Table 1 shows the baseline demographic, cognitive, health, activity, and physical performance characteristics of the 606 participants. The groups were similar, although the intervention group (n=305) contained more men and performed better in the timed up and go test.
Table 1 Characteristics of participants in intervention and control groups at entry to study. Values are numbers (percentages) unless stated otherwise
Intervention adherence and subject retention
The median time to the second glasses dispensing visit was 28 (interquartile range 19-43) days for the intervention group and 29 (14-53) days for the control group. Twenty-seven (9%) intervention group participants and 30 (10%) control group participants were provided with updated multifocal glasses at these visits.
Two hundred and seventy five (90%) intervention group participants received their single lens glasses within two months of randomisation. Nine intervention participants did not receive single lens glasses, two owing to diagnosed eye disease, one owing to dementia, and six owing to logistical difficulties. Thirty-five participants received additional counselling and support including follow-up telephone calls (n=24), replacement glasses (n=8), additional cords, glasses cases, reminder cards or “glasses use” information leaflets (n=8), and counselling sessions with optometrists (n=3).
The median number of months that the intervention participants reported satisfactory adherence with the recommendations for wearing single-lens glasses when walking in the street or in shopping centres was 7 (2-11) months. Ninety-six (32%) of 299 participants adhered for 0-3 months, 41 (14%) for 4-6 months, 40 (13%) for 7-9 months, and 122 (41%) for 10-12 months. Two control group participants purchased distance glasses during the follow-up period.
By the end of the 13 month intervention period, 47 participants had withdrawn from the trial (28 from the intervention group and 19 from the control group), giving completion rates of 90% and 94%. The figure shows the reasons for non-completion. The five deaths were unrelated to the intervention.
Effect of intervention on falls
Table 2 shows rates of falls and the number and proportion of intervention and control group participants who reported falls, falls outside the home, and injurious falls within the 13 month trial period; 299 intervention and 298 control participants completed one or more falls calendars. Blinding was maintained well throughout the study, although occasionally participants inadvertently revealed their group status (that is, mentioned new glasses) when research personnel made follow-up calls as part of the falls surveillance. In the 597 participants who completed calendars, fall rates did not differ significantly between the groups. In subgroup analysis, we found no differences in fall rates by baseline physiological falls risk (P=0.31) or number of reported falls in the previous year (P=0.19).
Table 2 Falls, outside falls, and injurious falls in 13 month trial period
Significant interactions were evident between the intervention and self reported baseline outdoor activity levels (assessed with the Adelaide activities profile) for the outcomes of all falls, falls outside the home, and injurious falls (P<0.001). In more active participants with Adelaide activities profile subtotals for outdoor activities above the median of 15, intervention group participants had significantly lower rates of all falls, falls outside the home, and injurious falls. In this subgroup, the numbers needed to treat to prevent one fall, one outdoor fall, and one injurious fall within 12 months were 1.1, 1.9, and 2.6. In contrast, in those who had Adelaide activities profile outdoor subtotal scores less than or equal to the median, intervention group participants had a non-significant increase in all falls and injurious falls and a significant increase in falls outside the home; the number needed to harm was 4.5. Sixteen (5%) intervention group participants had one or more fall related fractures compared with 10 (3%) control participants (χ2=1.43, df=1, P=0.23). In the intervention group, fall rates did not differ significantly among participants categorised into quarters of self reported adherence (P=0.60).
Regardless of outdoor Adelaide activities profile subtotals, we found no significant differences between the two groups in fall related usage of general practitioner (P=0.15), outpatient specialist (P=0.83), emergency department (P=0.66), or inpatient hospital resources (P=0.63).
Secondary outcome measures
Table 3 shows the baseline and 13 month re-test scores for physical activity levels, falls efficacy, and quality of life measures for the intervention and control groups. We found no statistically significant between group differences for any of the secondary outcome measures at the end of the intervention after adjustment for baseline scores. No significant differences were evident when we did analyses relating to the secondary outcome measures for the Adelaide activities profile subgroups.
Table 3 Secondary outcome variables at baseline and 13 month re-test. Values are means (SDs)
Two intervention participants and one control participant reported a fall while switching glasses. No intervention participants reported wearing single lens distance glasses when they had a fall related fracture: one reported wearing reading glasses, 12 multifocal glasses, and three no glasses; in two cases what type of glasses were worn, if any, was not clear. In the control group, nine participants reported wearing multifocal glasses and two reported wearing no glasses when a fall related fracture occurred; in one case what type of glasses were worn, if any, was unclear. Seventy-seven (26%) intervention participants had one or more non-fall injuries (laceration, lifting or twisting injury, burn/scald, eye injury, collision, pedestrian injuries) compared with 51 (17%) controls (χ2=6.61,df=1, P=0.01).