A dynamic social relations model for clustered longitudinal dyadic data with continuous or ordinal responses.

Social relations models allow the identification of cluster, actor, partner, and relationship effects when analysing clustered dyadic data on interactions between individuals or other units of analysis. We propose an extension of this model which handles longitudinal data and incorporates dynamic structure, where the response may be continuous, binary, or ordinal. This allows the disentangling of the relationship effects from temporal fluctuation and measurement error and the investigation of whether individuals respond to their partner's behaviour at the previous observation. We motivate and illustrate the model with an application to Canadian data on pairs of individuals within families observed working together on a conflict discussion task.

Association between surgical volume and failure of primary total hip replacement in England and Wales: findings from a prospective national joint replacement register.

OBJECTIVE: To investigate the association of volume of total hip arthroplasty (THA) between consultants and within the same consultant in the previous year and the hazard of revision using multilevel survival models. DESIGN: Prospective cohort study using data from a national joint replacement register. SETTING: Elective THA across all private and public centres in England and Wales between April 2003 and February 2017. PARTICIPANTS: Patients aged 50 years or more undergoing THA for osteoarthritis. INTERVENTION: The volume of THA conducted in the preceding 365 days to the index procedure. MAIN OUTCOME AND MEASURE: Revision surgery (excision, addition or replacement) of a primary THA. RESULTS: Of the 579 858 patients undergoing primary THA (mean baseline age 69.8 years (SD 10.2)), 61.1% were women. Multilevel survival found differing results for between and within-consultant effects. There was a strong volume-revision association between consultants, with a near-linear 43.3% (95% CI 29.1% to 57.4%) reduction of the risk of revision comparing consultants with volumes between 1 and 200 procedures annually. Changes in individual surgeons (within-consultant) case volume showed no evidence of an association with revision. CONCLUSION: Separation of between-consultant and within-consultant effects of surgical volume reveals how volume contributes to the risk of revision after THA. The lack of association within-consultants suggests that individual changes to consultant volume alone will have little effect on outcomes following THA.These novel findings provide strong evidence supporting the practice of specialisation of hip arthroplasty. It does not support the practice of low-volume consultants increasing their personal volume as it is unlikely their results would improve if this is the only change. Limiting the exposure of patients to consultants with low volumes of THA and greater utilisation of centres with higher volume surgeons with better outcomes may be beneficial to patients.

Intention, beliefs and mood assessed using electronic diaries predicts attendance at cardiac rehabilitation: An observational study.

BACKGROUND: Cardiac rehabilitationis effective in promoting physical/psychological recovery following acute coronary syndrome. Yet, rates of attendance at outpatient cardiac rehabilitation by eligible patients are low. OBJECTIVES: This study examined the determinants of attendance at outpatient cardiac rehabilitation in acute coronary syndrome patients following discharge until cardiac rehabilitation commencement. DESIGN: A weekly electronic diary measured cardiac-related cognitions and mood and examined their relation to attendance at outpatient cardiac rehabilitation. SETTINGS: Three United Kingdom National Health Service secondary care settings in two Health Board areas in Scotland. PARTICIPANTS: Acute coronary syndrome patients were recruited from March 2012 to June 2013 prior to hospital discharge. Of 488 eligible patients referred for cardiac rehabilitation, 214 consented. METHODS: Consecutive patients completed a pre-hospital discharge questionnaire targeting age, diagnosis, social class and smoking history. Acute coronary syndrome patients then completed a weekly electronic diary from the first week of discharge until the start of cardiac rehabilitation. Multilevel structural equation models estimated the effects of initial, i.e. baseline and rate of change in cardiac-related cognition and mood on attendance. Intention to attend cardiac rehabilitation was reflected, log transformed, reported thereafter as "do not intend". The role of "do not intend" was explored as a mediator of the relationship between cardiac-related cognition and mood on attendance. RESULTS: 166 participants provided, on average, 5 weeks of diary entries before cardiac rehabilitation commenced. High intention (i.e. low "do not intend") to attend CR and its rate of increase over time predicted attendance. Low negative emotional representation, high perceived necessity, high confidence in maintaining function, low negative affect, and high positive affect following discharge predicted attendance at cardiac rehabilitation. The rate of change in cardiac-related mood and these cognitions was not predictive. Baseline and rate of change in "do not intend" entirely mediated relationships between a) perceived necessity, b) negative affect and attendance at cardiac rehabilitation. CONCLUSIONS: Negative affect in the first weeks following discharge represents the key challenge to a patient maintaining their intention to attend cardiac rehabilitation. Intervention to improve attendance should focus on improving intention to attend following discharge and during recovery by improving patient understanding of cardiac rehabilitation and reducing negative affect.

Probabilistic record linkage.

Studies involving the use of probabilistic record linkage are becoming increasingly common. However, the methods underpinning probabilistic record linkage are not widely taught or understood, and therefore these studies can appear to be a 'black box' research tool. In this article, we aim to describe the process of probabilistic record linkage through a simple exemplar. We first introduce the concept of deterministic linkage and contrast this with probabilistic linkage. We illustrate each step of the process using a simple exemplar and describe the data structure required to perform a probabilistic linkage. We describe the process of calculating and interpreting matched weights and how to convert matched weights into posterior probabilities of a match using Bayes theorem. We conclude this article with a brief discussion of some of the computational demands of record linkage, how you might assess the quality of your linkage algorithm, and how epidemiologists can maximize the value of their record-linked research using robust record linkage methods.

Capturing the temporal sequence of interaction in young siblings.

We explored whether young children exhibit subtypes of behavioral sequences during sibling interaction. Ten-minute, free-play observations of over 300 sibling dyads were coded for positivity, negativity and disengagement. The data were analyzed using growth mixture modeling (GMM). Younger (18-month-old) children's temporal behavioral sequences showed a harmonious (53%) and a casual (47%) class. Older (approximately four-year-old) children's behavior was more differentiated revealing a harmonious (25%), a deteriorating (31%), a recovery (22%) and a casual (22%) class. A more positive maternal affective climate was associated with more positive patterns. Siblings' sequential behavioral patterns tended to be complementary rather than reciprocal in nature. The study illustrates a novel use of GMM and makes a theoretical contribution by showing that young children exhibit distinct types of temporal behavioral sequences that are related to parenting processes.

Adjusting for selection bias in longitudinal analyses using simultaneous equations modeling: the relationship between employment transitions and mental health.

BACKGROUND: Effects of labor force participation on mental health can be difficult to discern due to the possibility of selection bias. Previous research typically adjusts for direct selection (reverse causality) but ignores indirect selection (unmeasured confounders). METHODS: We investigate the relationship between men's employment transitions and mental health using a dynamic simultaneous equations model applied to data from the British Household Panel Survey (1991-2009). Outcome is self-reported distress and anxiety as summed on a 12-point scale. We allow for direct selection by allowing prior mental health to affect both subsequent mental health and employment transitions in the joint model. We adjust for indirect selection by allowing for residual correlation between mental health and employment. RESULTS: Moving from unemployment to employment was strongly associated with an improvement in mental health, whereas becoming unemployed was detrimental. However, these associations were attenuated by unmeasured confounders. After adjustment for indirect selection, the increased distress and anxiety associated with becoming unemployed decreased from 2.5 (95% confidence interval = 2.2 to 2.7) to 2.2 (2.0 to 2.5). (A change of 2.5 equates to half a standard deviation on the 12-point scale.) The improvement with moving from unemployment to employment was also weakened slightly (from -2.1 [-2.4 to -1.7] to -1.8 [-2.1 to -1.5]). CONCLUSIONS: There was strong evidence of indirect selection, but less support for direct selection. Nevertheless, the effects on psychological health of transitions between employment and unemployment, and between employment and economic inactivity, remained substantial after adjusting for selection.

Interrelationships between childbearing and housing transitions in the family life course.

Research has examined the effect of family changes on housing transitions and childbearing patterns within various housing types. Although most research has investigated how an event in one domain of family life depends on the current state in another domain, the interplay between them has been little studied. This study examines the interrelationships between childbearing decisions and housing transitions. We use rich longitudinal register data from Finland and apply multilevel event history analysis to allow for multiple births and housing changes over the life course. We investigate the timing of fertility decisions and housing choices with respect to each other. We model childbearing and housing transitions jointly to control for time-invariant unobserved characteristics of women, which may simultaneously influence their fertility behavior and housing choices, and we show how joint modeling leads to a deeper understanding of the interplay between the two domains of family life.

A multilevel simultaneous equations model for within-cluster dynamic effects, with an application to reciprocal parent-child and sibling effects.

There has been substantial interest in the social and health sciences in the reciprocal causal influences that people in close relationships have on one another. Most research has considered reciprocal processes involving only 2 units, although many social relationships of interest occur within a larger group (e.g., families, work groups, peer groups, classrooms). This article presents a general longitudinal multilevel modeling framework for the simultaneous estimation of reciprocal relationships among individuals with unique roles operating in a social group. We use family data for illustrative purposes, but the model is generalizable to any social group in which measurements of individuals in the social group occur over time, individuals have unique roles, and clustering of the data is evident. We allow for the possibility that the outcomes of family members are influenced by a common set of unmeasured family characteristics. The multilevel model we propose allows for residual variation in the outcomes of parents and children at the occasion, individual, and family levels and residual correlation between parents and children due to the unmeasured shared environment, genetic factors, and shared measurement. Another advantage of this method over approaches used in previous family research is it can handle mixed family sizes. The method is illustrated in an analysis of maternal depression and child delinquency using data from the Avon Brothers and Sisters Study.

Consequences of family disruption on children's educational outcomes in Norway.

Using high-quality data from Norwegian population registers, we examine the relationship between family disruption and children's educational outcomes. We distinguish between disruptions caused by parental divorce and paternal death and, using a simultaneous equation model, pay particular attention to selection bias in the effect of divorce. We also allow for the possibility that disruption may have different effects at different stages of a child's educational career. Our results suggest that selection on time-invariant maternal characteristics is important and works to overstate the effects of divorce on a child's chances of continuing in education. Nevertheless, the experience of marital breakdown during childhood is associated with lower levels of education, and the effect weakens with the child's age at disruption. The effects of divorce are most pronounced for the transitions during or just beyond the high school level. In models that do not allow for selection, children who experienced a father's death appear less disadvantaged than children whose parents divorced. After we control for selection, however differences in the educational qualifications of children from divorced and bereaved families narrow substantially and, at mean ages of divorce, are almost non-existent.

The use of simple reparameterizations to improve the efficiency of Markov chain Monte Carlo estimation for multilevel models with applications to discrete time survival models.

We consider the application of Markov chain Monte Carlo (MCMC) estimation methods to random-effects models and in particular the family of discrete time survival models. Survival models can be used in many situations in the medical and social sciences and we illustrate their use through two examples that differ in terms of both substantive area and data structure. A multilevel discrete time survival analysis involves expanding the data set so that the model can be cast as a standard multilevel binary response model. For such models it has been shown that MCMC methods have advantages in terms of reducing estimate bias. However, the data expansion results in very large data sets for which MCMC estimation is often slow and can produce chains that exhibit poor mixing. Any way of improving the mixing will result in both speeding up the methods and more confidence in the estimates that are produced. The MCMC methodological literature is full of alternative algorithms designed to improve mixing of chains and we describe three reparameterization techniques that are easy to implement in available software. We consider two examples of multilevel survival analysis: incidence of mastitis in dairy cattle and contraceptive use dynamics in Indonesia. For each application we show where the reparameterization techniques can be used and assess their performance.

Changing compatibility of cohabitation and childbearing between young British women born in 1958 and 1970.

We investigate the effect of parenthood on whether non-marital unions led to marriage or parting for two cohorts of British women when they were aged between 16 and 29. We compare the effect of conceptions leading to births and the presence and characteristics of children on the odds that a cohabitation was dissolved, or that it was converted to marriage, for women born in 1958 and 1970. A multilevel, multiprocess, competing-risks model allows for multiple cohabitation per woman and endogeneity of fertility status. We find that cohabiting couples' response to impending parenthood and the presence of children changed over time. In particular, the proportion of cohabiting couples who married before a birth decreased and, in the 1970 cohort only, the risk of dissolution declined during pregnancy. There is also evidence that the presence of a child cemented a cohabiting union for women from the 1970, but not the earlier, cohort.

The relationship between childbearing and transitions from marriage and cohabitation in Britain.

In this article, we describe a general framework for the analysis of correlated event histories, with an application to a study of partnership transitions and fertility among a cohort of British women. Using a multilevel, multistate competing-risks model, we examine the relationship between prior fertility outcomes (the presence and characteristics of children and current pregnancy) and the dissolution of marital and cohabiting unions and movements from cohabitation to marriage. Using a simultaneous-equations model, we model these partnership transitions jointly with fertility, allowing for correlation between the unobserved woman-level characteristics that affect each process. The analysis is based on the partnership and birth histories that were collected for the 1958 birth cohort (National Child Development Study) aged 16-42. The findings indicate that preschool children have a stabilizing effect on their parents 'partnership, whether married or cohabiting, but the effect is weaker for older children. There is also evidence that although pregnancy precipitates marriage among cohabitors, the odds of marriage decline to prepregnancy levels following a birth.

A discrete-time multilevel mixture model for event history data with long-term survivors, with an application to an analysis of contraceptive sterilization in Bangladesh.

Event history models typically assume that the entire population is at risk of experiencing the event of interest throughout the observation period. However, there will often be individuals, referred to as long-term survivors, who may be considered a priori to have a zero hazard throughout the study period. In this paper, a discrete-time mixture model is proposed in which the probability of long-term survivorship and the timing of event occurrence are modelled jointly. Another feature of event history data that often needs to be considered is that they may come from a population with a hierarchical structure. For example, individuals may be nested within geographical regions and individuals in the same region may have similar risks of experiencing the event of interest due to unobserved regional characteristics. Thus, the discrete-time mixture model is extended to allow for clustering in the likelihood and timing of an event within regions. The model is further extended to allow for unobserved individual heterogeneity in the hazard of event occurrence. The proposed model is applied in an analysis of contraceptive sterilization in Bangladesh. The results show that a woman's religion and education level affect her probability of choosing sterilization, but not when she gets sterilized. There is also evidence of community-level variation in sterilization timing, but not in the probability of sterilization.

Appropriate methods for analyzing the effect of method choice on contraceptive discontinuation.

The contraceptive method chosen is an important determinant of contraceptive discontinuation. However, method choice is endogenous to contraceptive discontinuation. Using data from the 1997 Indonesia Demographic and Health Survey, we apply a multilevel multi-process model to examine the impact of method choice on three types of contraceptive discontinuation. We confirm that method choice is endogenous to the processes of contraceptive abandonment and method switching, but not failure. Ignoring the endogeneity of contraceptive choice leads to various biases in the magnitude of estimated effects of method choice on abandonment and method switching, but the general conclusions are robust to these biases.

The effects of kin on child mortality in rural Gambia.

We analyzed data that were collected continuously between 1950 and 1974 from a rural area of the Gambia to determine the effects of kin on child mortality. Multilevel event-history models were used to demonstrate that having a living mother, maternal grandmother, or elder sisters had a significant positive effect on the survival probabilities of children, whereas having a living father, paternal grandmother, grandfather, or elder brothers had no effect. The mother's remarriage to a new husband had a detrimental effect on child survival, but there was little difference in the mortality rates of children who were born to monogamous or polygynous fathers. The implications of these results for understanding the evolution of human life-history are discussed.