You are viewing a free preview of this lesson.
Subscribe to unlock all 10 lessons in this course and every other course on LearningBro.
If genes confer vulnerability to mental disorder, then the children of parents with a disorder should be at raised risk — and that risk should be greatest of all for the children of two affected parents. Gottesman and colleagues' 2010 study tested exactly this, using one of the most powerful research resources in the world: the Danish national registers, which record the psychiatric history of an entire population. By linking parents to their children across millions of people, the study could ask, with a precision no small sample could match, how large the risk of schizophrenia and bipolar disorder is when both parents have a severe mental disorder. It is the prescribed key research for the medical-model topic of the OCR Issues in Mental Health section, and it is the empirical backbone of the genetic explanation you met in the previous lesson. This lesson works through it in the depth OCR's core-study format demands: background and aim, method (design, sample, procedure), results, conclusions, and a full evaluation. The citation is Gottesman, I. I., Laursen, T. M., Bertelsen, A. and Mortensen, P. B. (2010), Severe mental disorders in offspring with 2 psychiatrically ill parents, Archives of General Psychiatry, 67(3), 252–257.
| This lesson covers | OCR H567 Component 03, Section A topic | AO focus |
|---|---|---|
| Aim and background of Gottesman et al. (2010) | Medical model — KEY RESEARCH (genetic explanation) | AO1 knowledge of the study |
| Method: register-based design, sample, procedure | Key research — offspring of two ill parents | AO1; AO2 methodological understanding |
| Results: risk figures for offspring of two, one, and no affected parents | Key research — findings | AO1 |
| Conclusions: genetic vulnerability and heritability | Key research — conclusions | AO1; AO2 |
| Evaluation: sample power, generalisability, gene–environment | Key research — evaluation; issues and debates | AO3 |
The specification is referenced descriptively throughout; consult the official OCR H567 specification document for the exact published wording. This lesson develops AO1 (detailed knowledge of the study), AO2 (understanding its register-based methodology and applying its findings) and AO3 (evaluating its strengths, limitations and place in the nature–nurture debate). Where exact percentages are not certain they are given as approximate ranges; consult the primary paper for precise figures.
The genetic explanation of schizophrenia rests on family, twin and adoption studies, all of which point to inherited vulnerability. But most such studies face a practical problem: samples are small, and it is rare to find enough families in which both parents have a severe mental disorder to estimate the risk to their children reliably. The offspring of two ill parents are the "acid test" of the genetic hypothesis — the group in whom genetic loading should be at its maximum — yet they are precisely the hardest group to study, because dual-affected couples who go on to have children are uncommon.
Gottesman and colleagues set out to solve this by using national data on an entire population. Their aim was to estimate the risk of schizophrenia and of bipolar (affective) disorder in the offspring of couples where both, one, or neither parent had been admitted to hospital with such a disorder — and thereby to quantify how strongly this risk rises with the degree of genetic loading. In effect, they asked: how much does having two ill parents, rather than one or none, raise a child's risk of developing severe mental illness? A clear finding of markedly raised risk in the two-parent group would be strong evidence for the genetic (diathesis) component of the medical model.
The study is a longitudinal, register-based epidemiological study — a form of natural experiment using secondary data. There was no manipulation of variables and no contact with participants; instead the researchers analysed pre-existing national records. The independent variable of interest — how many parents were affected, and with what — was a naturally occurring feature of the families, making this in effect a quasi-experimental comparison of risk across groups defined by their parents' psychiatric status.
The sample is the study's most striking feature. The researchers used the Danish Civil Registration System together with the Danish Psychiatric Central Register. The Civil Registration System assigns every resident of Denmark a unique personal identification number and records family links (allowing children to be matched to both parents); the Psychiatric Central Register records all psychiatric hospital admissions in the country. By linking these, the researchers could construct the entire national population of couples and their children over several decades and identify who had ever been admitted with schizophrenia or bipolar disorder.
From this national resource they identified more than two million people and, within them, the relatively small but — because the base population was so vast — still substantial number of couples in which both partners had been admitted with schizophrenia (around 200 such couples) or with bipolar disorder, along with their children. This is the crucial pay-off of the register method: even a rare combination (two affected parents) yields a usable sample when you start from an entire nation rather than a clinic. The offspring were followed up to determine whether they too were admitted with a severe mental disorder.
For each child, the researchers recorded the psychiatric status of both parents and then the child's own eventual psychiatric outcome — specifically, whether the child was admitted to hospital with schizophrenia or with bipolar disorder. They then calculated the cumulative incidence (risk) of these disorders in offspring across the different parental categories:
Because Danish register data extend over long periods, offspring could be followed through the main age range of risk for these disorders, allowing a meaningful lifetime-risk estimate rather than a snapshot.
It is worth understanding why Denmark is such an unusually powerful place to run a study of this kind, because the method is much of what makes the study examinable. Denmark, like the other Nordic countries, maintains comprehensive national registers as an ordinary part of running the state. Every resident receives a unique personal identification number at birth or on immigration, and this number links together records held by different arms of government — including, crucially for this study, records of parentage (so that any individual can be connected to both of their biological parents) and records of every contact with the psychiatric hospital system. Because healthcare is universal and centralised, these registers capture essentially the whole population rather than a self-selected or clinic-based subset, and because they have been maintained for decades, they allow individuals to be followed across long stretches of their lives. For a researcher, this is close to a natural laboratory: it makes possible the study of questions — like the fate of the children of two mentally ill parents — that would be practically impossible to investigate by recruiting participants one at a time, because the relevant combinations are simply too rare to assemble in useful numbers any other way. The register method is thus not an incidental detail but the very thing that lets the study reach the acid-test group at all, and evaluating the study well means appreciating both the power this confers and the limits it imposes (the reliance on admissions, the Danish specificity) rather than treating "large sample" as a slogan.
The pattern in the results — a steep rise in offspring risk as the number of affected parents increases — is exactly the dose-response relationship the genetic hypothesis predicts, and it is worth being precise about why it counts as evidence. If a disorder were purely environmental and had nothing to do with inheritance, we would not necessarily expect risk to track the number of affected parents in this graded way; but if vulnerability is genetically transmitted, then a child of two affected parents inherits susceptibility from both sides and should be at markedly higher risk than a child of one affected parent, who in turn should exceed the child of none. The observed ordering — roughly 27 per cent, then 7 per cent, then about 1 per cent for schizophrenia — matches this prediction closely, which is why the study is treated as strong support for the genetic component of the medical model. Yet the very same gradient is also consistent, in part, with an environmental reading, because children of two affected parents also grow up in the most disrupted and stressful home environments; two affected parents means, on average, not just double the genetic loading but also a more difficult upbringing. This is the crux that the evaluation must confront: the elegant gradient supports familial transmission unambiguously, but familial transmission is genetic and environmental, and this design cannot prise the two apart. Holding that distinction clearly — the gradient is real and important, but it does not by itself prove the transmission is genetic rather than environmental — is the single most important interpretive skill this study demands.
The results show a clear and steep gradient of risk that rises with genetic loading — exactly what the genetic hypothesis predicts.
For schizophrenia, the risk to offspring rose dramatically with the number of affected parents. Where both parents had schizophrenia, the risk to their children was very high — of the order of 27 per cent (roughly one in four). Where one parent had schizophrenia, the risk was substantially lower — around 7 per cent. Where neither parent had been admitted, the risk was close to the general-population baseline of roughly 0.5 to 1 per cent. In other words, having two affected parents raised the risk many times over compared with one, and enormously compared with none.
A parallel pattern held for bipolar disorder. Where both parents had bipolar disorder, the risk to offspring of a severe mood disorder was again markedly elevated — of the order of 25 per cent — falling to around 4 per cent where one parent was affected, and to the low baseline where neither was.
The researchers also noted that the disorders were not entirely distinct in their transmission: offspring of parents with one disorder showed some raised risk of the other, suggesting a degree of shared genetic vulnerability across schizophrenia and bipolar disorder rather than two wholly separate inherited conditions.
This finding of overlapping transmission deserves emphasis because it quietly challenges a foundational assumption of the classification systems studied earlier. The diagnostic manuals treat schizophrenia and bipolar disorder as separate categories with distinct criteria, descending from Kraepelin's original decision to split dementia praecox from manic-depressive insanity. Yet if the genetic vulnerability to the two conditions substantially overlaps — if parents with one disorder are raising the risk of the other in their children — then the sharp categorical boundary between them looks less like a natural division and more like a convenient simplification of an underlying continuum of vulnerability. This connects the study directly to the validity debate from the second lesson: it is evidence that at least some psychiatric categories may not "carve nature at its joints", and it is part of why modern research increasingly explores dimensional and cross-cutting models of risk rather than assuming the traditional categories are biologically discrete. A candidate who spots that Gottesman's data bear not only on heritability but also on classification is displaying exactly the synoptic reach the top band rewards.
There is also a subtle point about what a risk figure does and does not tell us, which sharpens the study's interpretation. A cumulative risk of around 27 per cent is a statement about a population — about the proportion of a large group of similarly placed children who will, over their lifetimes, be admitted with the disorder. It is emphatically not a prediction about any individual child, who will either develop the disorder or not, and whose personal fate depends on the particular mix of genes, experiences and chance that no population average can capture. This distinction matters enormously in practice: it would be a serious misuse of the study to tell a specific family that "your child has a 27 per cent chance", as though the figure were a property of their child rather than an average over many. The population figure is genuinely useful for planning services and for informing families in general terms, but translating it into individual prediction is both statistically illegitimate and ethically hazardous — a point that connects the study to the socially-sensitive-research debate and that a careful answer will draw out.
| Parental psychiatric status | Approx. risk of the disorder in offspring |
|---|---|
| Both parents — schizophrenia | ~27% |
| One parent — schizophrenia | ~7% |
| Both parents — bipolar disorder | ~25% |
| One parent — bipolar disorder | ~4% |
| Neither parent affected | ~0.5–1% (baseline) |
The single most important interpretive point mirrors the twin-study lesson: even with two affected parents, the risk was around a quarter, not 100 per cent or even a majority. Most children of two parents with schizophrenia did not develop it. So the data demonstrate powerful genetic vulnerability and, in the same breath, that genes are not the whole story — the majority of even the highest-risk group stayed well.
Subscribe to continue reading
Get full access to this lesson and all 10 lessons in this course.