MedicaMundi

J. Case, T. Drew - Concord Hospital, Concord, NH, USA.

M. Jahrsdoerfer, G. Raber, K.K. Giuliano - Philips Healthcare, Andover, MA, USA.

The 1999 report issued by the Institute of Medicine, To Err is Human, was one of the first efforts to quantify the frequency and significant magnitude of variations in care with a particular focus on medical errors and adverse events that were associated with an estimated 98,000 deaths each year [1]. These results have since been confirmed by a numerous studies and reports, including the Community Quality Index Study (CTS) conducted by the Center for Studying Health System Change [2].

The CTS was a systematic evaluation of the quality of preventive, acute, and chronic health care services provided to adults in the United States (US) based on telephone interviews with 12,412 individuals who reported at least 1 visit to a health care provider in the two years preceding the survey. Medical records were reviewed and data abstracted by trained registered nurses for 6,712 of consenting study participants who resided in 12 major metropolitan areas. Overall, patients received 54.9% of recommended care (95% CI, 54.3-55.5) with similar rates for measures of preventive, acute, and chronic care. The majority of problems were associated with underutilization of recommended care, which was documented for 46.3% of patients (95% CI, 45.8-46.8) compared with only 11.3% of patients receiving excessive care that was not recommended and was considered potentially harmful (95% CI, 10.2-12.4) [2].

In response to these findings, numerous initiatives have been undertaken by state and federal agencies, private health care institutions, payers, and professional medical societies to address issues regarding the cost and quality of health care in the United States (US) [3]. It has been emphasized that a key element of interventions to achieve improvements in the quality of health care delivery in the US will include tools to ensure that information on specific patients is readily available to clinicians at all levels, with an emphasis on entry, retrieval, and presentation of data to support clinical decision making as well as measurement and feedback on the quality of care [2, 4, 5].

Increasingly, organizations and practitioners have focused on computerized clinical decision support systems (CDSS) to promote improvements in the quality of care, and manage resource utilization. Such systems have been defined as “an automated process for comparing patient-specific characteristics against a computerized knowledge base with resulting recommendations or reminders presented to the provider at the time of clinical decision making” [6]. Three features common to most CDSS tools include:

• an automated process for delivery of alerts or reminders
• patient-specific content resulting from the comparison of patient information against a set of knowledge ‘rules’ or guidelines
• delivery of alerts or reminders at the point of care [2, 6-9].

Assessments of the impact of CDSS on practitioner behaviors, patient outcomes, and medical costs suggest that proper implementation of CDSS tools has the potential to result in improvements in practitioner performance, more favorable clinical outcomes, enhanced quality of care, reductions in medical error rates, and decreases in resource utilization [7-11]. Two recent systematic reviews of randomized and non-randomized trials evaluated a wide variety of CDSS tools and reported that 64% to 66% of studies demonstrated improvements in the diagnosis, delivery of preventive care, disease management, and patterns of medication prescribing among health care practitioners [8, 9].

A systematic review of 70 studies including approximately 6,000 clinicians and 130,000 patients provides some insights about the specific features of CDSS that are essential to achieving improvements in clinical practice [9]. The investigators identified 15 features of CDSS tools that might contribute to the effectiveness of a CDSS (Table 1).

Table 1. Features of Clinical Decision Support Systems.

Overall, 68% (95% CI, 56%-78%) of 71 comparisons of a CDSS with a non-CDSS control significantly improved clinical practices. Multivariate analyses revealed four specific CDSS features that were associated with statistically significant improvements in clinical performance including [9]:

• automatic provision of decision support to clinicians as part of their workflow (P < 0.0001)
• provision of decision support at the time and location of decision making (P = 0.0263)
• use of a computer to generate the decision support (P = 0.0294)
• Provision of a recommendation for an specific intervention (P = 0.0187).

This pilot study was conducted to assess the impact of a CDSS tool incorporating these four features on clinicians’ ability to monitor critically ill patients receiving vasopressor BP support.

Materials and methods

Continuous measurement blood pressure (BP) measurement is an essential part of the clinical management of patients in ICU settings [12, 13], and the use of vasopressor therapy to treat hypotension is common clinical intervention for critically ill patients. However, the vasoconstriction that results in BP elevations in response to vasopressor therapy can also place increased metabolic demands on the myocardium, which can cause myocardial ischemia and irreversible damage to the myocardium if this oxygen supply/demand imbalance persists for some time [14, 15].

Clinicians usually depend on bedside monitors that issue audible alarms to indicate BP levels that are no longer within a therapeutic range. At the time of an alarm, however, the BP has already dropped below the therapeutic range, which is associated with unfavorable clinical outcomes. This pilot study was conducted to determine if a CDSS tool on patient bedside monitors was associated with improvements in the clinical management of critically ill patients receiving vasopressor support as measured by mean BP and amount of time spent in the target blood pressure range.

The Intellivue Horizon Trends Bedside Monitor (Philips Healthcare, Andover, MA, USA) was developed as a real-time CDSS tool available on patient monitors. Horizon Trends provide clinicians with a graphic visualization tool that presents a visual display of physiologic data that is intended to assist clinicians’ efforts to recognize significant changes in patients’ clinical status.

Figure 1. Key features of Intellivue Horizon Trend.

Figure 1 presents some of the key features of the Horizon Trends display. Baseline target values are established by clinicians and represent either the current value or goal values for specific patients. The upper and lower scales define a therapeutic range for the target value, which is also set by clinicians. The deviation bar is a bar graph which sits exactly at, above, or below the baseline target value and displays the degree of deviation between the currently measured value from the baseline or target value. Trend indicator arrows display how values for specific clinical parameters have changed during the preceding 10 minutes with updates provided every 12 seconds. When the arrow is horizontal to the baseline, it indicates that the value of the clinical parameter has changed 0% to 5% in the last 10 minutes. If the trend arrow points 45^o either up or down, the value has changed 5% to 15% in the last 10 minutes and changes greater than 15% from baseline during the previous 10 minutes are indicated by a trend arrow that points 90^o up or down.

Figure 2. SpO₂ display on Horizon Trends.

Figure 2 demonstrates a situation where the target value for oxygen perfusion levels (SpO₂) has been set at 95% and the patient is currently at 97%. Thus, the deviation bar is displayed above the baseline target level, which indicates that the patient’s SpO₂ value is currently above the target of 95%. In addition, the trend arrow is pointing 45^o up, which indicates that the SpO₂ value has increased 5% to 15% in the last 10 minutes, which could be a clinically meaningful dynamic change.

Figure 3. Typical Horizon Trend graphic on patient bedside monitor.

Figure 3 shows a typical Horizon Trend graphic on a patient bedside monitor including heart rate, ST-segment, SpO₂, mean non-invasive blood pressure, and mean arterial blood pressure (MAP).

This study was conducted in the 24-bed intensive care unit (ICU) at Concord Hospital in Concord, New Hampshire. The study received approval from the Institutional Review Board (IRB). Formal patient consent for study participation was not required by the IRB because none of the products used in this study were experimental and all are approved by the Food and Drug Administration. A convenience sample was used, and subjects included all patients 18 years or older who were receiving vasopressor BP support for at least 24 hours who were present in the ICU during the period of data collection.

Physiologic data were continuously collected and recorded with a laptop computer connected to Horizon Trends for the duration of time that each eligible patient received vasopressor BP support. These data included real-time BP measurements (systolic, diastolic, and mean), clinical BP target (the unit standard was 65 mm Hg for MAP), specific type of vasopressor medication, length of time on vasopressor medication, and ST-segment monitoring with a continuous 12-lead electrocardiogram (ECG). In addition to the use of Horizon Trends, this study evaluated the clinical usefulness of ST Map in combination with Horizon Trends to improve assessment of changes in ST-segments.

Patients were divided into three groups that were serially established. Group 1 involved no change from current practice with patients’ physiologic data displayed on standard bedside monitors showing continuous ECG waveforms, invasive BP waveforms, and continuous saturation of peripheral oxygen (SpO₂).

Upon completion of data collection for Group 1, Horizon Trends was introduced onto the standard resting display on the patient monitor, which comprised study Group 2. Training on how to interpret the display was provided to the nursing staff. The nurses were instructed to use the information provided by the display to manage patients’ BP in conjunction with information provided by the standard display.

Once enrollment for Group 2 was complete, both Horizon Trends and ST Map were introduced on to the standard resting display on all monitors in the unit, which defined study Group 3. The nursing staff was trained on the use of the ST Map feature and nurses were instructed to use both Horizon Trends and ST Map in conjunction with the standard display to manage patients’ BP and any ST-segment changes.

The statistical analysis for this pilot study included analysis of variance (ANOVA) to determine if there were statistically significant differences between groups with respect to mean arterial BP and percent of time BP was maintained within the target range. This analysis does not address changes in ST-segment between the three patient groups, which is currently under evaluation and will be submitted as a separate publication.

Results

The results of the ANOVA comparing differences in MAP and percent of time BP was maintained within the target range (the standard unit was 65 mm Hg) between the 3 groups are presented in Table 2.

*Study Group: Group 1, standard bedside monitor; Group 2, standard bedside monitor plus Horizon Trends; Group 3, standard bedside monitor plus Horizon Trends and ST Map. Abbreviations: SD, standard deviation.

Table 2. Comparison of Mean Blood Pressure and Time ≥ 65 mm Hg BP Target by Study Group*.

Patients in both Group 2 and Group 3 had higher MAP and spent a greater percentage of time within their target BP ranges compared with patients in the standard monitoring group (Group 1). These differences were significant between Group 3 and Group 1 with a MAP of 74.7 mm Hg in Group 3 compared with 68.1 mm Hg for patients in Group 1 (p = 0.001). Similarly, the percent of time spent at or above the target BP of 65 mm Hg was 81.1% for patients in Group 3 compared with 63.7% for Group 1 (P = 0.009).

Conclusions

This pilot study demonstrates that patients monitored with CDSS tools (Horizon Trends and ST MAP) in addition to a standard bedside monitor demonstrated improved BP control and spent significantly greater time within the therapeutic range for BP compared with patients monitored by the standard monitor. The Surviving Sepsis Campaign notes that maintaining mean arterial pressure at levels ≥ 65 mm Hg combined with other clinical support interventions is associated with a reduction in 28-day mortality rates [16].

Results from this pilot study suggest that the use of Horizon Trends and ST Map on patient bedside monitors help critical care nurses improve compliance with the clinical practice goal of maintaining patients who are receiving vasoactive medications to keep BP at or above a specified target of 65 mm Hg. This is likely to improve patients’ prognosis [16], and may impact the total medication dose needed to maintain the desired physiologic target.

The positive effects of the CDSS tools on nurses’ ability to monitor and manage patients were accomplished with no apparent disruption in processes of clinical care, which is an essential feature of CDSS tools. The results from this pilot study should be confirmed in larger, prospective randomized controlled trials.

References

[1]     Kohn K, Corrigan J, Donaldson M. To Err Is Human: Building A Safer Health System. Washington, DC: National Academy Press, 1999.

[2]     Mcglynn EA, Asch SM, Adams J, et al. The Quality of Health Care Delivered to Adults in The United States. New Engl J Med. 2003; 348(26): 2635-2645.

[3]     Bryan C, Boren SA. The Use and Effectiveness of Electronic Clinical Decision Support Tools in the Ambulatory/Primary Care Setting: A Systematic Review of the Literature. Inform Prim Care. 2008; 16(2): 79-91.

[4]     Berwick DM, James B, Coye MJ. Connections between Quality Measurement and Improvement. Med Care. 2003; 41(1 Suppl): 130-138.

[5]     Institute Of Medicine. Crossing The Quality Chasm: A New Health System for the  Twenty-First Century. Washington, DC: National Academy Press, 2001.

[6]     Hunt DL, Haynes RB, Hanna SE, Smith K. Effects of Computer-Based Clinical Decision Support Systems on Physician Performance and Patient Outcomes: A Systematic Review. JAMA. 1998; 280(15): 1339-1346.

[7]     Shiffman RN, Liaw Y, Brandt CA, Corb GJ. Computer-Based Guideline Implementation Systems: A Systematic Review of Functionality and Effectiveness. J Am Med Inform Assoc. 1999;6(2):104-114.

[8]     Garg AX Adhikari NKJ, Mcdonald H, et al. Effects of Computerized Clinical Decision Support Systems on Practitioner Performance and Patient Outcomes: A Systematic Review. JAMA. 2005; 293(10): 1223-1238.

[9]     Kawamoto K, Houlihan CA, Balas EA, Lobach DF. Improving Clinical Practice using Clinical Decision Support Systems: A Systematic Review of Trials to Identify Features Critical to Success. BMJ. 2005; 330(7494): 765.

[10]   Grimshaw JM, Thomas RE, Maclennan G, et al. Effectiveness and Efficiency of Guideline Dissemination and Implementation Strategies. Health Technol Assess. 2004; 8(6): III-IV, 1-72.

[11]   Grol R. Grimshaw J. From Best Evidence to Best Practice: Effective Implementation of Changes in Patients’ Care. Lancet. 2003; 362(9391): 1225-1230.

[12]   Franklin SS, Lopez VA, Wong ND, et al. Single Versus Combined Blood Pressure Components and Risk for Cardiovascular Disease: The Framingham Heart Study. Circulation. 2009; 119(2): 243-250.

[13]   Deedwania PC, Nelson JR. Pathophysiology of Silent Myocardial Ischemia during Daily Life: Hemodynamic Evaluation by Simultaneous Electrocardiographic and Blood Pressure Monitoring. Circulation, 1990; 82(4): 1296-1304.

[14]   Lim, W, Qushmaq I, Cook D, Crowther M, Heels-Ansdell D, Devereaux PJ. Elevated Troponin and Myocardial Infarction in the Intensive Care Unit: A Prospective Study. Critical Care 2005;9(6): 634-635.

[15]   Owens P, O’Brien E. Hypotension in Patients with Coronary Disease: Can Profound Hypotensive Events Cause Myocardial Ischemic Events? Heart. 1999; 82(4): 477-481.

[16]   Dellinger RP, Levy MM, Carlet JM, et al. Surviving Sepsis Campaign: International Guidelines for Management of Severe Sepsis and Septic Shock: 2008. Crit Care Med. 2008; 36(1): 296-327.