How to Find Reliable Sources for a Nursing Assignment

What "reliable" means

A reliable source is one where you can verify who produced it, how they produced it, what evidence it rests on, and whether it is current enough to inform practice. Peer-reviewed status helps, but it does not guarantee quality. Peer-reviewed journals publish weak studies, and some non-peer-reviewed documents, such as national clinical guidelines, are more rigorous than a single published trial.

Judge every source on five criteria:

A source that fails on method or authority is not rescued by being recent or relevant.

Common mistakes that cost marks

1. Starting with Google. Search engines rank by popularity and optimisation, not quality. You will spend an hour sorting out material your library database would have filtered in minutes.
2. Over-relying on textbooks. They are acceptable for background and definitions. They lag behind research by years, and markers expect primary research and current guidelines in the body of the essay.
3. Using health websites as evidence. Commercial health sites, blogs, and patient forums are not academic sources. Use them only to understand how patients receive information, and only if the assignment calls for it.
4. Citing a source you only read the abstract of. Abstracts omit limitations, sample details, and sometimes conflicting results. Read the full text of anything you cite.
5. Citing secondary sources as if they were primary. If Jones (2019) reports a finding from Lee (2015), find and read Lee.
6. Ignoring the date. A 2008 guideline on pressure injury prevention may have been replaced twice.
7. Collecting too many sources and reading none properly. Four to six well-appraised sources beat fifteen skimmed ones.
8. Trusting AI-generated references. Covered in section 8.

Turn the assignment into a searchable question

Searching without a defined question returns noise. Before opening a database, write your topic as a structured question.

For clinical or effectiveness questions, use PICO:

• Population: who are the patients?
• Intervention: what is being done?
• Comparison: compared with what?
• Outcome: what result matters?

Example: In adults over 65 on medical wards (P), does two-hourly repositioning (I) compared with four-hourly repositioning (C) reduce the incidence of pressure injuries (O)?

For experience or perception questions, use PICo:

• Population
• Phenomenon of Interest
• Context

Example: What are the experiences (I) of newly qualified nurses (P) working night shifts in rural hospitals (Co)?

If your assignment is a broad topic with no question, narrow it first. "Diabetes" is not searchable. "Nurse-led education to improve HbA1c in adults with type 2 diabetes in primary care" is.

Then build a keyword table. For each concept, list synonyms and alternative spellings, because authors do not use the same terms.

Missing synonyms is the most common reason students conclude "there's no research on this."

Step 2: Search the right places

Core databases for nursing

Use at least two databases. Searching only one will miss relevant studies, and your lecturer will know it.

Guidelines and official documents

For any assignment on clinical practice, include at least one current guideline or policy from a recognised body. Choose based on your country and topic:

• UK: NICE, SIGN, NHS England, Royal College of Nursing, NMC
• US: CDC, AHRQ, American Nurses Association, specialty bodies such as the American Heart Association
• International: World Health Organization, International Council of Nurses, Registered Nurses' Association of Ontario (RNAO) Best Practice Guidelines
• India: Ministry of Health and Family Welfare, ICMR, Indian Nursing Council, state health department protocols

Guidelines differ in rigour. Check what evidence the guideline cites and how recommendations were graded. A guideline built on expert consensus alone is weaker than one built on systematic reviews.

Where Google Scholar fits

Google Scholar is acceptable for two things: tracking down a full-text copy of a paper you've already identified, and using "Cited by" to find newer related studies. It is not suitable as your only search tool. It has no quality filter, limited search controls, and it indexes predatory journals and preprints alongside legitimate ones.

Your library

Use your university library's discovery tool and subject guide. Librarians run workshops on database searching and will check your search strategy. Most students never ask. It is the most underused resource available to you.

If you hit a paywall, check in this order:

1. Your library's access (log in through the library page, not directly)
2. Interlibrary loan (usually free, takes days, so plan ahead)
3. An open-access version via Unpaywall or the author's institutional repository
4. Emailing the corresponding author, who will often send the PDF

Do not pay for individual articles before trying these.

Search with technique

Boolean operators

• AND narrows: pressure injury AND repositioning
• OR broadens, and joins synonyms: "pressure injur" OR "pressure ulcer*" OR bedsore**
• NOT excludes, but use it sparingly because it can remove relevant results

Combine synonyms for one concept inside brackets with OR, then join concepts with AND:

(older adult OR elderly OR aged) AND (pressure injur* OR pressure ulcer*) AND (reposition* OR turning)*

Other tools

• Truncation (*): nurs* returns nurse, nurses, nursing.
• Quotation marks: "pressure injury" searches the exact phrase.
• Subject headings: MEDLINE uses MeSH terms and CINAHL uses CINAHL Headings. Searching by heading catches articles that use different wording for the same concept.
• Filters: Limit by publication date (last 5-10 years), peer-reviewed, and language. Be careful with "full text only" because it hides papers your library can still obtain by other routes.
• Study-type filters: Most databases let you filter to systematic reviews or randomised controlled trials. Start there for effectiveness questions.

Citation chasing

Once you find one strong, relevant paper:

• Backward chaining: read its reference list for foundational studies.
• Forward chaining: use "Cited by" in Google Scholar, Scopus, or Web of Science to find newer work.

Systematic reviews on your topic are especially useful. Their reference lists are essentially a pre-screened bibliography.

Keep a search log

Record the database, date, search string, filters, and number of results. It takes two minutes per search, lets you repeat or adjust the search, and some assignments require you to report your search strategy.

Screen and evaluate what you find

You will get hundreds of results. Reduce them in stages.

Stage 1: Title and abstract. Does it address your population, intervention or phenomenon, and outcome? Discard anything off-topic.

Stage 2: Quick quality checks. Before reading the full text, check:

• Study design: Does it suit the question? A cross-sectional survey cannot show that an intervention caused an outcome.
• Sample: How many participants, from where, and how selected?
• Journal: Is it indexed in MEDLINE, CINAHL, or Scopus? Does it have a named editorial board and a stated peer-review process?
• Date: Is there newer evidence?

Stage 3: Full-text read and appraisal. For the sources that survive, apply a validated appraisal tool (CASP, JBI, or whatever your module specifies). Appraisal is covered in detail in the critical analysis guide. The point here is that finding a source is not the same as qualifying it.

Spotting unreliable journals

Predatory journals charge fees but provide little or no genuine peer review. Warning signs:

• Unsolicited emails promising fast publication
• No identifiable editorial board, or board members who don't know they are listed
• Vague or missing peer-review policy
• Spelling errors on the journal site
• Claimed impact factors from unrecognised metrics companies
• Rapid acceptance with no revision requests

Check any unfamiliar journal against:

• NLM Catalog (is it indexed in MEDLINE?)
• Directory of Open Access Journals (DOAJ) for open-access titles
• The Think. Check. Submit. checklist

If you cannot confirm a journal is legitimate, don't cite its papers. Find a better source.

Check for retractions

Retracted papers still appear in searches and get cited by mistake. Before citing a key paper, search its title in PubMed, on the publisher's page, or in the Retraction Watch database. A retraction notice is visible on the article page, but you have to look.

Preprints

Preprints, such as those on medRxiv, have not been peer-reviewed. Don't use them as primary evidence in an assignment unless the brief allows it. If you do, label them clearly as preprints and note the limitation.

Source types: what to use and when

Match the source to the claim. A statement about treatment effectiveness needs a trial or systematic review. A statement about patient experience can rest on qualitative research. A statement about legal duty rests on professional codes or legislation. Citing a news article for a clinical effect is a mismatch markers notice immediately.

Using AI tools safely

AI chatbots can help you brainstorm keywords or understand a concept in plain language. They are not a source, and they are not safe for finding references.

• They fabricate references. Output can include real-sounding authors, plausible titles, and DOIs that point to nothing or to a different paper.
• They cannot be cited as evidence. There is no verifiable author, method, or accountability.
• You are responsible for every citation you submit. "The AI gave it to me" is not a defence in an academic integrity case.

Rules:

1. Never include a reference you have not located and read yourself.
2. Verify every reference by finding it in a database or on the publisher's site, and checking that the authors, year, journal, and findings match what you wrote.
3. Check your institution's AI policy. Some prohibit use entirely, some permit it with disclosure.

Currency: how recent is recent enough?

The default is the last 5 to 10 years. Adjust by topic:

• Fast-moving areas (infection control, COVID-19, medication guidelines, sepsis protocols): prefer the last 3 to 5 years, and always check for a newer guideline version.
• Stable areas (basic anatomy, ethical principles, foundational nursing theory): older sources are acceptable.
• Landmark studies: you may cite an older paper that shaped practice, but state why you are using it. Example: "Although published in 1990, the study remains the foundational account of..." Then pair it with recent evidence showing whether its findings still hold.

If your newest source is five years old, check whether something has since replaced it. Search for "update," "revised," or the latest version of the guideline.

 Step 5: Organise what you find

Set up a system before you have thirty PDFs in a folder named "Downloads."

1. Use a reference manager (Zotero is free; Mendeley and EndNote are alternatives). Save each source the moment you decide to keep it. This prevents a reference-list crisis the night before submission.
2. Create an appraisal grid per source: aim, design, sample, key findings, strengths, limitations, relevance to your assignment, quality rating.
3. Name files consistently: AuthorYear_ShortTitle.pdf.
4. Mark why you kept each source. One line is enough, such as "strongest RCT on repositioning frequency" or "qualitative, useful for patient experience section." You will not remember in a week.

Red flags: quick reference

Stop and reconsider a source if you see any of these:

Time plan: two to three hours for the first pass

Then move to full-text reading and appraisal. If you have less time, protect the question-building and database search. A poorly framed search wastes more time than any other step.

 Pre-submission source checklist

• At least two databases searched, with a log kept
• Mostly primary research, systematic reviews, and current guidelines
• Textbooks and websites limited to background use
• Every source read in full, not just the abstract
• Every reference located and verified (DOI or database record)
• No retracted papers
• No unverifiable journals
• Publication dates justified, and older sources explained
• Each source matches the type of claim it supports
• Sources fit the patient group and setting of your assignment
• Reference list matches in-text citations exactly, in the required style
• Citing guidelines and standards from your own country where relevant

Final advice

• Highest-return action: Build a PICO question and keyword table before you search. Most poor source selection starts with a vague search.
• Biggest risk: Citing something you haven't verified or read in full. This includes AI-suggested references and papers known only from other papers' citations.
• Unrealistic idea to drop: Finding the "perfect" source. Research has limitations. Your job is to identify them and say how they affect your argument, not to find a flawless study.
• What I can't tell you: Which databases your institution subscribes to, which referencing style you must use, and how many sources your module expects. Check the module guide and library page.