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 667 records · Page 37Linked to original sources
PubMed2007

HIV viral dynamic models with dropouts and missing covariates.

In recent years HIV viral dynamic models have received great attention in AIDS studies. Often, subjects in these studies may drop out for various reasons such as drug intolerance or drug resistance, and covariates may also contain missing data. Statistical analyses ignoring informative dropouts and missing covariates may lead to misleading results. We consider appropriate methods for HIV viral dynamic models with informative dropouts and missing covariates and evaluate these methods via simulations. A real data set is analysed, and the results show that the initial viral decay rate, which may reflect the efficacy of the anti-HIV treatment, may be over-estimated if dropout patients are ignored. We also find that the current or immediate previous viral load values may be most predictive for patients' dropout. These results may be important for HIV/AIDS studies.

Bias↗
PubMed2002

Advice to stay active as a single treatment for low back pain and sciatica.

BACKGROUND: Low back pain is one of the most common conditions managed in primary care. Restricted activity, rest, and symptomatic analgesics are the most commonly prescribed treatment for low back pain and sciatica. OBJECTIVES: To assess the effects of advice to stay active as single treatment for patients with low back pain. SEARCH STRATEGY: Computerised searches in MEDLINE, EMBASE, Sport, The Cochrane Controlled Trials Register, Musculoskeletal Group's Trials Register and Scisearch, and scanning of reference lists from relevant articles were undertaken. Relevant studies were also traced by contacting authors. Date of the most recent searches: December 1998. SELECTION CRITERIA: We included all randomised trials or quasi-randomised trials where the study population consisted of adult patients with low back pain or sciatica, in which one comparison group was advised to stay active. The main outcomes of interest were pain, functional status, recovery and return to work. DATA COLLECTION AND ANALYSIS: Two reviewers independently selected trials for inclusion, assessed the validity of included trials and extracted data. Investigators were contacted to collect missing data or for clarification when needed. MAIN RESULTS: Four trials, with a total of 491 patients, were included. Advice to stay active was compared to advice to rest in bed in all trials. Two trials were assessed to have low risk of bias and two to have moderate to high risk of bias. The results were heterogeneous. Results from one high quality trial of patients with acute simple LBP found small differences in functional status [Weighted Mean Difference (on a 0-100 scale) 6.0 (95% CI: 1.5, 10.5)] and length of sick leave [WMD 3.4 days (95% CI: 1.6, 5.2)] in favour of staying active compared to advice to stay in bed for two days. The other high quality trial compared advice to stay active with advice to rest in bed for 14 days for patients with sciatic syndrome, and found no differences between the groups. One of the high quality trials also compared advice to stay active with exercises for patients with acute simple LBP, and found improvement in functional status and reduction in sick leave in favour of advice to stay active. REVIEWER'S CONCLUSIONS: The best available evidence suggests that advice to stay active alone has small beneficial effects for patients with acute simple low back pain, and little or no effect for patients with sciatica. There is no evidence that advice to stay active is harmful for either acute low back pain or sciatica. If there is no major difference between advice to stay active and advice to rest in bed, and there is potential harmful effects of prolonged bed rest, then it is reasonable to advise people with acute low back pain and sciatica to stay active. These conclusions are based on single trials.

Bed Rest↗
PubMed2004

Exercise to improve self-esteem in children and young people.

BACKGROUND: Psychological and behavioural problems in children and adolescents are common, and improving self-esteem may help to prevent the development of such problems. There is strong evidence for the positive physical health outcomes of exercise, but the evidence of exercise on mental health is scarce. OBJECTIVES: To determine if exercise alone or exercise as part of a comprehensive intervention can improve self-esteem among children and young people. SEARCH STRATEGY: Computerised searches in MEDLINE, EMBASE, The Cochrane Controlled Trials Register (CENTRAL), CINAHL, PsycINFO and ERIC were undertaken and reference lists from relevant articles were scanned. Relevant studies were also traced by contacting authors. Dates of most recent searches: May 2003 in (CENTRAL), all others: January 2002. SELECTION CRITERIA: Randomised controlled trials where the study population consisted of children and young people aged from 3 to 20 years, in which one intervention arm was gross motor activity for more than four weeks and the outcome measure was self-esteem. DATA COLLECTION AND ANALYSIS: Two reviewers independently selected trials for inclusion, assessed the validity of included trials and extracted data. Investigators were contacted to collect missing data or for clarification when necessary. MAIN RESULTS: Twenty-three trials with a total of 1821 children and young people were included. Generally, the trials were small, and only one was assessed to have a low risk of bias. Thirteen trials compared exercise alone with no intervention. Eight were included in the meta-analysis, and overall the results were heteregeneous. One study with a low risk of bias showed a standardised mean difference (SMD) of 1.33 (95% CI 0.43 to 2.23), while the SMD's for the three studies with a moderate risk of bias and the four studies with a high risk of bias was 0.21 (95% CI -0.17 to 0.59) and 0.57 (95% CI 0.11 to 1.04), respectively. Twelve trials compared exercise as part of a comprehensive programme with no intervention. Only four provided data sufficient to calculate overall effects, and the results indicate a moderate short-term difference in self-esteem in favour of the intervention [SMD 0.51 (95% CI 0.15 to 0.88)]. REVIEWER'S CONCLUSIONS: The results indicate that exercise has positive short-term effects on self-esteem in children and young people. Since there are no known negative effects of exercise and many positive effects on physical health, exercise may be an important measure in improving children's self-esteem. These conclusions are based on several small low-quality trials.

Adolescent↗
PubMed2004

Glucocorticoids for acute viral bronchiolitis in infants and young children.

BACKGROUND: Systemic glucocorticoids have been widely prescribed for use in infants and young children with acute viral bronchiolitis but the actual benefit of this intervention requires clarification. OBJECTIVES: To systematically review the evidence on the effectiveness of systemic glucocorticoids for the treatment of infants and young children with acute viral bronchiolitis. SEARCH STRATEGY: Multiple strategies were incorporated to maximize identification of suitable studies. The following databases were searched: the Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library Issue 3, 2003); MEDLINE (January 1966 to September 2003); Current Contents (1998 to 2000); EMBASE (January 1990 to September 2003); and Sci Search. Handsearches through cited references and contacts with experts were also used. SELECTION CRITERIA: Only randomised controlled trials (RCT) were eligible for inclusion. Studies were included if participants were diagnosed with acute viral bronchiolitis and treated with systemic (oral, intramuscular or intravenous) corticosteroids. Three reviewers independently selected potentially relevant articles. Four reviewers evaluated these studies, determined eligibility and assessed the methodological quality of each RCT. DATA COLLECTION AND ANALYSIS: The primary outcome of interest was length of hospital stay (LOS). Secondary outcomes were: respiratory rate, haemoglobin oxygen saturation, and hospital admission and revisit rates. Data were extracted independently by the four reviewers and the results compiled and compared. Two reviewers reassessed studies to clarify points of discrepancy in the data extraction and database entry processes. Missing data were requested from the authors or calculated from other data presented in the study report. MAIN RESULTS: There was complete agreement on the inclusion of 13 trials and the exclusion of five studies. Two main study recruitment groups were identified: a) infants and young children within the first 48 hours of hospitalisation (10 trials), and b) outpatient infants and young children who were randomised from the emergency department and who may nor may not have required hospital admission (three trials).A total of 1,198 children aged 0 to 30 months were treated with the equivalent of 0.5 to 10 mg/kg of systemic prednisone for two to seven days. Outcomes of interest were not measured in each RCT. In the pooled analysis of seven trials, there was a decrease in LOS in treated children of 0.38 days (95% confidence interval (CI) -0.81 to 0.05), indicating no significant difference between treatment groups. In the pooled analysis of eight trials, the day three clinical score measured: a standard mean difference (SMD) of -0.20 (95% CI -0.73 to 0.32), indicating no difference between treatment groups. Subgroup analyses for base LOS and clinical score outcomes were performed on infants who were a) less than 12 months of age, b) all respiratory syncytial virus (RSV) positive, c) treated with less than 6 mg/kg of prednisone equivalent throughout the illness and d) first-time wheezers. These were limited by the small number of studies in each subgroup. Hospital admission rates were examined in three trials and no difference was seen between treatment groups (odds ratio (OR) 1.05 (95% CI 0.23 to 4.87). Readmission rates were reported in six studies; with no significant differences between treatment groups. Hospital revisit rates were reported in three studies, with a significant difference between treatment groups reported in one study only. The respiratory rate and haemoglobin oxygen saturation were reported descriptively in six RCTs; no differences were found between groups. Co-interventions (oxygen, supportive fluids and bronchodilators) were used similarly between treatment groups in all RCTs. REVIEWERS' CONCLUSIONS: No benefits were found in either LOS or clinical score in infants and young children treated with systemic glucocorticoids as compared to placebo. There were no differences in these outcomes between treatment groups; either in the pooled groups; either in the pooled analysis or in any of the sub analyses. Among the three studies evaluating hospital admission rates following the initial hospital visit there was no difference between treatment groups. There were no differences found in respiratory rate, haemoglobin oxygen saturation, hospital revisit or readmission rates. Subgroup analyses were significantly limited by the low number of studies in each comparison. Marked study heterogeneity and occasionally conflicting direction of benefit between trials suggests that these results should be interpreted with caution. Specific data on the harm of corticosteroid therapy in this patient population are lacking. Available evidence suggests that corticosteroid therapy is not of benefit in this patient group.

Acute Disease↗
PubMed2007

External validation of the San Francisco Syncope Rule.

STUDY OBJECTIVE: We externally validate the ability of the San Francisco Syncope Rule to accurately identify syncope patients who will experience a 7-day serious clinical event. METHODS: Patients who presented to a single academic emergency department (ED) between 8 am and 10 pm with syncope or near-syncope were prospectively enrolled. Treating physicians recorded the presence or absence of all San Francisco Syncope Rule risk factors. Patients were contacted by telephone at 14 days for a structured interview. A 3-physician panel, blinded to the San Francisco Syncope Rule score, reviewed ED medical records, hospital records, and telephone interview forms to identify predefined serious clinical events. The primary outcome was the ability of the San Francisco Syncope Rule to predict any 7-day serious clinical event. A secondary outcome was the ability of the San Francisco Syncope Rule to predict 7-day serious clinical events that were not identified during the initial ED evaluation. RESULTS: Of 592 eligible patients, 477 (81%) provided informed consent. Direct telephone contact or admission/outpatient records were successfully obtained for 463 (97%) patients. There were 56 (12%) patients who had a serious 7-day clinical event, including 16 (3%) who received a diagnosis after the initial ED evaluation. Sensitivity and specificity of the San Francisco Syncope Rule for the primary outcome were 89% (95% confidence interval [CI] 81% to 97%) and 42% (95% CI 37% to 48%), respectively, and 69% (95% CI 46% to 92%) and 42% (95% CI 37% to 48%), respectively, for the secondary outcome. Estimates of sensitivity were minimally affected by missing data and most optimistic assumptions for missing follow-up information. CONCLUSION: In this external validation cohort, the San Francisco Syncope Rule had a lower sensitivity and specificity than in previous reports.

Adult↗
PubMed2000

Using daily reports to measure drinking and drinking patterns.

Daily measurements of drinking are used to measure alcohol consumption, validate retrospective questionnaires, and examine associations between drinking and other behaviors. The advantages of using daily reports include reducing retrospective biases and forgetting of drinking occasions; disadvantages include increased costs and potential for reactivity. Methodological issues to consider in using daily reports include reporting method (written, telephone, electronic), reporting interval, data collection period, respondent attrition, and missing data. Compared to retrospective reports, daily reports result in more reported drinking occasions but similar average quantity consumed, and correlations between retrospective and daily measures are generally high. The utility of daily reports depends on the research question and the purpose of the measurements.

Alcohol Drinking↗
PubMed1996

Analysis of repeated measurements with dropouts among Alzheimer's disease patients using summary measures and meta-analysis.

A method of stratifying the data according to the patterns of missing observations, summarizing each subject's repeated measurements by a summary measure and then comparing the treatment groups with the help of a distribution-free test based on the summary measure, is used here to compare the efficacy of tacrine dose regimens with that of placebo in a recent trial in Alzheimer's disease patients. The usefulness of the method of meta-analysis for comparing the treatment groups, in the presence of missing data, is also investigated.

Alzheimer Disease↗
PubMed2002

Validity of breast cancer in the Danish Cancer Registry. A study based on clinical records from one county in Denmark.

Cancer registries are essential in order to monitor the incidence of cancer and, with proper follow-up, survival in a population. However, the usefulness of the registry depends upon the data quality. To validate the Danish Cancer Registry concerning breast cancer in female residents of Aarhus county 1983-1989, registry records were compared with clinical records. Completeness was validated in 2062 patients and correctness was determined for pathologically proven primary invasive breast cancer in 1949 patients. Incidence data were complete with no tumours missing. Data were coded according to ICD-7 and correctness of registry data in terms of basis of diagnosis and tumour malignancy was 99%. Information on extent of disease (stage) was not complete and there was a high disconcordance, in particular for bilateral breast tumours and cases with distant metastasis. Even crude staging into local and regional disease was inconsistent for 13% of cases, which probably hampered survival analysis by stage. The quality of registry data should be addressed when using variables, such as stage, not routinely reported by registries.

Adult↗
PubMed2004

Advances in statistical methodology and their application in critical care.

PURPOSE OF REVIEW: To review some of the major advances in statistical methodology of the past two decades and their application to investigations in critical care. RECENT FINDINGS: The introduction of new technologies and the ready availability of advanced computer resources have led to significant developments in statistical methodology. These include the development of computationally intensive methods, improved methods for modeling correlated outcome data, and new methods for handling missing data. Although many of these tools are available in standard statistical software, they remain underutilized in the medical literature. SUMMARY: By becoming familiar with advances in statistical methodology, researchers and clinicians can enhance collaboration with their statistical colleagues, toward the goal of better study design and analysis.

Critical Care↗
PubMed1999

A Bayesian approach to ordering gene markers.

A technique is presented whereby a marker map can be constructed using resource family data with an entire class of missing data. The focus is on a half-sib design where there is only information on a single parent and its progeny. A Bayesian approach is utilised with solutions obtained via a Markov chain Monte Carlo algorithm. Features of the approach include the capacity to determine parameters for the ungenotyped dam population, the ability to incorporate published information and its reliability, and the production of posterior densities and the consequent deduction of a wide range of inferences. These features are demonstrated through the analysis of simulated and experimental data.

Algorithms↗
PubMed2006

Missing stage and grade in Maryland prostate cancer surveillance data, 1992-1997.

BACKGROUND: Missing data in cancer surveillance records are common; however, little information exists on the types of cases most likely to have missing data, or how missing data influence research or policy. Two clinical elements often missing in surveillance data are histologic grade and stage of disease. Missing data are either not clinically ascertained or not successfully abstracted. METHODS: Prostate cancer cases (N=22,217) reported to the Maryland Cancer Registry during 1992-1997 were geocoded by residence and analyzed. Multi-level logistic regression was used to examine case attributes and area-level demographic, economic, and health services characteristics predictive of either missing stage or grade. A scanning statistic was used to explore geographic clustering of high and low rates of missing stage and grade within the state, before and after adjustment for significant variables from multi-level models. RESULTS: Older age, black race, missing grade, and higher county-level median income increased the likelihood of missing stage, whereas more recent year of diagnosis, higher blockgroup-level median income, and county-level rurality decreased the likelihood. Older age, missing or later stage, higher blockgroup-level median income, and more urologists per case in one's county of residence increased the likelihood of missing tumor grade, and more recent year of diagnosis, higher county-level median income, and rurality decreased the likelihood. Adjustment reduced statistically significant clusters of missing stage from six to two, and clusters of missing grade from three to zero. CONCLUSIONS: Results suggest systematic influences on missing stage and grade, which could be investigated with case-control follow-back studies.

Adolescent↗
PubMedDate not supplied

Challenges posed by non-random missing quality of life data in an advanced-stage colorectal cancer clinical trial.

Effects of variations in agent, dose, and route of treatment administration on patient reported quality of life (QOL) were examined for 279 patients enrolled on a seven-arm randomized clinical trial (S8905) of 5-FU and its modulation for advanced colorectal cancer. Patients completed QOL questionnaires at randomization and weeks 6, 11, and 21 post-randomization with five QOL endpoints considered primary: three treatment-specific symptoms (stomatitis, diarrhea, and hand/foot sensitivity); physical functioning; and emotional functioning. Patient compliance with the QOL assessment schedule was good, supporting the feasibility of including QOL measures in cooperative group trials. However, death and deteriorating health produced substantial missing data. Cross-sectional analyses indicated that the seven therapeutic arms did not differ in their impact on QOL. Unfortunately, longitudinal analyses of the QOL data were inappropriate given non-random missing data. Graphical presentation of non-random missing data identified the seriousness of this problem and its effect on potential conclusions about QOL during treatment. This problem appears to be particularly challenging in the context of advanced-stage disease. Failure to recognize the presence of non-random missing data can lead to serious overestimates of patient QOL over time.

Adult↗
PubMed2005

Methods to account for attrition in longitudinal data: do they work? A simulation study.

Attrition threatens the internal validity of cohort studies. Epidemiologists use various imputation and weighting methods to limit bias due to attrition. However, the ability of these methods to correct for attrition bias has not been tested. We simulated a cohort of 300 subjects using 500 computer replications to determine whether regression imputation, individual weighting, or multiple imputation is useful to reduce attrition bias. We compared these results to a complete subject analysis. Our logistic regression model included a binary exposure and two confounders. We generated 10, 25, and 40% attrition through three missing data mechanisms: missing completely at random (MCAR), missing at random (MAR) and missing not at random (MNAR), and used four covariance matrices to vary attrition. We compared true and estimated mean odds ratios (ORs), standard deviations (SDs), and coverage. With data MCAR and MAR for all attrition rates, the complete subject analysis produced results at least as valid as those from the imputation and weighting methods. With data MNAR, no method provided unbiased estimates of the OR at attrition rates of 25 or 40%. When observations are not MAR or MCAR, imputation and weighting methods may not effectively reduce attrition bias.

Bias↗
PubMed1993

Ignorability and coarse data: some biomedical examples.

Heitjan and Rubin (1991, Annals of Statistics 19, 2244-2253) define data to be "coarse" when one observes not the exact value of the data but only some set (a subset of the sample space) that contains the exact value. This definition covers a number of incomplete-data problems arising in biomedicine, including rounded, heaped, censored, and missing data. In analyzing coarse data, it is common to proceed as though the degree of coarseness is fixed in advance--in a word, to ignore the randomness in the coarsening mechanism. When coarsening is actually stochastic, however, inferences that ignore this randomness may be seriously misleading. Heitjan and Rubin (1991) have proposed a general model of data coarsening and established conditions under which it is appropriate to ignore the stochastic nature of the coarsening. The conditions are that the data be coarsened at random [a generalization of missing at random (Rubin, 1976, Biometrika 63, 581-592)] and that the parameters of the data and the coarsening process be distinct. This article presents detailed applications of the general model and the ignorability conditions to a variety of coarse-data problems arising in biomedical statistics. A reanalysis of the Stanford Heart Transplant Data (Crowley and Hu, 1977, Journal of the American Statistical Association 72, 27-36) reveals significant evidence that censoring of pretransplant survival times by transplantation was nonignorable, suggesting a greater benefit from cardiac transplantation than previous analyses had found.

Age Factors↗
PubMed2001

XML-based visual data mining in medicine.

Medical databases in general are characterized by a high degree of complexity in terms of quantity of items, number of parameter values and data types (free text, categorical, numerical and other). Substantial domain knowledge is required for adequate formalization of medical entities. In this context we developed medical database plot (mdplot), a data mining tool to visualize both structure and quality of data in medical databases to identify items suitable for evaluation. Data models are provided in XML format. Missing data is identified to enable targeted efforts to improve data quality prior to analysis. Database items are classified as 1:1- related to the patient (i.e. variables are collected once per patient) and 1:n related. mdplot provides a list of all classes contained in a database, the number of records each and a condensed bar chart for semi-quantitative description of completeness according to four types of items: categorical, numerical, text and other. All items in a category are grouped from left to right, the height of each bar represents the proportion of non-missing values with respect to the total number of records in the class; thus the amount of content in a specific class is visualized. By selection of a specific class, a detailed description of it is provided including mean completeness in each item category as well as number of values per item. The new methodology was applied to a cardiological research database consisting of 619 items on 88 patients.

Atrial Fibrillation↗
PubMed2003

Immunotherapy for IgM anti-Myelin-Associated Glycoprotein paraprotein-associated peripheral neuropathies.

BACKGROUND: Serum monoclonal anti-Myelin Associated Glycoprotein antibodies may be pathogenic in some patients with IgM paraprotein and demyelinating neuropathy. Immunotherapies aimed at reducing the level of these antibodies might be expected to be of benefit in the treatment of the neuropathy. Many potential therapies have been described in small trials, uncontrolled studies and case reports. OBJECTIVES: To examine the efficacy of any form of immunotherapy in reducing disability and impairment resulting from IgM anti-Myelin Associated Glycoprotein paraprotein-associated demyelinating peripheral neuropathy. SEARCH STRATEGY: We searched the Cochrane Neuromuscular Disease Group register (August 2002) and MEDLINE (January 1966 - August 2002) and EMBASE (January 1980 - August 2002) for controlled trials, checked the bibliographies to identify other controlled trials and contacted authors and other experts in the field. SELECTION CRITERIA: Types of studies: randomised or quasi-randomised controlled trials. TYPES OF PARTICIPANTS: patients of any age with anti-Myelin Associated Glycoprotein antibody associated demyelinating peripheral neuropathy with monoclonal gammopathy of undetermined significance of any severity. Types of interventions: any type of immunotherapy. Types of outcome measures: Primary: improvement in the Neuropathy Disability Score or Modified Rankin Scale six months after randomisation Secondary: Neuropathy Disability Score and/or the Modified Rankin Score 12 months after randomisation. Ten metre walk time, subjective clinical scores and electrophysiological parameters at six and 12 months after randomisation. IgM paraprotein levels and anti-Myelin Associated Glycoprotein antibody titres six months after randomisation. Adverse effects of treatments. DATA COLLECTION AND ANALYSIS: We identified six randomised controlled trials of which five were included after discussion between the authors. One author extracted the data and the other checked them. No missing data could be obtained from authors. MAIN RESULTS: The five eligible trials used four of the many available immunotherapy treatments. Only two had comparable interventions and outcomes but these were only short-term studies. There were no significant benefits of the treatments used in the predefined outcomes. However intravenous immunoglobulin showed benefits in terms of improved Modified Rankin Scale at two weeks and 10 metre walk time at four weeks. Serious adverse effects of intravenous immunoglobulin are known to occur from observational studies but none were encountered in these trials. REVIEWER'S CONCLUSIONS: There is inadequate reliable evidence from trials of immunotherapies in anti-Myelin Associated Glycoprotein paraproteinaemic neuropathy to recommend any particular immunotherapy treatment. Intravenous immunoglobulin is relatively safe and may produce some short-term benefit. Large well designed randomised trials are required to assess the efficacy of promising new therapies.

Demyelinating Diseases↗
PubMed2004

Coronary artery imaging using three-dimensional breath-hold steady-state free precession with two-dimensional iterative partial fourier reconstruction.

PURPOSE: To assess the feasibility of using a two-dimensional partial Fourier (PF) reconstruction scheme to reduce the acquisition time of magnetic resonance imaging (MRI) of coronary arteries. MATERIALS AND METHODS: Symmetric k-space data sets of coronary arteries were collected in seven volunteers using a three-dimensional breath-hold steady-state free precession (SSFP) sequence. Partial, asymmetric k-space data sets were generated by removing 25% of the data in the readout direction and 25% of the data in the phase encoding direction. The missing data were then estimated using a two-dimensional projection-onto-convex-sets (POCS) algorithm or filled with zeroes. Images were reconstructed from the full data set, the PF data set, and the zero-filled (ZF) data set, respectively. Coronary artery sharpness was evaluated quantitatively and qualitatively. RESULTS: Coronary artery sharpness in PF images was comparable to that in full k-space images and significantly better than that in ZF images. CONCLUSION: Two-dimensional POCS PF reconstruction is a potentially useful technique for reducing acquisition time or improving spatial resolution for breath-hold coronary MR angiography.

Adult↗
PubMed1994

Approaches to the analysis of quality of life data: experiences gained from a medical research council lung cancer working party palliative chemotherapy trial.

Standardization in the choice of quality of life (QOL) instruments and their application in randomised clinical trials have been advocated and generally accepted. However, there is now an urgent need to address the problems relating to the analysis and presentation of the data thus generated. There are intrinsic difficulties associated with QOL data, namely its multidimensional nature, attrition and missing data, and there is no consensus as to how these problems should be dealt with. This paper therefore considers these problems using interim data from a large Medical Research Council randomised trial in patients with small cell lung cancer and a poor prognosis, in which attrition and compliance are major concerns. Three possible approaches to the analysis of these data, which use different subsets of patients, are examined in detail. The strengths and weaknesses of these three methods are discussed, and examples of their use in the literature are given and compared with other reported approaches. The need for a standard definition of compliance is also emphasised, and a method of presentation suggested. The best current advice is that QOL data should be analysed in a number of different ways, and conclusions reached only when consistency is seen.

Antineoplastic Agents↗