Tables

Organise your tables efficiently into a borderless grid of rows and columns. Use horizontal rules sparingly, and no vertical rule for separation.

Each table caption begins with the word `Table' or the abbreviation `Tab.' followed by a number and a full stop (or a colon). The caption should allow the table to be understood by itself and should be placed above the table. Table captions should include a statement of the disease of interest, the number of observations in the study, the country, and details of the study period. For example:

Table 1: Breast cancer in New Zealand midwives, 1986 -- 2001. Frequencies of breast cancer cases and controls by level of alcohol consumption and 5-year age category, and age-specific relative risk estimates with 95% confidence intervals.

Strive for simplicity in organising tables, so that their information can be quickly absorbed. Some journals prefer details in footnotes below the table; others prefer more information in the title and fewer footnotes below the table. Remember that somewhere in your tables you should present some actual data, as opposed to derived statistics, such as the results from mathematical models. A paper that presents only derived results with no distribution of the data robs the reader of essential information. Whereas tables are useful to present detailed data or findings, well drawn figures can convey complicated patterns more effectively. For example, you might use a figure to convey the pattern of the relation between key study variables, especially if the pattern is derived from a locally smoothed function that cannot easily be presented in a table.

The ordering of tables must follow the order in which they are mentioned in the text.
Every table must be cited, by number, at least once within the text.
Each table caption begins with the word ‘Table’ followed by a number and a full stop (or a colon).
Align decimal points if it is the intention to make comparisons between numbers in a table. If there is no intention to compare numbers, left align the data. Left justification is preferable for non-numeric data.
If using superscripts to explain various aspects of the data within the table, use letters and sequence them horizontally down the table.
Assuming the type script is double-spaced, all table captions should be 1.5 spaced in order to distinguish them from the body of the text.

Table 1 is an example of appropriately presented tabular data.

Variable Subjects Failed Coefficient (SE) P Hazard (95% CI)

Prison: < 0.01 ^a

Prison absent 127 81 - 1.00

Prison present 111 69 0.3760 (0.1689) 1.46 (1.05 - 2.03) ^b

Dose: 238 150 -0.0350 (0.0064) < 0.01 0.97 (0.95 - 0.98)

^a Significance of the prison variable in the model.
^b Interpretation: compared with the reference category (patients without a prison record), after adjusting for the effect of maximum daily methadone dose, patients with a prison record had 1.46 (95\% CI 1.05 -- 2.03) times the daily hazard of relapse.
SE: standard error.
CI: confidence interval.

Table 1: Retention in two methadone treatment clinics for Australian heroin addicts, 1991 -- 1992. Stratified Cox proportional hazards regression model showing the effect of prison status and methadone dose on the daily hazard of relapse.

Variable	Subjects	Failed	Coefficient (SE)	P	Hazard (95% CI)
Prison:				< 0.01 ^a
Prison absent	127	81	-		1.00
Prison present	111	69	0.3760 (0.1689)		1.46 (1.05 - 2.03) ^b
Dose:	238	150	-0.0350 (0.0064)	< 0.01	0.97 (0.95 - 0.98)