified by purely morphological means. Out of Cajal's original 13 amacrine cell types described in the frog retina, 5 parallel unequivocally with neurons defined by neurochemistry. Three others have one close match each, but their exact identity is uncertain. The remaining amacrine cells have more than one potential matches. At the same time, on one hand the amacrine cell named two-layered by Cajal so far has no match among the neurochemically identified amacrine cells. On the other hand, the interplexiform subtype of the dopaminergic cell, the somatostatin-containing glycinergic interplexiform cell, the starburst cell, and the bistratified neuropeptide Y-immunoreactive cell have no match among Cajal's cells. All in all, the number of known amacrine and interplexiform cells now stands at at least 21 in the anuran retina. Physiological characterization of amacrine cells shows that their general features seem to be rather similar to those described in tiger salamander retina. In Xenopus retina, morphologically and physiologically identified amacrine cells responded to light stimulation most frequently with ON-OFF characteristics. Immunhistochemical identification of the recorded and dye injected cells showed that amacrine cells of the "same physiological type" might have different morphology. In other words, amacrine cells with different morphology can respond similarly to illumination. Even so, small differences between almost identical responses may reflect that the cell they stem from indeed belongs to different cell types.

Animals↗

PubMed1987

Local therapy in stage I clear cell adenocarcinoma of the vagina.

Of the 219 cases of Stage I vaginal clear cell adenocarcinoma reviewed, 176 had conventional therapy and 43 underwent local therapy. The two groups appear to be similar with respect to symptoms, stage, location of the lesion in the vagina, greatest tumor diameter, surface area, depth of invasion, predominant histologic pattern, grade, and number of mitoses; nonetheless, firm conclusions cannot be drawn because of missing data. Actuarial survival rates at 5 and 10 years for the local therapy group (92% and 88%, respectively) were essentially equivalent to those for the conventional therapy group (92% and 90%, respectively), however, the recurrence experience after local therapy was less favorable. Local therapy consisted of vaginectomy in nine cases, local excision alone in 17 cases, and local irradiation (with or without local excision) in 17 cases. The subgroup of patients receiving local irradiation had a recurrence experience as favorable as that of the conventional therapy group and more favorable than that of either the subgroup treated with vaginectomy or local excision alone. The data suggest that when employed, local therapy should include local irradiation.

Adenocarcinoma↗

PubMed1993

The American Joint Committee on Cancer. Criteria for prognostic factors and for an enhanced prognostic system.

The American Joint Committee on Cancer proposes the following criteria for evaluating putative prognostic factors: they must be (1) significant, (2) independent, and (3) clinically important. Furthermore, we suggest the criteria for selecting a prognostic system that includes TNM and new prognostic factors. These criteria are: (1) easy for physicians to use; (2) provides predictions for all types of cancer; (3) provides the most accurate relapse and survival predictions at diagnosis and for every year lived for each patient; (4) provides group survival curves, where the grouping can be by any variable including outcome and therapy; (5) accommodates missing data and censored patients and is tolerant of noisy and biased data; (6) makes no a priori assumptions regarding the type of data, the distribution of the variables, or the relationships among the variables; (7) can test putative prognostic factors for significance, independence, and clinical importance; (8) accommodates treatment information in the evaluation of prognostic factors; (9) accommodates new putative prognostic factors without changing the model; (10) accommodates emerging diagnostic techniques; (11) provides information regarding the importance of each predictive variable; and (12) is automatic.

Humans↗

PubMed2001

Guided tissue regeneration for periodontal infra-bony defects.

BACKGROUND: Conventional treatment of destructive periodontal (gum) disease arrests the disease but does not regain the bone support or connective tissue lost in the disease process. Guided tissue regeneration (GTR) is a surgical procedure that aims to regenerate the periodontal tissues when the disease is advanced and could overcome some of the limitations of conventional therapy. OBJECTIVES: To assess the efficacy of GTR in the treatment of periodontal infra-bony defects measured against the current standard of surgical periodontal treatment, open flap debridement. SEARCH STRATEGY: We conducted an electronic search of the Cochrane Oral Health Group specialised trials register and MEDLINE up to October 2000. Hand searching included Journal of Periodontology, Journal of Clinical Periodontology, Journal of Periodontal Research and bibliographies of all relevant papers and review articles up to October 2000. In addition, we contacted experts/groups/companies involved in surgical research to find other trials or unpublished material or to clarify ambiguous or missing data and posted requests for data on two periodontal electronic discussion groups. SELECTION CRITERIA: Randomised, controlled trials of at least 12 months duration comparing guided tissue regeneration (with or without graft materials) with open flap debridement for the treatment of periodontal infra-bony defects. Furcation involvements and studies specifically treating early onset diseases were excluded. DATA COLLECTION AND ANALYSIS: Screening of possible studies was conducted independently by two reviewers (RT & IN) and data abstraction by three reviewers (RT, IN & EGL). The methodological quality of studies was assessed in duplicate (RT & IN) using both individual components and a quality scale (Jadad 1998) and agreement determined by Kappa scores. Methodological quality was used in sensitivity analyses to test the robustness of the conclusions. The Cochrane Oral Health Group statistical guidelines were followed (HW) and the results expressed as weighted mean differences (WMD and 95% CI) for continuous outcomes and relative risk (RR and 95% CI) for dichotomous outcomes calculated using random effects models where significant heterogeneity was detected (P < 0.1). The final analysis was conducted using STATA 6 in order to combine both parallel group studies and intra-individual (split-mouth) studies. The primary outcome measure was gain in clinical attachment. Any heterogeneity was investigated. MAIN RESULTS: We initially included 23 trial reports. Twelve were subsequently excluded. Of these, seven presented six-months data only, three were not fully randomised controlled trials, one used a non-comparable radiographic technique. Eleven studies were finally included in the review, ten testing GTR alone and two testing GTR+bone substitutes (one study had both test treatment arms). For attachment level change, the weighted mean difference between GTR alone and open flap debridement was 1.11 mm (95% CI: 0.63 to 1.59), chi-square for heterogeneity 31.4 (df = 9), p<0.001) and for GTR+bone substitutes was 1.25 mm (95% CI: 0.89 to 1.61, chi-square for heterogeneity 0.01 (df = 1), p=0.91). GTR showed a significant benefit when comparing the numbers of sites failing to gain 2 mm attachment, with relative risk 0.58 (95% CI: 0.38, 0.88, chi-square for heterogeneity 5.72 (df = 3), p=0.13). The number needed to treat (NNT) for GTR to achieve one extra site gaining 2 mm or more attachment over open flap debridement was 8 (95% CI: 4, 33), based on an incidence of 32% of sites in the control group failing to gain 2 mm or more of attachment. For baseline incidences in the range of the control groups of 10% and 55% the NNTs are 24 and 3. Probing depth reduction demonstrated a small but statistically significant benefit for GTR, weighted mean difference 0.80 mm (95% CI: 0.14,1.46, chi-square for heterogeneity 10.0 (df = 4), p=0.04) or GTR+bone substitutes, weighted mean difference 1.24 mm (95% CI: 0.89, 1.59, chi-square for heterogeneity 0.03 (df = 1), p=0.85). No significant difference was noted for gingival recession between GTR and open flap debridement. Regarding hard tissue probing at surgical re-entry, a statistically significant greater gain was found for GTR compared with open flap debridement. This amounted to a weighted mean difference of 1.39 mm (95% CI: 1.08, 1.71, chi-square for heterogeneity 0.85 (df = 2), p=0.65). For GTR+bone substitutes the difference was greater, with mean difference 3.37 mm (95% CI: 3.14, 3.61). Heterogeneity between studies was highly statistically significant for all principal comparisons and could not be explained satisfactorily by sensitivity analyses. The quality of study reporting was poor with seven out of 11 studies graded as poor using the Jadad score. (ABSTRACT TRUNCATED)

Alveolar Bone Loss↗

PubMed2006

Guided tissue regeneration for periodontal infra-bony defects.

BACKGROUND: Conventional treatment of destructive periodontal (gum) disease arrests the disease but does not usually regain the bone support or connective tissue lost in the disease process. Guided tissue regeneration (GTR) is a surgical procedure that specifically aims to regenerate the periodontal tissues when the disease is advanced and could overcome some of the limitations of conventional therapy. OBJECTIVES: To assess the efficacy of GTR in the treatment of periodontal infra-bony defects measured against conventional surgery (open flap debridement (OFD)) and factors affecting outcomes. SEARCH STRATEGY: We conducted an electronic search of the Cochrane Oral Health Group Trials Register, MEDLINE and EMBASE up to April 2004. Handsearching included Journal of Periodontology, Journal of Clinical Periodontology, Journal of Periodontal Research and bibliographies of all relevant papers and review articles up to April 2004. In addition, we contacted experts/groups/companies involved in surgical research to find other trials or unpublished material or to clarify ambiguous or missing data and posted requests for data on two periodontal electronic discussion groups. SELECTION CRITERIA: Randomised, controlled trials (RCTs) of at least 12 months duration comparing guided tissue regeneration (with or without graft materials) with open flap debridement for the treatment of periodontal infra-bony defects. Furcation involvements and studies specifically treating aggressive periodontitis were excluded. DATA COLLECTION AND ANALYSIS: Screening of possible studies and data extraction was conducted independently. The methodological quality of studies was assessed in duplicate using individual components and agreement determined by Kappa scores. Methodological quality was used in sensitivity analyses to test the robustness of the conclusions. The Cochrane Oral Health Group statistical guidelines were followed and the results expressed as mean differences (MD and 95% CI) for continuous outcomes and risk ratios (RR and 95% CI) for dichotomous outcomes calculated using random-effects models. Any heterogeneity was investigated. The primary outcome measure was change in clinical attachment. MAIN RESULTS: The search produced 626 titles, of these 596 were clearly not relevant to the review. The full text of 32 studies of possible relevance was obtained and 15 studies were excluded. Therefore 17 RCTs were included in this review, 16 studies testing GTR alone and two testing GTR+bone substitutes (one study had both test treatment arms).No tooth loss was reported in any study although these data are incomplete where patient follow up was not complete. For attachment level change, the mean difference between GTR and OFD was 1.22 mm (95% CI Random Effects: 0.80 to 1.64, chi squared for heterogeneity 69.1 (df = 15), P < 0.001, I(2) = 78%) and for GTR + bone substitutes was 1.25 mm (95% CI 0.89 to 1.61, chi squared for heterogeneity 0.01 (df = 1), P = 0.91). GTR showed a significant benefit when comparing the numbers of sites failing to gain 2 mm attachment with risk ratio 0.54 (95% CI Random Effects: 0.31 to 0.96, chi squared for heterogeneity 8.9 (df = 5), P = 0.11). The number needed to treat (NNT) for GTR to achieve one extra site gaining 2 mm or more attachment over open flap debridement was therefore 8 (95% CI 5 to 33), based on an incidence of 28% of sites in the control group failing to gain 2 mm or more of attachment. For baseline incidences in the range of the control groups of 3% and 55% the NNTs are 71 and 4. Probing depth reduction was greater for GTR than OFD: 1.21 mm (95% CI 0.53 to 1.88, chi squared for heterogeneity 62.9 (df = 10), P < 0.001, I(2) = 84%) or GTR + bone substitutes, weighted mean difference 1.24 mm (95% CI 0.89 to 1.59, chi squared for heterogeneity 0.03 (df = 1), P = 0.85). For gingival recession, a statistically significant difference between GTR and open flap debridement controls was evident (mean difference 0.26 mm (95% CI Random Effects: 0.08, 0.43, chi squared for heterogeneity 2.7 (df = 8), P = 0.95), with a greater change in recession from baseline for the control group. Regarding hard tissue probing at surgical re-entry, a statistically significant greater gain was found for GTR compared with open flap debridement. This amounted to a weighted mean difference of 1.39 mm (95% CI 1.08 to 1.71, chi squared for heterogeneity 0.85 (df = 2), P = 0.65). For GTR + bone substitutes the difference was greater, with mean difference 3.37 mm (95% CI 3.14 to 3.61). Adverse effects were generally minor although with an increased treatment time for GTR. Exposure of the barrier membrane was frequently reported with a lack of evidence of an effect on healing. AUTHORS' CONCLUSIONS: GTR has a greater effect on probing measures of periodontal treatment than open flap debridement, including improved attachment gain, reduced pocket depth, less increase in gingival recession and more gain in hard tissue probing at re-entry surgery. However there is marked variability between studies and the clinical relevance of these changes is unknown. As a result, it is difficult to draw general conclusions about the clinical benefit of GTR. Whilst there is evidence that GTR can demonstrate a significant improvement over conventional open flap surgery, the factors affecting outcomes are unclear from the literature and these might include study conduct issues such as bias. Therefore, patients and health professionals need to consider the predictability of the technique compared with other methods of treatment before making final decisions on use. Since trial reports were often incomplete, we recommend that future trials should follow the CONSORT statement both in their conduct and reporting. There is therefore little value in future research repeating simple, small efficacy studies. The priority should be to identify factors associated with improved outcomes as well as investigating outcomes relevant to patients. Types of research might include large observational studies to generate hypotheses for testing in clinical trials, qualitative studies on patient-centred outcomes and trials exploring innovative analytic methods such as multilevel modelling. Open flap surgery should remain the control comparison in these studies.

Alveolar Bone Loss↗

PubMed2003

Ceramic inlays for restoring posterior teeth.

BACKGROUND: In recent decades ceramic inlays have been used with the increasing requirements from patients for tooth-coloured restorations in posterior teeth. Ceramic inlays can offer an excellent appearance, however, their long-term prognosis is uncertain, as only a few studies have reported the long-term clinical performance of these restorations. OBJECTIVES: To compare the effectiveness of ceramic inlays in posterior teeth with other posterior restorations. SEARCH STRATEGY: We conducted an electronic search of the Cochrane Oral Health Group Trials Register, the Cochrane Central Register of Controlled Trials ( The Cochrane Library Issue 1, 2002), MEDLINE, and EMBASE from 1990 to 2001. Handsearching included relevant journals and bibliographies of all relevant papers and review articles from 1990 up to 2001. In addition, we contacted experts and companies conducting clinical research on ceramic restorations to find other trials or unpublished materials or to clarify ambiguous or missing data. SELECTION CRITERIA: Randomized controlled trials, in which the longevity of ceramic inlays is compared with those of other posterior restorations. DATA COLLECTION AND ANALYSIS: Screening of possible studies and data extraction were independently conducted by two reviewers (MH and AY) using a specially designed chart. Authors of studies were contacted for additional information. The methodological quality of studies was assessed in duplicate using individual components. The Cochrane Oral Health Group statistical guidelines were followed and the results expressed as Odds Ratio (OR) and 95% confidence interval for dichotomous outcomes. MAIN RESULTS: Two studies fulfilled the criteria to be included in the review. However, one of them was later excluded from the review, as the study design was not clearly described. The remaining, included study, evaluated the clinical performance of 60 ceramic inlays and 20 gold inlays for five years. Seven of the 60 ceramic inlays and two of the 20 gold inlays failed at five-year review. No ceramic inlays resulted in postoperative pain/discomfort after the treatment, however, one gold inlay did. The power of the included study was not great enough to detect an important difference in longevity and postoperative pain/discomfort between ceramic and gold inlays. REVIEWER'S CONCLUSIONS: There is no strong evidence available to support any differences in the clinical performance of ceramic inlays and other posterior restorations. There is a limited number of well-designed clinical trials within this research area. Greater attention to the design and reporting of studies should be given to improve the study quality of ceramic restoration trials.

Ceramics↗

PubMed2000

Treatment with leflunomide slows radiographic progression of rheumatoid arthritis: results from three randomized controlled trials of leflunomide in patients with active rheumatoid arthritis. Leflunomide Rheumatoid Arthritis Investigators Group.

OBJECTIVE: To determine whether treatment with leflunomide (LEF), methotrexate (MTX), or sulfasalazine (SSZ) for 6-12 months retards progression of radiographic damage and to identify clinical variables that correlate with radiographic progression. METHODS: Radiographs of the hands and feet were performed at baseline and at the end of study or early exit in 3 randomized controlled trials. Protocol US301 was a 12-month controlled trial of LEF or MTX treatment compared with placebo in 482 patients randomized in a 3:3:2 ratio. Protocol MN301 compared 6 months of LEF or SSZ treatment with placebo in 358 patients, randomized in a 3:3:2 ratio, with continued blinded treatment in the active control arms for 12 months. Protocol MN302 compared 12 months of LEF treatment with MTX in 999 patients. Radiographs were blinded for sequence and treatment and were scored for erosions and joint space narrowing. All analyses were by intent-to-treat. Sensitivity analyses were performed to account for missing data. RESULTS: LEF, MTX, and SSZ treatment resulted in statistically significantly less radiographic progression compared with placebo at 6 and 12 months: for protocol US301, LEF versus placebo P = 0.0007 and MTX versus placebo P = 0.0196; for protocol MN301, LEF versus placebo P = 0.0004 and SSZ versus placebo P = 0.0484. The effect of LEF treatment was similar to that of MTX and SSZ. CONCLUSION: These are the first 6- and 12-month randomized placebo- and active drug-controlled trials to demonstrate retardation of radiographic progression by a new disease-modifying antirheumatic drug (DMARD), LEF, as well as 2 commonly used DMARDs, MTX and SSZ.

Anti-Inflammatory Agents, Non-Steroidal↗

PubMed2005

Nonsteroidal antiinflammatory drugs reduce radiographic progression in patients with ankylosing spondylitis: a randomized clinical trial.

OBJECTIVE: A 2-year randomized controlled trial was performed to test the hypothesis that long-term, continuous treatment with nonsteroidal antiinflammatory drugs (NSAIDs), in comparison with NSAID treatment on demand only, influences radiographic progression in patients with ankylosing spondylitis (AS). METHODS: Patients with AS (n = 215), who had previously participated in a 6-week, randomized, double-blind clinical trial that compared celecoxib, ketoprofen, and placebo, were randomly allocated to receive either continuous treatment with NSAIDs or on-demand treatment with NSAIDs for a period of 2 years. All patients began treatment with celecoxib, at a starting dosage of 100 mg twice daily; patients could increase this dosage to 200 mg twice daily or could switch to another NSAID while maintaining the same treatment strategy. Structural changes were assessed by radiographs of the lumbar and cervical spine and scored according to the modified Stoke Ankylosing Spondylitis Spine Score by one observer who was blinded to the treatment strategy and temporal order of the radiographs. Statistical analyses included a between-group comparison of 1) radiographic progression scores (by Mann-Whitney U test), 2) time-averaged values of variables reflecting signs and symptoms of AS (by linear regression analysis), and 3) the frequency of reported site-specific adverse events (by chi-square test or Fisher's exact test, as appropriate). RESULTS: Complete sets of radiographs were available for 76 of the 111 patients in the continuous-treatment group and for 74 of the 104 patients in the on-demand group. The mean +/- SD scores for radiographic progression were 0.4 +/- 1.7 in the continuous-treatment group and 1.5 +/- 2.5 in the on-demand treatment group (P = 0.002). Parameters reflecting signs and symptoms were not statistically significantly different between groups. The between-group difference in radiographic progression did not disappear after adjusting for baseline values of radiographic damage or disease activity variables and for time-averaged values of disease activity variables, nor after input of missing data. Relevant adverse events tended to occur more frequently in the continuous-treatment group than in the on-demand group (for hypertension, 9% versus 3%; for abdominal pain, 11% versus 6%; for dyspepsia, 41% versus 38%), but the differences were not statistically significant. CONCLUSION: A strategy of continuous use of NSAIDs reduces radiographic progression in symptomatic patients with AS, without increasing toxicity substantially.

Adult↗

PubMed2006

A simple imputation method for longitudinal studies with non-ignorable non-responses.

Missing data are a common problem in longitudinal studies in the health sciences. Motivated by data from the Muscatine Coronary Risk Factor (MCRF) study, a longitudinal study of obesity, we propose a simple imputation method for handling non-ignorable non-responses (i.e., when non-response is related to the specific values that should have been obtained) in longitudinal studies with either discrete or continuous outcomes. In the proposed approach, two regression models are specified; one for the marginal mean of the response, the other for the conditional mean of the response given non-response patterns. Statistical inference for the model parameters is based on the generalized estimating equations (GEE) approach. An appealing feature of the proposed method is that it can be readily implemented using existing, widely-available statistical software. The method is illustrated using longitudinal data on obesity from the MCRF study.

Algorithms↗

PubMed2006

Maximum likelihood estimation of marginal pairwise associations with multiple source predictors.

Researchers interested in the association of a predictor with an outcome will often collect information about that predictor from more than one source. Standard multiple regression methods allow estimation of the effect of each predictor on the outcome while controlling for the remaining predictors. The resulting regression coefficient for each predictor has an interpretation that is conditional on all other predictors. In settings in which interest is in comparison of the marginal pairwise relationships between each predictor and the outcome separately (e.g., studies in psychiatry with multiple informants or comparison of the predictive values of diagnostic tests), standard regression methods are not appropriate. Instead, the generalized estimating equations (GEE) approach can be used to simultaneously estimate, and make comparisons among, the separate pairwise marginal associations. In this paper, we consider maximum likelihood (ML) estimation of these marginal relationships when the outcome is binary. ML enjoys benefits over GEE methods in that it is asymptotically efficient, can accommodate missing data that are ignorable, and allows likelihood-based inferences about the pairwise marginal relationships. We also explore the asymptotic relative efficiency of ML and GEE methods in this setting.

Depression↗

PubMed2003

The correlation between patient characteristics and expectations of benefit from Phase I clinical trials.

BACKGROUND: Patients in Phase I clinical trials sometimes report high expectations regarding the benefit of treatment. The authors examined a range of patient characteristics to determine which factors were associated with greater expectations of benefit from Phase I trials. METHODS: Participants were adult patients with cancer who had been offered participation in Phase I studies and had decided to participate. Patients completed interviewer-administered surveys before initiation of treatment. Physicians assessed Eastern Cooperative Oncology Group performance status for each patient. Statistical analyses (Pearson product moment correlation and t tests) used multiple imputation to account for missing data. RESULTS: Overall, 593 patients who were offered participation in Phase I trials were contacted, and 328 patients agreed to participate in a study of decision making by cancer patients. Of these, 260 patients (79%) enrolled in a Phase I trial. Patients' expectations regarding the chance that their disease would be controlled with experimental therapy were unrelated to age, gender, living situation, education level, or functional status. Expectations were correlated positively with beliefs about the benefit of standard therapy and the maximum benefit patients may experience from experimental therapy. Greater expectations of benefit were associated with better health-related quality of life, stronger religious faith, optimism, relative health stock, monetary risk seeking, and poorer numeracy. CONCLUSIONS: Expectations expressed as beliefs in personal outcomes may be related more to quality of life and personality variables than to patients' knowledge or functional status. Whether such expectations are accurate reflections of knowledge has important implications for evaluating the informed consent process.

Adult↗

PubMed1991

Mixed model segregation analysis of LDL-C concentration with genotype-covariate interaction.

Mixed model complex segregation analyses have in the past ignored the possibility of genotype-covariate interaction. Only in the nonmixed model with polygenic heritability equal to zero have segregation analyses been performed that allowed for genotype specific regression of the phenotype on covariates. We present an extension of Hasstedt's [1982] mixed model likelihood approximation which does allow for genotype-covariate interaction in the mixed model. Following description of this approximation, we validate the likelihood calculation by a Monte Carlo procedure based on the actual pedigree and missing data structure used in a complex segregation analysis of low density plus very low density lipoprotein cholesterol (LDL-C + VLDL-C) in baboons. The observed averages of the bootstrap parameter estimates adequately recover the generating values, which included parameters specifying genotype-covariate interaction. We then applied both a traditional complex segregation analysis and an analysis with genotype-covariate interaction to test for the presence of a major locus affecting LDL-C levels in baboons. The model including genotype-covariate interaction was significantly different from the model without interactions, and strongly supported the hypothesis that there is a segregating Mendelian locus as opposed to a random environmental factor. This major locus accounts for approximately 46% of the variance in LDL-C levels, as compared to 40% explained by a locus with no genotype-covariate interaction.

Animals↗

PubMed2005

Analysis of case-control studies of genetic and environmental factors with missing genetic information and haplotype-phase ambiguity.

Case-control studies of unrelated subjects are now widely used to study the role of genetic susceptibility and gene-environment interactions in the etiology of complex diseases. Exploiting an assumption of gene-environment independence, and treating the distribution of environmental exposures as completely nonparametric, Chatterjee and Carroll recently developed an efficient retrospective maximum-likelihood method for analysis of case-control studies. In this article, we develop an extension of the retrospective maximum-likelihood approach to studies where genetic information may be missing on some study subjects. In particular, special emphasis is given to haplotype-based studies where missing data arise due to linkage-phase ambiguity of genotype data. We use a profile likelihood technique and an appropriate expectation-maximization (EM) algorithm to derive a relatively simple procedure for parameter estimation, with or without a rare disease assumption, and possibly incorporating information on the marginal probability of the disease for the underlying population. We also describe two alternative robust approaches that are less sensitive to the underlying gene-environment independence and Hardy-Weinberg-equilibrium assumptions. The performance of the proposed methods is studied using simulation studies in the context of haplotype-based studies of gene-environment interactions. An application of the proposed method is illustrated using a case-control study of ovarian cancer designed to investigate the interaction between BRCA1/2 mutations and reproductive risk factors in the etiology of ovarian cancer.

Algorithms↗

PubMed2006

Quantifying the amount of missing information in genetic association studies.

Many genetic analyses are done with incomplete information; for example, unknown phase in haplotype-based association studies. Measures of the amount of available information can be used for efficient planning of studies and/or analyses. In particular, the linkage disequilibrium (LD) between two sets of markers can be interpreted as the amount of information one set of markers contains for testing allele frequency differences in the second set, and measuring LD can be viewed as quantifying information in a missing data problem. We introduce a framework for measuring the association between two sets of variables; for example, genotype data for two distinct groups of markers, or haplotype and genotype data for a given set of polymorphisms. The goal is to quantify how much information is in one data set, e.g. genotype data for a set of SNPs, for estimating parameters that are functions of frequencies in the second data set, e.g. haplotype frequencies, relative to the ideal case of actually observing the complete data, e.g. haplotypes. In the case of genotype data on two mutually exclusive sets of markers, the measure determines the amount of multi-locus LD, and is equal to the classical measure r(2), if the sets consist each of one bi-allelic marker. In general, the measures are interpreted as the asymptotic ratio of sample sizes necessary to achieve the same power in case-control testing. The focus of this paper is on case-control allele/haplotype tests, but the framework can be extended easily to other settings like regressing quantitative traits on allele/haplotype counts, or tests on genotypes or diplotypes. We highlight applications of the approach, including tools for navigating the HapMap database [The International HapMap Consortium, 2003], and genotyping strategies for positional cloning studies.

Alleles↗