Table 2.

Summary of study characteristics and findings for attitudes to EAS (n = 16) 

Study/Setting	Sample size and characteristics (Total sample size, % female, mean age)	Study design, patient population & recruitment procedure	Social connectedness measure	Attitudes to EAS measure	Summary of whether exposure/outcome measures were validate (✔ or X)	Association between exposure and outcome (✔ or X)	Statistical analytic approach	Main findings	Quality assessment and notable methodological limitations
Social isolation (n = 1)
Wilson et al. (2007) [61] Canada	N = 379; mean age 67.2 (SD = 12.9); 55% female; majority were white, Protestant, were married, lived with their spouse and had more than a high school education; mean religiosity score was 9.6/15	Cross-sectional semi-structured interview study of patients with terminal cancer recruited from a national palliative care survey	Social isolation measured by summing social network size based on asking patients about number of children, close relatives and friends (unvalidated measure)	Attitude to EAS measured by asking patients if they would ask for EAS if it were legal, as part of a set of standardised questions developed previously by the authors in a previous study [62] (unvalidated measure)	Exposure X Attitudes outcome X	✔	Chi-square tests and logistic regressions to test associations	Social network size was associated with current interest in EAS I bivariate analyses (OR = 4.0, 95% CI = 1.6–10.5; p = .008)	High risk of bias No adjustment for potential confounders; no correction for multiple testing
Loneliness (n = 1)
Buiting et al. (2012) [63] Netherlands	N = 1,245 participants in the Longitudinal Ageing Study Amsterdam for 2008; a study that started in 1992 to recruit a nationally representative sample of people aged 55–84 years. In the sample used for this analysis (aged at least 70), more than half of the respondents were 74 years or younger, 55% were women and about 60% of the respondents were married or had a registered partnership	Longitudinal design, involving repeated cross-sectional surveys of separate samples of older adults to explore changes in attitudes over time. Each was a representative sample of older adults (age 55–85 years), stratified for age, sex, and level of urbanisation	Loneliness measured using the De Jong Gierveld Loneliness Scale (to capture social loneliness and emotional loneliness) (validated measure)	Attitudes to EAS captured by asking patients whether they could imagine themselves ever requesting EAS (unvalidated measure)	Exposure ✔ Attitudes outcome X	X	Logistic regression using Generalised Estimating Equations (GEE), adjusted for age group and time (multivariable analysis)	No differences in the proportion of people who reported no loneliness (40% vs 41%), loneliness (39% vs 36%). or severe loneliness (21 vs 23%) for people who could imagine requesting EAS vs people who could not image requesting EAS; text reported no significant association of loneliness with desire for EAS but no test statistic presented	Medium risk of bias no test statistic presented for this comparison (the lack of an association reported in text only); those sampled were adults aged over 70 and were not defined by life-limiting illness
Loneliness and perceived social support (n = 1)
Cicirelli V (1997) [64] USA (Indiana)	N = 388; mean age 73; 74% female; majority were widowed and had some college or vocational training; mean religiosity was 14.14/16; mixed socio-economic status and ethnicity (specifically recruited Black and White older adults)	Cross-sectional survey of older adults aged 60–100 years living in the community in Indiana state, recruited through older adults’ organisations	Loneliness measured using the 20 item UCLA Loneliness Scale (validated measure) Perceived social support measured using the Social Networks in Adult Life Survey to capture the number and closeness of people perceived by an individual to be in his or her support network (validated measure)	Attitudes to EAS using a rating scale of whether EAS was acceptable in 17 different personal situations, each with 7 end-of-life decision options (note that wording was authors’ own): “to strive to maintain life, to refuse medical treatment or request its withdrawal, to allow someone close to decide what is best in the situation, to commit suicide, to ask the doctor (or someone else) for assistance in committing suicide, to ask the doctor (or someone else) to end one's life, and to allow the doctor (or someone else) to decide to terminate life)”, rating each of the seven decision options whether they would or would not decide to do it using a 5-point Likert-type scale from would not do (1) to would do (5) to indicate their view, then collapsed to form a dichotomous variable indicating if the respondent would be likely to find the option acceptable or not (unvalidated measure)	Exposure ✔ Attitudes outcome X	X	Factor analysis reduced the 17 outcome variables to three outcomes: Endlife: included the decision options of suicide, assisted suicide, and voluntary euthanasia Others: included the two items in which the end-of-life decision was deferred to a significant other or to the physician Maintain:included the decision options of trying to live as long as possible and refusing or withdrawing treatment (although these were contradictory) Loneliness and perceived social support were tested among 14 different independent variables Hierarchical regression analysis adjusted for ethnicity, gender, socioeconomic status, marital status, age, health, functional dependency, quality of life values, religiosity, fear of death (fear of the dying process, fear of the destruction of the body, fear of the unknown & fear for significant others), wellbeing (self-esteem, depression, life satisfaction & loneliness), locus of control (internal, chance & powerful others), life events, social support (multivariable analysis)	Focussing on the outcome Endlife to represent direct desire for EAS, there was no significant association observed with loneliness (Beta coefficient = .08) or perceived social support (Beta coefficient = -0.02 for inner circle; 0.07 for middle circle; 0.00 for outer circle)	Medium risk of bias unvalidated measure of attitudes to EAS developed by the authors; risk of Type I error given exploratory nature of analysis; older adults sampled were not defined by whether or not they had a life-limiting illness
Social isolation and thwarted belongingness (n = 1)
Stolz et al. (2017) [65] Austria	N = 493 care-dependent older adults Mean age 74 58% female Majority did not live alone, had primary/ lower secondary education, and considered themselves ‘rather religious’	Cross-sectional survey of care-dependent older adults (age 50 and over) in the community sampled through a multistage random sampling procedure among the general population, with data collection by face-to-face computer-assisted interviews. Care-dependent status was defined as a having a physical or mental disability that was expected to last at least 6 months	Social isolation captured using a binary measure of living alone (objective measure) Thwarted belongingness measured using the thwarted belongingness sub-scale from the Interpersonal Needs Questionnaire (INQ-10) including 5 items such as ‘These days, I feel disconnected from other people’ or ‘These days, I feel like I belong’ (validated measure)	Attitudes towards EAS based on a set of 4 questions asking whether the participant would support EAS in different scenarios (worded as: “‘In case this older, care-dependent person would not want to live on, what of the following would you support?”) to derive four outcomes: Approval of availability of assisted suicide; Hypothetical utilisation of assisted Suicide; Approval of availability of euthanasia; Hypothetical utilisation of Euthanasia (unvalidated measures, based on questions used in previous study by the authors)	Exposure ✔ for both social isolation and for thwarted belongingness Attitudes outcome X	✔ social isolation (living alone) X thwarted belongingness	Multivariable logistic regression adjusted for gender, age, education, area (village/town/city), religiosity, physical illness (self-rated health, functional limitations, sensory functioning), psychological distress (fear of death, perceived burdensomeness, depressed affect, passive suicide ideation, active suicide ideation), living alone, social trust, trust in doctors (multivariable analysis)	Living alone was associated with approval of availability of euthanasia (OR =  2.54; 1.39–4.63; p = < 0.05) and hypothetical utilisation of euthanasia (OR = 3.99; 2.19–7.26; p = < 0.05) but was not associated with approval of availability of assisted suicide or hypothetical utilisation of assisted suicide thwarted belongingness was not associated with any of the four outcomes	Low risk of bias
Satisfaction with social relationships (n = 1)
Seidlitz et al. (1995) [66] US	N = 802 adults (541 women and 261 men); Mean age 71 67% female; Majority were white, married, and had religion as a major role in their lives.; mixed socio-economic status	Cross-sectional survey of older adults (aged 60 years and above) capable of independent living, selected through randomly generated telephone numbers via a Gallup telephone survey	Satisfaction with family relationships rated using a 5-point Likert scale (unvalidated measure)	Attitude to EAS captured using one standardised question (used in Gallup surveys for older people) with a 5 point Likert-type scale, alongside four other questions on attitudes to suicide more generally (unvalidated measure)	Exposure X Attitudes outcome X	X	Logistic regression adjusted for age, gender, race, marital status, religiousness, self-rated health, satisfaction with relationships, income (multivariable analysis)	No significant association of satisfaction with family relationships and attitude to EAS; No significant association of satisfaction with family relationships and other attitudes to suicide more generally e.g., that suicide is a personal decision and that others should not get involved; text reported that satisfaction with family relationships emerged as a predictor of the attitude that suicide is a personal decision in the multivariate analyses but p-value was 0.055 i.e., non- significant	High risk of bias no response rate recorded for telephone sampling; unvalidated measure of satisfaction with family relationships; older adults sampled were not defined by whether or not they had a life-limiting illness
Social isolation and social support (n = 1)
Berkman et al. (1999) [67] USA (Oregon, Michigan)	N = 511; mean age 50 80% female; majority were white, said religion was very helpful in their lives, were married, and had an undergraduate degree; mixed SES	Cross-sectional survey of patients with Multiple Sclerosis (MS) recruited from the National MS Society (Oregon and Michigan groups), using printed questionnaires (one participant completed on the telephone due to disability)	Social isolation measured by asking about number of close friends and relatives (unvalidated measure) Social support measured using the 19-item MOS-SSS, with items assessing the frequency of availability of social support in the domains of emotional/informational, tangible, affectionate, and positive social interaction (validated measure)	Attitudes to EAS measured using a question asking whether they would consider EAS as an option (yes/no) and rating five hypothetical scenarios when this would be acceptable: if you were experiencing unbearable pain, could no longer do anything that makes your life worth living; were causing a financial burden to caregivers and /or family members; could no longer enjoy anything that makes you happy; were feeling extreme emotional distress Also asked one direct question: Have you ever thought about assisted suicide as an option yourself? (unvalidated measures)	Exposure X and ✔ Attitudes outcome X	X for social isolation ✔ for social support	Chi-square tests for bivariate associations with a final multivariable regression model using all variables significantly associated with outcome (multivariable analysis)	Social isolation: People who had ever considered EAS as an option had a significantly lower number of family and friends (5.4; SD = 4.8 versus 6.8; SD = 4.9; p = < 0.001) in linear test for trend) In the multivariable regression model social isolation not significantly associated with positive attitude to EAS in any of the hypothetical scenarios, or to whether they had ever considered EAS Social support: People who had ever considered EAS as an option had a significantly lower score on social support scale (47.6 points, SD = 19.6 versus 56.8, SD = 16.9; p = < 0.001 using Pearson chi-square test) In the multivariable regression model higher scores on the Social Support Scale were associated with a lower likelihood of considering assisted suicide for all of the hypothetical scenarios (OR for experiencing unbearable pain = 0.98; 95% CI = 0.97,0.99; p = < 0.01; OR for no longer enjoying anything that makes life worth living = 0.99; 95% CI = 0.97, 0.99; p = < = 0.05; OR for could no longer do anything that makes one happy = 0.99; 95% CI = 0.97, 0.99; p = < = 0.05; OR for feeling extreme emotional distress = 0.97; 95% CI = 0.96, 0.99; p = < = 0.001) except for causing financial burden (non-significant). Note that two of these reported p values were of borderline significance Higher scores on the Social Support Scale were significantly negatively associated with having ever thought about assisted suicide (OR 0.98; 95% CI = 0.97, 0.99; p = < = 0.05) but again this reported p value was of borderline significance	Medium risk of bias membership of MS Society not representative of all patients with MS; specific confounders unclear
Social support (n = 10)
Achille & Ogloff (2004) [68] Canada, UK, & US	N = 44; mean age 52; 25% female Majority were Protestant, attended weekly services and prayed daily. Majority were married and lived with spouse, and had a college/university education	Cross-sectional survey of patients with Amyotrophic Lateral Sclerosis (ALS) recruited via advertisements in an ALS email newsletter sent weekly to a reported readership of 3,000 worldwide	Social support measured using the Social Support Questionnaire (SSQ-6) with 6 items capturing network availability and perceived satisfaction with the support available (validated measure)	Attitudes to EAS measured using Likert-type scales asking if participants could see themselves asking for EAS in the future, or whether they would have already asked for this if it were legal (unvalidated measures)	Exposure ✔ Attitudes outcome X	✔	Correlation	Negative correlation between availability of social support and whether patients were contemplating EAS (r = -.34, p < .05)	Medium risk of bias presented simple correlations rather than a multivariable model
Arnold (2004) [69] USA	N = 148; mean age 60; 57% female; majority were white and lived alone	Cross-sectional survey of patients with terminal illness recruited via national and state membership newsletters by a right-to-die organisation and local newspapers in a south-eastern USA state	Perceived social support from family, friends and significant others measured using the 12-item Multidimensional Scale of Perceived Social Support (MSPSS) (validated measure)	Attitude to EAS captured by asking if patient had considered hastening their death in the future (yes/no) (unvalidated measure)	Exposure ✔ Attitudes outcome X	✔	Discriminant function analysis; adjusted for depression, anxiety, pain, social support, hope (multivariable analysis)	Unadjusted effect size: Wilks lambda = .736 (p-value not reported) Adjusted effect size: Wilks lambda = .79 (p = .001)	High risk of bias Recruitment was via those who were members of a right-to-die organisation
Blank et al. (2001) [70] USA (Connecticut)	N = 158; mean age 74; 62% female; majority were white and unmarried	Cross-sectional study of older adults surveyed in an acute general medical hospital who had a life expectancy of over 6 months	Perceived social support measured using rating scales of perceived instrumental (i.e., practical), emotional, and financial support (standardised questions taken from the Established Populations for Epidemiologic Studies of the Elderly, but unvalidated measures)	Attitudes to EAS measured using responses to a range of hypothetical scenarios in which they might accept or refuse PAS: if they recovered after illness, if they had limited mobility after illness, if they had cognitive impairment, if they had a terminal illness, and if they required nursing home care. Also asked to rate their responses if they were under financial constraints (unvalidated measure)	Exposure X Attitudes outcome X	X	t-tests; chi-square tests; final logistic regression model including all those variables significantly associated with positive attitudes to PAS, including depression, marital status, religious coping, social support (multivariable analysis)	Social support was significantly associated with considering accepting PAS for only one of the hypothetical scenarios (recovering from illness and being restored to current condition; t = .199; p-value < .05) but not for those scenarios in which they had a imagine they had a terminal illness No significant association found when entered into multivariable model Adjusted effect size OR = 0.69; 95% CI = 0.34–1.43; p-value not shown) Those with lower social support were more likely to change their preferences and show more interest in PAS if in a situation of financial constraints (p = < 0.05)	High risk of bias Unvalidated measures; complex hypothetical scenarios to interpret where measuring outcome; no correction for multiple testing; older adults sampled were not defined by whether or not they had a life-limiting illness
Breitbart et al. (1996) [71] USA (New York)	N = 378; mean age 39; 23% female; mean education level 13 years; mixed ethnicity	Cross-sectional study of ambulatory HIV-infected patients recruited from three New York hospitals and other sites serving HIV patients as part of a questionnaire study of pain in patients with AIDS	Perceived adequacy of social support measured using the Social Support Questionnaire (SSQ), capturing two dimensions—number of social supports; quality of social supports (validated measure)	Attitudes to EAS measured using questions asking whether participants would consider EAS, using standardised questions from a previous questionnaire study (unvalidated measure)	Exposure ✔ Attitudes outcome X	✔	t-tests; discriminant function analysis, adjusted for death of friend/family member, suicidal ideation, race/ethnicity, quality of social supports, frequency of attending religious services (multivariable analysis)	Significant association between number of social supports and attitude to EAS ( t = -2.41; p = 0.02) Significant association between quality of social supports and attitude to EAS (t = -2.66; p = 0.009) The final multivariable model, developed using discriminant function analysis, included quality of social supports, and correctly predicted 60% of cases (p = < .0001)	Medium risk of bias not a representative sample of patients with HIV; attitudes to EAS in the mid-1990s will be influenced by public understanding of HIV prognosis at the time of sampling
Emanuel et al. (2000) [13] US (four randomly selected sites nationally)	N = 988; mean age 67 52% female Majority were white, Protestant, somewhat/ not religious, married living with their spouse and had graduated from high school. The SES was mixed	Longitudinal (prospective cohort) study of patients with terminal illness referred by their physicians and their primary caregivers and interviewed at two points (T0 = at recruitment, T1 = 2–6 months after the baseline interview)	Social support measured using the 19-item MOS-SSS, with items assessing the frequency of availability of social support in the domains of emotional/informational, tangible, affectionate, and positive social interaction (validated measure) Additional descriptive measure: Caregivers were asked whether they perceived they had adequate social support, but measure/wording not stated (unvalidated measure)	Attitudes towards EAS and personal interest in EAS measured at baseline and follow-up by asking patients a set of standardised questions about their views on EAS used in previous surveys (e.g., “When a person has a disease that cannot be cured, do you think doctors should be allowed by law to end a patient’s life by some painless means if a patient and his family request it?”) and whether they would consider EAS in a hypothetical situation of pain, and of feeling a burden on others, in the context of terminal illness. Also asked questions about whether they had ever considered EAS (unvalidated measures but some standardised questions on attitudes to EAS) Additional exposure measure (ineligible for inclusion in study but reported here): Family members /caregivers were also asked whether the patient had discussed EAS or hoarded drugs (unvalidated measures but some standardised questions on attitudes to EAS)	Exposure ✔ Attitudes outcome X	X	Multivariate regression models testing variables in groups, then selecting variables that were significantly associated in groups, then conducting a final stepwise regression (multivariable analysis)	No difference in overall social support for patients who did/did not discuss EAS or hoard drugs (p = .71) [Caregivers who perceived adequate social support were significantly more likely not to express support for EAS (OR = 0.63; 95% CI = 0.44–0.90)]	Medium risk of bias Sampling bias as physicians/carers were asked to suggest patients; small numbers for statistical analysis involving social support variable
Ganzini et al. (1998) [72] USA (Oregon; Washington)	N = 100; Mean age 54 39% female; majority were white and married. They had a mean of 14.4 years of education, and the mean religiosity score was 73.8/100	Cross-sectional questionnaire survey of patients with Amyotrophic Lateral Sclerosis (ALS) and their caregivers recruited from the ALS clinic at Oregon Health Sciences University in Portland, who had participated in previous studies of ALS at the university or had expressed an interest in participating in research on ALS in a prior survey conducted in 1996 on behalf of the university by the Muscular Dystrophy Associations of Oregon and Washington states	Social support measured using the Duke University of North Carolina Functional Support Questionnaire (FSSQ) (validated measure)	Attitudes towards EAS among patients measured using a set of questions asking whether they would ever consider taking a drug to end their life (unvalidated measures) Additional descriptive measure: Attitudes towards EAS among caregivers measured using questions on whether they would support or oppose the patient’s decision to take a drug to end their life, and the perceived likelihood that the patient would request such a prescription (unvalidated measures)	Exposure ✔ Attitudes outcome X	X	Chi-square tests for categorical variables, t-tests for continuous variables	Social support did not differ significantly between patients who would consider EAS and those who would not (finding provided in text but not tables as only the significant associations were tabulated)	Medium risk of bias unvalidated measure of attitudes to EAS; unclear whether group differences on social support were analysed in relation to social support as a continuous or categorical variable
Himchak (1997) [73] USA (New Jersey, New York, Pennsylvania)	N = 329; mean age 74 75% female The majority were white, Catholic, widowed and lived alone. The SES was mixed	Cross-sectional survey of older adults living in three US states (New Jersey, New York, Pennsylvania) recruited (unclear how) from a range of settings: 216 individuals recruited from group settings such as supported housing, lunch clubs, and senior social clubs, and 113 individuals who were more isolated due to being homebound or not participating in community projects	Social support measured using the Social and Family Contact Scale to capture two dimensions: social support experienced in relationships and social support experienced through decision-making, with categories addressing family structure, frequency of interaction, and mutual aid and assistance. (apparently unvalidated measure; unable to find further information on this measure) Social support also measured using the Family Contact Index, a 4-item scale measuring frequency of contacts for each friend/relative, including those living with them (apparently unvalidated measure; unable to find further information on this measure)	Attitudes towards EAS measured using nine questions taken from five national polls (e.g., Gallup), wording not provided, but using a 3 point Likert-type scale to rate agreement (standardised but unvalidated measures)	Exposure X Attitudes outcome X	X	ANCOVA, correlation, and multiple regression analysis (multivariable analysis)	ANCOCA showed that social support was not associated with attitudes to EAS (values not reported) Social support was not correlated with attitudes to EAS (Correlation coefficients for four dimensions of social support ranged from r = -0.08 to 0.02, p = 0.16 to 0.67) Multiple regression found that social support alone did not explain a significant portion of the 11% variance in attitudes to EAS contributed by all variables in the model Were association, values not reported ANCOVA No significant association, values not reported	Medium risk of bias unclear how sample recruited, sample were older adults regardless of life-limiting illness; no response rate stated
Lulé et al. (2014) [74] Germany (also see Table 3)	N = 93; Mean age 59; 58% female; majority were married; mean of 10.98 years in education	Longitudinal study of patients with Amyotrophic Lateral Sclerosis (ALS) recruited at an outpatient neurology clinic, interviewed at 3 timepoints: T1 baseline interview, T2 6 months, T3 12 months	Perceived social support measured using the emotional scale of the 14-item German version of the Social Support Questionnaire (F-SozU K-14) (validated for a German sample)	Attitudes towards EAS measured using two questions about behaviour/attitudes to end-of-life intervention: “Did you seek information how to shorten life?’’ and ‘‘Should euthanasia be allowed? ‘‘ (response options ‘‘yes’’ or ‘‘no’’) (unvalidated measures)	Exposure ✔ Attitudes outcome X	X	Repeated measures ANOVAs were conducted separately for all dependent and independent variables (including DHD, attitudes to EAS and perceived social support), providing comparisons at each of the three timepoints Kruskal–Wallis ANOVA to estimate associations of six factors and attitudes to life-prolonging treatment multivariate logistic regression to examine which variables predicted desire for hastened death, mutually adjusted for quality of life, depression, feeling of being a burden, physical function, age and perceived social support	No association of perceived social support with attitudes to EAS (but no test statistics reported)	Medium risk of bias unvalidated measures of attitudes to EAS
Marrie et al. (2017) [75] US (nationwide)	N = 7,534; Mean age 60; 80% female; majority were white and had an undergraduate degree	Cross-sectional survey of patients recruited from a national Multiple Sclerosis (MS) research register (the North American Research Committee on MS)	Perceived social support measured using the 5-item Modified Social Support Survey (MSSS), using the items from the 18-item MSSS that are most strongly correlated with the total score rated using a Likert-type scale (validated measure)	Attitudes towards EAS measured by patients rating the degree to which they would consider assisted suicide in five hypothetical scenarios (unbearable pain, financial burden to caregivers, extreme emotional distress, inability to do things that make you happy, inability to enjoy anything) using the options: definitely would consider, probably would consider, probably would not consider, definitely would not consider (as used in a previous study but unvalidated measures)	Exposure ✔ Attitudes outcome X	X	Logistic regression adjusted for sex, race, age, annual income, disease duration, disability, depression, anxiety, pain, social support, religion, use of an immune therapy (multivariable analysis)	Bivariate association of social support with having positive attitudes to PAS (p = < 0.0001); multivariable logistic regression found no association when other variables were adjusted for (adjusted OR = 0.82; 95% CI = 0.59–1.13) but reported as significant in the text. In a supplementary analysis, presented in online e-tables, low social support was not associated with “definitely considering physician-assisted death in all situations”, under any of the five hypothetical scenarios, in fully adjusted models. In a set of further supplementary analyses, there was a significant association between low social support and definitely or probably considering PAD in a hypothetical situation of unbearable pain, but it was not clear what covariates were adjusted for in these four sets of analyses for five different hypothetical situations	Medium risk of bias Unvalidated measure of attitudes to EAS; contradictory findings reported in text and tables; unclear whether the same covariates were used in all models; multiple analyses were presented in supplementary tables with high risk of Type I error)
Pacheco et al. (2003) [76] USA (Ohio)	N = 24 with baseline and follow-up data (of 38 at baseline). Mean age 60 0% female; majority were married and very/ moderately religious; mixed ethnicity	Longitudinal study of patients with terminal cancer recruited from an oncology clinic; T0 questionnaire upon enrolment, T1 three months later (or upon significant decline in health status), and every six months thereafter (or upon significant decline in health status), up to a maximum of five follow-up measures	Social support measured using two dimensions from the 15 dimensions of the Coping Orientations to Problems Experienced Scale (COPE): seeking (i.e., use of) social support for instrumental (practical) reasons, and seeking (i.e., use of) social support for emotional reasons (validated measure)	Attitudes towards EAS, measured using five questions (covering PAS legalization, Euthanasia legalization, PAS request, and Euthanasia request) with agreement captured using a 5-point Likert-type scale (unvalidated measure)	Exposure ✔ Attitudes outcome X	X	Biserial correlations to examine whether change in attitudes towards EAS was related to change in coping strategies	Use of support for emotional reasons was the only COPE dimension that increased significantly over follow-up (p = 0.04), but attitudes to EAS did not change significantly over follow-up, suggesting no association. No cross-sectional associations were tested The correlations between attitudes towards EAS and change in seeking social support for emotional reasons were negative (i.e., those who changed to or stayed with conservative views increased more in use of social support for emotional reasons as compared to the increase in those who changed to or stayed liberal), although significance levels were not stated The biserial correlations between attitudes towards EAS legislation and seeking social support for emotional reasons were significant (PAS legalization -.50; E legalization -.49), while the correlations between attitudes towards EAS request and seeking social support for emotional reasons were moderate (-.34 and -.37, respectively), but not significant	High risk of bias Small sample size; unclear whether all patients were invited to participate; unvalidated outcome measure; comparisons based on correlations