Harvard Nursing Literature Review Example: Nurse Staffing Ratios & Patient Outcomes

5 nursing-specific scoring dimensions (combined with Harvard)

Dimension Poor signal Strong signal

1 Exposure measure "More nurses = better" Distinguishes patient-to-nurse ratio, RN hours per patient-day, skill mix (RN vs LPN vs HCA)

2 Outcome tiering Mortality only Tiers mortality / failure-to-rescue / hospital-acquired complications (UTI, pneumonia, pressure ulcer) / nurse outcomes (burnout)

3 Design honesty Cross-sectional only, treated as causal Cross-sectional (Aiken 2002) + shift-level longitudinal (Needleman 2011) + systematic review/meta-analysis (Kane 2007) + evidence critique (Griffiths 2016) graded separately

4 Harvard mechanics APA &, double quotes, et al. italic and, single quotes, et al. roman, London: Wiley

5 Precise gap "Need more nurses" Names setting × ratio threshold × outcome × confounding (acuity, turnover, work environment)

Strong draft (Harvard / Cite Them Right 12th, annotated)

[Thesis / scope] Higher RN staffing is consistently associated with lower hospital mortality and fewer nursing-sensitive complications, but the evidence base is observational end-to-end—no RCTs exist and probably never will—so the defensible claim is association with moderate causal plausibility, not proven causation. [Narrative, classic cross-sectional] Aiken, Clarke, Sloane, Sochalski and Silber (2002) linked nurse surveys from 168 Pennsylvania hospitals to 232,342 surgical discharges and found each additional patient per nurse raised 30-day mortality odds by 7% (OR 1.07, 95% CI 1.03–1.12) and failure-to-rescue odds by 7%, alongside 23% higher nurse burnout odds. [Shift-level longitudinal] Needleman, Buerhaus, Pankratz, Leibson, Stevens and Harris (2011) addressed the cross-sectional critique by tracking 197,961 admissions and 176,696 eight-hour shifts in 43 units of one tertiary centre: each shift ≥8 hours below target RN staffing carried a hazard ratio for death of 1.02 (95% CI 1.01–1.03) and each high-turnover shift HR 1.04 (95% CI 1.02–1.06). [Systematic review / meta-analysis] Kane, Shamliyan, Mueller, Duval and Wilt (2007) pooled 96 reports and estimated that one additional RN per patient-day lowered mortality OR to 0.94 (medical), 0.84 (surgical), 0.94 (ICU), and cut hospital-acquired pneumonia odds by 19% overall. [Evidence critique] Griffiths, Ball, Drennan, Dall'Ora, Jones, Maruotti, Pope, Recio Saucedo and Simon (2016) argued that despite this volume, endogeneity (omitted variable bias, simultaneity, common-method variance) means effect sizes are biased and possibly qualitatively wrong for economic costing—which is why NICE (2014) still judged the direct RCT-grade evidence "insufficient" for hard thresholds. [Synthesis] The review-grade conclusion is therefore two-tiered: association = strong and consistent; causal effect size = uncertain. [Precise gap] What remains untested is whether mandated ratio changes (e.g., California-style legislation) produce mortality reductions net of concurrent work-environment reforms—a gap that matters because Aiken et al. (2010) and Needleman 2011 both show staffing effects are entangled with acuity, turnover, and Magnet-status confounders that single-institution or cross-state snapshots cannot separate.

[Evidence-quality subsection] Design hierarchy determines how much weight each number carries. [Narrative] Aiken et al. (2002) is a cross-sectional linkage—high external relevance (168 hospitals), low protection against confounding. [Parenthetical] Needleman et al. (2011) upgraded this to shift-level exposure with Cox models, removing the "which unit was the patient on" ambiguity but remaining observational and single-site. [Narrative] Kane et al. (2007) is the pooling anchor (Level II evidence): its dose–response curve (OR per +1 RN/patient-day) is the most cited quantification in policy briefs. [Critique] Griffiths et al. (2016) is not a new dataset but a methodological audit of exactly these three tiers, concluding that none escapes endogeneity. [Synthesis] A GRADE-style reading yields moderate certainty for the association, low certainty for the exact percentage effect size, very low for "ratio X saves Y lives" claims used in staffing mandates.

[Translation subsection] Policy translation is where reviews fail. [Narrative, 2 authors] Needleman and Buerhaus (2003) and later Aiken et al. (2010) showed California's mandated 5:1 medical / 4:1 surgical ratios coincided with lower mortality, but the natural experiment is non-random and concurrent with Magnet diffusion. [Precise gap] A literature review that converts Kane's OR 0.84 (surgical) into a blanket "hire one more nurse, save 16%" slogan ignores that the OR is per patient-day FTE, not per fixed roster, and ignores skill-mix substitution (HCA vs RN). The honest review states the association, names the confounding, and stops before pseudo-causality.

In-text citation audit (Harvard / CTR 12th, convention locked from 11a)

• 1 author: Needleman (2011) / (Needleman, 2011)

• 2 authors: Needleman and Buerhaus (2003) / (Needleman and Buerhaus, 2003) — and • 3+ authors, FIRST citation: (Aiken, Clarke, Sloane, Sochalski and Silber, 2002) — all five listed

• 3+ authors, SUBSEQUENT: (Aiken et al., 2002) — et al. roman • 6+ authors, FIRST citation: (Griffiths, Ball, Drennan, Dall'Ora, Jones, Maruotti et al., 2016) — first six + et al. is acceptable; this review lists all eight for transparency

• Multi-source: (Aiken et al., 2002; Kane et al., 2007; Needleman et al., 2011) — alphabetical

• Page number: (Aiken et al., 2002, p. 1989)

Reference list (Harvard / Cite Them Right 12th)

Reference list

Aiken, L. H., Clarke, S. P., Sloane, D. M., Sochalski, J. and Silber, J. H. (2002) 'Hospital nurse staffing and patient mortality, nurse burnout, and job dissatisfaction', Journal of the American Medical Association, 288(16), pp. 1987–1993. Available at: https://doi.org/10.1001/jama.288.16.1987

Griffiths, P., Ball, J., Drennan, J., Dall'Ora, C., Jones, J., Maruotti, A., Pope, C., Recio Saucedo, A. and Simon, M. (2016) 'Nurse staffing and patient outcomes: strengths and limitations of the evidence to inform policy and practice. A review and discussion paper based on evidence reviewed for the National Institute for Health and Care Excellence Safe Staffing guideline development', International Journal of Nursing Studies, 63, pp. 213–225. Available at: https://doi.org/10.1016/j.ijnurstu.2016.03.012

Kane, R. L., Shamliyan, T. A., Mueller, C., Duval, S. and Wilt, T. J. (2007) 'The association of registered nurse staffing levels and patient outcomes: systematic review and meta-analysis', Medical Care, 45(12), pp. 1195–1204. Available at: https://doi.org/10.1097/MLR.0b013e3181468ca3

Needleman, J., Buerhaus, P., Pankratz, V. S., Leibson, C. L., Stevens, S. R. and Harris, M. (2011) 'Nurse staffing and inpatient hospital mortality', New England Journal of Medicine, 364(11), pp. 1037–1045. Available at: https://doi.org/10.1056/NEJMsa1001025

Format checks (CTR 12th): • "Reference list", alphabetized A → G → K → N

• Surname, Initial(s). — Aiken, L. H. not Aiken, LH

• Single quotes on article titles, sentence case; journals italicized title case • 364(11), pp. 1037–1045 (en dash)

• et al. roman; all authors listed in list • Available at: https://doi.org/... — no access date

Note on source count: this page demonstrates the Harvard layer with 4 anchor sources (Aiken 2002, Needleman 2011, Kane 2007, Griffiths 2016) — the four that carry the argument. A production page on acade.ai would extend to 6–8 with Hodgson 2015 / Appleton 2015 (already in APA×Nursing) translated into Harvard, plus Aiken et al. 2014 Lancet European 9-country study for cross-national breadth. The 4-source core is sufficient to demonstrate structure + format; expansion is mechanical, not argument-changing.

What changed: poor draft → strong Harvard draft

Element Poor draft Strong Harvard draft (above)

Opening "Nurses help patients." Association-vs-causation thesis upfront

Exposure measure One number "staffing" Ratio / RN-hours-per-patient-day / skill mix separated

Outcome tiering Mortality only Mortality / FTR / HAP / burnout tiered

Effect sizes None OR 1.07 (Aiken), HR 1.02 (Needleman), OR 0.84 surgical (Kane)

Design grading All "studies" equal Cross-sectional → shift-longitudinal → meta-analysis → endogeneity critique

Griffiths 2016 mislabel "Lancet systematic review" Corrected to Int J Nurs Stud 63:213–225

Kane 2007 mislabel "Health Services Research" Corrected to Medical Care 45(12):1195–1204

Citation mechanics APA &/double-quote/et al. and/single-quote/roman et al./place retained

5 fatal signals (Harvard × Nursing)

1. Labeling Griffiths 2016 a "Lancet paper" — it is IJNS, not Lancet.

2. Putting Kane 2007 in Health Services Research — it is Medical Care.

3. Using & in citations ((Aiken & Clarke, 2002)) — Harvard uses and.

4. Italicizing et al. or using double quotes on article titles.

5. Claiming staffing ratios cause X% mortality reduction without naming endogeneity (Griffiths 2016) — overclaims a moderate-certainty association into a causal slogan.

Rewrite Checklist (Harvard × Nursing)

1. Open with association vs. causation framing — never "more nurses save lives" as a flat claim.

2. Name the exposure measure explicitly (ratio / RN-hours/patient-day / skill mix).

3. Tier outcomes: mortality → failure-to-rescue → HAP (UTI, pneumonia, pressure ulcer) → nurse burnout.

4. Anchor with Aiken 2002 (cross-sectional, OR 1.07), Needleman 2011 (shift-level, HR 1.02), Kane 2007 (meta, OR 0.84 surgical), Griffiths 2016 (endogeneity caveat).

5. Convert ≥40% parentheticals to narrative; et al. roman; 3+ authors full list on first citation.

6. Audit reference list: single quotes, italics journal, Place: Publisher only if book, Available at: https://doi.org/..., no access date for DOI.

7. Close gap with setting × threshold × outcome × confounder (acuity, turnover, Magnet status, concurrent policy).

8. Verify every DOI resolves.

FAQ

Why is Griffiths 2016 in IJNS and not Lancet? Because it isn't a Lancet paper. Griffiths et al. wrote it for the International Journal of Nursing Studies, vol. 63, pp. 213–225, 2016, as a discussion paper feeding NICE's safe-staffing guideline. The confusion is common because Griffiths co-authored the 2014 Aiken et al. Lancet 9-country paper — different article, different journal.

Is the Kane 2007 meta-analysis in Health Services Research? No. The AHRQ-funded Minnesota Evidence-based Practice Center report (Kane, Shamliyan, Mueller, Duval & Wilt, 2007) split into two outputs: the AHRQ Evidence Report/Technology Assessment No. 151 (nurse staffing review) and the peer-reviewed meta-analysis in Medical Care, 45(12), 1195–1204. The Medical Care paper is the citable meta-analysis; that's what this review uses.

How do I cite Needleman 2011's HR 1.02? Report it as (Needleman et al., 2011, p. 1041) with the hazard ratio and CI in your own prose: "each shift ≥8h below target carried HR 1.02 (95% CI 1.01–1.03)". Don't quote the number without the CI — nursing evidence reviews expect CIs on every effect size.

Can I translate APA×Nursing's Hodgson 2015 / Appleton 2015 into this page? Yes — those two are on a different sub-theme (e.g., dementia care / pressure-ulcer prevention), so they extend rather than replace the staffing-ratio core. In Harvard they become Hodgson, S., ... (2015) '...', Journal, vol(issue), pp. x–y. Available at: https://doi.org/.... Keep the same CTR rules.

Can I reuse this review's citations? No. Use the structure and Harvard mechanics as a model; replace sources with your own verified search, and confirm each DOI resolves.

/harvard-literature-review-example · /literature-review-example-annotated-sample#by-citation-style · /apa-nursing-literature-review-example · /harvard-education-literature-review-example

Continue the research workflow with Acade

Acade can help you search for literature, organize evidence notes, build a synthesis outline, and revise citation-supported academic writing. It cannot guarantee source accuracy or replace your reading, methodological judgment, institutional rules, or responsibility for the submitted work.

Tell Acade your topic, discipline, citation style, evidence requirements, and assignment constraints, then ask it to help organize the next verifiable research step.