High Rates Of Missing Data Research Articles

High rate of missing vital signs data at triage in a paediatric emergency department

Vital signs measurement is considered standard practice in paediatric emergency department triage assessment, but studies have shown variable incidence of missing data. To evaluate the rate of missing data for vital signs at triage and to determine clinical and environmental predictive factors. A retrospective cohort design was used to study a database of consecutive patients registered at a tertiary paediatric emergency department during randomly chosen shifts. Demographic and clinical data were collected. Univariate and multivariate logistic regression analyses were performed to evaluate the determinants of missing data for body temperature, heart rate, respiratory rate, blood pressure and pulse oximetry. There were 2081 patients triaged during the study periods. On multivariate logistic regression analysis, triage level (from 1 = priority to 4 = nonurgent) was an independent predictor of missing data for heart rate, respiratory rate, blood pressure and pulse oximetry (OR 1.48 to 2.05). Patients visiting the emergency department during the day shift (OR 1.08 to 4.72) and the evening shift (OR 1.38 to 9.24) had a higher rate of missing data than those visiting during the night shift. A decreased level of consciousness, an immunocompromised state and referral by a physician did not meet statistical significance as predictive factors. There was a high rate of missing data for vital signs. Factors related to patients' clinical characteristics, such as acuity of triage level, were associated with a higher rate of vital signs documentation at triage. An environmental factor, shift of presentation, was also independently associated with a higher rate.

Violation of the intent‐to‐treat principle and rate of missing data in superiority trials assessing structural outcomes in rheumatic diseases

To evaluate the methodologic quality of and identify methodologic issues in superiority trials assessing structural outcomes in rheumatic diseases. We searched Medline and the Cochrane Central Register of Controlled Trials for reports of randomized controlled trials assessing structural outcomes in osteoarthritis (OA), rheumatoid arthritis (RA), and osteoporosis (OP) published between January 1994 and December 2003 in high-impact factor general medical and specialty journals. One reader extracted data (quality assessment, intent-to-treat analysis [ITT analysis], rate of missing data, and methods of handling missing data), using a standardized form. A total of 81 reports were included in the analysis (37 on OP, 34 on RA, and 10 on OA). The mean +/- SD methodologic quality scores on the Jadad scale (possible range 0-5) and the Delphi list (possible range 0-9) were 2.9 +/- 1.2 and 6.4 +/- 1.3, respectively. Although it was reported in 54 articles (66.7%) that the analysis was done on an ITT basis, full ITT analysis was performed in only 6 of the studies (7.4%), modified ITT analysis in 11 (13.6%), and case-complete analysis in 48 (59.3%); the analysis was unclear in 16 articles (19.8%). The rate of missing data on structural outcomes could be determined in only 63 articles (77.8%) and was >20% in approximately one-third of these reports. Methods for handling missing data on structural outcomes were described in 19 articles (23.5%) and were, in general, inappropriate. Lack of ITT analysis and a high rate of missing data in superiority trials assessing structural outcomes may bias results from such trials. Our recommendations for improving these shortcomings may help researchers plan, analyze, and report the results of such trials.

Inference of missing SNPs and information quantity measurements for haplotype blocks

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