Journal of Medical Internet Research - International Scientific Journal for Medical Research, Information and Communication on the Internet #Preprint #PeerReviewMe: Warning: This is a unreviewed preprint. Readers are warned that the document has not been peer-reviewed by expert/patient reviewers or an academic editor, may contain misleading claims, and is likely to undergo changes before final publication, if accepted, or may have been rejected/withdrawn. Readers with interest and expertise are encouraged to sign up as peer-reviewer, if the paper is within an open peer-review period. Please cite this preprint only for review purposes or for grant applications and CVs (if you are the author). Background: Background: Postnatal depression adversely affects mothers and infants. There is evidence that caregiving difficulties associated with depressive symptoms also adversely affect later childhood development. In many countries, resources available to support mothers experiencing depressive symptoms during the postnatal period are limited. Cost-effective online group-based nurse-led interventions delivered by mobile phone ‘apps’ have the potential to help address this problem by providing large numbers of mothers with access to professional and peer support during the postnatal period. Objective: Objective: The aim of this study was to test the effectiveness of a 4-month online group-based nurse-led intervention delivered when infants were 2-6 months as compared to standard care outcomes. Methods: Methods: The study was a block randomised control trial. Mothers were recruited at the time they were contacted for the postnatal health check offered to all mothers in South Australia. Those who agreed to participate were randomly assigned to the intervention or standard care. The overall response rate was 63% (N=133/210). Primary outcomes were level of maternal depressive symptoms assessed with the Edinburgh Postnatal Depression Scale (EPDS); and quality of maternal caregiving assessed using the Parenting Stress Index (PSI competence and attachment subscales), the Parenting Sense of Competence Scale, and the Nursing Child Assessment Satellite Training Scale. Assessments were completed at baseline (mean child age=4.9 weeks, SD=1.4) and again when infants were aged 8, and 12 months. Results: Results: Outcomes were evaluated using linear generalized estimating equations adjusting for post-randomization group differences in demographic characteristics and the outcome score at baseline. There were no significant differences in the intervention and standard care groups in scores on the PSI competence subscale (P=.69) nor the Parenting Sense of Competence Scale (P=.11). While the group by time interaction suggested there were differences over time between the EPDS and PSI attachment subscales scores in the intervention and standard care groups (P=.001 and P=.04 respectively), these arose largely because the intervention group had stable scores over time while the standard care group showed some improvements between baseline to 12 months. Mothers engaged well with the intervention with 60% of mothers logging-in once per week during the first 11 weeks of the intervention. The majority of mothers also rated the intervention as helpful and user-friendly. Conclusions: Conclusions: Mothers reported that the intervention was helpful and the app was described as easy to use. As such, it appears that support for mothers during the postnatal period, provided using mobile phone technology, has the potential to be an important addition to existing services. Possible explanations for the lack of differences in outcomes for the two groups in this study are the failure of many mothers to use key components of the intervention and residual differences between the intervention and standard care groups post-randomization. Clinical Trial: Trial Registration: Australian New Zealand Clinical Trials Registry ACTRN 12616001732471

This study aims to demonstrate the use of the tree-based machine learning algorithms to predict the 3- and 5-year disease-specific survival of oral and pharyngeal cancers (OPCs) and compare their performance with the traditional Cox regression. A total of 21,154 individuals diagnosed with OPCs between 2004 and 2009 were obtained from the Surveillance, Epidemiology, and End Results (SEER) database. Three tree-based machine learning algorithms (survival tree (ST), random forest (RF) and conditional inference forest (CF)), together with a reference technique (Cox proportional hazard models (Cox)), were used to develop the survival prediction models. To handle the missing values in predictors, we applied the substantive model compatible version of the fully conditional specification imputation approach to the Cox model, whereas we used RF to impute missing data for the ST, RF and CF models. For internal validation, we used 10-fold cross-validation with 50 iterations in the model development datasets. Following this, model performance was evaluated using the C-index, integrated Brier score (IBS) and calibration curves in the test datasets. For predicting the 3-year survival of OPCs with the complete cases, the C-index in the development sets were 0.77 (0.77, 0.77), 0.70 (0.70, 0.70), 0.83 (0.83, 0.84) and 0.83 (0.83, 0.86) for Cox, ST, RF and CF, respectively. Similar results were observed in the 5-year survival prediction models, with C-index for Cox, ST, RF and CF being 0.76 (0.76, 0.76), 0.69 (0.69, 0.70), 0.83 (0.83, 0.83) and 0.85 (0.84, 0.86), respectively, in development datasets. The prediction error curves based on IBS showed a similar pattern for these models. The predictive performance remained unchanged in the analyses with imputed data. Additionally, a free web-based calculator was developed for potential clinical use. In conclusion, compared to Cox regression, ST had a lower and RF and CF had a higher predictive accuracy in predicting the 3- and 5-year OPCs survival using SEER data. The RF and CF algorithms provide non-parametric alternatives to Cox regression to be of clinical use for estimating the survival probability of OPCs patients.

This longitudinal study classified groups of children experiencing different trajectories of student-teacher relationship quality over the transition from preschool into school, and determined the strength of the association between different student-teacher relationship trajectories and childhood mental health problems in the second year of primary school. A community sample of 460 Australian children were assessed in preschool (age 4), the first school year (age 5), and second school year (age 6). Teachers at all three assessments reported on student-teacher relationship quality with the Student Teacher Relationship Scale. When the children were at preschool and in their second school year, parents and teachers rated children's mental health problems using the Strengths and Difficulties Questionnaire. Latent-class growth modelling identified two trajectories of student-teacher relationship quality: (1) a stable-high student-teacher relationship quality and (2) a moderate/declining student-teacher relationship quality trajectory. Generalised linear models found that after adjusting for family demographic characteristics, having a stable high quality student-teacher relationship trajectory was associated with fewer parent-rated and teacher-rated total mental health problems, and fewer conduct, hyperactivity, and peer problems, and greater prosocial behaviour at age 6. A stable high quality trajectory was also associated with fewer teacher-rated, but not parent-rated emotional symptoms. These effects remained after adjustment for levels of mental health problems at age 4. Findings suggest that early intervention and prevention strategies that focus on building stable high quality student-teacher relationships during preschool and children's transition into formal schooling, may help reduce rates of childhood mental health problems during the early school years.

Mediation analysis is concerned with the decomposition of the total effect of an exposure on an outcome into the indirect effect, through a given mediator, and the remaining direct effect. This is ideally done using longitudinal measurements of the mediator, which capture the mediator process more finely. However, longitudinal measurements pose challenges for mediation analysis, because the mediators and outcomes measured at a given time point can act as confounders for the association between mediators and outcomes at a later time point; these confounders are themselves affected by the prior exposure and outcome. Such posttreatment confounding cannot be dealt with using standard methods (e.g., generalized estimating equations). Analysis is further complicated by the need for so-called cross-world counterfactuals to decompose the total effect. This work addresses these challenges. In particular, we introduce so-called natural effect models, which parameterize the direct and indirect effect of a baseline exposure with respect to a longitudinal mediator and outcome. These can be viewed as a generalization of marginal structural mean models to enable effect decomposition. We introduce inverse probability weighting techniques for fitting these models, adjusting for (measured) time-varying confounding of the mediator-outcome association. Application of this methodology uses data from the Millennium Cohort Study, a longitudinal study of children born in the United Kingdom between September 2000 and January 2002.

Internet-based interventions moderated by community nurses have the potential to improve support offered to new mothers, many of whom now make extensive use of the Internet to obtain information about infant care. However, evidence from population-based randomized controlled trials is lacking. The aim of this study was to test the non-inferiority of outcomes for mothers and infants who received a clinic-based postnatal health check plus nurse-moderated, Internet-based group support when infants were aged 1-7 months as compared with outcomes for those who received standard care consisting of postnatal home-based support provided by a community nurse. The design of the study was a pragmatic, preference, non-inferiority randomized control trial. Participants were recruited from mothers contacted for their postnatal health check, which is offered to all mothers in South Australia. Mothers were assigned either (1) on the basis of their preference to clinic+Internet or home-based support groups (n=328), or (2) randomly assigned to clinic+Internet or home-based groups if they declared no strong preference (n=491). The overall response rate was 44.8% (819/1827). The primary outcome was parenting self-competence, as measured by the Parenting Stress Index (PSI) Competence subscale, and the Karitane Parenting Confidence Scale scores. Secondary outcome measures included PSI Isolation, Interpersonal Support Evaluation List-Short Form, Maternal Support Scale, Ages and Stages Questionnaire-Social-Emotional and MacArthur Communicative Development Inventory (MCDI) scores. Assessments were completed offline via self-assessment questionnaires at enrolment (mean child age=4.1 weeks, SD 1.3) and again when infants were aged 9, 15, and 21 months. Generalized estimating equations adjusting for post-randomization baseline imbalances showed that differences in outcomes between mothers in the clinic+Internet and home-based support groups did not exceed the pre-specified margin of inferiority (0.25 of a SD) on any outcome measure at any follow-up assessment, with the exception of MCDI scores assessing children's language development at 21 months for randomized mothers, and PSI Isolation scores at 9 months for preference mothers. Maternal and child outcomes from a clinic-based postnatal health check plus nurse-moderated Internet-based support were not inferior to those achieved by a universal home-based postnatal support program. Postnatal maternal and infant support using the Internet is a promising alternative to home-based universal support programs. Australian New Zealand Clinical Trials Registry Number (ANZCTR): ACTRN12613000204741; https://www.anzctr.org.au/Trial/Registration/TrialReview.aspx?id=363712&isReview=true (Archived by WebCite at http://www.webcitation.org/6rZeCJ3k1).

Pancreatic cancer (PC) has the worst survival of all periampullary cancers. This may relate to histopathological differences between pancreatic cancers and other periampullary cancers. Our aim was to examine the distribution and histopathologic features of pancreatic, ampullary, biliary and duodenal cancers resected with a pancreaticoduodenectomy (PD) and to examine local trends of periampullary cancers resected with a PD. A retrospective review of PD between January 2000 and December 2012 at a public metropolitan database was performed. The institutional ethics committee approved this study. There were 142 PDs during the study period, of which 70 cases were pre-2010 and 72 post-2010, corresponding to a recent increase in the number of cases. Of the 142 cases, 116 were for periampullary cancers. There were also proportionately more PD for PC (26/60, 43% pre-2010 vs 39/56, 70% post-2010, P = 0.005). There were 65/116 (56%) pancreatic, 29/116 (25%), ampullary, 17/116 (15%) biliary and 5/116 (4%) duodenal cancers. Nodal involvement occurred more frequently in PC (78%) compared to ampullary (59%), biliary (47%) and duodenal cancers (20%), P = 0.002. Perineural invasion was also more frequent in PC (74%) compared to ampullary (34%), biliary (59%) and duodenal cancers (20%), P = 0.002. Microvascular invasion was seen in 57% pancreatic, 38% ampullary, 41% biliary and 20% duodenal cancers, P = 0.222. Overall, clear margins (R0) were achieved in fewer PC 41/65 (63%) compared to ampullary 27/29 (93%; P = 0.003) and biliary cancers 16/17 (94%; P = 0.014). This study highlights that almost half of PD was performed for cancers other than PC, mainly ampullary and biliary cancers. The volume of PD has increased in recent years with an increased proportion being for PC. PC had higher rates of nodal and perineural invasion compared to ampullary, biliary and duodenal cancers.