Work

Cohort Profile: Stratifying Resilience and Depression Longitudinally (STRADL): a questionnaire follow-up of Generation Scotland: Scottish Family Health Study (GS:SFHS)

Year: 2017
Type: article
Study design: Observational Study
Abstract: Common health conditions such as heart disease, stroke and depression are a common cause of chronic suffering and economic burden worldwide.1 Scotland has a high prevalence of these conditions, and because of its comparatively stable population,2 it provides a useful citizenry to study their prevalence and impact. Generation Scotland (GS) is a multi-institutional, cross-disciplinary collaboration aiming to promote research into genetics and health care throughout Scotland.3 Between 2006 and 2011, GS undertook its first major study–Generation Scotland: Scottish Family Health Study1,4 (GS:SFHS). This large, family-based intensively phenotyped and genotyped population cohort was designed to examine a diverse range of illnesses such as those aforementioned. The work of GS is especially important to epidemiological research as it provides a means of separating genetic and shared environmental contributions to common non-communicable diseases. Furthermore, the ability to re-contact GS participants and obtain broad consent for future use of their data and samples is especially valuable for prospective studies associated with health outcomes. In 2015, a strategic award by the Wellcome Trust provided funding for ‘STRADL: Stratifying Resilience and Depression Longitudinally’. This project aimed to re-contact participants from GS:SFHS for a further assessment of mental health, specifically depression. Major depressive disorder (MDD) is a leading cause of global disease burden, with a lifetime prevalence of approximately 10%.5,6. In coming decades, the prevalence and impact of depression will likely increase,4 making the understanding of its aetiology of substantial importance to public health. STRADL was designed to investigate the aetiology of depression and its stratification, as it is hypothesized that the diagnosis may group together several causally distinct but symptomatically related syndromes.7 The increased kinship among STRADL participants created a rich dataset to conduct genetic studies of MDD aetiology, in addition to examining the complex genetic and environmental interactions which may increase risk for different depression phenotypes. STRADL was also designed to investigate psychological resilience8–10-the ability to ‘escape’ psychopathology despite exposure to known risk factors. Whereas the investigation of resilience may be expected to reveal similar results to studying MDD itself, it has the potential to elucidate protective factors in MDD, even in at-risk individuals. Indeed, examining the variability in response to known MDD risks may not just further our understanding of MDD; a better understanding of resilience mechanisms may also inform future interventions long before the development of the illness.11 Ultimately, the work of STRADL has the potential to identify causal mechanisms and pathways of depression sub-types and elucidate mechanisms which give rise to better than expected adjustment. The original GS:SFHS protocol and cohort characteristics have been described extensively elsewhere.1,4 Between 2006 and 2011, 24 084 participants were extensively phenotyped in addition to providing DNA samples for whole-genome genotyping. GS:SFHS was largely female (59%), and generally healthier and wealthier than the Scottish population. STRADL sought to recruit GS:SFHS participants for a follow-up assessment of mental health and resilience (N = 24 084). Individuals were eligible to participate if they had taken part in GS:SFHS, had a Community Heath Index (CHI) number, were alive and living in Scotland, and had given consent for re-contact. A total of 21 525 (89%) individuals from GS:SFHS were eligible for re-contact (Figure 1). STRADL recruitment flow diagram. Study packages were sent to potential participants (N = 21 525), which consisted of a study invitation letter, detailed information sheet and a paper booklet containing the STRADL questionnaires. Study packages were mailed by an independent party. Participants were given the option to complete the questionnaires on paper and return them via Freepost envelope, or use a given URL link for online submission. The opportunity to return an e-mail address for further contact was also provided. ‘Broad’ consent was obtained from participants, permitting the use of their data for ‘future medical research into health, illness and medical treatment’ without further specification. Participants were informed their data would remain anonymous and would be added to those already held by GS. A total of 9618 participants responded (45%) at follow-up. A total of 2460 families (N = 7158 individuals) of between 2 and 18 family members and 2460 unrelated individuals formed the STRADL cohort. The majority completed the paper version of the questionnaire (N = 8833), and 8% completed online (N = 785). Henceforth, individuals who participated in STRADL will be referred to as ‘respondents’, and those who did not reply will be referred to as ‘non-respondents’. All components of STRADL received formal, national ethical approval from the NHS Tayside committee on research ethics (reference 14/SS/0039). Age and sex information were collected in the STRADL, and all other demographic data were obtained from GS:SFHS (Table 1). The STRADL cohort was predominately female (62%) and were older [mean = 50.48 years, standard deviation (SD) = 13.41] than non-respondents [mean = 44.28, SD = 15.70, t(21457) = −31.25, P < 0.001, Cohen’s d = 0.42) at baseline. STRADL respondents were from less socioeconomically deprived areas compared with non-respondents (t(19812) = −15.15, P < 0.001, Cohen’s d = 0.21) in the Scottish Index of Multiple Deprivation (SIMD) 2009.12,13 The STRADL cohort was generally healthier and wealthier, with a different age-sex profile in comparison with GS:SFHS, although important similarities were apparent (Table 1). Although STRADL may not be truly representative of the Scottish population, the sample includes data on participants from all socioeconomic status strata. Some baseline comparisons between STRADL respondents and non-respondents and full baseline sample (GS:SFHS) Abbreviations: STRADL, Stratifying Resilience and Depression Longitudinally; GS, Generation Scotland; GS:SFHS, Generation Scotland: the Scottish Family Health Study; SIMD, the Scottish Index of Multiple Deprivation 2009. With the exception of age and SIMD, values represent percentage; SIMD represents mean rank (SD). Some baseline comparisons between STRADL respondents and non-respondents and full baseline sample (GS:SFHS) Abbreviations: STRADL, Stratifying Resilience and Depression Longitudinally; GS, Generation Scotland; GS:SFHS, Generation Scotland: the Scottish Family Health Study; SIMD, the Scottish Index of Multiple Deprivation 2009. With the exception of age and SIMD, values represent percentage; SIMD represents mean rank (SD). A summary of all data collected and their completeness is shown in Table 2. All data were anonymized using a barcode system which linked with a unique participant identification number at the time of data analysis. Paper questionnaires were scanned into an electronic database with detailed in-built validity checks. Summary of phenotype data available, and percentage providing valid/useable data (N = 9618) Summary of phenotype data available, and percentage providing valid/useable data (N = 9618) Although data collection was largely cross-sectional, repeated measures of GS:SFHS measures were collected which enabled longitudinal examination. Pearson’s chi-square (χ2) tests were conducted to illustrate group differences for categorical data during GS:SFHS. As an alternative to the independent t test, comparisons between respondents and non-respondents at baseline are reported using the Mann–Whitney–Wilcoxon (U) test. Wilcoxon signed rank (W) tests have been reported for differences between respondents across GS:SFHS and STRADL as a nonparametric test equivalent to the dependent ttest. Calculations for group differences and changes over time have ignored the relatedness of the sample, which will be appropriately controlled for in future publication of the data. These differences are shown in Table 3. Repeated measures between respondents and non-respondents during GS:SFHS (baselines) and STRADL (follow-up) Abbreviations: GS, Generation Scotland; GS:SFHS, Generation Scotland: the Scottish Family Health Study; STRADL, Stratifying Resilience and Depression Longitudinally; SCID, Structured Clinical Interview for DSM-IV; MDD, Major Depressive Disorder; CIDI-SF, Composite International Diagnostic Interview – Short Form; GHQ, General Health Questionnaire. Unless denoted by (%), results represent mean (SD). GHQ scores calculated using the Likert method. *Significantly different from non-respondents in Wave1 at P < 0.05; **significantly different from respondents in Wave 2 at P < 0.05. Repeated measures between respondents and non-respondents during GS:SFHS (baselines) and STRADL (follow-up) Abbreviations: GS, Generation Scotland; GS:SFHS, Generation Scotland: the Scottish Family Health Study; STRADL, Stratifying Resilience and Depression Longitudinally; SCID, Structured Clinical Interview for DSM-IV; MDD, Major Depressive Disorder; CIDI-SF, Composite International Diagnostic Interview – Short Form; GHQ, General Health Questionnaire. Unless denoted by (%), results represent mean (SD). GHQ scores calculated using the Likert method. *Significantly different from non-respondents in Wave1 at P < 0.05; **significantly different from respondents in Wave 2 at P < 0.05. At baseline 17% of all participants smoked, although respondents were less likely to smoke (13%) compared with non-respondents (21%) [χ2(1) = 260.58, P < 0.001, r = 0.11]. At follow-up, the percentage of respondents smoking increased to 17% [χ2(1) = 1563.60, P < 0.001, r = 0.60]. At baseline 90% of participants drank alcohol, with respondents more likely to drink alcohol than non-respondents [χ2(1) = 5.23, P = 0.02, r = 0.02]. Respondents consumed less units per week (mean = 9.64, SD = 10.79) than did non-respondents (mean = 10.32, SD = 12.25) at baseline [t(13221) = 3.39, P < 0.001, Cohen’s d = 0.06], although this difference was small in magnitude. The number of respondents who drank alcohol decreased at follow-up [χ2(1) = 2575.7, P < 0.001, r = 0.53]. Respondent alcohol consumption substantially increased at follow-up (W = 3149600, P < 0.001), although extreme values were reported (mean = 13.91, SD = 22.42, range = 0−914). Respondent’s alcohol consumption was moderately correlated between time points (r = 0.50). GS:SFHS participants were screened for a lifetime history of major depressive disorder (MDD) using the Structured Clinical Interview for DSM-IV Disorders (SCID).14 Although the SCID has the potential to make inferences on an array of Axis I disorders, only case/control classifications for MDD were ascertained. The threshold for lifetime prevalence of MDD follows Diagnostic and Statistical Manual of Mental Disorders15 (DSM) criteria. Where symptoms of either depressed mood or anhedonia have been endorsed, a minimum of four further symptoms must also be endorsed and their clinical significance confirmed (i.e. symptoms lasting nearly all day, every day for a minimum period of 2 weeks). STRADL participants completed the Composite International Diagnostic Interview–Short Form (CIDI-SF).16 The CIDI-SF is a self-report questionnaire measure of psychiatric symptoms, developed from the larger CIDI by the World Health Organization17 according DSM-IV15 criteria. The CIDI-SF uses a stem-branch logic in which two symptomatic screening questions (symptoms of depressive mood or anhedonia) must be endorsed and reach clinical significance (lasting nearly all day, every day for 2 weeks or more). A minimum of four other symptoms must also be endorsed in addition to at least one screening question. Respondents who meet criteria for lifetime history of MDD reliably meet full diagnostic criteria with excellent accuracy if given the long version of CIDI.16 As the CIDI-SF can be completed in a relatively short period (approximately 10 min), it is a scaleable and acceptable measure for epidemiological studies. At baseline, an identical proportion of respondents (13%) and non-respondents (13%) met criteria for a lifetime history of MDD as established using the SCID, [χ2(1) = 0.55, P = 0.457, r = 0.01]. In STRADL, 16% of respondents met the CIDI-SF criteria for lifetime MDD (N = 1506), of whom 16% reported being currently depressed. Lifetime history of MDD was moderately correlated between the two measures (r = 0.30). The General Health Questionnaire-2818 (GHQ-28) was administered across time points as a tool used to identify milder psychiatric problems in the general population.19 As psychological distress represents a cluster of emotional symptoms linked to depression, the GHQ-28 was used alongside clinical measures to make better distinctions between syndrome and sub-threshold symptoms.20,21 Responses were scored using the Likert method (0‐1‐2‐3) whereby higher scores represent higher levels of psychological distress. Individual domain scores gave information on somatic symptoms, anxiety, social dysfunction and depression. Non-respondents experienced more psychological distress at baseline than respondents (U = 44888000, P < 0.001) although levels of psychological distress increased over time in STRADL respondents (W = 9713600, P < 0.001). Total GHQ score of respondents across time points was moderately correlated (r = 0.46). Symptoms of GHQ depression appeared greater in non-respondents than respondents at baseline, (U = 44607000, P < 0.001). GHQ depression scores increased in respondents between time points (W = 1409700, P < 0.001). GHQ depression scores were moderately correlated (r = 0.46) in respondents between assessments. Symptoms of GHQ anxiety were higher in non-respondents than respondents at baseline, (U = 45042000, P < 0.001). Anxiety scores increased in respondents between time points (W = 8937100, P < 0.001), and were moderately correlated (r = 0.45). Questionnaire measures of psychological resilience,22 coping style,23 threatening life experiences24 and medical conditions were obtained in STRADL, and are summarized in Tables 4, 5 and 6. Respondent results of resilience, coping style and psychological distress testing in STRADL GHQ scores calculated using the Likert method. Abbreviations: STRADL, Stratifying Resilience and Depression Longitudinally; SD, Standard deviation; BRS, Brief Resilience Scale; CISS, Coping Inventory for Stressful Situations; GHQ-28, General Health Questionnaire-28. GHQ-28 domain scores give information on: (a) somatic symptoms; (b) anxiety/insomnia; (c) social dysfunction; and (d) depression. Respondent results of resilience, coping style and psychological distress testing in STRADL GHQ scores calculated using the Likert method. Abbreviations: STRADL, Stratifying Resilience and Depression Longitudinally; SD, Standard deviation; BRS, Brief Resilience Scale; CISS, Coping Inventory for Stressful Situations; GHQ-28, General Health Questionnaire-28. GHQ-28 domain scores give information on: (a) somatic symptoms; (b) anxiety/insomnia; (c) social dysfunction; and (d) depression. Results from respondents completing the List of Threatening Experiences summarizing the number of individuals who endorsed each event and their ratings of its impact n1, The number of participants who indicated they experienced the event within the past 6 months. n2, The number of individuals who subsequently gave criterion contextual threat ratings of the event. Results from respondents completing the List of Threatening Experiences summarizing the number of individuals who endorsed each event and their ratings of its impact n1, The number of participants who indicated they experienced the event within the past 6 months. n2, The number of individuals who subsequently gave criterion contextual threat ratings of the event. Self-reported diagnosis of common illnesses among respondents Self-reported diagnosis of common illnesses among respondents The Brief Resilience Scale22 (BRS) assessed the ability to ‘bounce back’ from stress. Six questions were answered on a five-point scale from ‘Strongly disagree’ to ‘Strongly agree’. A total score was calculated as the mean of the six items, with appropriate reverse scoring of even-numbered questions. The BRS has previously shown good internal consistency and test-retest reliability.22 The Coping Inventory for Stressful Situations (CISS)23 was a self-report questionnaire measuring three coping style scales: task-oriented, emotion-oriented and avoidance-oriented coping. Two sub-scales of avoidance-oriented coping were also derived: distraction and social diversion. Each item was rated on a five-point scale from (1) ‘Not at all’ to (5) ‘Very much’. The CISS has proven a robust measure of assessing situation-specific coping strategies, with a stable factor structure and high construct validity.23,25 The List of Threatening Experiences (LTE)24 was a self-report measure consisting of 12 common and threatening life events that may have occurred in the 6 months preceding completion. For each threatening event endorsed, criterion contextual threat ratings are measured on a scale from 3 (‘Very bad’) to 1 (‘Not too bad’). The LTE has been shown to have excellent test-retest reliability and high sensitivity.26 STRADL is a mental health questionnaire follow-up of GS:SFHS. Although data collection was cross-sectional, STRADL becomes a longitudinal cohort because of NHS data linkage using Community Health Index (CHI) numbers which are allocated to every individual registered with a GP in Scotland. The ability to link with routinely collected NHS data will allow validation of the self-reported illness recorded in the study, and provide information on clinical endpoints and follow-up. Furthermore, the use of NHS linkage converts this two-phase cohort study into a potentially lifelong study of resilience and depression. Future parts of the STRADL study will include DNA methylation analysis and depression-focused neuroimaging measures of brain structure, function and As with epidemiological study, a is participants are representative of the population from which they are As indicated in Table STRADL participants to be an wealthier, largely female of GS:SFHS. This is not as these characteristics are associated with higher response to follow-up Whereas results from STRADL may would be in the general Scottish population, the cohort results in of data which represent the full of sex and between time points has been described in is that the response to STRADL (45%) is to a response whereby individuals with mental health are more likely to to a mental health questionnaire the has been Although the use of paper questionnaires participant it is that the majority of potential participants to complete or return their questionnaire is the in the of paper and online differences between paper and online respondents are in Table have been hypothesized for these The older demographic of STRADL may have GS:SFHS participants who did not have to a with or those who were not or to complete an online Furthermore, the URL provided in the was also long and it is that a long URL in a address may have been or for especially if information was Some baseline (GS:SFHS) comparisons between STRADL paper and online respondents Abbreviations: GS:SFHS, Generation Scotland: Scottish Family Health Study; STRADL, Stratifying Resilience and Depression Longitudinally; SIMD, the Scottish Index of Multiple Deprivation 2009. With the exception of age and SIMD, values represent SIMD represents mean (SD). Some baseline (GS:SFHS) comparisons between STRADL paper and online respondents Abbreviations: GS:SFHS, Generation Scotland: Scottish Family Health Study; STRADL, Stratifying Resilience and Depression Longitudinally; SIMD, the Scottish Index of Multiple Deprivation 2009. With the exception of age and SIMD, values represent SIMD represents mean (SD). differences between GS:SFHS and STRADL are summarized 1 and and differences between online and paper respondents are given in Table This cohort represents a and potentially valuable data to examine depressive symptoms, longitudinal and mechanisms of psychological have been with these but the of this is and to these data will to and the of shared and environmental to These data may provide as to can to their risk of depression and psychological distress. Furthermore, STRADL will allow to conduct genetic epidemiological analysis on of mental health, data held by GS. The STRADL cohort includes important to allow genetic and epidemiological research on the of MDD and The of this cohort in the repeated assessment of mood disorders, psychological distress and making it a valuable dataset to investigate the mechanisms that in addition to making longitudinal on depression and As data can be linked to NHS STRADL can be from a analysis into a longitudinal cohort a range of outcomes. Furthermore, the for longitudinal is of in repeated measures of mental health and resilience that be by a of this cohort other population such as STRADL participants were more likely to be and to from less socioeconomically participants from all socioeconomic were in baseline (GS:SFHS) and follow-up in prevalence of MDD were in STRADL the compared with the use of the SCID at baseline. This may be because the SCID is administered by and may have better than the self-reported research that the diagnostic classifications obtained using the CIDI-SF those in the larger Composite International Diagnostic In the CIDI-SF will be compared with the SCID and with linked NHS that a comparison of each can be and potential of can be the GS:SFHS follow-up represents a valuable to investigate the of depression and mechanisms of psychological resilience in a large, cohort. information from this cohort will be to throughout the and to and are from the GS at in a GS Wave 2 is a re-contact study of GS:SFHS which aimed to longitudinal measures of depressive symptoms, major depressive disorder (MDD) and measures of psychological resilience to identify their 21 525 participants from the GS:SFHS cohort were and to part in a further mental health In 9618 GS:SFHS individuals to part and provided re-contact with a response of The re-contact data questionnaire of medical conditions, psychological resilience, threatening life events and repeated measures of mental health and for health linkage was in addition to whole-genome genotyping. This large, family-based cohort represents a valuable to investigate the aetiology and of depression, mechanisms of psychological resilience and the of STRADL is by the Wellcome Trust a (reference The of the Scottish Health and the Scottish provided for Generation Scotland. is by the and is by an by the and are members of the for and which also funding from the and and is would to to all individuals who have taken part in GS:SFHS and STRADL, and the project research and of The that they have
Language: English
Subfield: Genetics
Open Access status: hybrid
Estimated APC paid: $3,145
OpenAlex ID: W2737161124
FWCI: 3.996
Cites: 31
Cited by: 72
Related to: 9