Author [Country and Year of Publication]	Population	Study Characteristics	Outcome	Level of Evidence [MMAT 2018]
	Setting: Nursing homes	Study type: Quantitative non-randomized study Multicenter retrospective cohort study	Symptom definition:	-Sampling strategy is relevant for research question but there is a high number of patients dying in hospital and not withheld in this sample. -Outcome measurements restricted to nursing notes without access to medical notes. -Appropriate statistical analysis.
Schildmann [10]			Agitation, anxiety, delirium/hallucinations not further defined.
Germany	Number of participants: 512 residents	Study duration: January 2015–December 2017
	Age: Median 89 years [range 55–107]	Power analysis:	Symptom control [measures]: Use of sedatives Indications: agitation [53%], anxiety [32%], no indication noted [33%] Lorazepam median daily dose: 1 mg [range 0.5–6 mg] 10% of patients receiving sedatives and 2% of all deceased residents received total daily doses judged as moderately sedating] Symptoms and consciousness level were not systematically recorded. Proportion of residents receiving sedatives in the last week of life differed significantly between nursing homes [range 14–36%, p < 0.001] Group receiving sedatives residents were younger [p < 0.001], dementia was less present [p = 0.006], were more frequently followed by a SPC team [p = 0.011] Use of sedatives with continuous effect Indications: agitation [54%], anxiety [50%], no indication noted [30%] Median days of treatment was 3 [range 1–7] Lorazepam with continuous effect median daily dose: 2 mg [range 1–6] 2% [n = 9] received doses judged at least moderately sedating. Residents were younger and kidney disease was more prevalent [p < 0.001 and p = 0.015 respectively]
	Gender:
	Female: 70%
	Male: 30%
	Population characteristics: Majority had multiple diagnoses 54% [n = 273] had dementia 54% [n = 272] had cardiovascular disease	Intervention: Use of sedative Lorazepam 19% [n = 98] Lormetazepam 1% [n = 7] Haloperidol 0.8% [n = 4] Diazepam, midazolam, levomepromazine [0.2%, n = 1 [each]] Use of sedative with continuous effect 9% [n = 110] [or 42% of all prescriptions]		Comments: Notes of general practitioners and PCT were not available for examination. Only nurses notes. Stricter regulation about the use of sedatives in nursing homes can be a contributing factor to the relatively low percentage of use of sedatives in this setting] Oral lorazepam seemed to be the drug of choice in nursing homes as compared to midazolam in palliative care units. Missing data on symptoms and the level of sedation. High number dying in hospital so lost to follow up
	Only 19% [n = 95] had cancer 21% died in hospital 3% [n = 16] was followed by a specialist palliative care team
	Frailty characteristics	Comparison:	[Adverse] side effects: No information available.
		No use of sedatives
	No or little need for care 4% [n = 20]	No use of sedatives with continuous effect
	Medium to high need for care 55% [n = 282]
	Very high need for care 41% [n = 207]		Life span: No information available.
Author [country and year of publication]	Population	Study characteristics	Outcome	Level of evidence [MMAT 2018]
Tatokoro [11] Japan	Setting: In patient palliative care unit	Study type: Quantitative non-randomized study Retrospective cohort study	Symptom definition: No clear definitions stated.	Representative target population Outcome measures not clearly defined, lack of measurements of dosage for benzodiazepines and antipsychotics
	Number of participants: 1032 patients	Study duration: April 2012–October 2019
	Age: Median 79y [IQR 71–86]	Power analysis: Absent	Symptom control [measures]: Symptoms with the highest prevalence were appetite loss [94%], delirium [83%], fatigue [83%], oedema [81%], difficulty sleeping [79%], and pain [77%] As age increased, significant decreasing trends were observed in the prevalence of pain, dyspnoea, fatigue, constipation, nausea, drowsiness, difficulty sleeping, anxiety, and dysuria, but no trends were observed across age groups in appetite loss, oedema, sputum production, or delirium As age increased, a significant decreasing trend was observed in the dose of opioids [p < 0.0001]: The median doses of opioids for those aged <70 years, 70–79 years, 80–89 years, and ≥90 years were 118, 72, 48, and 48 mg OME/day, respectively As age increased, a significant decreasing trend was observed in the need for sedation [p < 0.0001] As age increased, a significant decreasing trend was observed in the need for benzodiazepines: for those aged <70 years, 70–79 years, 80–89 years, and ≥90 years 84.8%, 81.4%, 71.9% and 69.7%r respectively [p < 0.0001]. As age increased, a non-significant decreasing trend was observed in the need for antipsychotics: for those aged <70 years, 70–79 years, 80–89 years, and ≥90 years 27.4%, 35.1%, 34.9% 31.1% [p = 0.26]
	Gender: Male: 51.8%
	Population characteristics: Terminally ill cancer patients Most common cancer = pancreatic cancer	Intervention: Treatment of pain and dyspnoea with opioids. Use of benzodiazepines, antipsychotics, antiemetics, anticholinergics In specific age group		Comments: Somewhat atypical population with most prevalent cancer diagnosis being pancreatic cancer Absence of information about symptom control or medication dosage for use of benzodiazepines and antipsychotics.
	Frailty characteristics Functional state using ECOG PS and palliative performance status. 90% of all patients had relatively poor performance status: 3 or higher on ECOG.	Comparison: Same intervention in other age groups: <70 y 70–79 y 80–89 y 90 y or older	[Adverse] side effects: No information available Life span: No significant trends in survival time between age groups. Mean survival time: 20 days
Author [country and year of publication]	Population	Study characteristics	Outcome	Level of evidence [MMAT 2018]
Ellsworth [13] USA (Ellsworth [13] is a cohort study but the outcome measure of interest is purely descriptive)	Setting: Long term care hospice unit [often part of community living centers]	Study type: Quantitative non-randomized study Retrospective case- control study	Symptom definition: Delirium defined as a disturbance in attention that develops over a relatively short period with often an additional cognitive disturbance such as disorientation, memory deficit, language deficits, visuospatial deficit of deficit in perception. Terminal delirium is defined as delirium that occurs in the dying process and implies that reversal is less likely.	Participants representative for target population in this study, but not sure that they are representative for our target population. Treatment is taken as a surrogate for diagnosis making the measurements potentially inappropriate. Outcome data lacking dosage and indication to treat. Confounders are accounted for in the analysis.
	Number of participants: 307 cases analyzed, 31 excluded n = 276	Study duration: 1 October 2013–30 September 2015
	Age: Overall mean age: 75.5 y Cases: Mean 76 y [SD 12.1] Controls 74.8 y [SD 2.8] [p 0.46]	Power analysis: /	Symptom control [measures]: Primary outcome: Determination of risk factors for the development of delirium Secondary outcomes: •Non-pharmacological interventions •Indication for delirium •Current usage patterns in treatment for terminal delirium •Most used antipsychotic drug was haloperidol ▪4% [n = 175] in the case group received haloperidol at least once in the last 2 weeks ▪4.4% [n = 4] received haloperidol for another indication, namely nausea and vomiting •Atypical antipsychotics were infrequently used: risperidone, olanzapine, quetiapine, aripiprazole •52.2% [n = 97] veterans requiring any type of antipsychotics for the treatment of terminal delirium. •40.3% [n = 75] received only as-needed doses. Mean as-needed doses per veteran 5.8. •7.5% [n = 14] required only scheduled doses. •With as needed and scheduled doses combined, each veteran received a mean 14.9 doses. •Administration was split evenly throughout the day with about 30% during day-evening- night.
	Gender: Not explicitly stated
	Population characteristics: 2 weeks prior to death evaluation All veterans 24% and 18.9% of patients suffered from dementia Most common diagnosis was cancer	Intervention: Use of antipsychotic in the last 2 weeks of life		Comments: Lack of information about dosage administered. Lack of information about case- detection. Lack of transparency about case definition/identification
	Frailty characteristics: Absent	Comparison: No-use of antipsychotic in the last 2 weeks of life	[Adverse] side effects: No information available.
Author [country and year of publication]	Population	Study characteristics	Outcome	Level of evidence [MMAT 2018]
Gambles [12]	Setting: General hospitals	Study type: Quantitative descriptive study Retrospective epidemiological study	Symptom definition:	-Sample strategy relevant for research questions -Representative and large study population -Low risk of nonresponse bias. Acceptable statistical analysis
UK			Agitation and restlessness without further defining
	Number of participants: -3893 patients -30 consecutive patients from 155 hospitals	Study duration:
		1 October–31 December 2008
	Age: Median 81 y	Power analysis: /	Symptom control [measures]:
	Group 1 [receiving medication]:		51% of all patients were given medication for agitation/restlessness
	Median 79y [IQR 70–86] Group 2 [no medication]: Median 83 y [IQR 76–88]		Midazolam -PRN only [n = 612]: •n = 570 [93%] 2.5 mg median total dose last 24 h [90% CI: 2.5–10 mg]
	Gender: Female 55% Group 1: female n = 1047 [53%] Group 2: female n = 1094 [57%]		-CSCI only [n = 802]: •n = 696 [87%] •10 mg median total dose last 24 h in mg [90%CI: 5–20 mg] -PRN +CSCI [n = 568]: •n = 554 [98%] •15 median total dose last 24 h [90% CI: 7.5–40 mg]
	Population characteristics: 66% diagnosis other than cancer Frailty characteristics: absent	Intervention: Explore the use of medication for agitation and restlessness in the final days of life supported by the Liverpool Care pathway: midazolam, haloperidol, levomepromazine		Comments: -Large study population of predominantly older patients -Large non-cancer population -Absence of frailty characteristics -Lack of information about choice
			Haloperidol -PRN only [n = 612]: •n = 22 [4%] 1.5 mg median total dose last 24 h [90% CI: 0.5–2.85 mg]	of drug related to indication -Lack of a clear definition of agitation -Lack of response to medication
			-CSCI only [n = 802]: •n = 127 [16%] 3 mg median total dose last 24 h [90% CI: 1.5–5]
			-PRN +CSCI [n = 568]: •n = 94 [17% •3 mg median total dose last 24 h [90% CI: 1.5–10 mg]
			Levomepromazine -PRN only [n = 612]: •n = 17 [3%] 6.25 mg median total dose last 24 h [90% CI: 4.75–80 mg]
			-CSCI only [n = 802]: •n = 99 [12%] 6.25 mg median total dose last 24 h [90% CI: 5–25 mg]
			-PRN +CSCI [n = 568]: •N = 82 [14%]
			12.5 mg median total dose last 24 h [90% CI: 6.25–117.5]
		Comparison: Not receiving medication for agitation and restlessness.	[Adverse] side effects: No information available about adverse side effects