Your journal wants Vancouver and your table is in APA: what actually changes

You wrote the manuscript in APA because that is what your degree taught you, and the nursing journal you are targeting wants Vancouver. The normal reaction is to search for “Vancouver table rules” and find nothing definitive, because that manual does not exist in the shape you are imagining: the ICMJE governs numbered references and gives a handful of table rules, and every other decision belongs to the journal. Here is what actually changes, figure by figure, and what you can leave exactly as it is. If you are starting from clean APA formatting, the ground floor is how to report results in APA 7.

Vancouver is a citation style, not a table style

What we call Vancouver is the set of recommendations issued by the ICMJE, with the National Library of Medicine manual behind it for bibliographic detail. Its core is the numbered citation: sources are numbered in order of first appearance, the reference list follows that order rather than the alphabet, and journal names are abbreviated. APA does the opposite on all three counts: author-date citation, alphabetical list, journal names spelled out in full. That is where 90% of the conversion work sits, and your reference manager does it in two clicks.

On tables the ICMJE says little and says it precisely: number them in order of first citation in the text, give each a brief title, use no internal horizontal or vertical rules, keep column headings short, put explanatory matter in footnotes rather than in the heading, explain non-standard abbreviations there, use footnote symbols in the order *, †, ‡, §, ||, ¶, and identify the measure of variation you are reporting. Everything else — italics, decimals, capitalization, how a mean is written — is not Vancouver. It is house style, and in the health sciences the most common house is the AMA manual.

The practical consequence is that your source of truth is that journal's instructions for authors, and your second source is one of its papers from last year. Download one, look at its Table 1 and copy the shape. It is faster and more reliable than any general guide, this one included. It also settles the awkward cases quickly: a fair number of nursing journals published by the big houses still use APA even though the field around them is biomedical, so never assume the style from the discipline.

What changes in the table: title, notes and symbols

The title. APA 7 uses two lines: a bold “Table 1” on top and an italicized title underneath. A biomedical journal wants one plain line: “Table 1. Baseline characteristics of the sample”. This looks cosmetic until the manuscript comes back unreviewed for “table formatting”, which is one of the most avoidable desk rejections there is.

The notes. APA has three tiers in a fixed order: a general note beginning “Note.” for anything the whole table needs, specific notes flagged with superscript letters (a, b, c) for one cell or one column, and a probability note with asterisks. The ICMJE has no tiers, only a sequence: *, †, ‡, §, ||, ¶, where the asterisk is simply footnote one. Leave your “* p < .05. ** p < .01” untouched and that asterisk collides with your first real footnote, so the reader cannot tell whether you are flagging significance or warning that two participants did not answer. Converting is easy and improves the table anyway: delete the asterisks and add a P column with exact values.

The rules. The two schools agree more than you would expect: neither allows vertical rules. The difference is that APA does draw a horizontal rule under the column headings and one at the foot, whereas the ICMJE asks for no internal rules at all and leaves the typesetting to the journal. If your table came straight out of SPSS with its full grid, you have to strip it either way.

Where the table lives. APA 7 lets you embed each table in the text right after the paragraph that calls it, or place them all after the references. Many biomedical journals do not give you the choice: one table per page after the references, or each table as a separate file, with figures always uploaded separately. Check this before you lay the document out, because rebuilding it at ten o'clock on submission night is the expensive way to learn the rule.

What changes in the numbers: leading zeros, capital P and mean (SD)

The leading zero. APA forbids it for any statistic that cannot exceed 1: p = .001, r = .34, β = .21, α = .89. Biomedical practice has no single rule, and this is where most authors get stuck. Journals following the AMA manual — JAMA and its family — write a capital italic P and drop the zero as well: P = .001. Plenty of British and European nursing and public health journals do the opposite and write p = 0.001 with the zero in place. There is no way to deduce it; you look it up in a recent paper from the journal.

The means. In APA the mean is M and the standard deviation is SD, both italicized, in the text or inside the table. In a biomedical table those letters disappear: the label moves to the column heading, “Mean (SD)”, and the cell contains only “3.42 (0.79)”. When the variable is skewed you switch to “Median (IQR)”, and categorical variables are written as n (%): 61 (28.5) out of a sample of 214. The ICMJE explicitly asks you to identify the measure of variation, because “3.42 (0.79)” on its own does not say whether that 0.79 is a standard deviation or a standard error, and the two readings are not remotely the same. The canonical footnote is one line: “Values are mean (SD) unless otherwise stated”.

The decimals. APA asks for two decimal places for most statistics and exact p values to two or three, with p < .001 below that floor. The biomedical convention is the same idea in different clothes: percentages to one decimal when the sample supports it and no decimals at all when you have thirty people (with n = 28 the decimal is false precision: 8 cases are 28.6% and there is nothing between that and the 25.0% of 7 cases), and exact P values unless they are very small. What neither style tolerates is more precision than your data can carry.

What changes in the Results sentence

The APA sentence you already wrote probably reads like this: “Nurses working night shifts reported higher emotional exhaustion than those on day shifts, t(212) = 3.25, p = .001, d = 0.44.” Italicized statistic, degrees of freedom in brackets tight to the letter, effect size at the end. A biomedical journal wants the same sentence built around the estimate and its uncertainty: “Emotional exhaustion was higher on night shifts than on day shifts (mean 3.42 [SD 0.79] vs 3.06 [0.83]); mean difference 0.36 (95% CI 0.14 to 0.58); P = 0.001.” Same data, same analysis, different furniture.

The substantive change underneath is that the ICMJE asks authors not to rely solely on hypothesis testing, because a P value carries no information about the size of an effect or the precision of an estimate, and it asks you to quantify findings with indicators of uncertainty — in practice, confidence intervals. It also asks you to name the statistical software and its version, which psychology manuscripts forget about half the time. If your Results section is a string of p values, your problem is not the style sheet; no amount of reformatting will fix it.

The same pattern repeats across tests. A chi-square you would write in APA as χ²(1, N = 214) = 5.41, p = .020 — with the N inside the brackets, which is APA's one exception — usually appears in a biomedical table as the n (%) per category plus a P column. A regression that reads B = 0.24, SE = 0.08, β = .21, t(209) = 3.00, p = .003 in APA becomes B = 0.24 (95% CI 0.08 to 0.40), P = 0.003, often with no standardized beta at all: clinical readers want the effect in units they can act on — scale points, days of stay — rather than in standard deviations. The APA starting point is set out in how to report a regression in APA 7.

The Table 1 that psychology never has to write

In nursing and health research, Table 1 is nearly always the same object: participant characteristics by group, with n (%) for categorical variables and mean (SD) or median (IQR) for continuous ones. This is not a house habit. The STROBE checklist, which most journals ask you to complete for an observational study, requires you to describe participant characteristics and, under descriptive data, to report the number of participants with missing data for each variable. That missing-data count almost never appears in a psychology table, and it is the first thing a health reviewer looks for.

The P column in that table deserves a warning. In a randomized trial, CONSORT advises against significance testing of baseline differences: if allocation was random, any imbalance is by definition due to chance and the P value answers nothing. In observational work it is common, though still criticized. The working rule is simple — if the journal does not ask for it, describe the imbalance and adjust for it in the model rather than testing it.

Your submission also carries the reporting checklist for the design in hand — STROBE for observational studies, CONSORT for trials, PRISMA for reviews, COREQ for qualitative nursing work — and the flow diagram where one applies. In psychology this is required by some journals; in health it is close to universal. Filling the checklist in before you write, rather than after, saves you discovering at submission that you never recorded something you can no longer go back and collect.

How to convert the table without retyping it

Converting a twelve-row regression table by hand takes twenty minutes and, almost always, produces two typos: one leading zero you forgot, one interval still wearing square brackets. That is why the three table generators on this site carry an APA 7 / Vancouver switch that reformats the same output: the regression one, the correlation one and the mediation one. You paste your SPSS or R output once and pick the style at the end.

What the switch changes is exactly what this article has been describing: the leading zero returns in β, in R² and in the p column, lowercase p becomes capital P, the title loses its italics and joins the number after a full stop, the confidence interval swaps square brackets for parentheses, ΔR² is spelled out as “R² change”, the sentence separator moves from comma to semicolon, and the note drops its leading “Note.”. You also get the Results paragraph drafted in the style you chose, and the table downloads as a .docx so it lands in the manuscript already formatted.

What must not change is the numbers. Not the estimate, not the degrees of freedom, not the interval limits. If a figure moves when you change style, you have not changed the style — you have changed the analysis, and it is time to go back to the original software output. Keep that output filed next to the manuscript, because it is what you will be asked for when a reviewer doubts a table. A full worked Results section, tables and text together, is in this APA 7 Results section example.

Frequently asked questions

What is the difference between APA and Vancouver tables?

Vancouver governs few things in the table: numbering by order of mention, a brief title, no internal rules, short column headings, footnotes using the symbols *, †, ‡, §, ||, ¶, and a statement of the measure of variation. APA governs much more: bold table number above, italicized title below, three tiers of notes (general, specific with superscript letters, probability with asterisks) and italicized statistics without leading zeros. The other differences you will meet — leading zeros, capital P, mean (SD) — are journal house style, not Vancouver.

Do p values have a leading zero in Vancouver style?

It depends on the journal, not on Vancouver. Journals following the AMA manual use a capital italic P with no leading zero (P = .001), while many British and European nursing and public health journals write p = 0.001 with the zero. APA is the only unambiguous case: never a leading zero, because p cannot exceed 1. Check a recent paper from your target journal before you decide.

Is the table title italicized in Vancouver style?

No. APA 7 puts an italicized title on its own line under a bold “Table 1”. The usual biomedical format is a single unitalicised line with the number attached: “Table 1. Baseline characteristics of the sample”. It is one of the most visible differences and one of the fastest for an editor to spot.

Can I use asterisks to flag significance in a Vancouver journal?

It is a bad idea. In the ICMJE footnote order the asterisk is the first footnote symbol, so a significance asterisk gets confused with an ordinary note. Replace the asterisks with a column of exact P values: that is what the reviewer expects, and it also satisfies the recommendation to report the figure rather than a threshold.

Where do tables go in a Vancouver manuscript?

Most often at the end of the document, one table per page after the references, or each table as a separate file, with figures uploaded separately and their legends listed. APA 7 is more flexible and allows tables embedded in the text. As with everything else here, the journal's instructions for authors win.

Do I have to redo my tables when I resubmit to another journal?

The formatting only, never the content. The title, the notes, the leading zeros and the way means are written all change; estimates, degrees of freedom and confidence intervals do not. Keep one master version holding the numbers and generate each format from it, and switching journals takes five minutes instead of an afternoon.

If you are preparing a submission and would rather have someone go through the tables and the Results section before the editor does, that is part of what I do in statistical consulting: checking the analysis, leaving the tables in the exact format the journal asks for, and helping you answer the reviewers when they arrive.