Category: Literature Review

Write your search strategy as a reproducible record, not a narrative. It needs five things: a structured review question (usually PICO), the databases you searched with the dates you searched them, the exact search string including Boolean operators and truncation symbols, your inclusion and exclusion criteria with a justification for each, and a PRISMA flow diagram accounting for every record from initial hits to final included studies. If a second person could rerun your search and land on roughly the same papers, the section has done its job.

Most UK undergraduate nursing dissertations are literature-based rather than empirical, and the search strategy is where a mid-2:1 separates from a first. Markers cannot assess the quality of a synthesis if they cannot see how the papers arrived. This guide walks the section in the order you should write it, with the level of detail a Faculty of Health and Social Care marking rubric actually expects.

Why does the search strategy carry so many marks?

In an empirical dissertation, the methodology chapter explains how you produced data. In a literature-based nursing dissertation, the search strategy is your methodology chapter. It is the only evidence a marker has that your sample of studies was assembled systematically rather than by convenience.

Marking criteria across UK nursing programmes typically award credit under headings such as “rigour of the review process” or “systematic approach to literature retrieval”. The predictable failure is a paragraph reading: “A search was conducted using CINAHL and Medline using keywords relating to the topic.” That sentence is unrepeatable, unverifiable, and reads as though the papers were chosen because they were convenient. If you are still deciding whether your project is a narrative review or something more formal, settle the question of whether you need a systematic review or a literature review before you write a word of this section, because the two demand different levels of procedural detail.

How do you turn your topic into a PICO question?

PICO forces a vague clinical interest into a searchable structure. For most nursing dissertations the four elements are:

• P — Population: who, specifically. Not “patients” but “adults aged 65 and over admitted to acute medical wards”.
• I — Intervention: the thing being done. “Intentional rounding”, “early mobilisation”, “structured handover using SBAR”.
• C — Comparison: usual care, an alternative intervention, or nothing. Comparison is legitimately absent in qualitative reviews.
• O — Outcome: what changes. “Incidence of hospital-acquired pressure ulcers”, “patient-reported anxiety”, “call-bell frequency”.

A worked example: In adults aged 65 and over on acute medical wards (P), does hourly intentional rounding (I), compared with usual nursing observation (C), reduce the incidence of in-hospital falls (O)?

If your review is qualitative — nurses’ experiences of something, patients’ perceptions of something — PICO fits awkwardly. Use SPIDER instead (Sample, Phenomenon of Interest, Design, Evaluation, Research type), and say in the text that you selected SPIDER because it is better suited to retrieving qualitative evidence. Naming the framework and justifying the choice is itself worth marks.

Which databases should a UK nursing student search?

Search a minimum of three, and say why you chose each. The defensible core for nursing:

• CINAHL Complete — the primary nursing and allied health database. Non-negotiable for a nursing dissertation, and its absence will be noticed.
• Medline (via Ovid or EBSCO) — biomedical coverage, indexed with MeSH terms.
• British Nursing Index (BNI) — UK-focused nursing literature, useful when your question has NHS-specific relevance.
• PsycINFO — add it if your outcome is psychological: anxiety, coping, wellbeing, mental health nursing.
• Cochrane Library — search it for existing systematic reviews. If a Cochrane review already answers your exact question, you need to narrow your angle, and finding this out in week two is far better than in week ten.

Google Scholar is a supplementary tool, not a database in this sense. You may use it for citation chaining, but declare it as such rather than listing it alongside CINAHL. For a broader map of the platforms available and how their interfaces differ, see the comparison of databases worth searching for dissertation literature.

State the interface too. “CINAHL Complete via EBSCOhost, searched 14 October 2026” is what reproducibility looks like. Record the date of every search — results change as records are added, and an undated search is not repeatable.

How do you build the actual search string?

Take each PICO element, generate synonyms, join synonyms with OR, then join the blocks with AND. Add truncation and phrase searching.

Working from the falls example:

• Population block: (older adult* OR elderly OR geriatric* OR “aged 65”)
• Intervention block: (“intentional rounding” OR “hourly rounding” OR “comfort round*” OR “nursing round*”)
• Outcome block: (fall* OR “falls prevention” OR “patient safety incident*”)

Combined: (older adult* OR elderly OR geriatric*) AND ("intentional rounding" OR "hourly rounding" OR "nursing round*") AND (fall* OR "falls prevention")

Three mechanics to get right, because markers check them:

1. Truncation. The asterisk retrieves word variants. nurs* catches nurse, nurses, nursing. Be careful with short stems — car* returns care, caring, carer, but also cardiac, carcinoma and carbohydrate.
2. Phrase searching. Quotation marks keep multi-word concepts intact. Without them, intentional rounding may be read as two separate terms and your precision collapses.
3. Subject headings. CINAHL headings and Medline MeSH terms are the controlled vocabulary applied by indexers. Searching MH "Accidental Falls" alongside your free-text terms catches papers whose authors used wording you did not think of. Combining subject headings with free-text terms is the single clearest signal of a competent search, and it is worth a sentence in your text explaining that you did it.

Present the search in a table with one row per database: database, interface, date searched, full string as entered, and number of hits. Put the table in the chapter itself, not buried in an appendix — markers should not have to hunt for your central evidence.

What inclusion and exclusion criteria are defensible?

Every criterion needs a reason attached. A bare list looks arbitrary; a justified list looks methodological.

Two honesty points. “English only” is a real limitation — write it into your limitations section rather than hoping nobody notices. And do not exclude on geography as narrowly as “UK only” unless your question is genuinely NHS-specific; you will usually strand yourself with three papers and no review.

How do you build the PRISMA flow diagram?

PRISMA 2020 gives you a four-stage diagram, and every number in it must reconcile. The stages:

1. Identification: total records retrieved from each database, plus records identified through other methods such as reference-list searching. State each database’s contribution separately, then the total.
2. Screening: duplicates removed, then titles and abstracts screened. Record how many were excluded here.
3. Eligibility: full texts assessed. This is the stage students most often get wrong — you must give reasons for full-text exclusions with counts, for example “wrong population (n = 7), wrong outcome (n = 4), not primary research (n = 3)”.
4. Included: the final studies carried into your synthesis.

A worked reconciliation: 412 records identified (CINAHL 187, Medline 148, BNI 61, reference chaining 16). 68 duplicates removed, leaving 344 screened on title and abstract. 316 excluded, leaving 28 full texts assessed. 20 excluded with reasons stated. 8 studies included. The arithmetic must work in both directions — markers do check, and a diagram whose numbers do not add up undermines everything downstream of it.

Eight to twelve included studies is a normal, healthy yield for an undergraduate nursing dissertation. Three suggests your criteria were too tight. Forty suggests your question was too broad and you have more reading than the timetable allows.

What comes immediately after the search strategy?

Two things, and both belong in the same chapter.

First, quality appraisal. Having found your studies you must show you judged them rather than accepted them. In UK nursing dissertations that almost always means a CASP checklist matched to each study’s design — a different checklist for qualitative studies, cohort studies and randomised trials. If you have not chosen an appraisal tool yet, work through what a CASP checklist is and whether your nursing dissertation needs one before drafting your findings.

Second, your synthesis approach. Name the method. For a qualitative or mixed set of papers this is usually thematic synthesis, and you should describe how codes became themes across studies rather than merely asserting that themes “emerged”. The mechanics transfer directly from primary research — the sequence in the six phases of thematic analysis applies whether your data are interview transcripts or findings sections of published papers.

Do you need ethics approval for a literature-based nursing dissertation?

Almost never — but you still need to say so explicitly. A review of already-published literature involves no human participants and no identifiable data, so most UK schools of nursing either exempt it or require a short self-declaration form rather than full committee review. Write one sentence stating that ethical approval was not required because the study involved secondary analysis of published literature, and reference your school’s policy. The rules shift the moment your project touches patients, staff or NHS data, which is covered in detail in the guide to ethics approval for an undergraduate dissertation.

Five errors that cost marks in this section

1. Reporting hits without the string. “CINAHL returned 187 results” proves nothing without the search that produced them.
2. Free-text terms only. Ignoring CINAHL headings and MeSH is the most common reason a search misses obvious papers.
3. A PRISMA diagram that does not reconcile. Numbers that fail to add up read as fabricated even when they are simply careless.
4. No reasons for full-text exclusions. PRISMA 2020 requires them; omitting them is a straightforward compliance failure.
5. Undated searches. Without dates the search is not reproducible, which defeats the section’s entire purpose.

A realistic order of work

Draft the section in this sequence and it assembles itself: finalise the PICO question, build and test the string in CINAHL first, adjust once you see the hit count, replicate across your other databases, export everything to your reference manager and deduplicate there, screen titles and abstracts, screen full texts recording every exclusion reason as you go, then draw the PRISMA diagram from your own running tally rather than reconstructing it from memory afterwards.

Keep a search log from day one — a simple spreadsheet with database, date, string, hits. Students who reconstruct their search a fortnight before submission always lose numbers, and the diagram never reconciles. Because you will already be tracking references at this stage, having a realistic sense of how many references a dissertation needs helps you judge whether your yield is on target before you commit to the synthesis.

Get the section drafted while the search is fresh

The search strategy is the most mechanical chapter in a nursing dissertation, which makes it the one most worth drafting immediately rather than leaving to the end. Tesify can turn your PICO question, database list and screening numbers into a structured first draft of the section, formatted with the search table and PRISMA narrative in place, so you spend your remaining time on appraisal and synthesis rather than on prose scaffolding. The searching, screening and judgement stay yours — the drafting stops being the bottleneck.

Frequently asked questions