Evidence Digest

The Latest Evidence on Hourly Roundingand Rapid Response Teams in DecreasingAdverse Events in Hospitals

Bernadette Mazurek Melnyk, RN, PhD, CPNP/NPP, FAAN, FNAP

The purpose of Evidence Digest, a recurring columnin Worldviews, is to provide concise summaries of

well-designed and/or clinically important recent studiesalong with implications for practice, research, adminis-tration, and/or health policy. Articles highlighted in thiscolumn may include quantitative and qualitative studies,systematic and integrative reviews as well as consensusstatements by expert panels. Along with relevant implica-tions, the level of evidence generated by the studies or re-ports highlighted in this column (see Figure 1) is includedat the end of each summary so that readers can integratethe strength of evidence into their healthcare decisions.

________________________________________________________________________

• Level I: Evidence from a systematic review or meta-analysis of all relevant randomized controlled trials (RCTs), or evidence-based clinical practice guidelines based on systematic reviews of RCTs

• Level II: Evidence obtained from at least one well-designed RCT • Level III: Evidence obtained from well-designed controlled trials without randomization • Level IV: Evidence from well-designed case-control and cohort studies • Level V: Evidence from systematic reviews of descriptive and qualitative

studies • Level VI: Evidence from a single descriptive or qualitative study • Level VII: Evidence from the opinion of authorities and/or reports of expert committees

Modified from Guyatt & Rennie, 2002; Harris et al., 2001

Figure 1. Rating System for the Hierarchy of Evidence (fromMelnyk & Fineout-Overholt 2005).

NURSING ROUNDS AND PATIENT SAFETY

Meade C.M., Bursell A.L. & Ketelsen L. (2006). Effects ofnursing rounds on patients’ call light use, satisfaction, andsafety. American Journal of Nursing, 106(9), 58–70.

Purpose. The purpose of this study was to assess thefrequency of and reasons for patients’ call light use as wellas the effects of 1- and 2-hour nursing rounds on patients’use of the call light, patient satisfaction, and the rate ofpatient falls.

Copyright ©2007 Sigma Theta Tau International1545-102X1/07

Design. Quasi-experiment, with three study groups:(a) hourly patient rounding, (b) patient rounding every 2hours, and (c) control, with no regular rounding.

Sample. Twenty-seven nursing units in 14 hospitalsacross the United States, including small, large, rural andurban hospitals.

Methods. Nurses in the two experimental studygroups followed a specific protocol, either performing hos-pital rounds on their patients every hour or every twohours. Chief nursing officers and nurse managers madedecisions on what units in their hospital were assigned toone of the three study groups. After baseline assessment ofkey clinical variables for 2 weeks, rounding occurred fora time period of 4 weeks. After the 4 weeks of the inter-vention (i.e., rounding), follow-up assessments of the out-come variables (e.g., use of call lights, patient satisfaction,falls) were conducted. Nurses on the experimental round-ing units were instructed to inform their patients that theywere there to conduct rounds and perform key actions (e.g.,assess patient pain levels using a pain assessment scale, of-fer toileting assistance, make sure the call light and tele-phone were within the patients’ reach, inform the patientsthat staff would be back to round again in a set period oftime).

Findings. Of the 46 units in 22 hospitals that partic-ipated in the study, data from 19 units in 8 hospitals wereexcluded from the analyses because of poor reliability andvalidity of data collection. Findings indicated that hourlyrounding was more effective than 2-hour rounding and norounding on call light usage, patient satisfaction and num-ber of patient falls. Two-hour rounding was more effectivethan no rounding on patient satisfaction.

Commentary with Implications for Action in ClinicalPractice and Future Research. When assessing whetherfindings from an intervention study are valid (i.e., asclose to the truth as possible), it is important to answersome key questions, including whether: (1) randomassignment to study groups was used, (2) the studygroups were equal at baseline on key demographic and

220 Fourth Quarter 2007 �Worldviews on Evidence-Based Nursing

Evidence Digest

clinical variables, and (3) the all of the subjects wereaccounted for at the end of the study. This research useda quasi-experimental design that did not randomly as-sign hospital units to one of the three interventiongroups, which resulted in non-equivalent groups atthe beginning of the experiment (e.g., patient satisfac-tion and falls were not equal among the three groupsat the beginning of the study). When random assign-ment to intervention groups is used, there is a goodprobability that the study groups will be similar onkey demographic and clinical variables at the begin-ning of the study. Similarity in baseline clinical anddemographic variables at the outset of an experimentis important in order to support the internal validityof a study (i.e., being able to say at the end of a studythat it was the treatment or intervention that causeda change in the outcomes, not other extraneous vari-ables, such as a difference in baseline characteristicsbetween the study groups). In addition, there was ahigh attrition rate in this research (i.e., data from sev-eral units were excluded from the analyses), which alsothreatens the internal validity of the study. Despite sub-stantial threats to the validity of this study, the evidenceis compelling and suggests that hourly rounding maylead to increased patient satisfaction and a decreasein patient falls in hospitals. However, further researchis needed that uses randomized controlled trials be-cause they are the strongest design for testing causeand effect relationships. Cost benefit analyses also areneeded, which will be especially helpful to adminis-trators who are making decisions about implementingbest practices. In those hospitals that choose to imple-ment hourly rounding, it will be important to measureoutcomes of this practice within each of the institu-tions in order to generate internal evidence for supportof this practice.

Level of Evidence: III.

RAPID RESPONSE TEAM DECREASESCARDIAC ARRESTS

Offner P.J., Heit J., & Roberts, R. (2007). Implementationof a rapid response team decreases cardiac arrest outsideof the intensive care unit. The Journal of Trauma Injury,Infection and Critical Care, 62(5), 1223–1228.

Purpose. The purpose of this study was to assess theeffect of a rapid response team that evaluated and treatednon-intensive care unit patients with early signs of physi-cal instability on the number of patients who experiencedcardiac arrests.

Design. Prospective case study outcomes evalua-tion.

Sample/Setting. Level I trauma center in the Westernpart of the United States.

Methods. A rapid response team was developed thatconsisted of an intensivist, a critical care nurse, and a res-piratory therapist. The team was available 24 hours a dayfor 7 days a week.

Hospital staff were informed to activate the rapid re-sponse team for the following patient criteria: (1) heartrate greater than 130 or less than 40 beats per minute, (2)systolic blood pressure less than 90 mm Hg, (3) respiratoryrate more than 24 or less than 8 breaths per minute, (4)seizures, (5) acute mental status changes, and (6) nursingstaff concern for any other reason. An educational programwas conducted with the hospital staff so that all staff mem-bers were familiar with how to use the team. Data werecollected prospectively regarding the number of team acti-vations, reasons for activation, the intervention performed,and the outcome as well as the occurrence of cardiac arrest.

Findings. Between March and December of 2005, therapid response team was activated 76 times, an averageof eight times per month. The most common reason foractivation of the team was a change in mental status, fol-lowed by concern from the nursing staff that the patient’scondition was deteriorating. Most activations of the teamoccurred between 7 a.m. and 7 p.m. In over half of the ac-tivations, patients were moved to a higher level of care. Af-ter implementation of the rapid response team, there was a50% reduction in cardiac arrests in comparison to the sametime period the year before initiation of the rapid responseteam.

Commentary with Implications for Action in ClinicalPractice and Future Research. There continues to bemajor pressure placed upon hospitals to enhance pa-tient safety and quality of care. Thus, hospitals are in-tensifying their efforts to determine the most effica-cious strategies to decrease adverse events such as car-diac arrest, including the use of rapid response teams.Although this study was not a randomized controlledtrial, which is the strongest design for testing cause andeffect relationships, the findings indicate that use of arapid response team could possibly lead to a reduc-tion of cardiac arrests in hospitalized patients. Accu-mulating evidence has shown that patients show mul-tiple signs of deteriorating status before cardiac arrestoccurs, therefore, close monitoring of these signs bynurses at the bedside who can activate a rapid responseteam seasoned in intervening before cardiac arrest oc-curs can be lifesaving for patients. Future research isneeded to continue to document the positive outcomesof rapid response teams on a variety of patient and staffoutcomes.

Worldviews on Evidence-Based Nursing �Fourth Quarter 2007 221

Evidence Digest

CONSENSUS ON THE USE OFMEDICAL EMERGENCY TEAMS

DeVita M.A., Bellomo R., Hillman K., Kellum J., RotondiA., Teres D., Auerbach A., Chen W-J., Duncan K., KenwardG., Bell M., Buist M., Chen J., Bion J., Kirby A., Lighthall G.,Ovreveit J., Braithwaite S., Gosbee J., Milbrandt E., PeberdyM., Savitz L., Young L. & Galhotra S. (2006). Findingsof the first consensus conference on medical emergencyteams. Critical Care Medicine, 34(4), 2463–2478.

Purpose. To create a consensus document regardingthe use of medical emergency teams.

Methods. An international conference on medicalemergency teams (MET) was held in June of 2005 withexperts in hospital safety, hospital medicine, critical caremedicine and medical emergency teams. Pre-conferencetelephone and e-mail conversations were conducted alongwith an extensive literature review before the 2-day con-sensus conference. The review of studies conducted bythe experts did not include unpublished data or data pre-sented in abstract form only. The participants addressedfour key questions: (1) What is an MET response? (2) Isthere an MET syndrome? (3) What are barriers to METS?and (4) How should outcomes be measured? The expertpanelists for the consensus conference were faculty mem-bers of the First International Conference on Medical Emer-gency Team

Responses. Preventing Patient Crisis, Protecting Pa-tients in Crisis. The consensus panel of experts met follow-ing the international conference. At the consensus confer-ence, the experts were divided into four groups to addressthe four key questions. Each group generated a responseto their question and presented their progress to the entiregroup in order to receive additional input. In the weeksfollowing the conference, each of the four working groupsgenerated a report. These reports were distributed to allpanelists for final comments.

Results. The consensus panel defined an in-hospitalmedical emergency as a mismatch between patient needsand resources available (e.g., when a hospitalized patient’scondition deteriorates and resources are not available or be-ing implemented). An MET patient was defined as a personwho has deteriorated to the point where he or she is at riskfor serious harm and requires an urgent clinical response.Barriers to implementing Rapid Response Systems (RRSs)were identified, including such factors such as culture andprofessional role norms, structure and tendency to workin professional silos, resource constraints, and knowledgeof evidence regarding medical emergency teams, and train-ing/education. The panel achieved consensus that the newterm Rapid Response System should be used to describea whole system for implementing a safety net when pa-tients become critically ill and have a mismatch between

needs and resources. These RRSs should consist of fourcomponents: (1) a crisis detection and response trigger-ing mechanism, (2) a predetermined rapid response team,(3) a governance/administrative structure to supply andorganize resources, and (4) a mechanism to evaluate cri-sis antecedents and promote hospital process improvementto prevent future events. It was recommended that datashould be collected for a period of 3 months before imple-mentation of RRSs. Implementation should be conductedfor at least a year before the measurement of outcomes, in-cluding the frequency of adverse events, the quantity andtypes of resources used, and staff satisfaction. The expertsconcluded that there is strong evidence to support a rec-ommendation that hospitals implement and assess RRSsalthough they reported that there is insufficient evidencefor accrediting organizations as well as governmental andregulatory agencies to require hospitals to implement RRSs.Positive outcomes of RRSs include reduced deaths, cardiacarrests, hospital length of stay, ICU length of stay and cost.It also was recommended that more research is needed todetermine the extent of positive outcomes with RRSs andthe settings that benefit most from these systems.

Commentary with Implications for Action in ClinicalPractice, Administrators, and Future Research. Theconsensus meeting was thoughtfully planned and con-sisted of experts from multiple areas, which diversifiedthe perspectives of the panel. The process was care-fully planned and implemented, resulting in consensusrecommendations for hospitals surrounding whetherand how RRSs should be utilized. Based on multiplestudies that have documented the positive outcomes ofrapid response teams or medical emergency teams, itis recommended that hospital administrators stronglyconsider the implementation of RRSs. Measurement ofoutcomes once RRSs are implemented is critical to gen-erating additional evidence on the benefits of RRSs fora variety of hospital systems. Future research shouldassess moderators of the effects of RRSs on outcomes(i.e., under what types of systems and conditions thatRRSs work best).

Level of Evidence: I, since a thorough evidence reviewwas conducted as part of the consensus panel’s work.

MEDICAL EMERGENCY TEAM USEAMONG MEDICAL UNITS

Schmid A. (2007). Frequency and pattern of medical emer-gency team activation among medical cardiology patientcare units. Critical Care Nurse, 30(1), 81–84.

222 Fourth Quarter 2007 �Worldviews on Evidence-Based Nursing

Evidence Digest

Purpose. The purpose of this study was to deter-mine the frequency and pattern of medical emergency team(MET) activation among medical cardiology patient careunits.

Design. A retrospective descriptive study.Sample/Setting. Three medical cardiology units

within a 700-bed academic medical center, which houses54 patient beds. The 54 beds comprised 12% of the totalmedical surgical beds in the hospital.

Methods. Calls for a condition C team (i.e., an METteam) were reviewed in the log data base, which includedevent time, date and location. The log was assessed to deter-mine the frequency and timing of the events that occurredin the cardiology service, and the pattern was compared tothe total events for a 6-month period of time.

Findings. There were 886 events from the medicalsurgical nursing units of which 100, or 11%, of the activa-tions were to the medical cardiology patient units. Fifty-onepercent of the events happened on the day shift, with 21%occurring during 8 and 10 a.m., and 49% occurred on thenight shift.

Commentary with Implications for Action in Clini-cal Practice and Future Research. This descriptive in-vestigation supports prior studies that have shownthe majority of activations of METs occur during the

day hours. An interesting question that is generatedby this finding is whether patients are actually firstdeteriorating during these hours or it is more likelyto be discovered since this is the time when routinecare activities are taking place (during early morninghours) and patient assessments are occurring. In thefuture, it will be important to conduct further researchto determine whether hourly rounding is associatedwith more activations and other outcomes besides falls,call light uses and patient satisfaction.

Level of Evidence: VI

ReferencesGuyatt G. & Rennie D. (2002). Users’ guides to the medical

literature. Washington, DC: American Medical Associa-tion.

Harris R.P., Hefland M., Woolf S.H., Lohr K.N., MulrowC.D., Teutsch S.M. & Atkins D. (2001). Current meth-ods of the U.S. Preventive Services Task Force: A reviewof the process. American Journal of Preventive Medicine,20(3 Suppl.), 21–35.

Melnyk B.M. & Fineout-Overholt, E. (2005). Evidence-based practice in nursing & healthcare. A guide to bestpractice. Philadelphia: Lippincott, Williams & Wilkins.

Worldviews on Evidence-Based Nursing �Fourth Quarter 2007 223