e-Health interventions for healthy aging: a systematic review

Table 2 Certainty of the evidence

Outcome	Number of studies	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Certainty
Physical activity	11	Serious^a	Not serious^b	Serious^d	Serious^f	Serious^g	Very low
Healthy eating	2	Serious^a	Serious^c	Serious^d	Serious^f	Serious^g	Very low
Clinical parameters (body mass index, HbAIc, cholesterol)	5	Serious^a	Serious^c	Serious^d	Serious^f	Serious^g	Very low
Quality of life	3	Serious^a	Not serious^b	Not serious^e	Serious^f	Serious^g	Very low
Cognitive outcomes	2	Serious^a	Serious^c	Serious^d	Serious^f	Serious^g	Very low
Psychological outcomes (wellbeing, depression, loneliness)	4	Serious^a	Serious^c	Serious^d	Serious^f	Serious^g	Very low
Social outcomes (social support, social functioning)	2	Serious^a	Serious^c	Not serious^d	Serious^f	Serious^g	Very low

^aMost trials had unclear or high risk of bias in one or more domains, among which the lack of blinding of participants, rendering it necessary to downgrade the level of evidence
^bAlthough there was some variation in the direction of effect, we did not downgrade the level of evidence since most outcomes showed a positive trend of the effectiveness of eHealth interventions
^cFor these outcomes, there was important heterogeneity in the measures used across studies
^dMost studies used surrogate outcome measures, among which self-reported measures of physical activity
^eQuality of life was assessed directly using gold standard measures
^fWhen confidence intervals were available, they were usually large. Also, most studies had modest sample sizes
^gGiven the limited number of included studies, we did not compute a funnel plot to check for publication bias, but it is likely that such bias is present given that many studies have a modest sample size

ISSN: 2046-4053