Jarvis 2012.

Study characteristics
Methods	Assessment or screening intervention versus control, placebo or no intervention Design: pilot RCT Stratification: none Randomisation sequence: "randomized into one of two groups using a computer‐generated randomization table" Comparisons: 2 groups: Group A (control) and Group B (experimental) Allocation concealment: not stated ("randomization process was administered by a researcher at the University of Liverpool, who was not involved in data collection but was involved in the later data analysis") Blinding: participants were blinded ("participants were masked to group allocation"), "assessors were not masked" ("findings from the visual assessment were withheld from the therapy staff in group A (control group). In comparison the visual assessment details were made available to the therapy staff for participants in group B (experimental group)". Power calculation: feasibility study ("prospective observation cohort study in the UK suggested that of all stroke patients referred with suspected visual impairment, 85% were found to have an identifiable visual impairment (Rowe 2007). On the basis of this preliminary data, this pilot study aimed to screen 100 patients in order to recruit a minimum of 70 participants"). Intention‐to‐treat analysis: no ("All data analysis was conducted based on the recruited patients to each group with full FIM data collection for both baseline and 6‐week follow‐up assessment"). Other recruitment details: "Participants were prospectively recruited between February 2008 and July 2009". "Specified members of the health care team (nurses, stroke physicians, physiotherapists and occupational therapists) on the stroke unit were required to screen patients against these criteria to identify potential participants. Staff used a screening form with questions to identify visual signs (head turning, strabismus, ptosis) and symptoms (diplopia, loss of vision, field loss). This was adapted from the screening form used in the Visual In Stroke (VIS) study (Rowe 2010)"."If visual impairment was noted, the screening form was sent to the orthoptist and hence, it doubled as a referral form". "Where necessary, for example when a potential participant had communication difficulties, adaptations were made to the consent process". Patient and public involvement: not stated
Participants	Total study population: 64 Withdrawals: 13 died before full baseline assessment (Group A: 4; Group B: 9). Another 10 cases ( Group A: 6, Group B: 4) withdrew (death, early discharge, no follow‐up or DNA follow‐up) at follow‐up assessment. Method of diagnosing VFD: full visual assessment undertaken by orthoptist. Battery of routine tests used as part of a previous study (Rowe 2009). It comprised tests of visual acuity ‐ logMAR (Bailey 1976), ocular alignment ‐ cover test (Pediatric Eye Disease Investigator Group 2009), ocular motility ‐ saccadic, smooth pursuit and vergence assessment (Holmes 2001), stereopsis ‐ Frisby test (Rosner 1984), visual field ‐ confrontation (Cassidy 2001) and visual inattention ‐ line bisection, star cancellation, and clock drawing. "The orthoptist did not suggest possible adaptive strategies to be undertaken, the focus was on alterations made by the therapists based only on the visual assessment information". Characteristics of population: participant details are listed in Table 6. Type and severity of visual problems: participant details are listed in Table 7. Inclusion criteria: "stroke within 2–6 weeks of being recruited, had a decreased functional ability compared to pre‐stroke functioning, had a post‐stroke visual impairment, and were able to understand the research process" Exclusion criteria: "unable to consent due to cognitive impairment or communication difficulties, or if they had a visual field impairment pre‐existing their stroke" Baseline comparison of treatment groups: "no significant difference in the composition of both groups (x2 test) for gender (P = 0.846), age (P = 0.113), stroke type (P = 0.564), stroke area (P = 0.499), stroke laterality (P = 0.396) and handedness (P = 0.268)"
Interventions	Group 1: experimental (n = 24 full baseline assessment; n = 20 at follow‐up assessment) Intervention type: assessment or screening. Materials: modified screening form. Where can materials be accessed? Screening form (modified) available from the Visual In Stroke (VIS) study (Rowe 2010). Procedures: "all participants underwent a full visual assessment by an orthoptist. The findings from the visual assessment were ... made available to the therapy staff for participants in group B (experimental group)". "participants in both groups received occupational therapy and physiotherapy". "Therapy routinely included working to regain motor activity and increase ability to achieve valued functional tasks. The therapists used strategies such as visual scanning and cueing to the affected side as part of their practice". Treatment in the experimental group was informed by the results of the visual assessment. Provided by: orthoptists. No details provided about their training. Delivery: face‐to‐face, location Inpatient, stroke unit (Table 5). Regimen: not stated. Tailoring: yes. Details not supplied. Modification: not stated. Adherence: not stated Group 2: control (n = 27 full baseline assessment; n = 19 at follow‐up assessment) Intervention type: control/standard care. Materials: NA.Where can materials be accessed? NA. Procedures: described above. Treatment in the control group was not informed by results of the visual assessment ("all participants underwent a full visual assessment by an orthoptist. The findings from the visual assessment were withheld from the therapy staff in group A (control group)"). Provided by: OT, physiotherapy. No details provided about their training. Delivery: face‐to‐face, individual, location inpatient (Table 5). Regimen: NA. Tailoring: NA. Modification: NA Adherence: NA
Outcomes	See Table 8 Functional independence measure (FIM) Timed 10 m walk Non‐validated questionnaire ("Therapist with the most contact with each participant was asked to complete a non‐validated questionnaire giving qualitative information about their treatment approach." "Two versions of this questionnaire. The group A questionnaire asked the therapist to justify their treatment approach. The group B questionnaire required the therapist to comment on whether their treatment approach had been influenced by the visual assessment"). Time points when outcomes were assessed: "baseline and at 6 weeks after baseline (or on discharge if this occurred earlier)"
Notes	Review author Fiona Rowe was involved in this trial.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Low risk	"computer generated randomization table. The randomization process was administered by a researcher at the University of Liverpool, who was not involved in data collection but was involved in the later data analysis. Participants were masked to group allocation, but the assessors were not masked in this trial".
Blinding (performance bias and detection bias) All outcomes	High risk	"Participants were masked to group allocation, but the assessors were not masked in this trial." "The qualitative study indicated an inherent bias had been introduced to this trial, due to the inability to blind carers, and assessors to group allocation. The health care team perceived that the presence of a full baseline vision assessment enhanced their awareness of the effect of visual deficits following stroke. This was regardless of whether or not the full visual assessment details were available."
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	No intention‐to‐treat analysis. Dropouts were clearly reported, but reason for not attending was not clear and could be related to the intervention.
Other bias	Low risk	No other concerns noted