Proofs and post-publication care
The proofs of an accepted article are the last point at which an error can be caught before print. They are read against the manuscript, with particular care at the points where errors enter and for drug doses. Around publication, the author settles copies, circulation and media enquiries with the publisher and faces criticism of the work, and an archival author keeps its records.
Reading the proofs
Acceptance leaves three things to consider: proofreading, preventing errors, and what to do after publication. A proof is a copy of typeset work sent to authors or editors to review for errors. Galley proofs, or galleys, are the manuscript set in print but not yet formatted to the journal page, typeset matter without page numbers that the author reviews before it is converted into page proofs. Page proofs present typeset text as it will appear in print, including headings and page numbers. The author may or may not see the marked-up manuscript. Some journals always send galley proofs, while others send them only when they expect the author to want further changes.
A printer's proof shows exactly how the final document will look in print. One guide for clinicians holds that the author always receives proofs to review, and page proofs formatted to the journal page before publication, while another guide says that sending the author a printer's proof is not standard practice at every journal. It has to be checked thoroughly, because there is no going back once it is approved for printing. The proof stage is for correcting errors, not for major changes to the content. Small changes are still possible, but they can be more difficult and expensive, because the document has already been typeset, usually in a fixed format such as a PDF. An author tempted to add new material, such as a study published since submission or a newly introduced drug, should discuss it with the editorial office first, and some editors will approve a sentence or two or a reference. The printing schedule is often tight, leaving hours or days for proofreading, so the author should find out when the proofs will be sent and set the time aside.
Every word of the proofs should be read carefully. The proofs are compared with the edited manuscript to check that every insertion, change and deletion has carried through and, paragraph by paragraph, that every block of text has been retained. Text does get omitted or jumbled. The whole is then read from start to finish for sense and to see that nothing has been left out. Words can change between manuscript and proof, as for becomes of or antenatal becomes antinatal. Tables and figures deserve special attention, because errors like these lurk there. In one published article about medical publications, one of the two panels of a figure carried an incorrect series label. Numbers and totals should be recalculated, and every reference citation should appear in the text with numbers that match the reference list. A proofreading checklist adds the spelling, especially of specialist scientific terminology, the spelling of names and places, the end of each line, where words and figures should be split appropriately, tables that are complete with their rows and columns corresponding, cross-references, the numbering and content of footnotes, the scale, position, colour and content of every image, the page numbering, the headers and footers, and any mention of dates. A galley is checked carefully for errors in formatting, punctuation, terminology, spelling, typescript and headings.
Proofreader's marks are symbols for identifying errors or changes on proofs, used where the author receives a marked-up paper manuscript or galley proofs. They serve for hard-copy galley and page proofs, while digital or online proofs are marked by methods that vary with the program or platform. The marks ask for insertions, deletions, changes of spacing, superscripts and subscripts, straightened lines, new or run-in paragraphs, transpositions, moves left or right, changes of case and of type style, and spelling out, and one mark lets the original stand.
The proofs may carry queries to the author, in which the editor asks for a decision about an issue in the paper. Each must be answered precisely, without hedging, giving both sides or writing a long explanation, and the decision incorporated in the text. Adding a reference at proof stage need not mean renumbering all the others, because some journals let a reference added after reference 45 become 45a in both the text and the reference list. An author adding a reference should ask about numbering. The lettered number saves time and cost and avoids many later numbering errors. The proofs should be returned promptly, because the journal has probably already reserved space in an upcoming issue. A copy of the corrected proofs should be kept, since the version returned can be lost.
Where errors enter
Errors find their way into print in many ways, some beginning with the author, some with copyediting and some for no evident reason, and whatever their origin, it is the author's job to find them. The Lancet once apologised for printing a sour correspondent where it meant as our correspondent, and a JAMA table of contents repeated a phrase of an article's title, an error that was not the authors' fault.
Errors are commonly introduced at particular points in the publishing process, and those points call for particular vigilance. The common sources of error are the author's original writing, the copyediting stage, the typesetting stage, the conversion of the document into another format for desktop publishing software, spell-checking software whose suggested changes are not always correct, automatic text changes by software, a global find-and-replace whose individual changes go unchecked, more than one working version, so that a change reaches one and not the other, and last-minute changes that escape the review the rest of the text received.
Drug doses
Most errata are of minor significance, but errors in drug doses can be dangerous. One handbook of antimicrobial therapy printed the paediatric dose of doxycycline, combined with quinine sulfate, for chloroquine-resistant P. falciparum malaria as 30 mg/kg/d × 7 d, and came with a warning label correcting it to 2 mg/kg/d × 7 d. Such an error is much more worrying than the incorrect fractions of a university's advertisement for a mathematics research assistant in a 3/4 research and 1/3 teaching position. Every drug dose and frequency of administration should be checked very carefully. No dose should be split over two lines, because a line break between a number and its units could lead to dosing errors. At proof stage, the doses should be checked against the original source. It is good practice to have an independent checker confirm them against authoritative texts such as the British National Formulary or MIMS. Information not intended as a prescribing resource may still be used as one. Doses that will be used in a clinical setting warrant a second and even a third check before they are printed, because an error could harm a patient.
Copies, publicity and the media
Reprints, also called offprints, are separately printed copies of individual journal articles, which journals generally offer to authors for a fee. During production, the author may be asked whether they want to buy reprints of the paper. Offprints of articles and copies of whole journal issues usually cost more once the initial print run is over, so an author who needs a substantial number does best to ask for them before printing. In 2007, publishers often supplied authors with some free copies and usually offered further copies at cost or at a discount for professional use, such as distribution to colleagues. A contributing author to a book usually received at least one free copy and a sole author or editor probably around 10, and as a courtesy most journals supplied a dozen or so copies of the journal or article to contributing authors. Digital printing had much reduced the cost of offprints, and most printers could supply them in small numbers on demand. Before the internet, reprints were an important way to disseminate findings. By 2018, one experienced author found reprints going out of fashion and expensive, with journals using them to generate income. By his account, most reprint requests then arrived online and were answered with a PDF of the article. In 2018, JAMA advised its authors that they may acquire reprints and e-prints for personal use, such as classroom teaching or sharing with a research colleague. A century earlier, the editor of Medical Libraries had reported an inborn craving among medical men for reprints of their articles, which he attributed to every writer's wish to give the work the stamp of individuality.
Whether the printer's proofs may be circulated to colleagues before publication depends on the publication's policy. The publisher may require a strict embargo on an article with sensitive or controversial information. The copyright often belongs to the publisher after initial submission, so the author should check before circulating the article to anyone before its official release. Copies issued ahead of release usually carry an embargo, under which whoever receives them agrees to publish nothing on the topic before the official publication date. The embargo gives journalists time to prepare their stories for the launch day.
Before publication, the author should ask the editor or publisher how media enquiries are to be handled. The copyright may have passed to the publisher at acceptance, and the publisher may then have the right to decide whether anything is disclosed before the official date, so in doubt, media or professional enquiries should be referred to the editor. The publisher should usually tell the author about a press release and may ask the author to review it or contribute to it. A press release is a good way for publishers to notify a wide audience of a publication, and can bring publicity and help to generate sales, but for controversial work it can also draw public and media debate. News of a research article whose outcome favours a drug may considerably affect the share price of the company or its competitors. News of a new side effect can lead to alarmist reports in the popular press that could have serious implications for public health, as the publicity given to reports linking the MMR vaccine to autism affected the number of children receiving the vaccine.
Unless the author publishes the work themselves, the publisher's marketing department usually promotes and sells it. Its means include selling advertising space within a journal, promotional point-of-sale material, promotional launches, advertising in related publications and on related websites, catalogues of the publisher's titles, a press release to coincide with a book's launch, a review of the book arranged in another publication, targeted email campaigns, and promoting licences for online viewing of journals published electronically.
Criticism of the published work
Publication exposes the work to readers and reviewers, who may praise it in letters to the editor and in reviews but more often criticise it. Authors of research reports are not exempt: in one case, a printed comment challenged the solution the authors of a journal paper had proposed. Letters that only praise a work are unlikely to be published, because they generate no controversy. A writer has to be able to withstand that criticism, and the only preparation is to check everything carefully and have a colleague review the manuscript before submitting it. Once the work is in print, the writer should be prepared to accept the criticism or the applause, and can take pride in having had the work published and in the critic having been one of its readers.
The record of the work
An archival author is an author who assumes responsibility for maintaining the research and other records related to a published article, in case questions arise later.