Total N (%) | Spatial N1 | Non-spatial N2 | |
Number of time points | |||
1 | 21 (75.0) | 7 | 14 |
2 | 4 (14.3) | 0 | 4 |
≥3 | 3 (10.7) | 1 | 2 |
Whether spatial distribution mentioned | |||
Yes | 15 (53.6) | 6 | 9 |
No (but sector-wise measurements mentioned) | 13 (46.4) | 2 | 11 |
Aim of investigating spatial data | |||
To identify clinical risk factors of a retina-related disease | 10 (35.7) | 1 | 9 |
To identify the topographic profile of one risk factor and its relationship with other risk factors | 8 (32.1) | 2 | 6 |
To investigate progression/detection/diagnosis of a retina-related disease | 7 (25.0) | 5 | 2 |
To establish normative data | 2 (7.1) | 0 | 2 |
To propose a spatial statistical approach | 1 (3.6) | 1 | 0 |
Clinical variables included | 19 (67.9) | 6 | 13 |
Method used | |||
Statistical hypothesis tests | 18 (64.3) | 0 | 18 |
Linear/GLM | 3 (10.7) | 2 | 1 |
Linear/generalised linear mixed-effect model | 3 (10.7) | 2 | 1 |
Generalised estimating equation | 3 (10.7) | 2 | 1 |
Empirical estimates | 2 (7.1) | 1 | 1 |
Point process model | 1 (3.6) | 1 | 0 |
Spatial correlation considered in analysis | 12 (42.9) | 5 | 7 |
Correlation of one variable between different locations | 7 (25.0) | 3 | 4 |
Correlation between two variables in separate locations | 5 (17.9) | 2 | 3 |
How spatial information/distribution reported | |||
Line graph | 12 (42.9) | 5 | 7 |
Table | 10 (35.7) | 0 | 10 |
Colour intensity map | 9 (32.1) | 4 | 5 |
Grayscale intensity map | 5 (17.9) | 2 | 3 |
Contour plot | 1 (3.6) | 0 | 1 |
Not reported | 3 (10.7) | 1 | 2 |
Software | |||
SPSS | 13 (46.4) | 3 | 10 |
R | 6 (21.4) | 5 | 1 |
R (nlme)79 | 2 (7.1) | 2 | 0 |
R (spatstat)80 | 1 (3.6) | 1 | 0 |
R (unspecified packages) | 3 (10.7) | 2 | 1 |
STATA (no available code) | 3 (10.7) | 1 | 2 |
MedCalc | 3 (10.7) | 0 | 3 |
MATLAB (no available code) | 2 (7.1) | 1 | 1 |
SAS (GENMOD)81 | 1 (3.6) | 1 | 0 |
Unclear | 5 (17.9) | 0 | 5 |
Note for ‘Method used’, ‘How spatial information/distribution reported’ and ‘Software’ more than one method, way of reporting or software could be recorded per included study giving total >28.
GLM, general linear model.