Using electronic medical record notes to measure ICU telemedicine utilization

Abstract

Given the complexity of high-acuity health care, designing an effective clinical note template can be beneficial to both document patient care and clarify how telemedicine is used. We characterized documented interactions via a standardized note template between bedside intensive care unit (ICU) providers and teleintensivists in 2 Veterans Health Administration ICU telemedicine support centers. All ICUs linked to support centers and providing care from October 2012 through September 2014 were considered. Interactions were assessed based on initiation site, bedside initiator, contact type, and patient care change. Of 14 511 ICU admissions with teleintensivist access, teleintensivist interaction was documented in 21.6% (N = 3136). In particular, contacts were primarily initiated by bedside staff (74.4%), use increased over time, and of contacts resulting in changes in patient care, most were initiated by a bedside nurse (84.3%). Given this variation, future research necessitates inclusion of utilization in evaluation of Tele-ICU and patient outcomes.

INTRODUCTION

Although research suggests that patients in intensive care units (ICUs) do better under the care of an intensivist,^1–5 ICU telemedicine (Tele-ICU) programs show mixed results. Several studies demonstrate reductions in hospital mortality, ICU mortality, and hospital and ICU length of stay, plus increased adherence to best practice protocols and lower rates of preventable complications.^6–10 Others find no statistical differences in these measures,^11–13 even among studies with the same telemedicine technology in similar hospital systems.^6,11 Kahn¹⁴ hypothesized that these conflicting results may be due to differences in physician buy-in, synchronization of services, and linkage to quality improvement programs. Complex relationships between staff acceptance and varying degrees of patient co-management among Tele-ICU and bedside providers further confound the issue.^14–16

Tele-ICU is a complex intervention with social and technological implications. Varying definitions, uses, and implementation processes of this clinical tool complicate assessment.^15,17–19 Further, studies describe the role of Tele-ICU programs in various ways, including for efficient deployment of critical-care staffing, monitoring of patients and identification of potential problems, diagnostic consultation, ICU staff education, data collection and outcome reporting, and dissemination of best practices and standardization of care.^6,14 These broad categories lack detailed descriptions of program organization and day-to-day interactions between Tele-ICU staff (critical care–trained nurses and physicians) and bedside staff (attending physicians, nurses, residents, and other personnel), which can vary greatly between Tele-ICU networks. These specifics can impact the effect of telemedicine on the care of critically ill patients and must be better understood.

Characterizing telemedicine utilization in a complex, high-acuity health care setting is a challenging task. Few studies describe the interactions between bedside and Tele-ICU staff.^16,20 To our knowledge, none present a comprehensive review of Tele-ICU documentation in administrative records. The study objective here is to systematically describe and compare interactions between bedside staff and Tele-ICU intensivists (hereafter referred to as teleintensivists), as documented in a clinical note template in 2 independently operated Tele-ICU support centers located in geographically separated Veterans Integrated Service Networks (VISNs) of the Veterans Health Administration and serving primarily ICUs in their designated VISNs. We intentionally did not evaluate the impact of Tele-ICU on specific patient outcomes, in order to comprehensively describe the teleintensivist clinical note template and examine how critical-care staff uses the system.

METHODS

Setting

We obtained data from 2 ICU telemedicine programs in the Veteran Affairs Midwest Health Care Network (VISN 23) and the Veteran Affairs Healthcare System of Ohio (VISN 10) from October 2012 through September 2014. The teleintensivist note template was designed at the Veterans Health Administration (VHA), where ICUs are classified according to multiple criteria as highly complex (level 1) to basic (level 4).²¹ Each telemedicine center serves multiple facilities (Figure 1 ) with a range of ICU complexities, rurality, and residency program affiliations. Most ICUs contain a mixture of medical and surgical patients.

Figure 1.

VA Tele-ICU networks and characteristics of partnered ICUs.

Both programs connected support center staff with bedside ICUs using the same advanced technological system, eCaremanager^® (Philips), which provides bidirectional audiovisual connectivity and allows Tele-ICU staff to virtually round on all patients at any time or frequency. This system integrates bedside patient data and iconographic portrayals of patient acuity and alerts Tele-ICU staff about abnormal laboratory studies or changes in physiologic status.¹⁵ Full access to bedside electrocardiographic waveforms and the VA Computerized Patient Record System were available. Tele-ICU staff contacted bedside staff by initiating audiovisual communication, and bedside staff activated contact by pressing a Tele-ICU call button within the patient’s room.

These programs had different standards of operation. Notably during the study period, support center hours varied. Specifically, in Minneapolis the teleintensivist was available 24 h and in Cincinnati 12 h (7 pm to 7 am). We thus considered “exposure,” which was calculated as the amount of time a patient was in the ICU with teleintensivist support available throughout the entire stay.

This research was reviewed and approved by the Institutional Review Boards and Research and Development Committee at the Iowa City Veteran Affairs Health Care System, IRB 200910772. Patient consent was not required.

Outcomes

Primary outcomes were measures of telemedicine utilization during ICU visits. Specifically, measures included the percentage of ICU admissions with teleintensivist involvement, who initiated the interaction, the reason for contact, and whether a patient care change resulted. This information originated as predefined options from a standardized clinical note template recorded only by teleintensivists in the computerized patient record system; interactions triaged before reaching the teleintensivist were not recorded. The note template was designed in consultation with the 2 Tele-ICU support centers (by LJ) and integrated into VHA’s electronic medical record system. There were 53 predefined fields and space for free text (Figure 2 ). Selections were dependent on teleintensivist judgment. Teleintensivists, similar to bedside staff, interacted with patients on multiple occasions and recorded multiple interactions within a single note or throughout a patient’s ICU stay. Telemedicine interactions occurring in non-ICU settings (N = 51) were excluded.

Figure 2.

Screenshot of ICU telemedicine (template in the computerized patient record system).

Statistical analysis

Three categorical variables were used to classify admission cases: first, by initiation site (ie, bedside unit or Tele-ICU center); second, by initiator at bedside (ie, attending physician, registered nurse, resident, or other personnel); and third, by contact type (ie, critical, routine, or urgent intervention). Routine interventions included guideline adherence, line placement review, medication management, and order renewals. Fever management, initial admission and status change assessments, critical lab values, invasive monitor placement, and pain, agitation, delirium, and insomnia management comprised urgent interventions. Critical interventions involved urgent intubation, cardiopulmonary arrest, bradycardia, tachycardia, unstable hyper-/hypotension, sepsis resuscitation, or respiratory distress. As multiple interactions often occurred within a patient’s ICU stay, we summarized how many admissions contained 0, 1, 2–4, or >5 total interactions, as well as the maximum, mean, and standard deviation of interactions. Mean (SD) of contacts per week of Tele-ICU exposure were reported. Plots of overall trends in the daily rate of utilization categories using 3-month moving averages were created. Contacts yielding a patient care change were assessed via individual interactions.

All analyses were generated using SAS Enterprise Guide, © 2014 SAS Institute Inc., Cary, NC, USA.

RESULTS

In total, 6470 teleintensivist notes were linked to 3136 of 14 511 (21.6%) ICU admissions occurring from October 2012 through September 2014, using dates of admission and discharge and recorded date of telemedicine interaction.

Among admissions with teleintensivist contact, there was a mean (SD) of 2.06 (2.34) communications per admission (Table 1). Contacts were initiated by a bedside resident, attending physician, or bedside nurse in 55.1%, 12.4%, and 12.1% of eligible admissions, respectively. Routine interventions (59.4%) were the most common contact type. Among admissions with telemedicine contact, the majority (60.3%) had 1 contact, 31.7% had 2–4 contacts, and 8.0% had ≥5. The bedside unit initiated twice as often as the Tele-ICU support center, on average. The most frequent bedside initiators were residents, with 4.23 contacts/week of teleintensivist exposure. Overall, the mean (SD) rate of contacts per week of teleintensivist exposure was 7.54 (8.60).

Table 1.

Overview of intensive care unit telemedicine admissions by utilization counts

ICU telemedicine utilization typesa	Eligible admissions with any contacts (%)	Number of contacts within admission (%)				Contacts per admission	Contacts/week of TICUb exposurec
	(N = 3136)	1	2–4	5+	Max	Mean (SDd)	Mean (SD)
Any interaction	3136 (100.0)	1892 (60.3)	995 (31.7)	249 (8.0)	28	2.06 (2.34)	7.54 (8.60)
Initiation site
Bedside unit	2332 (74.4)	1522 (48.5)	664 (21.2)	146 (4.7)	20	1.37 (1.82)	5.31 (7.40)
ICU telemedicine	1132 (36.1)	796 (25.4)	301 (9.6)	35 (1.1)	20	0.56 (1.06)	1.83 (5.01)
Initiator at bedside
Attending physician	389 (12.4)	288 (9.2)	90 (2.9)	11 (0.4)	16	0.19 (0.70)	0.65 (3.24)
Registered nurse	381 (12.1)	307 (9.8)	68 (2.2)	6 (0.2)	8	0.16 (0.53)	0.42 (2.00)
Resident	1727 (55.1)	1119 (35.7)	497 (15.9)	111 (3.5)	18	1.01 (1.63)	4.23 (6.97)
Other personnel	25 (0.8)	25 (0.8)	0 (0.0)	0 (0.0)	1	0.01 (0.09)	0.02 (0.25)
Contact type
Routine intervention	1863 (59.4)	1280 (40.8)	501 (16.0)	82 (2.6)	20	0.99 (1.40)	3.63 (7.05)
Urgent intervention	558 (17.8)	481 (15.3)	75 (2.4)	2 (0.1)	7	0.21 (0.50)	1.06 (4.34)
Critical intervention	253 (8.1)	206 (6.6)	45 (1.4)	2 (0.1)	8	0.11 (0.43)	0.39 (3.05)
Contact outcome
Change in patient care	1709 (54.5)	1174 (37.4)	460 (14.7)	75 (2.4)	19	0.90 (1.34)	3.34 (6.96)

^aCategories within initiation site, initiator at bedside, and contact type are not mutually exclusive, as they are summarized by admission, eg, contact can occur in multiple ways for multiple reasons during a single patient stay.

^bTICU = intensive care unit telemedicine.

^cOperation hours differed by support center. Overall TICU exposure is therefore calculated as the amount of time during the patient’s admission when the teleintensivist was available.

^dSD = standard deviation.

We noted increasing usage trends in overall interaction rates, changes in patient care, and contact initiation by bedside unit and teleintensivist (Figure 3 ). Initiation rates by bedside physicians, nurses, and other personnel remained steady during our study period, while bedside resident use increased considerably by study end. Similarly, urgent and critical intervention rates remained level, while routine interventions were used increasingly over time.

Figure 3.

(A–D) Three-month moving average plots of ICU telemedicine rates, overall and by utilization category.

Of all teleintensivist contacts, 2819 interactions (43.6%) resulted in a patient care change, with Cincinnati reporting 1277 (19.7%) compared to 1542 (23.8%) in Minneapolis (Table 2). The percentage of patient care changes among all contacts varied by initiation site: 44.8% for bedside unit initiation compared to 51.3% by teleintensivist. Percentages of contacts resulting in patient care changes were 84.3%, 62.0%, 60.0%, and 35.2% when initiated by a registered nurse, an attending physician, other personnel, or a resident, respectively. The contact type most likely to result in a patient care change was routine (90.9%), followed by critical (82.8%) and urgent (71.4%) interventions.

Table 2.

Teleintensivist contacts resulting in a change in patient care

ICU telemedicine	Overall (N = 6470)		VISN 23 (N = 3112)		VISN 10 (N = 3358)
Utilization types N (%)	Total contacts	Contacts resulting in patient care change	Total contacts	Contacts resulting in patient care change	Total contacts	Contacts resulting in patient care change	P-value
Initiation site
Bedside unit	4293 (66.4)	1924 (44.8)	1873 (60.2)	1063 (56.8)	2420 (72.1)	861 (35.6)	<.001
ICU telemedicine	1744 (27.0)	895 (51.3)	925 (29.7)	479 (51.8)	819 (24.4)	416 (50.8)	.68
Initiator at bedside
Attending physician	587 (9.1)	364 (62.0)	542 (17.4)	336 (62.0)	45 (1.3)	28 (62.2)	.98
Registered nurse	509 (7.9)	429 (84.3)	353 (11.3)	288 (81.6)	156 (4.7)	141 (90.4)	.005
Resident	3172 (49.0)	1116 (35.2)	961 (30.9)	428 (44.5)	2,211 (65.8)	688 (31.1)	<.001
Other personnel	25 (0.4)	15 (60.0)	17 (0.6)	11 (64.7)	8 (0.2)	4 (50.0)	.51
Contact type
Routine interventions	3100 (47.9)	2819 (90.9)	1647 (52.9)	1542 (93.6)	1453 (43.3)	1277 (87.9)	<.001
Urgent interventions	657 (10.2)	469 (71.4)	207 (6.7)	165 (79.7)	450 (13.4)	304 (67.6)	<.001
Critical interventions	332 (5.1)	275 (82.8)	111 (3.6)	95 (85.6)	221 (6.6)	180 (81.5)	.33

DISCUSSION

Interactions between bedside staff and support center teleintensivists showed that overall use increased over time; when Tele-ICU was utilized, multiple interactions were common. Comparing utilization across support centers, we found significant differences in bedside initiator and urgent intervention contacts during night shift hours (Supplementary Table S1). Teleintensivists experienced more nighttime contacts (Supplementary Table S2), while teleintensivist-initiated contacts were more likely to result in a patient care change.

Only 21.6% of the 14 511 qualifying ICU admissions had an associated teleintensivist note in the electronic medical record. Typically, Tele-ICU staffing was nurse-intensive.^20,22,23 In both support centers, calls were fielded and transferred to the teleintensivist if the Tele-ICU nurse and initiating bedside staff decided the issue needed teleintensivist attention. The teleintensivist then determined whether the issue warranted documentation in the patient’s record. In these 2 support centers, 1 in 5 admissions rose to this level. Qualitative research in the VISN 23 ICU telemedicine program indicated that a considerable amount of undocumented Tele-ICU interaction occurred with telemedicine critical-care nurses.¹⁵ Additional factors, including patient severity, skilled triaging by telemedicine critical-care nurses, and infrequent documentation, may have contributed to this seemingly low rate.

Trend data indicated that utilization rates increased during the study, suggesting that bedside ICUs and support centers remained in the implementation phase. It is likely that telemedicine center use is still becoming a part of the ICU culture. Over time, we expect interactions to stabilize, though research is needed to understand the length of time it takes for this complex sociotechnological system to reach a plateau of effective utilization.

Qualitative research showed that bedside nurses appreciated access to Tele-ICU programs in order to quickly respond to changes in patient status and reported contacting teleintensivists for orders when they could not reach a local physician.^15,23 We found that 84.3% of contacts resulting in a change of care involved a bedside ICU nurse. This is evidence of the importance of bedside nurse and teleintensivist interaction, which could impact patient outcomes as well as nurse satisfaction rates.

We observed routine interventions by teleintensivists more frequently than urgent or critical interventions and note that these more commonly led to changes in patient care. It is possible that teleintensivists were able to detect gaps in routine critical care, such as deep vein thrombosis or gastrointestinal prophylaxis, and intervened accordingly. Further, residents require supervision by an attending physician to change a care plan. If bedside physicians were unavailable, teleintensivists could fill this need.

This study had several limitations. First, we relied on data collected using a standardized template in the VHA Tele-ICU program, which may not be generalizable. Second, teleintensivists may have charted interactions with bedside staff differently,¹⁵ and we could not validate the consistency of their note template use. Third, other “utilization” measures were not considered when operationalizing use: number of calls in and out of the support center, video initiations, or nurse interaction. However, this was the first study we are aware of to describe Tele-ICU utilization from a standardized physician note template and represents a first step in developing robust process measures for how and how often Tele-ICU should be used.

CONCLUSION

Previous Tele-ICU studies suggested that utilization might be an important moderating variable. We hypothesize that Tele-ICU utilization is driven by differing facility/support center protocols, implementation of quality improvement programs, and discretion afforded to Tele-ICU staff in interactions with bedside providers. These variations in use could impact Tele-ICU program assessment in terms of quality measurement, execution, and patient outcomes. Development of the teleintensivist note template is an important step in characterizing utilization of this clinical tool in critical-care settings; however, further research is needed to inform development of a standardized template for Tele-ICU nurses, who are involved in the majority of Tele-ICU interactions, and to validate template documentation. The validation research may result in standardized process measures to improve documentation of how Tele-ICU is used. With this research, more Tele-ICU networks could integrate these templates and the impact of utilization on patient outcomes could be further elucidated. By paying careful attention to Tele-ICU utilization measures and their association with improved patient outcomes, effective standardized operating procedures for Tele-ICU programs can be developed.

SUPPLEMENTARY MATERIAL

Supplementary material is available at Journal of the American Medical Informatics Association online.