NURS 8710 Week 5 Critical Appraisal Paper Example

What the Design Can Support and What This Study Showed: A Critical Appraisal of a Prospective Cohort Evaluating an ICU Mobility Team

Student Name

College of Nursing, Walden University

NURS 8710: Doctoral Seminar I

Instructor Name

Month Day, Year

What the Design Can Support and What This Study Showed: A Critical Appraisal of a Prospective Cohort Evaluating an ICU Mobility Team

The Claim Under Appraisal

This appraisal examines one study in the review's evidence table: a prospective cohort study of a mobility protocol for patients with acute respiratory failure in a university medical intensive care unit (Morris et al., 2008). The study's own conclusion is the target. Its authors conclude that a mobility team using a protocol initiated earlier physical therapy that was feasible, safe and cost-neutral, and that it was associated with decreased ICU and hospital length of stay in survivors who received physical therapy in the ICU, compared with usual care.

The appraisal keeps two questions apart throughout. The first is what a prospective cohort comparison of this kind can establish in principle. The second is what happened when these investigators carried it out. A strong design can be badly executed, and a limited design can be carried out carefully; mixing the two questions produces verdicts that are either too generous or too harsh. The appraisal draws on the JBI checklist for cohort studies (Moola et al., 2020) and on the standard account of threats to validity in nonrandomized designs (Shadish et al., 2002). It is based on the published report as summarized in its abstract, and claims are limited to what the report states.

Design: What a Prospective Cohort Comparison Can Establish

The study compared 165 patients managed under the mobility protocol with 165 receiving usual care. Because patients were not randomly assigned, the design can establish that outcomes differed between the groups, and it can describe what the protocol delivered, but it cannot by itself establish that the protocol caused the differences in outcomes. Three threats are specific to this kind of comparison.

Selection. The central design question is how patients came to be in each group. If the protocol group differed systematically, for example by admission period, bed location or the clinicians caring for them, those differences could explain the outcomes. The abstract reports that baseline characteristics were similar between groups, which reduces but does not remove this concern, since similarity on measured characteristics says nothing about unmeasured ones.

Confounding. Patients who are well enough to get out of bed early are, for that reason alone, likely to leave the ICU sooner. A design without randomization must deal with this by adjustment. According to the abstract, the length of stay comparison was adjusted for three variables, namely weight relative to height, severity score on admission and whether vasopressors were needed, which addresses some but not all of the ways sicker patients differ.

Secular trend. If the protocol and usual care periods were not concurrent, other changes in the unit over time, in sedation practice, staffing or discharge processes, could affect length of stay. The design is stronger if the comparison was concurrent, and an appraiser needs the methods section to confirm which it was.

Execution: What Happened in This Study

The primary outcome was a process measure. The study's primary outcome was the proportion of surviving patients who received physical therapy, and the difference was large: 80% of protocol patients against 47% of usual care patients. Protocol patients were out of bed earlier, at a median of 5 rather than 11 days, and therapy was started in the ICU far more often, 91% against 13%. For a process outcome, the cohort design is adequate. The protocol was designed to change who received therapy and when, and it did; selection or confounding is unlikely to produce differences of this size in a process the protocol directly controlled.

Length of stay was secondary and restricted to survivors. The length of stay findings, 5.5 against 6.9 days in the ICU and 11.2 against 14.5 days in hospital, are reported for survivors. This matters. If more patients died in one group, the survivors in that group are a selected subset, and comparing survivors' length of stay can produce a difference even when the intervention has no effect. The abstract states that outcome data reflect survivors but does not report deaths by group, so the size of this bias cannot be judged from the summary.

Safety and cost were reported in a limited way. No untoward events occurred during mobility sessions, and there was no cost difference between arms including the mobility team. These are useful findings for feasibility. The safety finding refers to events during sessions, which is a narrow window, and the cost analysis included survivors and nonsurvivors, which is a strength.

Verdict

The study supports, with reasonable confidence, the claim that a nurse, nursing assistant and physical therapist team using a protocol can start therapy earlier and more often in ventilated medical ICU patients without adverse events during sessions and without raising costs. It supports the length of stay claim only weakly, and the reasons stack on one another: the outcome was secondary, adjusted for a limited set of confounders and measured in survivors only, in a design that cannot rule out selection. For the review, the study contributes most to the question of feasibility and least to the question of benefit. Its findings were later tested by the same research group in a randomized trial of standardized rehabilitation, in which hospital stays were the same length in both arms (Morris et al., 2016). That result is consistent with the appraisal's caution about the cohort's length of stay finding.

What the Appraisal Changes in the Review

The cohort study will be cited in the review for feasibility and process change, not for length of stay, and the sentences that use it will say so. Its place in the evidence table is unchanged, but the appraisal note will say that its outcomes reflect survivors. If the full methods confirm a concurrent comparison, the selection concern will be noted as reduced; if not, secular trend will be added as a limitation. Either way, the review will not use this study to support a claim that mobilization shortens length of stay.

References

Moola, S., Munn, Z., Tufanaru, C., Aromataris, E., Sears, K., Sfetcu, R., Currie, M., Lisy, K., Qureshi, R., Mattis, P., & Mu, P.-F. (2020). Systematic reviews of etiology and risk. In E. Aromataris & Z. Munn (Eds.), JBI manual for evidence synthesis. JBI.

Morris, P. E., Berry, M. J., Files, D. C., Thompson, J. C., Hauser, J., Flores, L., Dhar, S., Chmelo, E., Lovato, J., Case, L. D., Bakhru, R. N., Sarwal, A., Parry, S. M., Campbell, P., Mote, A., Winkelman, C., Hite, R. D., Nicklas, B., Chatterjee, A., & Young, M. P. (2016). Standardized rehabilitation and hospital length of stay among patients with acute respiratory failure: A randomized clinical trial. JAMA, 315(24), 2694-2702. https://doi.org/10.1001/jama.2016.7201

Morris, P. E., Goad, A., Thompson, C., Taylor, K., Harry, B., Passmore, L., Ross, A., Anderson, L., Baker, S., Sanchez, M., Penley, L., Howard, A., Dixon, L., Leach, S., Small, R., Hite, R. D., & Haponik, E. (2008). Early intensive care unit mobility therapy in the treatment of acute respiratory failure. Critical Care Medicine, 36(8), 2238-2243. https://doi.org/10.1097/CCM.0b013e318180b90e

Shadish, W. R., Cook, T. D., & Campbell, D. T. (2002). Experimental and quasi-experimental designs for generalized causal inference. Houghton Mifflin.