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 1,351 records · Page 75Linked to original sources

PubMed2006

Tobacco cessation interventions for young people.

BACKGROUND: Teenage smoking prevalence is around 15% in developing countries (with wide variation from country to country), and around 26% in the UK and USA. Although most tobacco control programmes for adolescents are based around prevention of uptake, there are also a number of initiatives to help those who want to quit. Since those who do not smoke before the age of 20 are significantly less likely to start as adults, there is a strong case for programmes for young people that address both prevention and treatment. OBJECTIVES: To evaluate the effectiveness of strategies that help young people to stop smoking tobacco. SEARCH STRATEGY: We searched the Cochrane Central Register of Controlled Trials (CENTRAL) and the Cochrane Tobacco Addiction Group's Specialized Register, MEDLINE, EMBASE, PsyclNFO, ERIC, CINAHL, and the bibliographies of identified trials. We also searched the 'grey' literature (unpublished materials), and contacted authors and experts in the field where necessary. SELECTION CRITERIA: Types of studies: Randomized controlled trials, cluster-randomized controlled trials and controlled trials. TYPES OF PARTICIPANTS: Young people, aged less than 20, who are regular tobacco smokers. Types of interventions: The interventions ranged from simple ones such as pharmacotherapy, targeting individual young people, through complex programmes targeting people or organizations associated with young people (for example, their families or schools), or the community in which young people live. We included cessation programmes but excluded programmes primarily aimed at prevention of uptake. Types of outcome measures: The primary outcome was smoking status at six months follow up, among those who smoked at baseline. We report the definition of cessation used in each trial (e.g seven- or thirty-day point prevalence abstinence, or sustained or prolonged abstinence), and we preferred biochemically verified cessation when that measure was available. DATA COLLECTION AND ANALYSIS: Both authors independently assessed the eligibility of candidate trials identified by the searches, and extracted data from them. We categorized included trials as being at low, medium or high risk of bias, based on concealment of allocation, blinding (where applicable) and the handling of attrition and losses to follow up. We conducted limited meta-analyses of some of the trials, provided that it was appropriate to group them and provided that there was minimal heterogeneity between them. We estimated pooled odds ratios using the Mantel-Haenszel method, based on the quit rates at longest follow up for trials with at least six months follow up from the start of the intervention. MAIN RESULTS: We found 15 trials, covering 3605 young people, which met our inclusion criteria (seven cluster-randomized controlled trials, six randomized controlled trials and two controlled trials). Three trials used or tested the transtheoretical model (stages of change) approach, two tested pharmacological aids to quitting (nicotine replacement and bupropion), and the remaining trials used various psycho-social interventions, such as motivational enhancement or behavioural management. The trials evaluating TTM interventions achieved moderate long-term success, with a pooled odds ratio (OR) at one year of 1.70 ( 95% confidence interval (CI) 1.25 to 2.33) persisting at two-year follow up with an OR of 1.38 (95% CI 0.99 to 1.92). Neither of the pharmacological intervention trials achieved statistically significant results (data not pooled), but both were small-scale, with low power to detect an effect. The three interventions (5 trials) which used cognitive behavioural therapy interventions did not individually achieve statistically significant results, although when the three Not on Tobacco trials were pooled the OR 1.87; (95% CI 1.00 to 3.50) suggested some measure of effectiveness. Although the three trials that incorporated motivational interviewing as a component of the intervention achieved a pooled OR of 2.05 (95% CI 1.10 to 3.80), the impossibility of isolating the effect of the motivational interviewing in these trials meant that we could not draw meaningful inferences from that analysis. AUTHORS' CONCLUSIONS: Complex approaches show promise, with some persistence of abstinence (30 days point prevalence abstinence at six months), especially those incorporating elements sensitive to stage of change. There were few trials with evidence about pharmacological interventions (nicotine replacement and bupropion), and none demonstrated effectiveness for adolescent smokers. Psycho-social interventions have not so far demonstrated effectiveness, although pooled results for the Not on Tobacco trials suggest that that this approach may yet prove to be effective; however, their definition of cessation (one or more smoke-free days) may not adequately account for the episodic nature of much adolescent smoking. There is a need for well-designed adequately powered randomized controlled trials for this population of smokers, with a minimum of six months follow up and rigorous definitions of cessation (sustained and biochemically verified). Attrition and losses to follow up are particularly problematic in trials for young smokers, and need to be kept to a minimum, so that management and interpretation of missing data need not compromise the findings.

Adolescent↗

PubMed2002

Detection of high-risk HPV (16, 18, 33) in situ cancer of the cervix by PCR technique.

OBJECTIVE: The purpose of this study was to collect data about the incidence of high-risk HPV (16, 18, 33) types in in situ cervical cancers, and to evaluate the reliability of the morphological signs of HPV infection by comparing the presence of these signs to the PCR-proven HPV virus infection. METHODS: Fifty patients who underwent conisation at the Department of Obstetrics and Gynecology of Semmelweis University, Budapest, Hungary because of in situ cervical cancer were examined retrospectively for the presence of HPV infection by the PCR technique. The direct and indirect morphological signs of HPV infection identified in the histological and cytological samples were compared to the actual results of virus DNA amplification by PCR in the identical histological sections. The evaluation of the cytological smears and the histological sections was accomplished independently by two different pathologists. RESULTS: E6 open reading frame of HPV 16, 18 or 33 was detected by PCR in 56% (28 cases) of the histological sections of the 50 examined patients with in situ cancer. In 92% (26 patients) of the 28 HPV positive patients one HPV type was detected, while in one of the remaining two cases two HPV types (16/33), or all three types could be detected. The direct morphological signs for HPV infection proved to be 75% sensitive and 50% specific when compared to the results of PCR. Their predictive value for HPV infection was 65%. For the indirect HPV signs the sensitivity was 64% and specificity 31%. The predictive value, prognosticating the presence of HPV 16, 18, 33 infection was 54% in the same sections. Using significance analysis no significant relationship (p = 0.7728) could be detected between the positivity of indirect signs and the presence of HPV 16, 18, 33 infection, while in case of direct signs the relationship was almost significant (p = 0.0675). The joint testing of the direct and indirect signs did not improve the results (p = 0.1338). During the review of the cytological smears the specificity of the cytology in predicting true HPV infections was found to be 68% and sensitivity was 20%. The predictive value was only 50%. A significance analysis was not accomplished by this diagnostic method because of the missing data (see text). CONCLUSION: The method of Nawa et al. seems to be a reliable approach for the detection of HPV DNA in paraffin-embedded material. The three main types of HPV (16, 18, 33) are probably represented in lower percentages in CIN III in Hungary, but a larger survey is needed to obtain reliable data. The direct and indirect morphological signs of HPV infection failed to show a significant relationship with the PCR proven presence of HPV 16, 18, 33.

Biopsy, Needle↗

PubMed2000

Extended-release physostigmine in Alzheimer disease: a multicenter, double-blind, 12-week study with dose enrichment. Physostigmine Study Group.

BACKGROUND: The efficacy of extended-release physostigmine salicylate, an acetylcholinesterase inhibitor, was evaluated in 850 subjects with mild-to-moderate Alzheimer disease (AD) in a multicenter trial. METHODS: Subjects initially entered a dose-enrichment phase in which they received 1 week each of physostigmine salicylate, 24 mg/d and 30 mg/d, and daily placebo. Among the subjects who completed this phase, 35.9% responded to physostigmine treatment, whereas 62.4% were considered nonresponders, and 1.6% could not be evaluated because of missing data. After a 4-week placebo-washout phase, 176 responder subjects were randomized to receive their best dose of physostigmine or placebo in a 12-week double-blind phase. Primary efficacy measures included the cognitive subscale of the Alzheimer's Disease Assessment Scale (ADAS-Cog), the Clinician's Interview-Based Impression of Change With Caregiver Input (CIBIC+), and the Clinical Global Impression of Change (CGIC). RESULTS: In the intent-to-treat analysis of the double-blind phase, physostigmine-treated subjects scored -2.02 points better than placebo-treated subjects on the ADAS-Cog (F1,167 = 6.42 [P = .01]) and 0.33 points higher on the CIBIC+ (F1,150 = 5.68 [P = .02]). No significant improvement was observed on the CGIC or the secondary outcome measures. Nausea and vomiting were experienced by 47.0% of all physostigmine-treated subjects during the double-blind phase. CONCLUSIONS: Physostigmine demonstrated a statistically significant benefit compared with placebo on a clinical global rating of change and an objective test of cognitive function. Given the frequency of gastrointestinal side effects, the role of this agent in clinical use remains to be determined.

Aged↗

PubMed1995

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

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

Analysis of Variance↗

PubMed1999

Prognostic value of a treadmill exercise score in symptomatic patients with nonspecific ST-T abnormalities on resting ECG.

CONTEXT: Exercise testing of patients with ST-T abnormalities on the resting electrocardiogram (ECG) is problematic because in the presence of pre-existing ST-T abnormalities, the exercise test is less specific for the diagnosis of coronary artery disease. The prognostic capability of the Duke treadmill score in patients with ST-T abnormalities vs those with normal findings on resting ECG has, to our knowledge, not been evaluated. OBJECTIVE: To compare the prognostic accuracy of the Duke treadmill score in patients with nonspecific ST-T abnormalities vs those with normal results on resting ECG. DESIGN: Inception cohort study with 7 years of follow-up. SETTING: Nuclear cardiology laboratory of a US referral center. PATIENTS: All symptomatic patients who underwent exercise thallium testing between 1989 and 1991,939 of whom had nonspecific ST-T abnormalities and 1466 of whom had normal findings on resting ECG. Exclusion criteria included congenital, valvular, or cardiomyopathic heart disease; prior coronary artery revascularization; resting ECG with secondary ST-T abnormalities; or missing data. MAIN OUTCOME MEASURES: Rates of overall mortality and cardiac death for subjects classified by Duke treadmill score risk group. RESULTS: For the end point cardiac death, 7-year survival in the study population in the low-, intermediate-, and high-risk groups was 97%, 92%, and 76%, respectively (P<.001). Compared with the control group, the study group had lower 7-year survival (94% vs 98%; P<.001), fewer low-risk patients (426 [45%] vs 811 [55%]; P<.001) with worse 7-year survival (97% vs 99%; P= .008), and more high-risk patients (49 [5%] vs 34 [2%];P<.001) with a nonsignificant trend toward worse 7-year survival (76% vs 93%; P= .36). CONCLUSIONS: The Duke treadmill score can effectively risk-stratify patients with ST-T abnormalities on the resting ECG. In classified risk categories, patients with ST-T abnormalities have a worse prognosis than those with normal results on resting ECG.

Aged↗

PubMed2000

Broad-spectrum sunscreen use and the development of new nevi in white children: A randomized controlled trial.

CONTEXT: High nevus density is a risk factor for cutaneous malignant melanoma. Melanocytic nevi originate in childhood and are largely caused by solar exposure. OBJECTIVE: To determine whether use of broad-spectrum, high-sun protection factor (SPF) sunscreen attenuates development of nevi in white children. DESIGN: Randomized trial conducted June 1993 to May 1996. SETTING AND PARTICIPANTS: A total of 458 Vancouver, British Columbia, schoolchildren in grades 1 and 4 were randomized in 1993. After exclusion of nonwhite children and those lost to follow-up or with missing data, 309 children remained for analysis. Each child's nevi were enumerated at the start and end of the study in 1996. INTERVENTION: Parents of children randomly assigned to the treatment group (n=222) received a supply of SPF 30 broad-spectrum sunscreen with directions to apply it to exposed sites when the child was expected to be in the sun for 30 minutes or more. Children randomly assigned to the control group (n=236) received no sunscreen and were given no advice about sunscreen use. MAIN OUTCOME MEASURE: Number of new nevi acquired during the 3 years of the study, compared between treatment and control groups. RESULTS: Children in the sunscreen group developed fewer nevi than did children in the control group (median counts, 24 vs 28; P=.048). A significant interaction was detected between freckling and study group, indicating that sunscreen use was much more important for children with freckles than for children without. Modeling of the data suggests that freckled children assigned to a broad-spectrum sunscreen intervention would develop 30% to 40% fewer new nevi than freckled children assigned to the control group. CONCLUSIONS: Our data indicate that broad-spectrum sunscreens may attenuate the number of nevi in white children, especially if they have freckles. JAMA. 2000.

Child↗

PubMed2004

Coronary artery calcium score combined with Framingham score for risk prediction in asymptomatic individuals.

CONTEXT: Guidelines advise that all adults undergo coronary heart disease (CHD) risk assessment to guide preventive treatment intensity. Although the Framingham Risk Score (FRS) is often recommended for this, it has been suggested that risk assessment may be improved by additional tests such as coronary artery calcium scoring (CACS). OBJECTIVES: To determine whether CACS assessment combined with FRS in asymptomatic adults provides prognostic information superior to either method alone and whether the combined approach can more accurately guide primary preventive strategies in patients with CHD risk factors. DESIGN, SETTING, AND PARTICIPANTS: Prospective observational population-based study, of 1461 asymptomatic adults with coronary risk factors. Participants with at least 1 coronary risk factor (>45 years) underwent computed tomography (CT) examination, were screened between 1990-1992, were contacted yearly for up to 8.5 years after CT scan, and were assessed for CHD. This analysis included 1312 participants with CACS results; excluded were 269 participants with diabetes and 14 participants with either missing data or had a coronary event before CACS was performed. MAIN OUTCOME MEASURE: Nonfatal myocardial infarction (MI) or CHD death. RESULTS: During a median of 7.0 years of follow-up, 84 patients experienced MI or CHD death; 70 patients died of any cause. There were 291 (28%) participants with an FRS of more than 20% and 221 (21%) with a CACS of more than 300. Compared with an FRS of less than 10%, an FRS of more than 20% predicted the risk of MI or CHD death (hazard ratio [HR], 14.3; 95% confidence interval [CI]; 2.0-104; P =.009). Compared with a CACS of zero, a CACS of more than 300 was predictive (HR, 3.9; 95% CI, 2.1-7.3; P<.001). Across categories of FRS, CACS was predictive of risk among patients with an FRS higher than 10% (P<.001) but not with an FRS less than 10%. CONCLUSION: These data support the hypothesis that high CACS can modify predicted risk obtained from FRS alone, especially among patients in the intermediate-risk category in whom clinical decision making is most uncertain.

Aged↗

PubMed2004

Efficacy and safety of tenofovir DF vs stavudine in combination therapy in antiretroviral-naive patients: a 3-year randomized trial.

CONTEXT: Tenofovir disoproxil fumarate (DF) is a once-daily nucleotide analogue reverse transcriptase inhibitor. OBJECTIVE: To evaluate the efficacy and safety of tenofovir DF compared with stavudine in antiretroviral-naive patients. DESIGN, SETTING, AND PARTICIPANTS: A prospective, randomized, double-blind study conducted at 81 centers in the United States, South America, and Europe from June 9, 2000, to January 30, 2004. A total of 753 patients infected with HIV who were antiretroviral naive were screened and 602 patients entered the study. INTERVENTION: Patients were randomized to receive either tenofovir DF (n = 299) or stavudine (n = 303), with placebo, in combination with lamivudine and efavirenz. MAIN OUTCOME MEASURE: Proportion of patients with HIV RNA levels of less than 400 copies/mL at week 48. RESULTS: In the primary intent-to-treat analysis in which patients with missing data or who added or switched antiretroviral medications before week 48 were considered as failures, the proportion of patients with HIV RNA of less than 400 copies/mL at week 48 was 239 (80%) of 299 in patients receiving tenofovir DF and 253 (84%) of 301 in patients receiving stavudine (95% confidence interval, -10.4% to 1.5%), exceeding the predefined -10% limit for equivalence. However, equivalence was demonstrated in the secondary analyses (HIV RNA <50 copies/mL) at week 48 and through 144 weeks. Virologic failure was associated most frequently with efavirenz and lamivudine resistance. Through 144 weeks, the K65R mutation emerged in 8 and 2 patients in the tenofovir DF and stavudine groups, respectively (P =.06). A more favorable mean change from baseline in fasting lipid profile was noted in the tenofovir DF group at week 144: for triglyceride levels (+1 mg/dL for tenofovir DF [n = 170] vs +134 mg/dL for stavudine [n = 162], P<.001), total cholesterol (+30 mg/dL [n = 170] vs +58 mg/dL [n = 162], P<.001), direct low-density lipoprotein cholesterol (+14 mg/dL [n = 169] vs +26 mg/dL [n = 161], P<.001), and high-density lipoprotein cholesterol (+9 mg/dL [n = 168] vs +6 mg/dL [n = 154], P =.003). Investigator-reported lipodystrophy was less common in the tenofovir DF group compared with the stavudine group (9 [3%] of 299 vs 58 [19%] of 301, P<.001). The number of bone fractures and the renal safety profile were similar between the 2 groups. CONCLUSIONS: Through 144 weeks, the combination of tenofovir DF, lamivudine, and efavirenz was highly effective and comparable with stavudine, lamivudine, and efavirenz in antiretroviral-naive patients. However, tenofovir DF appeared to be associated with better lipid profiles and less lipodystrophy.

Adenine↗

PubMed2004

Functional decline in peripheral arterial disease: associations with the ankle brachial index and leg symptoms.

CONTEXT: Among individuals with lower-extremity peripheral arterial disease (PAD), specific leg symptoms and the ankle brachial index (ABI) are cross-sectionally related to the degree of functional impairment. However, relations between these clinical characteristics and objectively measured functional decline are unknown. OBJECTIVE: To define whether PAD, ABI, and specific leg symptoms predict functional decline at 2-year follow-up. DESIGN, SETTING, AND PARTICIPANTS: Prospective cohort study among 676 consecutively identified individuals (aged > or =55 years) with and without PAD (n = 417 and n = 259, respectively), with baseline functional assessments occurring between October 1, 1998, and January 31, 2000, and follow-up assessments scheduled 1 and 2 years thereafter. PAD was defined as ABI less than 0.90, and participants with PAD were categorized at baseline into 1 of 5 mutually exclusive symptom groups. MAIN OUTCOME MEASURES: Mean annual changes in 6-minute walk performance and in usual-paced and fast-paced 4-m walking velocity, adjusted for age, sex, race, prior-year functioning, comorbid diseases, body mass index, pack-years of cigarette smoking, and patterns of missing data. RESULTS: Lower baseline ABI values were associated with greater mean (95% confidence interval) annual decline in 6-minute walk performance (-73.0 [-142 to -4.2] ft for ABI <0.50 vs -58.8 [-83.5 to -34.0] ft for ABI 0.50 to <0.90 vs -12.6 [-40.3 to 15.1] ft for ABI 0.90-1.50, P =.02). Compared with participants without PAD, PAD participants with leg pain on exertion and rest at baseline had greater mean annual decline in 6-minute walk performance (-111 [-173 to -50.0] ft vs -8.67 [-36.9 to 19.5] ft, P =.004), usual-pace 4-meter walking velocity (-0.06 [-0.09 to -0.02] m/sec vs -0.01 (-0.03 to 0.003] m/sec, P =.02), and fastest-pace 4-meter walking velocity (-0.07 [-0.11 to -0.03] m/sec vs -0.02 [-0.04 to -0.006] m/sec, P =.046). Compared with participants without PAD, asymptomatic PAD was associated with greater mean annual decline in 6-minute walk performance (-76.8 (-135 to -18.6] ft vs -8.67 (-36.9 to 19.5] ft, P =.04) and an increased odds ratio for becoming unable to walk for 6 minutes continuously (3.63; 95% confidence interval, 1.58-8.36; P =.002). CONCLUSIONS: Baseline ABI and the nature of leg symptoms predict the degree of functional decline at 2-year follow-up. Previously reported lack of worsening in claudication symptoms over time in patients with PAD may be more related to declining functional performance to than lack of disease progression.

Activities of Daily Living↗

PubMed2026

Bilateral Conversion Risk in Unilateral Retinoblastoma Using Age and Genetic Testing.

IMPORTANCE: Metachronous bilateral conversion in initially unilateral retinoblastoma is uncommon but clinically consequential, potentially requiring intensified treatment and carrying worse prognosis. Clarifying how age at diagnosis refines genetic-risk stratification could enable safer, more efficient surveillance protocols. OBJECTIVE: To estimate the incidence and timing of metachronous bilateral conversion in unilateral retinoblastoma and assess whether age at diagnosis and RB1 testing are associated with bilateral conversion risk. DESIGN, SETTING AND PARTICIPANTS: This was a retrospective cohort study at a tertiary center in Shanghai, China, including 1108 consecutive children with initially unilateral retinoblastoma diagnosed from July 2010 to October 2024 (after exclusions for short follow-up [n = 139], missing data [n = 53], or synchronous bilateral disease [n = 10]). The median (IQR) follow-up was 43.4 (24.2-67.6) months. EXPOSURES: Age at diagnosis and RB1 genetic status/subtypes assessed by next-generation sequencing and multiplex ligation-dependent probe amplification, including penetrance class (high vs low) and mosaic vs germline categorization. MAIN OUTCOMES AND MEASURES: Time to metachronous bilateral conversion; cumulative incidence functions with death as a competing risk; spatial distribution of fellow-eye tumors. RESULTS: Among 1108 patients (median [IQR] age at diagnosis, 22.2 [12.0-31.4] months; 591 [53.3%] male), 24 (2.2%) developed metachronous bilateral disease. At 24 months, cumulative incidence was 2.2% (95% CI, 1.3-3.1) overall. By genetic status, the 24-month cumulative incidence was 24.8% (95% CI, 13.8-35.9) in RB1 variant-positive vs 1.6% (95% CI, 0.0-3.1) in RB1 variant-negative patients. Among RB1 variant-positive patients, risk clustered among those diagnosed before 9 months, whereas no conversions were observed among those diagnosed at older than 9 months. Four RB1 variant-negative patients who were initially diagnosed at notably late ages (20.9, 42.7, 79.6, and 118 months) subsequently converted; these cases likely represent undetected low-level mosaicism, somatic variants below detection thresholds, or rare genomic events not captured by standard sequencing panels. Fellow-eye tumors did not involve macula and showed a nasal-predominant distribution. CONCLUSIONS AND RELEVANCE: The findings in this study suggest that age at diagnosis may refine genetic risk stratification for metachronous bilateral conversion. RB1 variant-positive patients diagnosed at 9 months or later represent a very low-risk subgroup that may warrant surveillance deescalation, while rare late conversions in RB1 variant-negative patients necessitate continued long-term monitoring.

Humans↗

PubMed1997

Effect of image artifacts, organ motion, and poor segmentation on the reliability and accuracy of three-dimensional chamfer matching.

Our objective was to investigate the influence of various image artifacts on three-dimensional chamfer matching. A number of artificial and natural artifacts (for instance, as a model for CT-MR matching) were introduced or suppressed in pairs of pelvic CT scans, and a perturbation study was used to determine reliability and accuracy in a well known ground truth situation. In general, chamfer matching is extremely robust against missing data, low resolution, and poor segmentation of the images. In the presence of artifacts, minimization of the average distance outperformed minimization of the root-mean-square distance. Outliers in the scan from which the point list is obtained must be avoided. For example, rotation of the femurs reduces CT-CT registration accuracy by 1-2 mm. The robustness of chamfer matching is confirmed by a limited perturbation study of CT-MR registration for the pelvic region. In conclusion, chamfer matching is extremely accurate and reliable if outliers are avoided in the scan from which the point list is derived, and the average distance is used as a cost function.

Artifacts↗

PubMed1996

A tree-based method of analysis for prospective studies.

Prospective studies often involve rare events as study outcomes, and a primary concern is to identify risk factors and risk groups associated with the outcomes. We discuss practical solutions to risk factor analyses in prospective studies and address strategies to determine tree structures, to estimate relative risks, and to manage missing data in connection with some important epidemiologic problems. Some of the basic ideas for our strategies follow from work of Breiman, Friedman, Olshen, and Stone, although we propose extensions to their methods to resolve some practical problems that arise in implementation of these methods in epidemiologic studies. To illustrate these ideas, we analyse low birthweight associated risk factors with use of a data set from the Yale Pregnancy Outcome Study.

Bias↗

PubMed1996

On design considerations and randomization-based inference for community intervention trials.

This paper discusses design considerations and the role of randomization-based inference in randomized community intervention trials. We stress that longitudinal follow-up of cohorts within communities often yields useful information on the effects of intervention on individuals, whereas cross-sectional surveys can usefully assess the impact of intervention on group indices of health. We also discuss briefly special design considerations, such as sampling cohorts from targeted subpopulations (for example, heavy smokers), matching the communities, calculating sample size, and other practical issues. We present randomization tests for matched and unmatched cohort designs. As is well known, these tests necessarily have proper size under the strong null hypothesis that treatment has no effect on any community response. It is less well known, however, that the size of randomization tests can exceed nominal levels under the 'weak' null hypothesis that intervention does not affect the average community response. Because this weak null hypothesis is of interest in community intervention trials, we study the size of randomization tests by simulation under conditions in which the weak null hypothesis holds but the strong null hypothesis does not. In unmatched studies, size may exceed nominal levels under the weak null hypothesis if there are more intervention than control communities and if the variance among community responses is larger among control communities than among intervention communities; size may also exceed nominal levels if there are more control than intervention communities and if the variance among community responses is larger among intervention communities. Otherwise, size is likely near nominal levels. To avoid such problems, we recommend use of the same numbers of control and intervention communities in unmatched designs. Pair-matched designs usually have size near nominal levels, even under the weak null hypothesis. We have identified some extreme cases, unlikely to arise in practice, in which even the size of pair-matched studies can exceed nominal levels. These simulations, however, tend to confirm the robustness of randomization tests for matched and unmatched community intervention trials, particularly if the latter designs have equal numbers of intervention and control communities. We also describe adaptations of randomization tests to allow for covariate adjustment, missing data, and application to cross-sectional surveys. We show that covariate adjustment can increase power, but such power gains diminish as the random component of variation among communities increases, which corresponds to increasing intraclass correlation of responses within communities. We briefly relate our results to model-based methods of inference for community intervention trials that include hierarchical models such as an analysis of variance model with random community effects and fixed intervention effects. Although we have tailored this paper to the design of community intervention trials, many of the ideas apply to other experiments in which one allocates groups or clusters of subjects at random to intervention or control treatments.

Clinical Trials as Topic↗

PubMedDate not supplied

Toward a curse of dimensionality appropriate (CODA) asymptotic theory for semi-parametric models.

We argue, that due to the curse of dimensionality, there are major difficulties with any pure or smoothed likelihood-based method of inference in designed studies with randomly missing data when missingness depends on a high-dimensional vector of variables. We study in detail a semi-parametric superpopulation version of continuously stratified random sampling. We show that all estimators of the population mean that are uniformly consistent or that achieve an algebraic rate of convergence, no matter how slow, require the use of the selection (randomization) probabilities. We argue that, in contrast to likelihood methods which ignore these probabilities, inverse selection probability weighted estimators continue to perform well achieving uniform n 1/2-rates of convergence. We propose a curse of dimensionality appropriate (CODA) asymptotic theory for inference in non- and semi-parametric models in an attempt to formalize our arguments. We discuss whether our results constitute a fatal blow to the likelihood principle and study the attitude toward these that a committed subjective Bayesian would adopt. Finally, we apply our CODA theory to analyse the effect of the 'curse of dimensionality' in several interesting semi-parametric models, including a model for a two-armed randomized trial with randomization probabilities depending on a vector of continuous pretreatment covariates X. We provide substantive settings under which a subjective Bayesian would ignore the randomization probabilities in analysing the trial data. We then show that any statistician who ignores the randomization probabilities is unable to construct nominal 95 per cent confidence intervals for the true treatment effect that have both: (i) an expected length which goes to zero with increasing sample size; and (ii) a guaranteed expected actual coverage rate of at least 95 per cent over the ensemble of trials analysed by the statistician during his or her lifetime. However, we derive a new interval estimator, depending on the Randomization probabilities, that satisfies (i) and (ii).

Algorithms↗

PubMed1997

Testing for treatment differences with dropouts present in clinical trials--a composite approach.

A major problem in the analysis of clinical trials is missing data from patients who drop out of the study before the predetermined schedule. In this paper we consider the situation where the outcome measure is a continuous variable and the final outcome at the end of the study is the main interest. We argue that the hypothetical complete-data marginal mean averaged over the dropout patterns is not as relevant clinically as the conditional mean of the completers together with the probability of completion or dropping out of the trial. We first take the pattern-mixture modelling approach to factoring the likelihood function, then direct the analysis to the multiple testings of a composite of hypotheses that involves the probability of dropouts and the conditional mean of the completers. We review three types of closed step-down multiple-testing procedures for this application. Data from several clinical trials are used to illustrate the proposed approach.

Bias↗

PubMed1997

Predicting in-hospital mortality of patients with AIDS-related Pneumocystis carinii pneumonia: an example of hierarchically optimal classification tree analysis.

A non-linear statistical classification methodology known as hierarchically optimal classification tree analysis (CTA) offers promise of outperforming linear alternatives. We present the first example of CTA in medicine, for an application that uses four attributes (age, body mass index, alveolar-arterial oxygen gradient, prior history of AIDS) to predict in-hospital mortality for a sample of 1339 patients with AIDS-related Pneumocystis carinii pneumonia. We also illustrate use of a hold-out (cross-generalizability) sample that suggests a three-attribute model (alveolar-arterial oxygen gradient, total lymphocyte count, private insurance). Both CTA models achieved approximately 25 per cent of the total possible theoretical improvement beyond chance in classification accuracy (for comparative purposes, logistic regression and regression-tree models obtained in prior research returned less than 10 per cent). We illustrate how one can use CTA models to construct staging systems for assessing severity of illness, and for identifying important directions for future research. In the context of the example, we discuss additional advantages of CTA versus linear alternatives that include treatment of missing data, control of experimentwise type I error, richness of the substantive findings, and ease of use.

AIDS-Related Opportunistic Infections↗

PubMed1998

Semi-parametric models for mismeasured exposure information in vaccine trials.

Exposure to infection information is important for estimating vaccine efficacy, but it is difficult to collect and inherently prone to missingness and mismeasurement. It is, therefore, generally not feasible to collect good exposure information on all participants in a large vaccine trial. We discuss study designs that collect detailed exposure information for only a small subset of trial participants, while collecting crude exposure information on all participants, and treat estimation of vaccine efficacy in the missing data/measurement error framework. We demonstrate with the example of an HIV vaccine trial the improvements in bias and efficiency when we combine the different levels of exposure information to estimate vaccine efficacy for reducing both susceptibility and infectiousness. We compare the performance of recently developed semi-parametric missing data methods of Pepe and Fleming and Carroll and Wand, Robins, Hsieh and Newey, and Reilly and Pepe.

AIDS Vaccines↗

PubMed1999

A GEE moving average analysis of the relationship between air pollution and mortality for asthma in Barcelona, Spain.

Several studies have assessed the association between air pollution and hospital admissions or emergency room visits for asthma. Because of both the presence of missing data and the small number of observations, the relationship between air pollution and mortality for respiratory causes has been rarely analysed, and when it has, the results are very inconclusive or even inconsistent. The objective of this study is to assess the relation between levels of air pollutants (black smoke, sulphur dioxide, nitrogen dioxide and ozone), meteorological variables (24th average temperature and relative humidity) and daily mortality for asthma (ICD-9 493, 2 to 45 years old) in Barcelona, Spain, during the period 1986-1989. Since the range of daily mortality for asthma (2 to 45 years old) during the period 1986-1989 was 0-1), we have preferred to consider this variable as dichotomous. First, the relationship between air pollutants, meteorological variables and daily mortality (controlled for the occurrence of asthma epidemics) was estimated using logistic regression models. As was expected, the residuals from this regression were autocorrelated, showing a complex moving average (MA) structure. If covariates were not time dependent the so-called generalized linear mixed models, could be applied. In our case the covariates vary. As a consequence the likelihood is numerically intractable because it involves the evaluation of n-fold integral. An alternative method that avoids these numerical problems is the generalized estimating equations method (GEE). It is a multivariate analogue of quasi-likelihood estimation. In the absence of a likelihood function the parameters can be estimated by solving a multivariate analogue of the quasi score function. We have modified the GEE method in this paper, allowing for a different structure in the error covariance matrix (MA). Both air pollutants and meteorological variables are related with the occurrence of a death for asthma. In this sense, nitrogen dioxide, NO(2) (ss=0.037, p<0. 05), ozone, O(3) (ss=0.021, p<0.06) and high temperature (the ss's were in the range (0.098-0.182), p<0.05) increased the probability of dying for asthma in Barcelona during the period 1986-1989.

Adolescent↗