Are social workers turning a deaf ear? A study of social services to the deaf.

This study assessed the quality and quantity of social services received by deaf persons residing in Salt Lake County, Utah. It was found that although a majority of human service agencies in the county have had some contact with deaf clients, only those agencies that have made an effort to develop programs especially for the deaf and employ someone who is skilled in sign language could deliver services effectively to this population.     

Effect of outdoor airborne particulate matter on daily death counts.

To investigate the possible relationship between airborne particulate matter and mortality, we developed regression models of daily mortality counts using meteorological covariates and measures of outdoor PM10. Our analyses included data from Cook County, Illinois, and Salt Lake County, Utah. We found no evidence that particulate matter < or = 10 microns (PM10) contributes to excess mortality in Salt Lake County, Utah. In Cook County, Illinois, we found evidence of a positive PM10 effect in spring and autumn, but not in winter and summer. We conclude that the reported effects of particulates on mortality are unconfirmed.     

The effect of a no-cost contraceptive initiative on method selection by women with housing insecurity

ObjectivesTo compare the sociodemographic characteristics of participants in a contraceptive initiative by housing security and determine the association between housing insecurity on contraceptive method selection before and after the removal of cost.Study designThis cross-sectional assessment includes 4,327 reproductive-aged participants in the HER Salt Lake Contraceptive Initiative who sought new contraceptive services and reported housing status at enrollment. HER Salt Lake prospectively explored the impact of improved contraceptive access on socioeconomic outcomes in Salt Lake County (USA). For six months (September 2015–March 2016) we collected control data, which included clinic standard-of-care cost-sharing. The intervention started March 2016, and provided no-cost contraception services and unlimited opportunities for method switching over the subsequent three years.ResultsThere were 964 (22%) housing-insecure participants. Compared to those with stable housing, housing-insecure individuals more commonly identified as a sexual minority, received public assistance and lacked health insurance. Housing-insecure women preferentially selected long-acting reversible contraception during the control period (aOR 1.60; 95%CI 1.01–2.56), but method selection equalized across housing status during the intervention.ConclusionsWhen cost is not a barrier, all women desire a comprehensive selection of contraceptive methods, regardless of housing security. Contraceptive clients in this vulnerable population need interventions which address access barriers to all methods to support reproductive planning.ImplicationsUnintended pregnancy during housing insecurity may result in homelessness. This study found housing-insecure women desire access to all contraceptive methods, not just long acting reversible contraception. Integration of comprehensive family planning initiatives into efforts to address homelessness is essential to support this vulnerable population in their reproductive planning.     

Missing data in a multi-item instrument were best handled by multiple imputation at the item score level

ObjectivesRegardless of the proportion of missing values, complete-case analysis is most frequently applied, although advanced techniques such as multiple imputation (MI) are available. The objective of this study was to explore the performance of simple and more advanced methods for handling missing data in cases when some, many, or all item scores are missing in a multi-item instrument.Study Design and SettingReal-life missing data situations were simulated in a multi-item variable used as a covariate in a linear regression model. Various missing data mechanisms were simulated with an increasing percentage of missing data. Subsequently, several techniques to handle missing data were applied to decide on the most optimal technique for each scenario. Fitted regression coefficients were compared using the bias and coverage as performance parameters.ResultsMean imputation caused biased estimates in every missing data scenario when data are missing for more than 10% of the subjects. Furthermore, when a large percentage of subjects had missing items (>25%), MI methods applied to the items outperformed methods applied to the total score.ConclusionWe recommend applying MI to the item scores to get the most accurate regression model estimates. Moreover, we advise not to use any form of mean imputation to handle missing data.     

Validation of administrative data to estimate vaccine impact: Audit of the Fiji hospital admissions electronic database, 2007–2011 & 2014–2015

ObjectivesPost-licensure studies to evaluate vaccine impact are an important component of introducing new vaccines. Such studies often rely on routinely collected data but the limitations to these data must be understood. To validate administrative data for use in 10-valent pneumococcal conjugate and rotavirus vaccine impact evaluations we have audited the two electronic database capturing hospital admissions in Fiji for completeness and consistency.MethodsHospital admission data for one week per year between 2007–2011 and 2014–2015 was collected from ward registers for selected hospitals. Ward registers were defined as the reference standard and compared to data captured in electronic databases. Data quality was assessed for completeness of admissions data (percentage of admissions in the electronic database, expressed as sensitivity), consistency of complete reporting (determined by identifying variables associated to complete reporting), and completeness of coding (percentage of admissions in the electronic database with an assigned ICD-10-AM code).ResultsOver all hospitals and years, the sensitivity for completeness of admission data was 83% (95% CI: 81.3, 84.6). Consistency of complete reporting varied and was highest at tertiary hospitals using the electronic database (sensitivity: 89.1%, 95% CI: 87.4, 90.7). The overall completeness of coding at tertiary hospitals was 90.8% (95% CI: 90.5, 91.1) with annual and hospital variation.ConclusionThe administrative data in the electronic databases in Fiji are of reasonable quality for the vaccine impact evaluation. This quantification of the missing data can be used to adjust the vaccine impact estimates.     

Three challenges described for identifying participants with missing data in trials reports,and potential solutions suggested to systematic reviewers

ObjectiveTo categorize the challenges in determining the extent of missing participant data in randomized trials and suggest potential solutions for systematic review authors.Study Design and SettingDuring the process of updating a series of Cochrane systematic reviews on the topic of anticoagulation in patients with cancer, we identified challenges and used an iterative approach to improve, and a consensus process to agree on the challenges identified, and to suggest potential ways of dealing with them. The five systematic reviews included 58 trials and 75 meta-analyses for patient-important dichotomous outcomes with 27,037 randomized participants.ResultsWe identified three categories of challenges: (1) Although systematic reviewers require information about missing data to be reported by outcome, trialists typically report the information by participant; (2) It is not always clear whether the trialists followed up participants in certain categories (e.g., noncompliers), that is, whether some categories of participants did or did not have missing data; (3) It is not always clear how the trialists dealt with missing data in their analysis (e.g., exclusion from the denominator vs. assumptions made for the numerator). We discuss potential solutions for each one of these challenges and suggest further research work.ConclusionCurrent reporting of missing data is often not explicit and transparent, and although our potential solutions to problems of suboptimal reporting may be helpful, reliable and valid characterization of the extent and nature of missing data remains elusive. Reporting of missing data in trials needs further improvement.     

A comparison of methods to address item non-response when testing for differential item functioning in multidimensional patient-reported outcome measures

Purpose

Item non-response (i.e., missing data) may mask the detection of differential item functioning (DIF) in patient-reported outcome measures or result in biased DIF estimates. Non-response can be challenging to address in ordinal data. We investigated an unsupervised machine-learning method for ordinal item-level imputation and compared it with commonly-used item non-response methods when testing for DIF.

Methods

Computer simulation and real-world data were used to assess several item non-response methods using the item response theory likelihood ratio test for DIF. The methods included: (a) list-wise deletion (LD), (b) half-mean imputation (HMI), (c) full information maximum likelihood (FIML), and (d) non-negative matrix factorization (NNMF), which adopts a machine-learning approach to impute missing values. Control of Type I error rates were evaluated using a liberal robustness criterion for α?=?0.05 (i.e., 0.025–0.075). Statistical power was assessed with and without adoption of an item non-response method; differences?>?10% were considered substantial.

Results

Type I error rates for detecting DIF using LD, FIML and NNMF methods were controlled within the bounds of the robustness criterion for?>?95% of simulation conditions, although the NNMF occasionally resulted in inflated rates. The HMI method always resulted in inflated error rates with 50% missing data. Differences in power to detect moderate DIF effects for LD, FIML and NNMF methods were substantial with 50% missing data and otherwise insubstantial.

Conclusion

The NNMF method demonstrated comparable performance to commonly-used non-response methods. This computationally-efficient method represents a promising approach to address item-level non-response when testing for DIF.

     

Late Pregnancy Ultrasound to Screen for and Manage Potential Birth Complications in Nulliparous Women: A Cost-Effectiveness and Value of Information Analysis

BackgroundFetal growth restriction is a major risk factor for stillbirth. A routine late-pregnancy ultrasound scan could help detect this, allowing intervention to reduce the risk of stillbirth. Such a scan could also detect fetal presentation and predict macrosomia. A trial powered to detect stillbirth differences would be extremely large and expensive.ObjectivesIt is therefore critical to know whether this would be a good investment of public research funds. The aim of this study is to estimate the cost-effectiveness of various late-pregnancy screening and management strategies based on current information and predict the return on investment from further research.MethodsSynthesis of current evidence structured into a decision model reporting expected costs, quality-adjusted life-years, and net benefit over 20 years and value-of-information analysis reporting predicted return on investment from future clinical trials.ResultsGiven a willingness to pay of £20 000 per quality-adjusted life-year gained, the most cost-effective strategy is a routine presentation-only scan for all women. Universal ultrasound screening for fetal size is unlikely to be cost-effective. Research exploring the cost implications of induction of labor has the greatest predicted return on investment. A randomized, controlled trial with an endpoint of stillbirth is extremely unlikely to be a value for money investment.ConclusionGiven current value-for-money thresholds in the United Kingdom, the most cost-effective strategy is to offer all pregnant women a presentation-only scan in late pregnancy. A randomized, controlled trial of screening and intervention to reduce the risk of stillbirth following universal ultrasound to detect macrosomia or fetal growth restriction is unlikely to represent a value for money investment.     

Disparities in hospital utilization in an urban county: use of local consumer marketing data for priority setting.

National data sets are often insufficient for priority setting by local public health systems and the communities they serve. We used marketing data to conduct an ecological analysis of hospital discharge rates in DeKalb County, Georgia, during 1996. Persons living in poorer areas had significantly higher discharge rates for the following conditions: hypertensive disease, blood-related conditions, pneumonia/influenza, diabetes, and chronic obstructive pulmonary diseases. Local marketing data helped identify conditions associated with higher hospital utilization in poorer areas of this urban county. This identification of priority issues informs plans for behavior modification, access to primary care and a healthy environment.     

Cross-Cultural Validation of the Scales for Outcomes in Parkinson's Disease-Psychosocial Questionnaire (SCOPA-PS) in Four Latin American Countries

ObjectiveTo conduct a cross-national validation of the Scales for Outcomes in Parkinson's Disease-PsychoSocial questionnaire (SCOPA-PS) in four Latin American Countries.MethodsData quality (missing items), scale assumptions (item–test correlation), internal consistency (Cronbach's alpha, item homogeneity), factor structure, content validity, and precision (standard error of measurement, SEM) of the scale were explored, as was convergent validity with motor symptoms (Clinical Impression of Severity Index [CISI-PD], Scales for Outcomes in Parkinson's Disease-Motor Scale), emotional status (Hospital Anxiety and Depression Scale) and health-related quality of life (Parkinson Disease Questionnaire-39). Known-groups validity was studied by category of severity, based on Hoehn and Yahr staging (HY), CISI-PD, and disease duration.ResultsThree hundred thirty-one Parkinson's disease (PD) patients with usable data participated (mean age 64.7 years; 42.3% female; mean PD duration 8.5 years; HY, 1 to 5). Data quality (missing items <10%), scale assumptions (item–total correlation = 0.43 ? 0.71) and internal consistency of SCOPA-PS (Cronbach's alpha = 0.87; item homogeneity = 0.38) were satisfactory. Factor analysis suggested a unifactorial structure. High convergent validity was found for depression (r_S = 0.61), anxiety (r_S = 0.62), and health-related quality of life (r_S = 0.82). Known-groups validity analyses indicated a gradual influence of severity category and disease duration on SCOPA-PS scores (P < 0.0001). SEM value was 8.24 (7 to 12 in previous studies). These magnitudes may be indicative of the threshold for a real change and a minimum important difference.ConclusionsThe Latin American versions of the SCOPA-PS displayed appropriate psychometric attributes.     

Standards should be applied in the prevention and handling of missing data for patient-centered outcomes research: a systematic review and expert consensus

ObjectivesTo recommend methodological standards in the prevention and handling of missing data for primary patient-centered outcomes research (PCOR).Study Design and SettingWe searched National Library of Medicine Bookshelf and Catalog as well as regulatory agencies' and organizations' Web sites in January 2012 for guidance documents that had formal recommendations regarding missing data. We extracted the characteristics of included guidance documents and recommendations. Using a two-round modified Delphi survey, a multidisciplinary panel proposed mandatory standards on the prevention and handling of missing data for PCOR.ResultsWe identified 1,790 records and assessed 30 as having relevant recommendations. We proposed 10 standards as mandatory, covering three domains. First, the single best approach is to prospectively prevent missing data occurrence. Second, use of valid statistical methods that properly reflect multiple sources of uncertainty is critical when analyzing missing data. Third, transparent and thorough reporting of missing data allows readers to judge the validity of the findings.ConclusionWe urge researchers to adopt rigorous methodology and promote good science by applying best practices to the prevention and handling of missing data. Developing guidance on the prevention and handling of missing data for observational studies and studies that use existing records is a priority for future research.     

Methodological note: Allocation of disability items in the American Community Survey

BackgroundDetermining the prevalence and correlates of disability requires the use of sample surveys in data analysis. In an effort to generate complete datasets, allocation procedures (i.e., the assignment of values to missing or illogical responses) are frequently used for missing or inconsistent responses.ObjectivesThe goal of this investigation was to explore how six disability-related questions vary in their degree of allocation and how research results may be sensitive to this procedure. This is important because many researchers using large disability information banks are not survey methodologists and may be unaware of how the Census Bureau's editing procedures can influence research findings.MethodsWe use 2010 1-year Public Use Microdata Sample files from the American Community Survey (ACS). We investigated the allocation rates of the following disability items: self-care; hearing; vision; independent living; ambulatory; and cognitive ability. We also asked how allocation rates varied by demographic characteristics and whether the allocated values could influence multivariate results.ResultsDisability item allocation in ACS data have detectable patterns, where the rate of disability allocation is higher for mail surveys, males, older people, groups who speak English not well or not at all, US citizens, Latinos(as), and for people living in or near poverty. Multivariate models may be sensitive to how these allocated values are treated.ConclusionsThe rate of allocations varies as a function of demographic variables because of methodological procedures and survey participation behaviors. Because allocation rates may affect research and policy about the disabled population, more research is required.     

Application of Rasch analysis in health care is increasing and is applied for variable reasons in mobility instruments

ObjectiveTo identify the frequency of Rasch analysis use in health instrument development or refinement and the characteristics of Rasch application in mobility scales.Study Design and SettingThe entire databases of Medline, CINAHL, PEDro, EMBASE, Cochrane Central Register of Controlled Trials, and Cochrane Database of Systematic Reviews were searched until January 2009. Articles that reported the development or refinement of health instruments using Rasch analysis were included. Of the 234 articles that met inclusion, 10 were categorized as “mobility” instruments. Data were extracted relating to each instrument and the use of Rasch analysis in the development or refinement of the instruments.ResultsThe number of articles reporting the use of Rasch analysis of health instruments is increasing, from 1 article in 1987 to 48 articles in 2007. Of the 10 mobility instruments examined, the primary reason Rasch was used varied. Reasons included assessing instrument unidimensionality, differential item functioning, rating categories, item hierarchy, and redundant items.ConclusionThe application of Rasch analysis in health instrument development has markedly increased in recent years. However, few mobility instruments have been developed or refined using Rasch analysis. The reasons that the Rasch model was used varied across mobility instruments.     

Analyzing weight loss intervention studies with missing data: Which methods should be used?

ObjectiveMissing data due to study dropout is common in weight loss trials and several statistical methods exist to account for it. The aim of this study was to identify methods in the literature and to compare the effects of methods of analysis using simulated data sets.MethodsLiterature was obtained for a 1-y period to identify analytical methods used in reporting weight loss trials. A comparison of methods with large or small between-group weight loss, and missing data that was, or was not, missing randomly was conducted in simulated data sets based on previous research.ResultsTwenty-seven studies, some with multiple analyses, were retrieved. Complete case analysis (n = 17), last observation carried forward (n = 6), baseline carried forward (n = 4), maximum likelihood (n = 6), and multiple imputation (n = 2) were the common methods of accounting for missing data. When comparing methods on simulated data, all demonstrated a significant effect when the between-group weight loss was large (P < 0.001, interaction term) regardless of whether the data was missing completely at random. When the weight loss interaction was small, the method used for analysis gave considerably different results with mixed models (P = 0.180) and multiple imputations (P = 0.125) closest to the full data model (P = 0.033).ConclusionThe simulation analysis showed that when data were not missing at random, treatment effects were small, and the amount of missing data was substantial, the analysis method had an effect on the significance of the outcome. Careful attention must be paid when analyzing or appraising studies with missing data and small effects to ensure appropriate conclusions are drawn.     

Validity and Reliability of Stillbirth Data Using Linked Self-Reported and Administrative Datasets

Background

A high rate of stillbirth was previously observed in the Australian Longitudinal Study of Women’s Health (ALSWH). Our primary objective was to test the validity and reliability of self-reported stillbirth data linked to state-based administrative datasets.

Methods

Self-reported data, collected as part of the ALSWH cohort born in 1973–1978, were linked to three administrative datasets for women in New South Wales, Australia (n = 4374): the Midwives Data Collection; Admitted Patient Data Collection; and Perinatal Death Review Database. Linkages were obtained from the Centre for Health Record Linkage for the period 1996–2009. True cases of stillbirth were defined by being consistently recorded in two or more independent data sources. Sensitivity, specificity, positive predictive value, negative predictive value, percent agreement, and kappa statistics were calculated for each dataset.

Results

Forty-nine women reported 53 stillbirths. No dataset was 100% accurate. The administrative datasets performed better than self-reported data, with high accuracy and agreement. Self-reported data showed high sensitivity (100%) but low specificity (30%), meaning women who had a stillbirth always reported it, but there was also over-reporting of stillbirths. About half of the misreported cases in the ALSWH were able to be removed by identifying inconsistencies in longitudinal data.

Conclusions

Data linkage provides great opportunity to assess the validity and reliability of self-reported study data. Conversely, self-reported study data can help to resolve inconsistencies in administrative datasets. Quantifying the strengths and limitations of both self-reported and administrative data can improve epidemiological research, especially by guiding methods and interpretation of findings.Key words: data, kappa, linkage, reliability, self-report, sensitivity, specificity, stillbirth, validity     

Multivariable adjustments counteract spectrum and test review bias in accuracy studies

ObjectivesThe STAndards for Reporting studies of Diagnostic accuracy (STARD) for investigators and editors and the Quality Assessment of Diagnostic Accuracy Studies (QUADAS) for reviewers and readers offer guidelines for the quality and reporting of test accuracy studies. These guidelines address and propose some solutions to two major threats to validity: spectrum bias and test review bias.Study Design and SettingUsing a clinical example, we demonstrate that these solutions fail and propose an alternative solution that concomitantly addresses both sources of bias. We also derive formulas that prove the generality of our arguments.ResultsA logical extension of our ideas is to extend STARD item 23 by adding a requirement for multivariable statistical adjustment using information collected in QUADAS items 1, 2, and 12 and STARD items 3–5, 11, 15, and 18.ConclusionWe recommend reporting not only variation of diagnostic accuracy across subgroups (STARD item 23) but also the effects of the multivariable adjustments on test performance. We also suggest that the QUADAS be supplemented by an item addressing the appropriateness of statistical methods, in particular whether multivariable adjustments have been included in the analysis.     

L’imputation multiple des données manquantes aléatoirement : concepts généraux et présentation d’une méthode Monte-Carlo

BackgroundStatistical analysis of a data set with missing data is a frequent problem to deal with in epidemiology. Methods are available to manage incomplete observations, avoiding biased estimates and improving their precision, compared to more traditional methods, such as the analysis of the sub-sample of complete observations.MethodsOne of these approaches is multiple imputation, which consists in imputing successively several values for each missing data item. Several completed data sets having the same distribution characteristics as the observed data (variability and correlations) are thus generated. Standard analyses are done separately on each completed dataset then combined to obtain a global result. In this paper, we discuss the various assumptions made on the origin of missing data (at random or not), and we present in a pragmatic way the process of multiple imputation. A recent method, Multiple Imputation by Chained Equations (MICE), based on a Monte-Carlo Markov Chain algorithm under missing at random data (MAR) hypothesis, is described. An illustrative example of the MICE method is detailed for the analysis of the relation between a dichotomous variable and two covariates presenting MAR data with no particular structure, through multivariate logistic regression.ResultsCompared with the original dataset without missing data, the results show a substantial improvement of the regression coefficient estimates with the MICE method, relatively to those obtained on the dataset with complete observations.ConclusionThis method does not require any direct assumption on joint distribution of the variables and it is presently implemented in standard statistical software (Splus, Stata). It can be used for multiple imputation of missing data of several variables with no particular structure.