Online dissemination of Cochrane reviews on digital health technologies: a cross-sectional study

Table 2 Association between dissemination (Altmetric scores) and review characteristics

Variables in binary logistic regression	Variable definition	Variable coding	Odds ratio [95% confidence interval]
Dependent	Dissemination	0 = lower Altmetric scores in < top 5% of all research outputs; 1 = higher Altmetric scores in top 5% of all research outputs	-
Independent: Bibliographic characteristics	Publication year	2005–2023	0.83 [0.70–0.99]*
Independent: Bibliographic characteristics	PLS languages	3–14	1.28 [1.05–1.56]*
Independent: Evidence quality characteristics	Studies in review	0 = few (1–10 studies); 1 = many (11–132 studies)	1.69 [0.59–4.79]
	Meta-analysis performed in review	0 = no; 1 = yes	1.24 [0.40–3.88]
	Certainty of evidence	0 = low (review with no outcomes rated as strong or moderate); 1 = high (review with at least one outcome rated as strong or moderate)	1.56 [0.60–4.06]

Note: Based on visual inspection, variables with a skewed distribution of scores were dichotomised using median values. Other review characteristics were not included as independent variables due to the potential overlap in categories (e.g. ‘any digital interventions’ partially overlapped with ‘mobile interventions’ or ‘non-mobile interventions’). A total of n = 87 reviews were included in the binary logistic regression analysis because certainty of evidence ratings was not reported in 13 reviews. Reference categories were coded as 0. Overall model details are as follows: chi-square = 12.8, df = 5, p = 0.025. *p < 0.05

ISSN: 2046-4053