Parental mental illness: estimating prevalence to inform policy and practice

html xmlns=”http://www.w3.org/1999/xhtml” xmlns:mml=”http://www.w3.org/1998/Math/MathML” xmlns:epub=”http://www.idpf.org/2007/ops”>


Chapter 3 Parental mental illness: estimating prevalence to inform policy and practice


Darryl Maybery, Joanne Nicholson, and Andrea Reupert


Many of us have been faced with the challenge of convincing others – policymakers, providers, and research colleagues – of the significance of parental mental illness in the lives of children and adults alike. We use the phrase “mental illness” to refer to specific diagnosable psychiatric disorders such as depression or schizophrenia. An understanding of the prevalence of mental illnesses in parents and of the factors often associated with psychiatric conditions that moderate outcomes for children (e.g., “risk” or “protective” factors such as minority status, marital status, education level, employment, and economic status) is essential to inform policy and practice. Prevalence data are useful to individuals framing policy at the national and state levels; to agency administrators who are developing agency operating procedures, allocating resources, and planning programs; and to providers coming into contact with family members on a day-to-day basis. The dearth of existing prevalence data in many countries or locales, and the challenges in compiling the data that do exist could well frustrate researchers otherwise eager to inform policy and practice. Determining the prevalence of parental mental illness and understanding the extant circumstances of these families (i.e., correlates of risk and resilience) are important to inform policy and ensure that the appropriate level and type of services are offered.


In Chapter 1, various theoretical models were outlined that address ways in which parental mental illness contributes to outcomes for children. In this chapter we focus on issues in estimating the prevalence of parental mental illness and the numbers of adults and children potentially affected. We will use data from several countries to illustrate some of the methodological approaches to assess prevalence and the inherent complexities associated with these strategies. Previously unpublished data will be presented to extend the knowledge base further. Our goal is to provide “best case” examples of ways in which prevalence data are compiled to inform the field.



Extrapolating from population-level data, or the “top down” approach


The approach we are calling “top down” typically involves analyzing existing, large-scale, population-level data to estimate the prevalence of parenthood among adults living with mental illness and the prevalence of children living with parents who have mental illness. The numbers of parents and children living in the community in these situations is extrapolated from population parameters.


An example of this type of approach is the work of Nicholson and colleagues in the USA, who estimated the prevalence of parenthood among adults meeting psychiatric diagnostic criteria for mental illnesses with data from the US National Comorbidity Survey (NCS) obtained in the early 1990s (Nicholson et al., 2004). Cross-sectional interview data were obtained from a nationally representative household sample of individuals living in the community (i.e., not hospitalized, in the military, or in jail or prison). Psychiatric diagnoses were assigned on the basis of a structured research interview, a modification of the Composite International Diagnostic Interview. Those individuals who screened positive for psychotic symptoms were reinterviewed by experienced clinicians, again using a structured interview measure (the Structured Clinical Interview for DSM-III-R).


In secondary analyses of the NCS data, women (68%) and men (54%) meeting criteria for mental illness were slightly more likely to be parents than those reporting no psychiatric or substance use disorder (62.4% and 52.9% respectively). Over the course of their lifetime, the majority of women and men falling into the representative diagnostic categories were parents: affective disorders (66.8% of the women; 58% of the men), anxiety disorders (68.2% women; 56.4% men) and post-traumatic stress disorder (72.8% women; 67.9% men), and nonaffective psychosis (61.8% women; 55.2% men). Among NCS participants who were parents at the time of the survey, almost half of the mothers and almost a third of the fathers had a lifetime prevalence of mental illness.


In terms of moderators that contribute to outcomes for children, a significantly larger proportion of fathers in the psychiatric disorder group were White than in the “no disorder” group; they reported significantly fewer years of education as well. Fewer mothers in the psychiatric disorder group were married; significantly more mothers in this group were living below the poverty level.


Luciano et al. (2014) conducted similar analyses more recently, using data from the 2009 and 2010 US National Survey on Drug Use and Health. Well-established measures were used to distinguish participants with no, mild, moderate, and serious mental illness. The prevalence of motherhood among women was similar across mental illness categories, ranging from 42% among women with no mental illness to 38% among women with serious mental illness. Rates of fatherhood decreased with increasing mental illness severity, ranging from 35% among men with no mental illness to 23% among men with serious mental illness. Rates of employment were low for parents with serious mental illness (38% full-time and 17% part-time among mothers; 60% full-time and 9% part-time among fathers) compared with parents with no mental illness (50% full-time and 19% part-time among mothers; 85% full-time and 5% part-time among fathers). Parents living with serious mental illness were twice as likely to fall below the US Census poverty threshold as their well counterparts. In addition, mothers with serious mental illness were significantly more likely to be younger, White, less well-educated, and never married; to have had a substance use disorder; and to be in fair or poor health. Fathers with serious mental illness were more likely to lack a college education, have a substance use disorder, and be in fair or poor health.


The many reasons why the overall prevalence rates of parenthood vary between these two studies will be discussed below. An important note is that while participants in each of these studies may have met diagnostic criteria for mental illness, only a small percentage were likely to be receiving treatment. However, data regarding variables considered as moderators of child outcomes (i.e., risk or protective factors) support consistent conclusions across these large-scale, population-based studies. Parents meeting criteria for mental illness and, particularly, serious mental illness are more likely than well parents to be living with limited resources (e.g., limited education, greater unemployment, fewer emotional supports), circumstances that convey risk to children.



Using clinical or treatment sample data: the “bottom up” approach


There is also a “bottom up” approach to estimating prevalence; this determines the numbers of clients in the psychiatric setting (sometimes referred to as adult mental health services) who are parents, or determines the prevalence of parental mental illness in samples of children who are receiving services. In a 1-day, cross-sectional survey of psychiatric in- and outpatients in Sweden, 36% (n = 137) of adult clients had children under 18 years of age (Östman and Eidevall, 2005). The Australian Infant, Child, Adolescent, and Family Mental Health Association (2001) summarized several Australian studies that employed this approach and found that between 29% and 35% of psychiatric unit clients were female with dependent children under the age of 18.


Other strategies for exploiting samples receiving services of families living with parental mental illness can be found in the US literature. Estimates of the prevalence of parental mental illness or the circumstances of these families have been derived from samples of children receiving services for behavioral and emotional difficulties, or receiving child welfare services. In systems of care model community programs for children with serious behavioral and emotional difficulties, 52% reported a history of mental illness in the biological family; 58% reported a family history of substance use, and 36.5% reported histories of both mental illness and substance abuse (Hinden et al., 2006).


In another example, in a sample of Medicaid-eligible mothers, more than 14% of mothers with serious mental illness were receiving child welfare services compared with just 4.2% of those without a diagnosis. Mothers with serious mental illness were almost three times as likely to have had involvement in the child welfare system or to have children who had an out-of-home placement (Park et al., 2006). In an earlier study of US fathers, data were obtained from state Department of Mental Health (DMH) case managers regarding the prevalence of fatherhood and characteristics of fathers (Nicholson et al., 1999). Among the men with the most severe mental illnesses, eligible for case-management services due to their inability to meet basic life needs of shelter, food, clothing, and self-care, and histories of psychiatric hospitalization and homelessness, 21% were reported by case managers as having children. This is understandably a gross underestimate of the prevalence of parenthood among men living with mental illnesses in general given the nature of the population (i.e., DMH case-management clients) and the source of the data (i.e., case managers who may not have asked their clients about their parenting status).


In the USA, the challenge in estimating the prevalence of parenthood through the use of mental health treatment samples is made more complicated by the fact that fewer than half of those individuals meeting criteria for psychiatric disorder report that they are receiving treatment (Wang et al., 2005). In analyses of National Co-Morbidity Survey Replication study data from the early 2000s, among individuals meeting criteria for anxiety, mood, or impulse-control disorders, about one-third received treatment from mental health specialists. Larger percentages, for example, received treatment from general medical practitioners. And unmet treatment need is greatest in traditionally underserved groups – the elderly, racial-ethnic minorities, those with low incomes and without health insurance, and rural residents. Consequently, estimates based on the characteristics of samples of individuals receiving mental health treatment may not accurately reflect the larger population of parents living with mental health conditions.


An important consideration in these types of studies, further elaborated below, is the issue of the denominator. That is, any estimate of the prevalence of parental mental illness or risk and protective factors is based on a total sample that is different in some way from the general population (e.g., clinic population, child welfare population, case-managed clients). Conclusions must be carefully drawn. For example, while mothers with serious mental illness may be overrepresented in the child welfare population (e.g., Park et al., 2006), it is not possible to conclude from these data that mothers with serious mental illness are more likely to neglect or abuse their children. A further serious consideration is the identity of the reporter. In the study of fatherhood cited above (Nicholson et al., 1999), providers provided data that were quite likely an underestimate of the prevalence of fatherhood given that providers generally do not routinely ask this question of male clients, and men may not have frequent or regular contact with their children.



Integrating approaches: top down and bottom up


Both bottom up and top down approaches were employed by Maybery and colleagues to estimate the prevalence of families and children living in families with a parental mental illness (Maybery et al., 2005, 2009). The top down approach combined two population studies from the public domain. The first, using the Mental Health and Wellbeing: Profile of Adults, Australia, 1997 (Australian Bureau of Statistics (ABS), 1998), determined by telephone interview the number of Australians with a mental health condition. While discussed in detail elsewhere (Maybery et al., 2005), this profile examined the prevalence and severity of various mental disorders in a representative sample of 10,600 Australians over 18 years of age in the previous 12 months, using a modified computerized version of the Composite International Diagnostic Survey to assess mental health. This was combined with a similarly obtained population-based study, the Family Characteristics, Australia (ABS, 2003), which surveyed “household and family composition including demographic, labour force and family. A particular focus of the survey is families with children aged 0–17 years” (ABS, 2003, p. 2). This report was particularly useful, as it provided information on numbers and types of households in Australia in conjunction with numbers of children. The strength of the estimation is that in both studies participants were randomly chosen from the population. The bringing together of Australian adult mental health status with family characteristics allowed a broad estimation based upon all Australian children likely to be living in families with a parent with a mental illness (see details of estimates in Maybery et al., 2005, 2009). Table 3.1 shows the estimate of 23.3% of all Australian families having a parent that has experienced mental health problems in the last year.


Mar 18, 2017 | Posted by in PSYCHIATRY | Comments Off on Parental mental illness: estimating prevalence to inform policy and practice

Full access? Get Clinical Tree

Get Clinical Tree app for offline access