Table 1.

Summary of study characteristics and findings for requested/actual EAS (n = 9) 

Study/Setting	Sample size and characteristics (Total sample size, % female, mean age)	Study design, patient population & recruitment procedure	Social connectedness measure	EAS measure	Summary of whether exposure/outcome measures were validated (✔ or X)	Association between exposure and outcome (✔ or X)	Statistical analytic approach	Main findings	Quality assessment and notable methodological limitations
Broad measures of social connectedness (n = 2)
Smith et al. (2011) [52] USA (Oregon)	N = 149; 43% female; Mean age (SD): 61.0 (13.3) for the group who requested and received PAD 61.2 (12.2) for the group who requested but did not receive PAD 60.1 (14.2) for the comparison group of deceased family members who did not pursue physician assisted death	Cross-sectional survey of the relatives of patients with terminal illness recruited from medical centres, hospices, organisations for patients with Amyotrophic Lateral Sclerosis (ALS), and euthanasia advocacy organisations in Oregon (where physician assisted death has been legal since 1997); 52 who received lethal prescriptions, 34 who requested but did not receive lethal prescriptions, and 63 deceased individuals who did not pursue physician assisted death	Social connectedness measured using 3 items from the Quality of Death and Dying Questionnaire (QODD): time spent with family and friends, attendance at important events, and physical expressions of affection (validated as a full measure but not individual items; note that this was completed by relatives retrospectively rating their relative’s experience)	Request for and actual EAS measured using categorical measures of: death by EAS, request for EAS, or not considering EAS, based on analysis of records kept in hospices, medical centres, and voluntary sector organisations representing those with terminal illness (objective measures based on routine records)	Exposure X EAS outcome ✔	X	ANCOVA	No significant group differences in scores on the three social connectedness measures between those who died by EAS, requested EAS, or were not considering EAS (effect sizes & p-values not reported)	Medium risk of bias No adjustment for baseline characteristic; control group was a convenience sample of family members recruited from the hospices, medical centres, and the same voluntary sector organisation as the other two groups
Ruijs et al. (2014) [53] Netherlands	N = 76; Mean age 70 52% female Majority were, single and had a middle/lower education. 51% were religious; all patients were Caucasian	Cross-sectional study of all patients on GP lists in Utrecht with terminal cancer who were expected to die within half a year and who were expected to live at home (most of the time) until death, as nominated by GPs	Measures of loneliness, Unsatisfactory social contacts, subjective isolation (capturing a sense of isolation or of ‘no one being present for you’ but termed ‘practical loneliness’); and insufficient support (from family/relatives) from the State-of-Suffering V (SOS-V) [54] measure, developed by the authors as a comprehensive instrument for quantitative and qualitative assessment of unbearable suffering related to 69 physical, psychological, and social symptoms for end-of-life patients, divided into five domains (validated measure)	Explicit request for EAS, as documented by GP (objective measure based on routine records)	Exposure ✔ EAS outcome ✔	X	Fisher’s exact test for nominal variables (using Bonferroni correction for multiple testing)	No associations between request for EAS and measures of loneliness, unsatisfactory social contacts, insufficient support, or subjective isolation (effect sizes not reported, p values ranged from p = .420 to p = .995 for social connectedness measures)	Medium risk of bias Whilst overall SOS-V was validated, the 4 measures of social connectedness were of uncertain validity used in isolation
Social isolation (n = 1)
Virik & Glare (2002) [55] Australia	N = 6 patients from a sample of 490 patients referred to the palliative care service over one year; 1 female (age 44) and 5 male (age range 58–78 years) patients Mean age 65	Cross-sectional study of the notes of palliative care patients who had requested euthanasia, as recorded in clinical audit records within a palliative care service	Social isolation measured based on subjective researchers’ judgement (binary measure) as to whether there was little or no contact with or support from family or friends, based on analysis of clinical audit notes (unvalidated measure)	Request for EAS, as documented as a request for euthanasia in the clinical notes (objective measure based on routine records)	Exposure X EAS outcome ✔	X	Frequency counts	67% (4/6) patients were judged to be socially isolated; No association described as numbers too small to conduct statistical analysis	High risk of bias No tests of statistical significance used in the analysis of the patient profiles, as the number of patients was too small for any meaningful comparisons
Loneliness (n = 1)
Snijdewind et al. (2015) [11] Netherlands	N = 645; Modal age was 80 + ; 62% female; Majority were single	Cross-sectional study of patients requesting EAS in 2013, by analysing content of applications to an end-of-life clinic offering euthanasia or physician-assisted suicide that were concluded during that year	Loneliness measured using frequency counts of mentions of loneliness within application forms (unvalidated measure)	Request for and actual EAS based on records of applications to an end-of-life clinic, categorised based on whether the patient’s request for EAS was: granted (n = 162) rejected (n = 300), withdrawn by the patient (n = 59), or the patient died before a final decision was reached (n = 124) (objective measures based on routine records)	Exposure X EAS outcome ✔	✔	Frequency counts were compared; bivariate associations of specific variables with EAS application outcome (granted versus rejected) presented in table, text lists variables that were independently associated with outcome after being entered into a multivariable regression model (multivariable analysis)	Loneliness was mentioned within the EAS request form for 49% of cases in which EAS was granted versus 72% of cases in which it was rejected 213 (p = < .001). Also reported, but did not compare statistically, that loneliness was mentioned in the application forms of 44% of those where the patient died before a decision was made, ad 70% of cases in which the application was withdrawn. Loneliness was independently associated with an EAS application being rejected (O = 0.52; 95% CI = 0.31–0.87; p = .01) as was being single, being single, having a solely psychological condition, and having loss of mental capacity. At least 13 other variables were not associated with EAS outcome	Medium risk of bias Exploratory analysis in which loneliness was one of at least 20 variables tested, with the potential for Type I error; analysis described in text as a multivariable regression analysis but no specific confounders listed
Loss of social role (n = 1)
Comby & Filbet (2005) [56] France	N = 13; mean age 63; range 15-/82 years; 54% female	Cross-sectional survey of palliative care patients requesting EAS from all those in four eligible palliative care units, representing 1.6% of the 611 patients under their care, with measures completed by clinical staff	Loss of social role captured by palliative care staff judgement over whether a patient had experienced loss of social roles e.g., within the family, or within their professional responsibilities because of illness (unvalidated measure)	Request for EAS based on palliative care registry of requests for EAS expressed by terminally ill patients or their families (objective measures based on routine records)	Exposure X EAS outcome ✔	X	Simple frequency counts; no associations investigated	12/13 patients experienced loss of social roles, and this (along with physical changes and existential suffering) was the most common reason for requesting EAS	High risk of bias highly subjective measure of loss of social role; no comparison with those who did not request EAS; no formal statistical analysis
Social support (n = 4)
Ganzini et al. (2006) [57] USA (Oregon) (also see Table 3)	N = 161; Mean age 62; 27% female; majority were white and married; mean of 13.5 years in education	Longitudinal study of patients with advanced cancer recruited from oncology clinics in Oregon (where physician assisted death has been legal since 1997), interviewed at baseline in 1998 and followed-up every three months for up to two years	Social support measured using the Duke University of North Carolina Functional Support Questionnaire (FSSQ) (validated measure)	Request for EAS measured as follows; those who indicated they might consider or were planning to request PAS were followed up every three months for up to two years; once referred to hospice, patients were followed monthly, as tolerated. After the patient’s death, each oncologist was asked if the patient had initiated discussion about PAS and any details of discussion (objective measures based on routine records)	Exposure X EAS outcome  ✔	n/a	Bivariate models to estimate associations at baseline using a proportional odds regression model For 42/161 patients followed up, 9 patient factors were regressed against outcomes	Insufficient power to investigate this outcome (but see Table 3 for DHD). Only 2 patients requested PAS so formal statistical analysis was not possible	High risk of bias analyses not adjusted for potential confounders; small sample size for longitudinal analysis and insufficient power to analyse data for this outcome
Ganzini et al. (2008) [58] USA (Oregon)	N = 83; 43% female; majority white	Cross-sectional survey of family members of deceased patients from the state of Oregon (where physician assisted death has been legal since 1997) who had requested assisted death, recruited from an Amyotrophic Lateral Sclerosis (ALS) support organisation and an assisted suicide advocacy organisation	Perception of the deceased having felt a lack of social support prior to death, as a potential reason for them having requested PAS, rated by family members using a 5 point Likert-type scale (unvalidated measure)	Request for EAS, based on whether patients had considered, requested or died by EAS prior to their death, based on the centralized registry of patients who explicitly requested a lethal prescription from a physician kept at two medical centres in Portland, Oregon (objective measures based on routine records)	Exposure X EAS outcome ✔	X	Descriptive statistics; ranking of the median score for each factor rated by family members	Family members did not identify social support as an important reason for a PAD request (median = 1/5, IQR = 1.1), and this was ranked at the bottom of the list (alongside depressed mood and financial concerns)	High risk of bias exposure measure highly subjective and subject to recall bias; statistical analysis limited
Ganzini et al. (2009) [59] USA (Oregon)	N = 56; Mean age 66; 52% female; majority were white and single/ divorced/ widowed; mean of 16.1 years in education	Cross-sectional survey of people who had requested EAS or contacted a euthanasia information organisation to request advice on PAD	Perception of a lack of social support, rated by patients using a 5 point Likert-type scale (unvalidated measure)	Patient ranking of the importance of their reasons for having requested PAS, rating each between 1 and 5, where 1 was “reason not at all important in decision to request a lethal prescription” and 5 was “reason very important in decision to request a lethal prescription”, in which one option was “Lack of support” (unvalidated measure)	Exposure X EAS outcome X	X	Descriptive statistics ranking the median score for each factor rated by patients, also presenting inter-quartile range	Patients did not identify social support as an important reason for a PAD request (median = 1/5, IQR = 1.1), and this was ranked at the bottom of the list (alongside depressed mood, current mental confusion, and current loss of continence)	High risk of bias exposure measure unvalidated; mixed sample of those who requested EAS formally and those who requested information about it; statistical analysis limited
Smith et al. (2015) [60] USA (Oregon)	N = 94; 50% female The majority were married. The ethnicity was mixed	Cross-sectional questionnaire study of patients with terminal illness recruited from end-of-life advocacy organisations and hospices, and nominated by ethics consultants, palliative medicine and oncology physicians at large medical centres	Social support measured using the Duke University of North Carolina Functional Support Questionnaire (FSSQ) (validated measure)	Request for EAS based on having contacted a large end-of-life advocacy organisation for information about accessing PAD, or made an explicit request for PAD to their physician (objective measures based on routine records)	Exposure ✔ EAS outcome ✔	X	Student’s t-tests to test for between-group differences on 6 measures (including social support) by PAS request; those significantly correlated at a threshold of p = < 0.20 were entered into a final logistic regression model (multivariable analysis)	No difference in social support for those who requested and did not request PAS (t = -0.67, p = 0.51) and so was not entered into the final logistic regression model	Medium risk of bias possibility of misclassification if those in the control group were in fact intending to pursue PAD but had not yet actioned or articulated it