Your Systematic Review Search Is Six Months Old: Running an Update Search Before Submission

You ran your searches in March. It is now September, the submission portal closes Friday, and your supervisor has just written one line in the margin of your draft: "When was this search last run?" That question is not administrative housekeeping. A search is a claim about what the literature contained at a moment in time, and if that moment has passed, your conclusions may rest on an incomplete record.

The fix is not to start over. It is to run a structured update search, deduplicate the new records against the library you already screened, screen only what is genuinely new, and then decide — explicitly — whether your conclusions still hold. This article walks through that repair using a concrete student case, then ends with the submission decision you actually have to make.

Imagine a master's student, Priya, writing a systematic-style review for a research methods course. Her rubric awards marks for a documented search strategy, a screening log, and a discussion that reflects the current state of the evidence. Her search was planned from her eligibility criteria, run across multiple sources, and combined free-text terms with subject headings — the approach the Cochrane Handbook describes in its guidance on searching for and selecting studies. She documented the process, deduplicated records, and screened in stages.

All of that was correct in March. By September, three things are true at once:

1. New papers have almost certainly been published in her topic area.
2. Her library still contains the March record set, already screened and coded.
3. Her conclusions were written against that March set.

The Cochrane Handbook explicitly recommends rerunning searches before publication. That recommendation is the anchor for everything below: an update search is a planned step in the review process, not an admission that the first search was flawed.

Before touching a search box, write down the question the update search is meant to settle. For Priya, it is narrow: Have any eligible studies appeared since the March search date that would change how I describe the evidence or the strength of my conclusion?

That framing produces three practical decisions:

• Keep the eligibility criteria fixed. Changing inclusion criteria mid-update turns a refresh into a new review. If the criteria must change, say so and treat it as a protocol amendment.
• Record the exact cut-off. The update window runs from the date of the last search to the date of the new one. Both dates belong in your methods section.
• Decide in advance what counts as a conclusion-changing finding. A single small study with a consistent direction is different from a large study that contradicts your main result. Naming this threshold before you screen protects you from motivated reasoning.

Rerunning means reusing the strategy, not improvising a new one. The Cochrane Handbook's guidance on planning searches from eligibility criteria, using multiple sources, and combining free-text terms with subject headings is what makes a rerun comparable to the original.

If your original search used a database with field tags, filters, and a search-details view that shows how terms were translated and combined — as PubMed's user guide describes — reuse those exact strings and filters, changing only the date limit. That search-details view is your audit trail: it lets you confirm the rerun matches the original rather than drifting.

If your original strategy included citation chasing, the update should too. Google Scholar's documented features — searching by author and title, date limits and date sorting, Cited by and Related articles, and email alerts — support exactly this kind of forward and backward citation check on your seed papers. Set an alert now so the next update is smaller.

For the rerun itself, you can execute the search through WisPaper search, which gives you a scholarly search interface for finding academic papers. If you want a public entry point that explains the search capability before you begin, the WisPaper Scholar Search page serves that purpose. Keep your original strategy document open beside you and paste the same terms.

The most common waste of a deadline week is re-screening papers you already excluded in March. Deduplication comes first, and it has two layers.

Layer one: record-level deduplication. The new export will contain records already in your March set, because databases index the same articles. Match on title, authors, year, and DOI. Cochrane's guidance treats deduplication as a standard step before screening, and it applies just as much to an update as to a first search.

Layer two: study-level linking. Cochrane also advises linking multiple reports of one study. A follow-up analysis of a trial you already included is not a new study — it is a second report. If you count it as new evidence, you will double-count the same participants.

This is where a working library earns its keep. If your March records live in My Library, you can save or upload papers and organize them there, which gives you a single place to compare the new export against the old set rather than juggling spreadsheets. Build a simple three-column log as you go: new record, already screened (with reason), possible duplicate of included study.

Once duplicates are removed, screen in the same stages you used originally: title and abstract first, then full text. Cochrane's staged screening model exists because it is faster and more consistent than reading everything in full.

For Priya, the new export after deduplication contains a manageable set. She screens titles and abstracts against her fixed criteria, then retrieves full text for the survivors. Two practical habits make this stage defensible under deadline:

• Screen in one sitting per stage. Consistency drops when screening is interleaved with other tasks.
• Log exclusion reasons at the full-text stage. "Did not meet population criterion" is a record; "didn't seem relevant" is not.

If you need to interrogate the papers you have already added to your library — for example, to check how an included study defined its outcome before deciding whether a new record is comparable — Library QA lets you ask questions based on papers uploaded to My Library or added from search results. Use it to locate the relevant outcome definition, then record the decision from the new full text in the screening log.

Your methods section needs to show the update happened and what it produced. A compact, honest record includes:

1. The original search date and the update search date.
2. The sources rerun and confirmation that the strategy was unchanged (or a note on what changed and why).
3. The number of records retrieved in the update.
4. The number removed as duplicates, with the split between record-level duplicates and additional reports of included studies.
5. The number screened at title/abstract and at full text, with exclusion reasons.
6. The number of new studies included.

This is the same documentation discipline Cochrane recommends for the original search, applied to the refresh. It also answers your supervisor's margin note directly, which is worth more than a paragraph of reassurance.

This is the step students skip, and it is the one the rubric is really testing. After screening, you have one of three outcomes, and each demands a different response.

No new eligible studies. Your conclusions stand. Add one sentence to the methods noting the update date and that no new eligible records were identified. Do not silently leave the search date in the past.

New studies that are consistent with your existing evidence. Your conclusions stand, but the strength of the statement may improve. Update the synthesis and note the addition.

New studies that contradict or substantially qualify your conclusion. Now you revise. This is not a failure — it is the update search doing its job before a reviewer finds the paper you missed.

Write the decision down even when nothing changed. A short "update search log" appended to your project files turns a stressful question into a documented answer. If you are managing several pieces of a larger project, the Research Projects area is where you can create and keep that work organized.

Priya's revised submission now contains a methods paragraph stating the original and update search dates, the sources rerun, and the deduplication and screening counts. Her results section includes one newly identified study that is consistent with her existing synthesis. Her discussion notes that the conclusion is unchanged but now rests on a slightly larger evidence base. Her appendix holds the update log.

The deliverable is not a longer paper. It is the same paper with a defensible, dated evidence claim — which is exactly what the rubric's search-strategy and currency criteria reward.

Here is the concrete call to make. If the update search is complete, deduplicated, screened, and logged, submit. If the update search is complete but you have not yet written the methods sentence and the conclusion decision, spend the time on that before submitting — an undocumented update is nearly as vulnerable as no update. If the update search surfaced a contradicting study you have not yet integrated, do not submit the current draft; revise the synthesis first, even if it means using an extension or a late-submission allowance.

The transferable lesson is simple: treat the search date as a variable in your manuscript, not a fixed fact. Set an alert, diarize a rerun before every submission, and keep your library deduplicated so the next update takes an afternoon instead of a week.