Cohen‐Mansfield 1997

Methods	Method of allocation: non‐randomised cross‐over design. Cross‐over controlled study with 4 interventions: 1) family‐generated videotape (SPT); 2) audiotape recording of music; 3) one‐to‐one social interaction; and 4) control no‐intervention phase. Participant blinding: unblinded due to the nature of the intervention. Assessor blinding: not reported. Loss to follow‐up: 5 of the remaining 53 participants died before completing the interventions, 11 became quiet, 3 refused to continue their participation, and 2 were physically restrained and thus excluded. Intention‐to‐treat analysis: per protocol. Post‐program follow‐up: not reported.
Participants	Location: nursing home. N for each group: 32 residents in total (cross‐over trial with 4 groups). Sample: 60 participants (32 analysed); 4 groups. Age: 86.8 ± 1.16. Sex: 81% female. Baseline: Activities of Daily Living (ADL) assessed by Lawton and Brody's scale; Physical Self Maintenance Scale (PSMS): 4.0; Brief Cognitive Rating Scale (BCRS): 5.5. Rest of demographic characteristics (marital status, length of stay in the NH, diagnosis of dementia) are shown in table 1. Inclusion criteria: nursing home residents with VDB at least several times a day (in the results the authors state 97% of the patients were diagnosed as suffering from dementia). Exclusion criteria: physically restrained.
Interventions	Participants were assigned into 4 groups, each containing 3 interventions: 1) Medical examination. Special attention to identify underlying causes of physical discomfort; 2) Music half‐hour audio; tapes according to the participants' preferences were prepared (9 participants = classical music, 8 = Jewish music, 4 = big band music, 4 = old popular music, 2 = opera, 3 = show tunes, and 2 = country music); 3) Video tapes: the relatives of all the participants were asked to prepare a videotape. For these participants, a research assistant, a nursing assistant, and a volunteer prepared the videotapes. Family members received general guidelines to prepare the videotapes but the specific content and manner of conveying it were left to them; 4) One‐to‐one social interaction: an extensive list of possible activities was provided. Application of intervention: unclear. Equipment: not reported. Frequency: the interventions were performed at the time of the day at which the resident was most verbally disruptive and were performed only if the resident was verbally disruptive. Reps: not reported. Duration: each intervention lasted for 30 consecutive minutes. Each intervention lasted 2 weeks, followed by a washout week. Setting: nursing home. Recording: "Participants were observed for an hour a day: 15 min before the intervention, during the 30‐min intervention, and for 15 min after the intervention. Each observation lasted 3 consecutive mi. Total = nine 3‐min observations/hour". Supervision: unclear. Adherence: unclear. Participants were assigned into 4 groups, each containing 3 intervention conditions and 1 control “no‐intervention” condition. In each group the order of treatments was alternated, so that by the end of the evaluation period all groups were exposed to all the treatments and to the no‐intervention phase. Each treatment continued for 2 consecutive weeks and was followed by a washout week. Before exposure to any of the interventions, all participants underwent a detailed medical examination, by their attending physician, to identify any underlying causes of pain or physical discomfort. 1) SPT videotape. The relatives of all the participants were asked to prepare a videotape. Family members received general guidelines to prepare the videotapes but the specific content and manner of conveying it were left to them; 2) Music. 30‐minute audiotapes based on the participants' preferences were prepared and played to the participants in their rooms. 9 participants were exposed to classical music, 8 participants to Jewish music, 4 to big band music, 4 to old popular music, 2 to opera, 3 to show tunes, and 2 to country music; 3) One‐to‐one social interaction. A manual was prepared to guide the social interaction, which included information on how to interact with a nursing home resident, how to start an activity and a broad list of possible activities ordered by the degree of involvement, from high to low, required of the resident. A "box of activities", comprising games, pictures, balls, books, etc. was assembled. Alternatively, the following activities were provided: (i) conversation with a research assistant; (ii) motion exercise (e.g., tossing a foam ball, hand and arm movement); (iii) sensory stimulation (a sensory kit including different fabrics, school supplies, health aids, make‐up, spices, soaps, etc. was available to be presented to the residents to stimulate touch, sight, and smell); (iv) manual activities (such as making a collage, completing a simple puzzle, clipping coupons). Application of intervention: unclear. Equipment: not reported. Frequency: the intervention was performed at the time of the day during which the resident was most verbally disruptive (based on nurses' reports) and performed only if the participant was verbally disruptive. Reps: not reported. Duration: each intervention lasted 30 minutes per day for 2 consecutive weeks and was followed by a 1‐week washout. Setting: nursing homes. Recording: participants were observed 1 hour/day (15 minutes before, 30 minutes during, and 15 minutes after, an intervention), composed of nine, 3‐minute staggered observations. Supervision: unclear. Adherence: unclear.
Outcomes	1. Overall behavioural and psychological symptoms: 2. VDB: 3 different methods were used to assess VDB: tape recordings; standardised observations; and Informant ratings. Tape recordings: VDB duration was registered with a tape recorder located near a participant. The tape was coded by a trained research assistant using a computer program specifically developed for coding video and audiotapes. An episode of VDB was considered to be finished if the resident was silent for 1 full minute. An overall VDB score was calculating by averaging complaining (weight = 1), disruptive vocalisations (weight = 2) and shouting (weight = 4). Tape recordings of VDB were performed 1 hour/day (15 min before an intervention, 30 min during an intervention, and 15 mins after an intervention), for 10 days for each of the interventions (music, videotape, social interaction, and no‐intervention). Standardised observations: the frequency and context of screaming was examined with the Screaming Behavioural Mapping Instrument (SBMI) containing 9 types of VDB (shouting, screaming, or howling; constant requests for attention; repeating words; complaining or inappropriate verbalizations; cursing; verbal aggression; nonsense talk; hallucinations; and other disruptive verbal behaviours (e.g. groaning and singing)). Informant ratings: nursing staff well‐acquainted with the subjects rated their VDB using the Cohen‐Mansfield Agitation Inventory (CMAI) on a 7‐point scale (1: never manifests the behaviour; 7: manifests the behaviour several times/hour). Information was collected at the end of each intervention. Duration and frequency of appropriate verbal behaviours were also assessed. 3. Depression: not assessed. 4. Anxiety: not assessed. 5. Quality of life: not assessed. 6. Activities of daily living: functional independence (Lawton and Brody's Physical Self‐Maintenance Scale (PSMS)). 7. Caregiver burden: not assessed. 8. Dropouts: 5 of the remaining 53 participants (9%) died before completing the interventions, 11 (21%) became quiet, 3 (6%) refused to continue their participation, and 2 (4%) were physically restrained and excluded based on previous findings showing that the manifestation of VDB increased when nursing home residents were physically restrained.
Notes	Funding: this study was supported by National Institute on Aging Grant R01 AG‐10642. Declarations of interest: not stated.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	"Participants were assigned to four groups, each containing three intervention conditions and one "no‐treatment" condition. In each group, the order of treatments was alternated, so that by the end of the evaluation period all groups were exposed to all the treatments and to the no‐treatment period".
Allocation concealment (selection bias)	High risk	"Participants were assigned to four groups, each containing three intervention conditions and one "no‐treatment" condition. In each group, the order of treatments was alternated, so that by the end of the evaluation period all groups were exposed to all the treatments and to the no‐treatment period".
Blinding of participants and personnel (performance bias) All outcomes	High risk	Not possible due to the nature of the intervention.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not stated.
Incomplete outcome data (attrition bias) All outcomes	High risk	28 participants (47%) were excluded from analysis, 11 participants (18.3%) were excluded because they became quiet (they had a had a significantly higher cognitive status than those who continued manifesting VDB (3.8 and 5.5 on the BCRS respectively, tm = 2.8, P < 0.01).
Selective reporting (reporting bias)	Unclear risk	Protocol not available. Outcomes not clearly specified in the aim section. 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 analysis for VDB was based on number of observations rather than number of participants. The frequency of VDB was examined using the SBMI. A total of 11,520 3‐min observations were made (including observation before, during, and after the interventions). The percentage of observations in which VDB occurred was calculated.