Practical considerations for sensitivity analysis after multiple imputation applied to epidemiological studies with incomplete data

Effectiveness of cladribine compared to fingolimod, natalizumab, ocrelizumab and alemtuzumab in relapsing-remitting multiple sclerosis

BACKGROUND

Comparisons between cladribine and other potent immunotherapies for multiple sclerosis (MS) are lacking.

OBJECTIVES

To compare the effectiveness of cladribine against fingolimod, natalizumab, ocrelizumab and alemtuzumab in relapsing-remitting MS.

METHODS

Patients with relapsing-remitting MS treated with cladribine, fingolimod, natalizumab, ocrelizumab or alemtuzumab were identified in the global MSBase cohort and two additional UK centres. Patients were followed for ⩾6/12 and had ⩾3 in-person disability assessments. Patients were matched using propensity score. Four pairwise analyses compared annualised relapse rates (ARRs) and disability outcomes.

RESULTS

The eligible cohorts consisted of 853 (fingolimod), 464 (natalizumab), 1131 (ocrelizumab), 123 (alemtuzumab) or 493 (cladribine) patients. Cladribine was associated with a lower ARR than fingolimod (0.07 vs. 0.12, p = 0.006) and a higher ARR than natalizumab (0.10 vs. 0.06, p = 0.03), ocrelizumab (0.09 vs. 0.05, p = 0.008) and alemtuzumab (0.17 vs. 0.04, p < 0.001). Compared to cladribine, the risk of disability worsening did not differ in patients treated with fingolimod (hazard ratio (HR) 1.08, 95% confidence interval (CI) 0.47-2.47) or alemtuzumab (HR 0.73, 95% CI 0.26-2.07), but was lower for patients treated with natalizumab (HR 0.35, 95% CI 0.13-0.94) and ocrelizumab (HR 0.45, 95% CI 0.26-0.78). There was no evidence for a difference in disability improvement.

CONCLUSION

Cladribine is an effective therapy that can be viewed as a step up in effectiveness from fingolimod, but is less effective than the most potent intravenous MS therapies.

Self-help mobile messaging intervention for depression among older adults in resource-limited settings: a randomized controlled trial

Scalable solutions to treat depression in older adults in low-resourced settings are urgently needed. The PRODIGITAL-D pragmatic, single-blind, two-arm, individually randomized controlled trial assessed the effectiveness of a mobile messaging psychosocial intervention in improving depressive symptomatology among older adults in socioeconomically deprived areas of Guarulhos, Brazil. Older adults (aged 60+ years) registered with 24 primary care clinics and identified with depressive symptomatology (9-item Patient Health Questionnaire (PHQ-9) scores ≥ 10) received the 6-week Viva Vida intervention based on psychoeducation and behavioral activation (n = 298) or a single message (n = 305). No health professional support was offered. The primary outcome was improvement from depressive symptomatology (PHQ-9 < 10) at 3 months. Of the 603 participants enrolled (mean age = 65.1 years; 451 (74.8%) women), 527 (87.4%) completed the follow-up assessment. In the intervention arm, 109 of 257 (42.4%) participants had an improved depressive symptomatology, compared with 87 of 270 (32.2%) participants in the control arm (adjusted odds ratio = 1.57; 95% confidence interval = 1.07-2.29; P = 0.019). No severe adverse events related to trial participation were observed. These results demonstrate the usefulness of a digital messaging psychosocial intervention in the short-term improvement from depressive symptomatology that can potentially be integrated into primary care programs for treating older adults with depression. Brazilian Registry of Clinical Trials registration: ReBEC ( RBR-4c94dtn ).

Random forest classification as a tool in epidemiological modelling: Identification of farm-specific characteristics relevant for the occurrence of Fasciola hepatica on German dairy farms

Fasciola hepatica is an internal parasite of both human and veterinary relevance. In order to control fasciolosis, a multitude of attempts to predict the risk of infection such as risk maps or forecasting models have been developed. These attempts mainly focused on the influence of geo-climatic and meteorological features. Predicting bovine fasciolosis on farm level taking into account farm-specific settings yet remains challenging. In the present study, a new methodology for this purpose, a data-driven machine learning approach using a random forest classification algorithm was applied to a cross-sectional data set of farm characteristics, management regimes, and farmer aspects within two structurally different dairying regions in Germany in order to identify factors relevant for the occurrence of F. hepatica that could predict farm-level bulk tank milk positivity. The resulting models identified farm-specific key aspects in regard to the presence of F. hepatica. In study region North, farm-level production parameters (farm-level milk yield, farm-level milk fat, farm-level milk protein), leg hygiene, body condition (prevalence of overconditioned and underconditioned cows, respectively) and pasture access were identified as features relevant in regard to farm-level F. hepatica positivity. In study region South, pasture access together with farm-level lameness prevalence, farm-level prevalence of hock lesions, herd size, parity, and farm-level milk fat appeared to be important covariates. The stratification of the analysis by study region allows for the extrapolation of the results to similar settings of dairy husbandry. The local, region-specific modelling of F. hepatica presence in this work contributes to the understanding of on-farm aspects of F. hepatica appearance. The applied technique represents a novel approach in this context to model epidemiological data on fasciolosis which allows for the identification of farms at risk and together with additional findings in regard to the epidemiology of fasciolosis, can facilitate risk assessment and deepen our understanding of on-farm drivers of the occurrence of F. hepatica.

Perivascular spaces mediate a relationship between diabetes and other cerebral small vessel disease markers in cerebrovascular and neurodegenerative diseases

Type 2 diabetes mellitus (T2DM) and hypertension are risk factors for cerebral small vessel disease (SVD); however, few studies have characterised their relationships with MRI-visible perivascular spaces (PVS). MRI was used to quantify deep (d) and periventricular (p) white matter hyperintensities (WMH), lacunes, PVS in the white matter (wmPVS) or basal ganglia (bgPVS), and diffusion metrics in white matter. Patients with T2DM had greater wmPVS volume and there were greater wmPVS volumes in patients with T2DM and hypertension together. Counterfactual moderated mediation models found indirect effects of T2DM on volumes of other SVD and diffusion markers that were mediated by wmPVS: pWMH, dWMH, periventricular lacunes, and deep lacunes, and progression of deep lacunes over 1 year, in patients with hypertension, but not in patients without hypertension. Studying the regulation of cortical perivascular fluid dynamics may reveal mechanisms that mediate the impact of T2DM on cerebral small vessels.

Risk factors and mortality associated with undertriage after major trauma in a physician-led prehospital system: a retrospective multicentre cohort study

PURPOSE

To assess the incidence of undertriage in major trauma, its determinant, and association with mortality.

METHODS

A multicentre retrospective cohort study was conducted using data from a French regional trauma registry (2011-2017). All major trauma (Injury Severity Score ≥ 16) cases aged ≥ 18 years and managed by a physician-led mobile medical team were included. Those transported to a level-II/III trauma centre were considered as undertriaged. Multivariable logistic regression was used to identify factors associated with undertriage.

RESULTS

A total of 7110 trauma patients were screened; 2591 had an ISS ≥ 16 and 320 (12.4%) of these were undertriaged. Older patients had higher risk for undertriage (51-65 years: OR = 1.60, 95% CI [1.11; 2.26], p = 0.01). Conversely, injury mechanism (fall from height: 0.62 [0.45; 0.86], p = 0.01; gunshot/stab injuries: 0.45 [0.22; 0.90], p = 0.02), on-scene time (> 60 min: 0.62 [0.40; 0.95], p = 0.03), prehospital endotracheal intubation (0.53 [0.39; 0.71], p < 0.001), and prehospital focussed assessment with sonography [FAST] (0.15 [0.08; 0.29], p < 0.001) were associated with a lower risk for undertriage. After adjusting for severity, undertriage was not associated with a higher risk of mortality (1.22 [0.80; 1.89], p = 0.36).

CONCLUSIONS

In our physician-led prehospital EMS system, undertriage was higher than recommended. Advanced aged was identified as a risk factor highlighting the urgent need for tailored triage protocol in this population. Conversely, the potential benefit of prehospital FAST on triage performance should be furthered explored as it may reduce undertriage. Fall from height and penetrating trauma were associated with a lower risk for undertriage suggesting that healthcare providers should remain vigilant of the potential seriousness of trauma associated with low-energy mechanisms.

Measuring bereavement prevalence in a complex sampling survey: the 2019 Georgia Behavioral Risk Factor Surveillance System (BRFSS)

BACKGROUND

The Behavioral Risk Factor Surveillance System (BRFSS) is an annual survey designed to identify trends in the public's health. In its 2019 field survey, the U.S. state of GA tested a new 3 - item module to measure the numbers of bereaved, resident adults aged 18 years and older. Participants were eligible if they answered 'Yes' to the item 'Have you experienced the death of a family member or close friend in the years 2018 or 2019?'. This analysis explores two research questions. Can estimates for bereavement prevalence be derived without large sampling errors, low precision, and small subsamples? Can multiple imputation techniques be applied to overcome non-response and missing data to support multivariate modeling?

METHODS

BRFSS is a non-institutionalized sample of adults aged 18 years and older living in the U.S. state of Georgia. Analyses in this study were conducted under two scenarios. Scenario 1 applies the complex sample weights created by the Centers for Disease Control and imputes values for missing responses. Scenario 2 treats the data as a panel - no weighting combined with removal of persons with missing data. Scenario 1 reflects the use of BRFSS data for public health and policy, while Scenario 2 reflects data as it is commonly used in social science research studies.

RESULTS

The bereavement screening item has a response rate (RR) of 69.1% (5206 of 7534 persons). Demographic subgroups and categories of health have RR of 55% or more. Under Scenario 1, the estimated prevalence of bereavement is 45.38%, meaning that 3,739,120 adults reported bereaved in 2018 or 2019. The estimated prevalence is 46.02% with Scenario 2 which removes persons with any missing data (4,289 persons). Scenario 2 overestimates the bereavement prevalence by 1.39%. An illustrative logistic model is presented to show the performance of exposure to bereavement under the two data scenarios.

CONCLUSIONS

Recent bereavement can be ascertained in a surveillance survey accounting for biases in response. Estimating bereavement prevalence is needed for measuring population health. This survey is limited to one US state in a single year and excludes persons aged 17 years and younger.

The Effect of a Liver Transplant Program on the Outcomes of Resectable Hepatocellular Carcinoma: A Nationwide Multicenter Analysis

OBJECTIVE

To evaluate the effect of a liver transplantation (LT) program on the outcomes of resectable hepatocellular carcinoma (HCC).

BACKGROUND

Surgical treatment of HCC includes both hepatic resection (HR) and LT. However, the presence of cirrhosis and the possibility of recurrence make the management of this disease complex and probably different according to the presence of a LT program.

METHODS

Patients undergoing HR for HCC between January 2005 and December 2019 were identified from a national database of HCC. The main study outcomes were major surgical complications according to the Comprehensive Complication Index, posthepatectomy liver failure (PHLF), 90-day mortality, overall survival, and disease-free survival. Secondary outcomes were salvage liver transplantation (SLT) and postrecurrence survival.

RESULTS

A total of 3202 patients were included from 25 hospitals over the study period. Three of 25 (12%) had an LT program. The presence of an LT program within a center was associated with a reduced probability of PHLF (odds ratio=0.38) but not with overall survival and disease-free survival. There was an increased probability of SLT when HR was performed in a transplant hospital (odds ratio=12.05). Among transplant-eligible patients, those who underwent LT had a significantly longer postrecurrence survival.

CONCLUSIONS

This study showed that the presence of a LT program was associated with decreased PHLF rates and an increased probability to receive SLT in case of recurrence.

Clinical correlates of early onset antipsychotic treatment resistance

BACKGROUND

There is evidence of heterogeneity within treatment-resistant schizophrenia (TRS), with some people not responding to antipsychotic treatment from illness onset and others becoming treatment-resistant after an initial response period. These groups may have different aetiologies.

AIM

This study investigates sociodemographic and clinical correlates of early onset of TRS.

METHOD

Employing a retrospective cohort design, we do a secondary analysis of data from a cohort of people with TRS attending the South London and Maudsley. Regression analyses were conducted to identify the correlates of the length of treatment to TRS. Predictors included the following: gender, age, ethnicity, problems with positive symptoms, problems with activities of daily living, psychiatric comorbidities, involuntary hospitalisation and treatment with long-acting injectable antipsychotics.

RESULTS

In a cohort of 164 people with TRS (60% were men), the median length of treatment to TRS was 3 years and 8 months. We observed no cut-off on the length of treatment until TRS presentation differentiating between early and late TRS (i.e. no bimodal distribution). Having mild to very severe problems with hallucinations and delusions at the treatment start was associated with earlier TRS (~19 months earlier). In sensitivity analyses, including only complete cases (subject to selection bias), treatment with a long-acting injectable antipsychotic was additionally associated with later TRS (~15 months later).

CONCLUSION

Our findings do not support a clear separation between early and late TRS but rather a continuum of the length of treatment before TRS onset. Having mild to very severe problems with positive symptoms at treatment start predicts earlier onset of TRS.

Handling missing data in clinical research

Because missing data are present in almost every study, it is important to handle missing data properly. First of all, the missing data mechanism should be considered. Missing data can be either completely at random (MCAR), at random (MAR), or not at random (MNAR). When missing data are MCAR, a complete case analysis can be valid. Also when missing data are MAR, in some situations a complete case analysis leads to valid results. However, in most situations, missing data imputation should be used. Regarding imputation methods, it is highly advised to use multiple imputations because multiple imputations lead to valid estimates including the uncertainty about the imputed values. When missing data are MNAR, also multiple imputations do not lead to valid results. A complication hereby is that it not possible to distinguish whether missing data are MAR or MNAR. Finally, it should be realized that preventing to have missing data is always better than the treatment of missing data.

ROCCA cohort study: Nationwide results on safety of Gam-COVID-Vac vaccine (Sputnik V) in the Republic of San Marino using active surveillance

BACKGROUND

Gam-COVID-Vac is the world's first registered vector vaccine against COVID-19 based on a combination of two heterologous adenoviruses. It was chosen by the Republic of San Marino as the main tool in its vaccination campaign, which started on 25 February 2021. Our aim was to build up on the ROCCA study, focused on the older population, by describing adverse effects following immunisation (AEFIs) rates and characteristics in all age groups for the first time in a real-world context.

METHODS

An active surveillance study on recipients of at least one dose of the Gam-COVID-Vac vaccine was conducted. Participants were administered online questionnaires through live/phone interviews with physicians, by e-mail or by scanning a QR code at different points in time after the first dose: one week (Q1) one month (Q2), and three months (Q3) between March and August 2021.

FINDINGS

Overall, 6190 vaccine recipients were recruited. Mean age was 52·4 ± 18·2 years. After the first dose, systemic reactions were reported by 57·5% of the participants, while injection site reactions were reported by 46·7%. The most common AEFIs were pain at the injection site, fatigue and headache. Grade 3 or 4 AEFIs were reported by 0·8% and 0·3% of the participants, respectively. After the second dose, systemic reactions were reported by 63·1% of the participants, while injection site reactions by 54·7%. The most common AEFIs were malaise, pain at injection site and myalgia. Grade 3 or 4 AEFIs were reported by 2·7% and 1·1% of the participants, respectively. Multivariate analysis showed younger age, being a woman and food allergies are risk factors for more severe AEFIs.

INTERPRETATION

Our results confirm a good tolerability profile for the population aged 18 and over providing useful data for vaccination campaigns ongoing in countries planning to use Gam-COVID-Vac.

FUNDING

None.

Comparative effectiveness of hysteroscopic and laparoscopic sterilization for women: a retrospective cohort study

OBJECTIVE

To compare real-world effectiveness of hysteroscopic to laparoscopic sterilization.

DESIGN

Retrospective cohort of Medicaid claims for hysteroscopic or laparoscopic sterilization procedures performed in California, 2008-2014. After excluding postpartum procedures, we applied log-linear (Poisson) event-history regression models for clustered person-period data, weighted for propensity to receive either sterilization procedures, and adjusted for sociodemographic and clinical variables to examine the poststerilization pregnancy rates.

SETTING

Clinics, hospitals.

PATIENT(S)

Women aged 18-50 years with Medicaid claims between January 1, 2008, and August 31, 2014.

INTERVENTION(S)

Hysteroscopic or laparoscopic sterilization procedure.

MAIN OUTCOME MEASURE(S)

Poststerilization pregnancy measured by pregnancy-related claims.

RESULT(S)

Among women with hysteroscopic (n = 5,906) or laparoscopic (n = 23,965) sterilization, poststerilization pregnancy claims were identified for 4.74% of women after hysteroscopic sterilization and 5.57% after laparoscopic sterilization. The pregnancy rates decreased over time after either procedure. Twelve months after the procedure, the crude incidence of pregnancy claims was higher for hysteroscopic sterilization than for laparoscopic sterilization (3.26 vs. 2.61 per 100 woman-years), but the propensity-weighted adjusted incidence rate ratio was 1.06 (95% confidence interval [CI], 0.85-1.26). Between 13 and 24 months after the procedure, there were fewer pregnancies for women after hysteroscopic sterilizations than for those after laparoscopic sterilizations (adjusted incidence rate ratio, 0.63 [95% CI, 0.45-0.88]), with no statistically significant differences in later years. The cumulative pregnancy rates 5 years after sterilization were lower with hysteroscopic sterilization than with laparoscopic sterilization (6.26 vs. 7.22 per 100 woman-years; propensity-weighted, adjusted risk ratio, 0.76 [95% CI, 0.62-0.90]). The poststerilization pregnancy rates varied by age and race/ethnicity.

CONCLUSION(S)

The pregnancy rates after female sterilization are higher than expected, whether performed hysteroscopically or laparoscopically. These findings are reassuring that the effectiveness of hysteroscopic sterilization was not inferior to laparoscopic sterilization.

CLINICAL TRIAL REGISTRATION NUMBER

NCT03438682.

First-line therapies in late-onset multiple sclerosis: An Italian registry study

BACKGROUND AND PURPOSE

The diagnosis of late-onset (age ≥50 years old) relapsing remitting multiple sclerosis (LORRMS) has been increasingly described in clinical practice, whereas data focusing on the specific therapeutic management of LORRMS are scarce. Our objective was to compare the effectiveness of injectable and oral first-line disease-modifying therapies (DMTs) in a cohort of LORRMS patients with time to first relapse, time to confirmed disability progression (CDP), and time to discontinuation.

METHODS

This is a multicenter, observational, retrospectively acquired cohort study on LORRMS-naïve patients from the Italian MS Register who started either injectable or oral first-line DMTs between January 1, 2013 and December 31, 2017. LORRMS patients were divided into two groups, namely the injectable group (IG) and oral group (OG). Cox models adjusted with inverse probability-weighted propensity score were built for the investigated outcomes.

RESULTS

Of a cohort of 3989 patients, 302 were enrolled (203 in the IG and 99 in the OG). The two cohorts did not differ in baseline characteristics. Time to first relapse did not show any difference between the two groups (hazard ratio [HR]: 1.10; 95% confidence interval [CI]: 0.50-2.46, p = 0.797). Furthermore, no differences were found between the two groups with respect to the risk of CDP (HR: 1.04; 95% CI: 0.35-3.06, p = 0.939), nor for the risk of DMT discontinuation (HR: 0.90; 95% CI: 0.17-2.08, p = 0.425).

CONCLUSIONS

Real-world data from the Italian MS Register suggested that both injectables and oral first-line DMTs similarly controlled the investigated outcomes in LORRMS.

Injectable Versus Oral First-Line Disease-Modifying Therapies: Results from the Italian MS Register

The current study aims to compare injectable and oral first-line disease-modifying therapies (DMTs) for time to first relapse, time to confirmed disability progression (CDP), and time to discontinuation using a cohort of relapsing remitting multiple sclerosis (RRMS) patients, with data extracted from the Italian MS Register. This multicenter, observational, retrospectively acquired, and propensity-adjusted cohort study utilized RRMS-naïve patients from the Italian MS Register who started either injectable or oral first-line DMTs between January 1, 2010, and December 31, 2017, to evaluate the impact on disability outcomes in patients. Enrolled patients were divided into two groups, namely the injectable group (IG) and the oral group (OG). Of a cohort of 11,416 patients, 4602 were enrolled (3919 in the IG and 683 in the OG). The IG had a higher rate of women (67.3% vs 63.4%, p < 0.05) and a lower mean age (36.1 ± 10.9 vs 38.9 ± 11.8, p < 0.001). The event time to first relapse demonstrated a lower risk in the OG (HR = 0.58; CI 95% 0.48-0.72, p < 0.001). However, no differences were found between the two groups with respect to the risk of CDP (HR = 0.94; CI 95% 0.76-1.29, p = 0.941), while a lower risk of DMT was found in the OG (HR = 0.72; CI 95% 0.58-0.88, p = 0.002) for the event time to discontinuation. Real-world data from the Italian MS Register suggests that first-line oral DMTs are associated with a lower risk of experiencing a new relapse and of therapy discontinuation compared to injectable DMTs.

A practical guide to multiple imputation of missing data in nephrology

Health data are often plagued with missing values that can greatly reduce the sample size if only complete cases are considered for analysis. Furthermore, analyses that ignore missing data have the potential to introduce bias in the parameter estimates. Multiple imputation techniques have been developed to recover the information that would otherwise be lost when excluding observations with missing data and to help minimize bias. However, the validity of analyses using imputed data relies on the imputation model having been correctly specified. The aim of this guide is to aid the reader in the decision-making process when conducting an analysis with multiply imputed data in the context of nephrology research. We discuss (i) missing mechanism assumption, (ii) imputation method, (iii) imputation model, (iv) derived variables, (v) the number of imputed data sets, (vi) diagnostic checks, (vii) analysis and pooling of results, and (viii) reporting the results. This process is demonstrated using data from the National Health and Nutrition Examination Survey to explore the association between hypertension and kidney disease in adults from the general population. Example code is provided for SAS software and the mice package in R.

Dimethyl fumarate vs Teriflunomide: an Italian time-to-event data analysis

BACKGROUND

The introduction of oral disease-modifying therapies (DMTs) for relapsing-remitting multiple sclerosis (RRMS) changed algorithms of RRMS treatment.

OBJECTIVES

To compare the effectiveness of treatment with dimethyl fumarate (DMF) and teriflunomide (TRF) in a large multicentre Italian cohort of RRMS patients.

MATERIALS AND METHODS

Patients with RRMS who received treatment with DMF and TRF between January 1st, 2012 and December 31st, 2018 from twelve MS centers were identified. The events investigated were "time-to-first-relapse", "time-to-Magnetic-Resonance-Imaging (MRI)-activity" and "time-to-disability-progression".

RESULTS

1445 patients were enrolled (1039 on DMF, 406 on TRF) and followed for a median of 34 months. Patients on TRF were older (43.5 ± 8.6 vs 38.8 ± 9.2 years), with a predominance of men and higher level of disability (p < 0.001 for all). Patients on DMF had a higher number of relapses and radiological activity (p < .05) at baseline. Time-varying Cox-model for the event "time-to-first relapse" revealed that no differences were found between the two groups in the first 38 months of treatment (HR_t < 38DMF = 0.73, CI = 0.52 to 1.03, p = 0.079). When the time-on-therapy exceeds 38 months patients on DMF had an approximately 0.3 times lower relapse hazard risk than those who took TRF (HR_t>38DMF = 3.83, CI = 1.11 to 13.23, p = 0.033). Both DMTs controlled similarly MRI activity and disability progression.

CONCLUSIONS

Patients on DMF had higher relapse-free survival time than TRF group after the first 38 months on therapy.

Handling missing data in modelling quality of clinician-prescribed routine care: Sensitivity analysis of departure from missing at random assumption

Missing information is a major drawback in analyzing data collected in many routine health care settings. Multiple imputation assuming a missing at random mechanism is a popular method to handle missing data. The missing at random assumption cannot be confirmed from the observed data alone, hence the need for sensitivity analysis to assess robustness of inference. However, sensitivity analysis is rarely conducted and reported in practice. We analyzed routine paediatric data collected during a cluster randomized trial conducted in Kenyan hospitals. We imputed missing patient and clinician-level variables assuming the missing at random mechanism. We also imputed missing clinician-level variables assuming a missing not at random mechanism. We incorporated opinions from 15 clinical experts in the form of prior distributions and shift parameters in the delta adjustment method. An interaction between trial intervention arm and follow-up time, hospital, clinician and patient-level factors were included in a proportional odds random-effects analysis model. We performed these analyses using R functions derived from the jomo package. Parameter estimates from multiple imputation under the missing at random mechanism were similar to multiple imputation estimates assuming the missing not at random mechanism. Our inferences were insensitive to departures from the missing at random assumption using either the prior distributions or shift parameters sensitivity analysis approach.

Methodological standards for the development and evaluation of clinical prediction rules: a review of the literature

Clinical prediction rules (CPRs) that predict the absolute risk of a clinical condition or future outcome for individual patients are abundant in the medical literature; however, systematic reviews have demonstrated shortcomings in the methodological quality and reporting of prediction studies. To maximise the potential and clinical usefulness of CPRs, they must be rigorously developed and validated, and their impact on clinical practice and patient outcomes must be evaluated. This review aims to present a comprehensive overview of the stages involved in the development, validation and evaluation of CPRs, and to describe in detail the methodological standards required at each stage, illustrated with examples where appropriate. Important features of the study design, statistical analysis, modelling strategy, data collection, performance assessment, CPR presentation and reporting are discussed, in addition to other, often overlooked aspects such as the acceptability, cost-effectiveness and longer-term implementation of CPRs, and their comparison with clinical judgement. Although the development and evaluation of a robust, clinically useful CPR is anything but straightforward, adherence to the plethora of methodological standards, recommendations and frameworks at each stage will assist in the development of a rigorous CPR that has the potential to contribute usefully to clinical practice and decision-making and have a positive impact on patient care.

Comparison of fingolimod, dimethyl fumarate and teriflunomide for multiple sclerosis

OBJECTIVE

Oral immunotherapies have become a standard treatment in relapsing-remitting multiple sclerosis. Direct comparison of their effect on relapse and disability is needed.

METHODS

We identified all patients with relapsing-remitting multiple sclerosis treated with teriflunomide, dimethyl fumarate or fingolimod, with minimum 3-month treatment persistence and disability follow-up in the global MSBase cohort study. Patients were matched using propensity scores. Three pairwise analyses compared annualised relapse rates and hazards of disability accumulation, disability improvement and treatment discontinuation (analysed with negative binomial models and weighted conditional survival models, with pairwise censoring).

RESULTS

The eligible cohorts consisted of 614 (teriflunomide), 782 (dimethyl fumarate) or 2332 (fingolimod) patients, followed over the median of 2.5 years. Annualised relapse rates were lower on fingolimod compared with teriflunomide (0.18 vs 0.24; p=0.05) and dimethyl fumarate (0.20 vs 0.26; p=0.01) and similar on dimethyl fumarate and teriflunomide (0.19 vs 0.22; p=0.55). No differences in disability accumulation (p≥0.59) or improvement (p≥0.14) were found between the therapies. In patients with ≥3-month treatment persistence, subsequent discontinuations were less likely on fingolimod than teriflunomide and dimethyl fumarate (p<0.001). Discontinuation rates on teriflunomide and dimethyl fumarate were similar (p=0.68).

CONCLUSION

The effect of fingolimod on relapse frequency was superior to teriflunomide and dimethyl fumarate. The effect of the three oral therapies on disability outcomes was similar during the initial 2.5 years on treatment. Persistence on fingolimod was superior to the two comparator drugs.

Do Maternal Dietary Antioxidants Modify the Relationship Between Binge Drinking and Small for Gestational Age? Findings from a Longitudinal Cohort Study

BACKGROUND

Vitamin C, vitamin E, and carotenoids are potent dietary antioxidants that have been shown to attenuate ethanol-induced harm in animal models of fetal alcohol spectrum disorders. A diet low in antioxidant-rich foods may induce a state of oxidative stress in the context of maternal alcohol consumption during pregnancy, potentially causing growth restriction in the developing fetus.

METHODS

We conducted a secondary analysis of a longitudinal U.K. birth cohort. The sample comprised 9,699 women and their babies in Avon, U.K., with an estimated delivery date between April 1, 1991 and December 31, 1992. Alcohol consumption data were self-reported at 18 weeks' gestation via a postal questionnaire. Women reported any binge drinking (≥4 U.K. units/occasion) during the past month. Dietary data were self-reported at 32 weeks' gestation using a food frequency questionnaire. Estimated intakes of vitamins C and E and carotenoids were categorized into quartiles. Logistic regression models with interaction terms were used to investigate relationships between maternal binge drinking, dietary antioxidants, and fetal growth. Models were adjusted for maternal sociodemographic and lifestyle characteristics. Small for gestational age (SGA; <10th percentile) was defined using customized birth centiles.

RESULTS

In the unadjusted models, binge drinking was associated with higher risk of SGA birth (odds ratio [OR] 1.38, 95% confidence interval [CI] 1.10, 1.72, p = 0.005), and higher maternal intakes of vitamin C (OR = 0.90, 95% CI 0.84, 0.96, p = 0.002) and vitamin E (OR = 0.90, 95% CI 0.84, 0.95, p < 0.0001) were associated with lower risk of SGA birth. However, addition of potentially confounding variables attenuated these relationships. Likelihood ratio tests indicated that interaction terms were not significant for vitamin C (p = 0.116), vitamin E (p = 0.059), or carotenoid intakes (p = 0.174).

CONCLUSIONS

There was no evidence of maternal intake of dietary antioxidants modifying the relationship between maternal binge drinking and SGA birth.

Characterizing and Managing Missing Structured Data in Electronic Health Records: Data Analysis

BACKGROUND

Missing data is a challenge for all studies; however, this is especially true for electronic health record (EHR)-based analyses. Failure to appropriately consider missing data can lead to biased results. While there has been extensive theoretical work on imputation, and many sophisticated methods are now available, it remains quite challenging for researchers to implement these methods appropriately. Here, we provide detailed procedures for when and how to conduct imputation of EHR laboratory results.

OBJECTIVE

The objective of this study was to demonstrate how the mechanism of missingness can be assessed, evaluate the performance of a variety of imputation methods, and describe some of the most frequent problems that can be encountered.

METHODS

We analyzed clinical laboratory measures from 602,366 patients in the EHR of Geisinger Health System in Pennsylvania, USA. Using these data, we constructed a representative set of complete cases and assessed the performance of 12 different imputation methods for missing data that was simulated based on 4 mechanisms of missingness (missing completely at random, missing not at random, missing at random, and real data modelling).

RESULTS

Our results showed that several methods, including variations of Multivariate Imputation by Chained Equations (MICE) and softImpute, consistently imputed missing values with low error; however, only a subset of the MICE methods was suitable for multiple imputation.

CONCLUSIONS

The analyses we describe provide an outline of considerations for dealing with missing EHR data, steps that researchers can perform to characterize missingness within their own data, and an evaluation of methods that can be applied to impute clinical data. While the performance of methods may vary between datasets, the process we describe can be generalized to the majority of structured data types that exist in EHRs, and all of our methods and code are publicly available.

Effects of women's groups practising participatory learning and action on preventive and care-seeking behaviours to reduce neonatal mortality: A meta-analysis of cluster-randomised trials

BACKGROUND

The World Health Organization recommends participatory learning and action (PLA) in women's groups to improve maternal and newborn health, particularly in rural settings with low access to health services. There have been calls to understand the pathways through which this community intervention may affect neonatal mortality. We examined the effect of women's groups on key antenatal, delivery, and postnatal behaviours in order to understand pathways to mortality reduction.

METHODS AND FINDINGS

We conducted a meta-analysis using data from 7 cluster-randomised controlled trials that took place between 2001 and 2012 in rural India (2 trials), urban India (1 trial), rural Bangladesh (2 trials), rural Nepal (1 trial), and rural Malawi (1 trial), with the number of participants ranging between 6,125 and 29,901 live births. Behavioural outcomes included appropriate antenatal care, facility delivery, use of a safe delivery kit, hand washing by the birth attendant prior to delivery, use of a sterilised instrument to cut the umbilical cord, immediate wrapping of the newborn after delivery, delayed bathing of the newborn, early initiation of breastfeeding, and exclusive breastfeeding. We used 2-stage meta-analysis techniques to estimate the effect of the women's group intervention on behavioural outcomes. In the first stage, we used random effects models with individual patient data to assess the effect of groups on outcomes separately for the different trials. In the second stage of the meta-analysis, random effects models were applied using summary-level estimates calculated in the first stage of the analysis. To determine whether behaviour change was related to group attendance, we used random effects models to assess associations between outcomes and the following categories of group attendance and allocation: women attending a group and allocated to the intervention arm; women not attending a group but allocated to the intervention arm; and women allocated to the control arm. Overall, women's groups practising PLA improved behaviours during and after home deliveries, including the use of safe delivery kits (odds ratio [OR] 2.92, 95% CI 2.02-4.22; I2 = 63.7%, 95% CI 4.4%-86.2%), use of a sterile blade to cut the umbilical cord (1.88, 1.25-2.82; 67.6%, 16.1%-87.5%), birth attendant washing hands prior to delivery (1.87, 1.19-2.95; 79%, 53.8%-90.4%), delayed bathing of the newborn for at least 24 hours (1.47, 1.09-1.99; 68.0%, 29.2%-85.6%), and wrapping the newborn within 10 minutes of delivery (1.27, 1.02-1.60; 0.0%, 0%-79.2%). Effects were partly dependent on the proportion of pregnant women attending groups. We did not find evidence of effects on uptake of antenatal care (OR 1.03, 95% CI 0.77-1.38; I2 = 86.3%, 95% CI 73.8%-92.8%), facility delivery (1.02, 0.93-1.12; 21.4%, 0%-65.8%), initiating breastfeeding within 1 hour (1.08, 0.85-1.39; 76.6%, 50.9%-88.8%), or exclusive breastfeeding for 6 weeks after delivery (1.18, 0.93-1.48; 72.9%, 37.8%-88.2%). The main limitation of our analysis is the high degree of heterogeneity for effects on most behaviours, possibly due to the limited number of trials involving women's groups and context-specific effects.

CONCLUSIONS

This meta-analysis suggests that women's groups practising PLA improve key behaviours on the pathway to neonatal mortality, with the strongest evidence for home care behaviours and practices during home deliveries. A lack of consistency in improved behaviours across all trials may reflect differences in local priorities, capabilities, and the responsiveness of health services. Future research could address the mechanisms behind how PLA improves survival, in order to adapt this method to improve maternal and newborn health in different contexts, as well as improve other outcomes across the continuum of care for women, children, and adolescents.

mStroke: "Mobile Stroke"-Improving Acute Stroke Care with Smartphone Technology

OBJECTIVE

This study aimed to evaluate the effect of method and time of system activation on clinical metrics in cases utilizing the Stop Stroke (Pulsara, Inc.) mobile acute stroke care coordination application.

METHODS

A retrospective cohort analysis of stroke codes at 12 medical centers using Stop Stroke from March 2013 to May 2016 was performed. Comparison of metrics (door-to-needle time [DTN] and door-to-CT time [DTC], and rate of DTN ≤ 60 minutes [goal DTN]) was performed between subgroups based on method (emergency medical service [EMS] versus emergency department [ED]) and time of activation. Effects were adjusted for confounders (age, sex, National Institutes of Health Stroke Scale [NIHSS] score) using multiple linear and logistic regression.

RESULTS

The final dataset included 2589 cases. Cases activated by EMS were more severe (median NIHSS score 8 versus 4, P < .0001) and more likely to receive recombinant tissue plasminogen activator (20% versus 12%, P < .0001) than those with ED activation. After adjustment, cases with EMS activation had shorter DTC (6.1 minutes shorter, 95% CI [-10.3, -2]) and DTN (12.8 minutes shorter, 95% CI [-21, -4.6]) and were more likely to meet goal DTN (OR 1.83, 95% CI [1.1, 3]). Cases between 1200 and 1800 had longer DTC (7.7 minutes longer, 95% CI [2.4, 13]) and DTN (21.1 minutes longer, 95% CI [9.3, 33]), and reduced rate of goal DTN (OR .3, 95% CI [.15, .61]) compared to those between 0000 and 0600.

CONCLUSIONS

Incorporating real-time prehospital data obtained via smartphone technology provides unique insight into acute stroke codes. Activation of mobile electronic stroke coordination in the field appears to promote a more expedited and successful care process.

Childhood Language Disorder and Social Anxiety in Early Adulthood

Language disorder is associated with anxiety and with social problems in childhood and adolescence. However, the relation between language disorder and adult social anxiety is not well known. This study examines social anxiety in early adulthood in a 26-year prospective longitudinal study following individuals identified with a communication disorder at age 5 and a control group. Social anxiety diagnoses and subthreshold symptoms were examined at ages 19, 25, and 31 using a structured diagnostic interview; social anxiety symptoms related to social interaction and social performance were also assessed dimensionally at age 31. Multiple imputation was used to address attrition. Compared to controls, participants with childhood language disorder had higher rates of subthreshold social phobia at ages 19 and 25 and endorsed higher levels of social interaction anxiety symptoms at age 31, with particular difficulty talking to others and asserting their perspectives. Childhood language disorder is a specific risk factor for a circumscribed set of social anxiety symptoms in adulthood, which are likely associated with communication challenges.

Survival Trends After Surgery for Acute Subdural Hematoma in Adults Over a 20-year Period

OBJECTIVE

We sought to determine 30-day survival trends and prognostic factors following surgery for acute subdural hematomas (ASDHs) in England and Wales over a 20-year period.

SUMMARY OF BACKGROUND DATA

ASDHs are still considered the most lethal type of traumatic brain injury. It remains unclear whether the adjusted odds of survival have improved significantly over time.

METHODS

Using the Trauma Audit and Research Network (TARN) database, we analyzed ASDH cases in the adult population (>16 yrs) treated surgically between 1994 and 2013. Two thousand four hundred ninety-eight eligible cases were identified. Univariable and multiple logistic regression analyses were performed, using multiple imputation for missing data.

RESULTS

The cohort was 74% male with a median age of 48.9 years. Over half of patients were comatose at presentation (53%). Mechanism of injury was due to a fall (<2 m 34%, >2 m 24%), road traffic collision (25%), and other (17%). Thirty-six per cent of patients presented with polytrauma. Gross survival increased from 59% in 1994 to 1998 to 73% in 2009 to 2013. Under multivariable analysis, variables independently associated with survival were year of injury, Glasgow Coma Scale, Injury Severity Score, age, and pupil reactivity. The time interval from injury to craniotomy and direct admission to a neurosurgical unit were not found to be significant prognostic factors.

CONCLUSIONS

A significant improvement in survival over the last 20 years was observed after controlling for multiple prognostic factors. Prospective trials and cohort studies are expected to elucidate the distribution of functional outcome in survivors.

Impact of home visit capacity on genetic association studies of late-onset Alzheimer's disease

INTRODUCTION

Findings for genetic correlates of late-onset Alzheimer's disease (LOAD) in studies that rely solely on clinic visits may differ from those with capacity to follow participants unable to attend clinic visits.

METHODS

We evaluated previously identified LOAD-risk single nucleotide variants in the prospective Adult Changes in Thought study, comparing hazard ratios (HRs) estimated using the full data set of both in-home and clinic visits (n = 1697) to HRs estimated using only data that were obtained from clinic visits (n = 1308). Models were adjusted for age, sex, principal components to account for ancestry, and additional health indicators.

RESULTS

LOAD associations nominally differed for 4 of 21 variants; CR1 and APOE variants were significant after Bonferroni correction.

DISCUSSION

Estimates of genetic associations may differ for studies limited to clinic-only designs. Home visit capacity should be explored as a possible source of heterogeneity and potential bias in genetic studies.

A randomised controlled trial of dietary improvement for adults with major depression (the 'SMILES' trial)

BACKGROUND

The possible therapeutic impact of dietary changes on existing mental illness is largely unknown. Using a randomised controlled trial design, we aimed to investigate the efficacy of a dietary improvement program for the treatment of major depressive episodes.

METHODS

'SMILES' was a 12-week, parallel-group, single blind, randomised controlled trial of an adjunctive dietary intervention in the treatment of moderate to severe depression. The intervention consisted of seven individual nutritional consulting sessions delivered by a clinical dietician. The control condition comprised a social support protocol to the same visit schedule and length. Depression symptomatology was the primary endpoint, assessed using the Montgomery-Åsberg Depression Rating Scale (MADRS) at 12 weeks. Secondary outcomes included remission and change of symptoms, mood and anxiety. Analyses utilised a likelihood-based mixed-effects model repeated measures (MMRM) approach. The robustness of estimates was investigated through sensitivity analyses.

RESULTS

We assessed 166 individuals for eligibility, of whom 67 were enrolled (diet intervention, n = 33; control, n = 34). Of these, 55 were utilising some form of therapy: 21 were using psychotherapy and pharmacotherapy combined; 9 were using exclusively psychotherapy; and 25 were using only pharmacotherapy. There were 31 in the diet support group and 25 in the social support control group who had complete data at 12 weeks. The dietary support group demonstrated significantly greater improvement between baseline and 12 weeks on the MADRS than the social support control group, t(60.7) = 4.38, p < 0.001, Cohen's d = -1.16. Remission, defined as a MADRS score <10, was achieved for 32.3% (n = 10) and 8.0% (n = 2) of the intervention and control groups, respectively (χ 2 (1) = 4.84, p = 0.028); number needed to treat (NNT) based on remission scores was 4.1 (95% CI of NNT 2.3-27.8). A sensitivity analysis, testing departures from the missing at random (MAR) assumption for dropouts, indicated that the impact of the intervention was robust to violations of MAR assumptions.

CONCLUSIONS

These results indicate that dietary improvement may provide an efficacious and accessible treatment strategy for the management of this highly prevalent mental disorder, the benefits of which could extend to the management of common co-morbidities.

TRIAL REGISTRATION

Australia and New Zealand Clinical Trials Register (ANZCTR): ACTRN12612000251820 . Registered on 29 February 2012.

Childhood Language Disorder and Social Anxiety in Early Adulthood

Language disorder is associated with anxiety and with social problems in childhood and adolescence. However, the relation between language disorder and adult social anxiety is not well known. This study examines social anxiety in early adulthood in a 26-year prospective longitudinal study following individuals identified with a communication disorder at age 5 and a control group. Social anxiety diagnoses and subthreshold symptoms were examined at ages 19, 25, and 31 using a structured diagnostic interview; social anxiety symptoms related to social interaction and social performance were also assessed dimensionally at age 31. Multiple imputation was used to address attrition. Compared to controls, participants with childhood language disorder had higher rates of subthreshold social phobia at ages 19 and 25 and endorsed higher levels of social interaction anxiety symptoms at age 31, with particular difficulty talking to others and asserting their perspectives. Childhood language disorder is a specific risk factor for a circumscribed set of social anxiety symptoms in adulthood, which are likely associated with communication challenges.

Polysubstance use patterns and trajectories in vocational students--a latent transition analysis

BACKGROUND

The transition from late adolescence to early adulthood (16-20 years) represents a time of exploration and self-discovery for many young people. As such, it is often associated with experimentation in substance use. Vocational students in particular report high substance use. Thus, the aim of this study is to examine patterns and trajectories of their substance use behavior.

METHODS

On two occasions (interval 18 months), we investigated 5214 students (M=19.39 years; 54% male) from 49 vocational schools in seven German federal states. We identified classes of substance use and trajectories via latent transition analysis, controlling for gender, age, and socio-economic status. Additionally, we investigated work-related (job demands/stress/satisfaction) and psychopathological (depressive symptoms) predictors of substance use via multinomial regressions.

RESULTS

We found three latent stages of substance use: low use (baseline: 43%/follow-up: 44%), mainly alcohol use (50%/45%), and polysubstance use (7%/11%). Over time, 10% of alcohol users at baseline transitioned to polysubstance use at follow-up, while there were smaller transition rates (2-9%) between the other stages. Compared to low use, polysubstance use at follow-up was predicted by high job stress (aOR=1.45, 1.07-1.96) at baseline.

CONCLUSIONS

High alcohol use is associated with bidirectional transitions in young adults. Hence, future research needs to identify mechanisms of change to identify protective factors. Regarding vocational practice, early stress management seems to be a viable path to prevent polysubstance use in vocational students.

Mental health trajectories from adolescence to adulthood: Language disorder and other childhood and adolescent risk factors

Longitudinal research on mental health development beyond adolescence among nonclinical populations is lacking. This study reports on psychiatric disorder trajectories from late adolescence to young adulthood in relation to childhood and adolescent risk factors. Participants were recruited for a prospective longitudinal study tracing a community sample of 5-year-old children with communication disorders and a matched control cohort to age 31. Psychiatric disorders were measured at ages 19, 25, and 31. Known predictors of psychopathology and two school-related factors specifically associated with language disorder (LD) were measured by self-reports and semistructured interviews. The LD cohort was uniquely characterized by a significantly decreasing disorder trajectory in early adulthood. Special education was associated with differential disorder trajectories between LD and control cohorts, whereas maltreatment history, specific learning disorder, family structure, and maternal psychological distress were associated with consistent trajectories between cohorts. From late adolescence to young adulthood, childhood LD was characterized by a developmentally limited course of psychiatric disorder; maltreatment was consistently characterized by an elevated risk of psychiatric disorder regardless of LD history, whereas special education was associated with significantly decreasing risk of psychiatric disorder only in the presence of LD.

Evaluation of a weighting approach for performing sensitivity analysis after multiple imputation

Background

Multiple imputation (MI) is a well-recognised statistical technique for handling missing data. As usually implemented in standard statistical software, MI assumes that data are ‘Missing at random’ (MAR); an assumption that in many settings is implausible. It is not possible to distinguish whether data are MAR or ‘Missing not at random’ (MNAR) using the observed data, so it is desirable to discover the impact of departures from the MAR assumption on the MI results by conducting sensitivity analyses. A weighting approach based on a selection model has been proposed for performing MNAR analyses to assess the robustness of results obtained under standard MI to departures from MAR.

Methods

In this article, we use simulation to evaluate the weighting approach as a method for exploring possible departures from MAR, with missingness in a single variable, where the parameters of interest are the marginal mean (and probability) of a partially observed outcome variable and a measure of association between the outcome and a fully observed exposure. The simulation studies compare the weighting-based MNAR estimates for various numbers of imputations in small and large samples, for moderate to large magnitudes of departure from MAR, where the degree of departure from MAR was assumed known. Further, we evaluated a proposed graphical method, which uses the dataset with missing data, for obtaining a plausible range of values for the parameter that quantifies the magnitude of departure from MAR.

Results

Our simulation studies confirm that the weighting approach outperformed the MAR approach, but it still suffered from bias. In particular, our findings demonstrate that the weighting approach provides biased parameter estimates, even when a large number of imputations is performed. In the examples presented, the graphical approach for selecting a range of values for the possible departures from MAR did not capture the true parameter value of departure used in generating the data.

Conclusions

Overall, the weighting approach is not recommended for sensitivity analyses following MI, and further research is required to develop more appropriate methods to perform such sensitivity analyses.

Electronic supplementary material

The online version of this article (doi:10.1186/s12874-015-0074-2) contains supplementary material, which is available to authorized users.

Using Observational Data to Estimate the Effect of Hand Washing and Clean Delivery Kit Use by Birth Attendants on Maternal Deaths after Home Deliveries in Rural Bangladesh, India and Nepal

BACKGROUND

Globally, puerperal sepsis accounts for an estimated 8-12% of maternal deaths, but evidence is lacking on the extent to which clean delivery practices could improve maternal survival. We used data from the control arms of four cluster-randomised controlled trials conducted in rural India, Bangladesh and Nepal, to examine associations between clean delivery kit use and hand washing by the birth attendant with maternal mortality among home deliveries.

METHODS

We tested associations between clean delivery practices and maternal deaths, using a pooled dataset for 40,602 home births across sites in the three countries. Cross-sectional data were analysed by fitting logistic regression models with and without multiple imputation, and confounders were selected a priori using causal directed acyclic graphs. The robustness of estimates was investigated through sensitivity analyses.

RESULTS

Hand washing was associated with a 49% reduction in the odds of maternal mortality after adjusting for confounding factors (adjusted odds ratio (AOR) 0.51, 95% CI 0.28-0.93). The sensitivity analysis testing the missing at random assumption for the multiple imputation, as well as the sensitivity analysis accounting for possible misclassification bias in the use of clean delivery practices, indicated that the association between hand washing and maternal death had been over estimated. Clean delivery kit use was not associated with a maternal death (AOR 1.26, 95% CI 0.62-2.56).

CONCLUSIONS

Our evidence suggests that hand washing in delivery is critical for maternal survival among home deliveries in rural South Asia, although the exact magnitude of this effect is uncertain due to inherent biases associated with observational data from low resource settings. Our findings indicating kit use does not improve maternal survival, suggests that the soap is not being used in all instances that kit use is being reported.

Multiple Imputation by Fully Conditional Specification for Dealing with Missing Data in a Large Epidemiologic Study

Missing data commonly occur in large epidemiologic studies. Ignoring incompleteness or handling the data inappropriately may bias study results, reduce power and efficiency, and alter important risk/benefit relationships. Standard ways of dealing with missing values, such as complete case analysis (CCA), are generally inappropriate due to the loss of precision and risk of bias. Multiple imputation by fully conditional specification (FCS MI) is a powerful and statistically valid method for creating imputations in large data sets which include both categorical and continuous variables. It specifies the multivariate imputation model on a variable-by-variable basis and offers a principled yet flexible method of addressing missing data, which is particularly useful for large data sets with complex data structures. However, FCS MI is still rarely used in epidemiology, and few practical resources exist to guide researchers in the implementation of this technique. We demonstrate the application of FCS MI in support of a large epidemiologic study evaluating national blood utilization patterns in a sub-Saharan African country. A number of practical tips and guidelines for implementing FCS MI based on this experience are described.

Age 31 mental health outcomes of childhood language and speech disorders

OBJECTIVE

Language disorders are associated with emotional and behavioral problems in childhood and adolescence. Although clinical studies with small samples suggest that psychosocial difficulties continue into adulthood, adult mental health outcomes of childhood language disorders are not well known. The objective of this prospective longitudinal study is to determine whether the age 31 mental health outcomes of individuals who had childhood language disorders differ from the outcomes of typically developing controls.

METHOD

A 26-year cohort study followed up children with language or speech disorders from age 5 to age 31. The children were selected from a 1-in-3 random sample of 5-year-olds using a 3-stage screening and assessment process. A control group matched by sex, age, and classroom or school was also selected. Diagnoses were assigned with the Composite International Diagnostic Interview with the additional criterion that Global Assessment of Functioning scores indicated at least mild impairment. Dimensional psychosocial self-report measures were also administered.

RESULTS

Rates of diagnosis at age 31 years were equivalent between participants who had childhood language disorders and controls, with and without multiple imputation to estimate missing outcomes. Differences in rates of affective and substance use disorders could not be ruled out because of attrition in the cohort with language disorders, who were less likely to participate at age 31. Psychosocial scores for both cohorts were in the normal range. The cohort with language disorders had poorer self-rated physical health than controls.

CONCLUSION

Mild/moderate language disorders may not have significant long-term mental health consequences in early adulthood.

Mindfulness-based cognitive therapy for recurrent depression: A translational research study with 2-year follow-up

OBJECTIVE

While mindfulness-based cognitive therapy (MBCT) has demonstrated efficacy in reducing depressive relapse/recurrence over 12-18 months, questions remain around effectiveness, longer-term outcomes, and suitability in combination with medication. The aim of this study was to investigate within a pragmatic study design the effectiveness of MBCT on depressive relapse/recurrence over 2 years of follow-up.

METHOD

This was a prospective, multi-site, single-blind trial based in Melbourne and the regional city of Geelong, Australia. Non-depressed adults with a history of three or more episodes of depression were randomised to MBCT + depression relapse active monitoring (DRAM) (n=101) or control (DRAM alone) (n=102). Randomisation was stratified by medication (prescribed antidepressants and/or mood stabilisers: yes/no), site of usual care (primary or specialist), diagnosis (bipolar disorder: yes/no) and sex. Relapse/recurrence of major depression was assessed over 2 years using the Composite International Diagnostic Interview 2.1.

RESULTS

The average number of days with major depression was 65 for MBCT participants and 112 for controls, significant with repeated-measures ANOVA (F(1, 164)=4.56, p=0.03). Proportionally fewer MBCT participants relapsed in both year 1 and year 2 compared to controls (odds ratio 0.45, p<0.05). Kaplan-Meier survival analysis for time to first depressive episode was non-significant, although trends favouring the MBCT group were suggested. Subgroup analyses supported the effectiveness of MBCT for people receiving usual care in a specialist setting and for people taking antidepressant/mood stabiliser medication.

CONCLUSIONS

This work in a pragmatic design with an active control condition supports the effectiveness of MBCT in something closer to implementation in routine practice than has been studied hitherto. As expected in this translational research design, observed effects were less strong than in some previous efficacy studies but appreciable and significant differences in outcome were detected. MBCT is most clearly demonstrated as effective for people receiving specialist care and seems to work well combined with antidepressants.

Evaluation of imputation methods for microbial surface water quality studies

Longitudinal studies of microbial water quality are subject to missing observations. This study evaluates multiple imputation (MI) against data deletion, mean or median imputation for replacing missing microbial water quality data. The specific context is data collected in Chicago Area Waterway System (2007-2009), where 45% of Escherichia coli and 53% of enterococci densities were missing owing to sample analysis deficiencies. Imputation methods were compared performing a simulation study using complete observations with introduced missing values and subsequently compared with the original data with missing observations. Coefficients for E. coli densities in linear regression models predicting somatic coliphages density show that MI introduces the least bias among other methods while controlling Type I error. Further exploration of utilizing different MI implementations is recommended to address the influence of missing percentage on MI performance and to explore sensitivity to the degree of violation of the missing completely at random assumption.