Camberg 1999

Methods	A Latin‐Square double‐blind cross‐over design. Method of allocation: random allocation. Date: not reported. Assessor blinding: yes. Participant blinding: nursing staff and observers were blinded. Participants were not blinded due to the nature of the intervention. Loss to follow‐up: unclear. Intention‐to‐treat analysis: yes. Subsequently an intervention‐specific analysis was made. Post‐program follow up: not reported.
Participants	Location: 9 nursing homes in Eastern Massachusetts and Southern New Hampshire (5 for‐profit facilities, 3 nonprofit, 1 government owned. 15 nursing units (9 Alzheimer's special care unit, 4 dementia units, 2 mixed units). Geographical area: USA. Sample: 54 subjects; 94.7% white. Age: mean age 82.7 years (SD 7.5 years). Sex: 48F, 77.2%. Baseline: average baseline cognitive measures indicated severe cognitive impairment (MMSE: 5.1, SD ± 4.4; TSI:13.1 SD ± 6.9) and low functional ability (BANS: 13.2, SD ± 4.0; ADL Self‐Performance Scale: 2.7, SD ± 1.2) Inclusion criteria: a documented diagnosis of Alzheimer's disease and related dementias (ADRD); age 50 years or older; the presence of at least 1 agitated behaviour per day of those listed on the short form of the Cohen‐Mansfield Agitation lnventory Scale (SCMAl) or 1 indication of withdrawn behaviour defined either by "sounding sad" or "crying" that occurred at least "often"; or "seldom" interested in activities, social interaction, or in their immediate environment, from the Multidimensional Observation Scale for Elderly Subjects (MOSES); medically stable; resident in the current nursing home for at least 3 months; and no planned discharge. Subjects must have a verbal interactive capacity judged by the appropriateness of verbal responses and pauses during a conversation when 5 questions and comments were introduced. Individuals were included if they responded verbally, if their verbal response was appropriate at least some of the time, and if they paused to listen, as in a conversation. Exclusion criteria: nonverbal patients, potential subjects who did not tolerate listening through headphones for 5 minutes to a nonpersonalised interactive conversation tape, or who were known to have a severe hearing impairment or premorbid history of psychiatric illness.
Interventions	The interventions were: (1) simulated presence therapy, (2) placebo and (3) usual care. Interventions: simulated presence: an interactive audio tape that contains 1 side of a live telephone conversation with a family member or surrogate, rich in selected memories and positive emotions. Placebo: "The placebo audio tape was a recording of a person reading emotionally neutral articles from the newspaper". Usual care: the routine interventions nursing home staff used to manage behaviour, such as staff interactions, redirection, or physical restraints. Application of intervention: nursing staff on the study units. Equipment: the audio tape was played using a headset and an auto‐reverse tape recorder enclosed in a hip pack. Frequency: twice a day (Monday through Friday) in place of their usual intervention. Reps: whenever a study subject exhibited an agitated or withdrawn behaviour. Duration: each treatment continued for 17 days over a 4‐week course with a 10‐day washout period following each treatment. Setting: nursing home. Supervision: a study monitor was assigned to spend 20 hours per week at each facility to ensure adherence to the protocol and to provide feedback to staff. Adherence: subject acceptance of the study intervention was high. 23 subjects (43%) accepted the intervention 100% of the time, and only 5 subjects (9%) refused it more than 50% of the time.
Outcomes	Several measures and 3 data collection methods were used as validity checks to identify negative and positive aspects of well‐being: (1) direct observations by trained nonparticipant observers; (2) daily staff observation logs of participants' responses to interventions used; and (3) weekly behavioural rating surveys by nursing staff. Direct observation of subject behaviour was done by trained non‐participant observers using a structured time‐sampling technique. Observers were blinded to the study intervention: each subject was observed at least 3 hours and 20 minutes per week between 9 a.m. and 7 p.m. Each observation block was 20 minutes long, divided into four 5‐minute observation intervals. Daily staff observation logs: when the staff decided to intervene, staff documented the target behaviour, i.e., agitation or withdrawal; the specific intervention (tape, restraint, staff interaction); the duration of the intervention; and the participant's response (improved, worsened, or remained the same). Nursing home staff documented daily (Monday to Friday) when an intervention was used for agitated or withdrawn behaviours. Weekly staff survey (administered to a nursing staff member who had cared for the subject for at least 3 days during the preceding week) used 2 behavioural rating scales: (1) the short form of the Cohen‐Mansfìeld Agitation Inventory (SCMAI) to record staff recall of the frequency of agitated behaviours; and (2) selected, modified items from the Multidimensional Observation Scale for Elderly Subjects (MOSES) to record the subject's mood and interest. 1. Overall behavioural and psychological symptoms. 2. Agitation: direct observations by trained nonparticipant observers: (1) Scale for the Observation of Agitation in Persons with Dementia (SOAPD), a 7‐item Observed Agitation Scale (developed by the researchers); (2) an agitation visual analog scale (AVAS). Daily staff observation logs. Weekly staff survey: (1) the short form of the Cohen‐Mansfield Agitation Inventory (SCMAI) to record the frequency of agitated behaviours. 3. Depression: direct observations by trained nonparticipant observers: (1) 2 Positive Affect items, "interest" and "pleasure", from the Philadelphia Geriatric Center Affect Rating Scale (PARS); (2) a withdrawal visual analog scale (WVAS); and (3) facial diagrams of mood (FACE). Daily staff observation logs. Weekly staff survey: (1) selected, modified items from the Multidimensional Observation Scale for Elderly Subjects (MOSES) to record the subject's mood and interest. 4. Anxiety: not assessed. 5. Quality of life: not assessed. 6. Activities of daily living: not assessed. 7. Caregiver burden: not assessed. 8. Dropouts: not assessed.
Notes	Funding: government (VA) and private (Harvard University) Declarations of interest: not stated
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	The method of randomisation was not clearly reported. "The treatments were applied in a restrictively randomised manner to each facility grouping".
Allocation concealment (selection bias)	Unclear risk	No statement was provided about allocation concealment.
Blinding of participants and personnel (performance bias) All outcomes	High risk	"The placebo audio tape was made to look like the SimPres treatment in routine use. Staff were asked not to listen to the audio tapes". Participants were not blinded due to the nature of the intervention.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	"Observers were blinded to the study intervention".
Incomplete outcome data (attrition bias) All outcomes	Low risk	Withdrawals are clearly described. "During the course of the study five subjects were too ill to participate in at least one of the treatment phases". "Their available information was retained for all analyses".
Selective reporting (reporting bias)	Unclear risk	Protocol not available. Details of the results especially regarding agitated behaviour were not clearly reported: the study reported the prevalence of non‐agitated participants (25%); however data were not completely available to perform a meta‐analysis.
Other bias	High risk	The study was a cross‐over study and it was considered suitable for evaluating the intervention. However, the study failed to report paired data and the analyses were performed by observation rather than by participant in at least 2 of the 3 measurements used (Staff observation logs and Direct observation).