A Novel Use of Bar Code Medication Administration Data to Assess Nurse Staffing and Workload

Abstract

Objective The aim of the study is to introduce an innovative use of bar code medication administration (BCMA) data, medication pass analysis, that allows for the examination of nurse staffing and workload using data generated during regular nursing workflow.

Methods Using 1 year (October 1, 2014–September 30, 2015) of BCMA data for 11 acute care units in one Veterans Affairs Medical Center, we determined the peak time for scheduled medications and included medications scheduled for and administered within 2 hours of that time in analyses. We established for each staff member their daily peak-time medication pass characteristics (number of patients, number of peak-time scheduled medications, duration, start time), generated unit-level descriptive statistics, examined staffing trends, and estimated linear mixed-effects models of duration and start time.

Results As the most frequent (39.7%) scheduled medication time, 9:00 was the peak-time medication pass; 98.3% of patients (87.3% of patient-days) had a 9:00 medication. Use of nursing roles and number of patients per staff varied across units and over time. Number of patients, number of medications, and unit-level factors explained significant variability in registered nurse (RN) medication pass duration (conditional R²  = 0.237; marginal R²  = 0.199; intraclass correlation = 0.05). On average, an RN and a licensed practical nurse (LPN) with four patients, each with six medications, would be expected to take 70 and 74 minutes, respectively, to complete the medication pass. On a unit with median 10 patients per LPN, the median duration (127 minutes) represents untimely medication administration on more than half of staff days. With each additional patient assigned to a nurse, average start time was earlier by 4.2 minutes for RNs and 1.4 minutes for LPNs.

Conclusion Medication pass analysis of BCMA data can provide health systems a means for assessing variations in staffing, workload, and nursing practice using data generated during routine patient care activities.

Keywords

nursing care - bar code medication administration - assessment - registered nurse - licensed practical nurse - workload

Background and Significance

Understanding the drivers and effects of variation in individual nurse practice is a key research question for nursing health services research.[1] Progress toward this goal to date has been hampered by limited availability of granular data at the individual and unit level.[2] While nurses interact with many technologies and data systems during the course of patient care, data generated as part of regular workflow are often not structured or standardized to accommodate nurse and unit-level assessment.[3] [4] [5] [6] [7] However, with nurses spending up to one-fifth of their time overall on medication tasks, staff activity recorded in bar code medication administration (BCMA) systems offers a promising source for granular nursing data.[8] [9] [10] [11] [12] [13] The nursing unit morning medication pass is a known period of high workload with known time constraints for accomplishing the work. The nurse medication pass has been explored qualitatively, through in-person observation and through surveys, but system-wide analysis of variation has not been possible using these resource-intensive methods.[14] [15] [16] [17] [18] [19] Leveraging high-quality electronic health data, such as that generated from BCMA systems, for research is a key goal in the National Health Information Technology Priorities for Research Policy and Development Agenda.[20] BCMA data create the possibility of large-scale quantitative exploration of variation in the medication pass between staff, across locations, and over time.

BCMA systems are in place at 98% of U.S. hospitals and have been in place for over two decades at Veterans Health Administration (VA) hospitals, where BCMA originated.[21] [22] [23] [24] [25] Adoption of BCMA systems has been associated with improvements in medication safety.[26] [27] [28] [29] [30] [31] Only recently have researchers started to explore BCMA data for insights into nursing practice.[32] [33]

Objective

When BCMA tools are used as intended, a nurse will scan the barcode on a patient's wristband to confirm their identity and scan the barcodes for the medications to be administered, all while at the patient bedside.[34] A scan time is recorded for each medication, and the patient's location is associated with the medication scan. BCMA data can be used to place a staff member with a patient at a specific time in a particular place. In this way, a nurse's progress through the medication pass and from patient to patient is documented via BCMA. Most medications in hospitals are scheduled for standard times, and U.S. national guidelines state that medications should be administered within ± 60 minutes from the scheduled time (± 30 minutes for time critical medications).[35] [36]

Examining each staff member's administration of medications scheduled for a single time allows for comparison of individual workload and process for the same time period within a hospital or unit. The scheduled time with the highest volume of medications defines the peak-time medication pass and this high medication workload time period is a prime place to examine variation in staffing, workload, and nursing practice.

Using data from one hospital for 1 year, this paper will introduce an innovative use of BCMA data, medication pass analysis, that allows for examination of nurse staffing, workload, and practice using data generated during regular nursing workflow.

Methods

Data Sources

Data were obtained from the BCMA system for fiscal year 2015 (FY15; October 1, 2014–September 30, 2015) for acute care units at one large, urban, academic, VA medical center. FY15 reflects a time when this hospital employed both registered nurses (RNs) and licensed practical nurses (LPNs) on inpatient units, allowing for examination of differences in utilization and practice of these nursing roles across units and over time, an important consideration as staffing challenges with coronavirus disease 2019 (COVID-19) prompt reconsideration of alternative staffing models. At this hospital, LPNs administered oral, subcutaneous, and piggyback intravenous (IV) medications but not cardiac IV, emergency, experimental, or chemotherapeutic medications, or any medications delivered via IV push, patient-controlled analgesic pump, central line, or peripherally inserted central catheter. By January 2018, the hospital no longer employed LPNs in the acute inpatient setting, necessitating the use of this older data. We used human resources data to link staff with their occupations, and the VA's Nursing Unit Mapping Application (NUMA) was used to link patient bed locations to inpatient nursing units.[37]

Approach

The patient bed location associated with each dose in the BCMA data was matched with an inpatient nursing unit using NUMA. Doses were excluded when the patient bed location could not be mapped to an inpatient nursing unit. The most frequent time for scheduled medications was taken to represent the peak-time (highest volume) medication pass period. The set of peak-time medications scanned by each staff member constitutes the staff member's scheduled BCMA medication workload for the peak medication pass period. Medications scanned greater than 2 hours from the scheduled time were excluded from analysis.

Staff members were identified as RNs, LPNs, or another role. For each staff member on each day, workload- and time-related features of their peak-time medication pass were assessed: number of patients, number of scanned medications, start time of medication pass (scan time for the first medication administered to the staff member's first patient) and, for staff with two or more patients, medication pass duration (number of minutes from time of scan for first patient's first medication until time of scan for last patient's last medication). Daily and yearly summaries were calculated by role, by unit, and hospital wide, and trends in staffing and workload over time were examined. Median values are reported in the text and means are reported in the tables ([Table 1] and [Table 2]).

A linear mixed effects model was estimated using restricted maximum likelihood with medication pass duration and start time as dependent variables. Analyses were conducted in R using the lme4 package.[38] Effects of units were treated as fixed thus controlling for all unit-level differences in patient burden and staff resources. Variability in outcome across staff was treated as random. Weekdays and months were included to account for temporal variability. Staff-level predictors for each medication pass included number of patients and number of medications administered. Corresponding means of peak-time patients-per-staff and medications-per-staff were included as predictors at the unit-day level. The daily counts of the unit-level peak-time number of patients and number of staff were included as covariates. Identical models were estimated separately across RNs and LPNs, for units where the respective roles had peak-time BCMA activity >150 days.

Results

In FY15, 1,373,596 acute-care medication doses were recorded in the BCMA system at the study hospital. Of these, 99.2% were associated with a patient bed location mapped to an acute-care nursing unit. Scheduled medications constituted 83.8% of all doses mapped to a unit.

As the highest frequency time (39.7%; 453,310) for scheduled medication doses mapped to a unit ([Fig. 1]), 9:00 was the peak time for scheduled medications; 98.3% of patients (87.3% of patient-days) had a 9:00 medication. 96.0% (435,207) of peak-time medications had a scan time within ± 2 hours of the scheduled time, and these constituted the final dataset for analysis ([Table 1]).

Table 1 Summary of FY15 BCMA activity for medications scanned by staff within 2 h of 9:00 scheduled time

Location/Unit	Role	Days with peak-time BCMA activity count	Medication doses scanned Count (% of location total, % of role total)	Patient days Count (% of location total, % of role total)	Unique patients Count (% of location total, % of role total)	Unique daily patients Mean (SD)	Unique staff Count (% of location total, % of role total)	Unique daily staff on days when role is present Mean (SD)	Unique daily staff across 365 d Mean
Hospital	All staff	365	435,207 (100, 100)	63,238 (100, 100)	8,587 (100, 100)	173.3 (23.0)	444 (100, 100)	70.1 (4.7)	70.1
	RN	365	363,821 (84, 100)	55,387 (88, 100)	8,181 (95, 100)	151.7 (21.6)	366 (82, 100)	58.9 (4.6)	58.9
	LPN	360	55,611 (13, 100)	8,055 (13, 100)	3,113 (36, 100)	22.4 (9.6)	34 (8, 100)	3.8 (1.5)	3.7
	Other staff	365	15,775 (4, 100)	8,814 (14, 100)	1,661 (19, 100)	24.1 (5.7)	44 (10, 100)	7.5 (1.0)	7.5
Medical-1	All staff	365	53,342 (100, 12)	7,562 (100, 12)	1,584 (100, 18)	20.7 (4.0)	132 (100, 30)	7.0 (1.0)	7.0
	RN	365	47,297 (89, 13)	6,892 (91, 12)	1,536 (97, 19)	18.9 (4.6)	89 (67, 24)	5.6 (0.9)	5.6
	LPN	161	4,684 (9, 8)	689 (9, 9)	429 (27, 14)	4.3 (1.9)	10 (8, 29)	1.0 (0.2)	0.5
Medical-2	All staff	365	51,329 (100, 12)	7,261 (100, 11)	1,692 (100, 20)	19.9 (4.0)	90 (100, 20)	7.2 (0.9)	7.2
	RN	364	43,395 (85, 12)	6,312 (87, 11)	1,608 (95, 20)	17.3 (4.1)	47 (52, 13)	5.6 (0.8)	5.6
	LPN	244	6,426 (13, 12)	941 (13, 12)	532 (31, 17)	3.9 (1.5)	8 (9, 24)	1.1 (0.3)	0.7
Medical-3	All staff	365	51,168 (100, 12)	7,188 (100, 11)	1,708 (100, 20)	19.7 (3.9)	73 (100, 16)	6.2 (1.2)	6.2
	RN	361	39,037 (76, 11)	5,701 (79, 10)	1,558 (91, 19)	15.8 (5.8)	34 (47, 9)	4.5 (1.3)	4.5
	LPN	217	10,882 (21, 20)	1,591 (22, 20)	812 (48, 26)	7.3 (4.8)	5 (7, 15)	1.4 (0.6)	0.8
Medical-4	All staff	365	50,208 (100, 12)	6,859 (100, 11)	1,637 (100, 19)	18.8 (3.4)	85 (100, 19)	6.1 (1.6)	6.1
	RN	336	21,577 (43, 6)	3,183 (46, 6)	1,222 (75, 15)	9.5 (5.4)	40 (47, 11)	4.5 (1.6)	4.2
	LPN	306	27,365 (55, 49)	3,826 (56, 47)	1,253 (77, 40)	12.5 (4.7)	11 (13, 32)	1.3 (0.5)	1.1
Surgical-1	All staff	365	35,926 (100, 8)	5,852 (100, 9)	1,533 (100, 18)	16.0 (3.8)	76 (100, 17)	6.1 (1.1)	6.1
	RN	365	34,735 (97, 10)	5,744 (98, 10)	1,523 (99, 19)	15.7 (3.8)	44 (58, 12)	5.4 (1.0)	5.4
	LPN	27	619 (2, 1)	105 (2, 1)	91 (6, 3)	3.9 (1.5)	5 (7, 15)	1.0 (0.0)	0.1
Surgical-2	All staff	365	33,978 (100, 8)	5,570 (100, 9)	1,571 (100, 18)	15.3 (4.0)	84 (100, 19)	5.6 (1.3)	5.6
	RN	364	28,289 (83, 8)	4,718 (85, 9)	1,451 (92, 18)	13.0 (4.6)	45 (54, 12)	4.7 (1.4)	4.7
	LPN	151	5,229 (15, 9)	859 (15, 11)	514 (33, 17)	5.7 (1.6)	10 (12, 29)	1.0 (0.1)	0.4
Step Down-1	All staff	365	66,566 (100, 15)	8,692 (100, 14)	1,575 (100, 18)	23.8 (3.3)	145 (100, 33)	9.4 (1.2)	9.4
	RN	365	62,711 (94, 17)	8,630 (99, 16)	1,571 (100, 19)	23.6 (3.3)	99 (68, 27)	7.5 (1.0)	7.5
	LPN	34	396 (1, 1)	42 (0, 1)	21 (1, 1)	1.2 (1.2)	10 (7, 29)	1.0 (0.2)	0.1
Step Down-2	All staff	365	39,188 (100, 9)	5,480 (100, 9)	1,147 (100, 13)	15.0 (3.2)	98 (100, 22)	6.5 (0.8)	6.5
	RN	365	36,664 (94, 10)	5,457 (100, 10)	1,145 (100, 14)	15.0 (3.3)	66 (67, 18)	5.6 (0.8)	5.6
Critical Care-1	All staff	365	22,530 (100, 5)	4,134 (100, 7)	1,089 (100, 13)	11.3 (3.2)	97 (100, 22)	8.4 (1.8)	8.4
	RN	365	21,072 (94, 6)	4,119 (100, 7)	1,086 (100, 13)	11.3 (3.2)	69 (71, 19)	7.3 (1.5)	7.3
Critical Care-2	All staff	365	15,819 (100, 4)	2,541 (100, 4)	610 (100, 7)	7.0 (1.5)	57 (100, 13)	5.8 (0.9)	5.8
	RN	365	14,452 (91, 4)	2,533 (100, 5)	609 (100, 7)	6.9 (1.5)	35 (61, 10)	4.8 (0.7)	4.8
Critical Care-3	All staff	364	15,153 (100, 3)	2,102 (100, 3)	635 (100, 7)	5.8 (1.8)	59 (100, 13)	4.5 (1.1)	4.4
	RN	364	14,592 (96, 4)	2,100 (100, 4)	633 (100, 8)	5.8 (1.8)	31 (53, 8)	3.9 (0.8)	3.9

Abbreviations: LPN, licensed practical nurse; RN, registered nurse.

Note: LPN data not reported for units with no LPN data or LPN data for less than 11 d. Other staff data reported only at the hospital level.

Fig. 1 Frequency of scheduled medications by scheduled time.

On 99.8% (63,099) of peak-time patient-days, the patient received a medication from a nurse (RN and/or LPN). On step down units and critical care units, >99% of peak-time patient-days included medication administered by an RN. On medical and surgical units, 79.3 to 98.2% of patient-days included medication administered by an RN, except for unit Medical-4 (46.4%). The median number of medications per patient was 7 on all medical units, 6 on all surgical units, 6 to 7 on stepdown units, and 5 to 7 on critical care units.

Staff administering at least one peak-time medication included RNs (n = 366; 82.4%), LPNs (n = 34; 7.7%), and other staff who were primarily respiratory therapists or respiratory technicians (n = 44; 10.0%).

Sorting the peak-time medication data by staff member and day, there were 25,592 staff days, or medication pass instances, with 84.0% performed by RNs, 5.3% by LPNs, and 10.7% by other staff ([Table 2]). In nearly all cases, RNs and LPNs administered medications on only one unit during the peak-time medication pass, while other staff often administered medications across multiple units. Therefore, data are reported at the unit-level for RNs and LPNs and hospital wide for other staff. The median number of patients per RN across all acute care units was three, including one to two patients per RN for critical care units and two to four patients per RN on medical, surgical, and step-down units. Patients per LPN ranged from median four to six on units with LPN activity >150 days, except for Unit Medical-4 with ten patients per LPN, double that on other units. Hospital wide, other staff had a median of three patients per medication pass.

Table 2 Summary of individual staff peak-time medication pass instances for FY15

Location/unit	Role	Staff medication pass instances with one or more patients (M1) Count (% of location total; % of role total)	Patients per medication pass Mean (SD)	Medication doses per medication pass Mean (SD)	Difference in minutes between scheduled time and scan time for the first medication for the first patient of a staff medication pass instance (Start) Mean (SD)	Staff medication pass instances with two or more patients (M2) Count (% of M1)	Difference in minutes between the scan time of the first and last medications of a staff medication pass instance across all patients (Duration) Mean (SD)
Hospital	All staff	25,592 (100, 100)	2.8 (1.6)	17.0 (12.6)	–33.6 (26.0)	20,901 (82)	57.5 (39.4)
	RN	21,503 (84, 100)	2.6 (1.1)	16.9 (9.9)	–32.8 (25.8)	17,447 (81)	58.2 (38.5)
	LPN	1,353 (5, 100)	6.0 (3.1)	41.1 (24.2)	–38.7 (22.5)	1,280 (95)	87.1 (44.0)
	Other staff	2,736 (11, 100)	3.2 (1.9)	5.8 (3.5)	–36.6 (28.9)	2,174 (79)	34.4 (29.7)
Medical-1	All staff	2,541 (100, 10)	3.3 (1.1)	21.0 (11.0)	–38.9 (24.8)	2,344 (92)	63.2 (38.9)
	RN	2,055 (81, 10)	3.4 (1.0)	23.0 (9.3)	–41.4 (22.9)	1,953 (95)	67.7 (37.3)
	LPN	168 (7, 12)	4.1 (1.5)	27.9 (12.0)	–33.7 (22.5)	163 (97)	68.2 (37.2)
Medical-2	All staff	2,627 (100, 10)	3.1 (1.1)	19.5 (10.2)	–33.9 (23.7)	2,444 (93)	60.8 (38.1)
	RN	2,028 (77, 9)	3.1 (0.9)	21.4 (8.9)	–34.6 (22.1)	1,948 (96)	65.9 (37.5)
	LPN	270 (10, 20)	3.5 (1.0)	23.8 (9.2)	–31.9 (25.4)	259 (96)	57.9 (30.3)
Medical-3	All staff	2,263 (100, 9)	3.6 (1.5)	22.6 (13.8)	–38.5 (26.2)	2,026 (90)	63.9 (42.2)
	RN	1,632 (72, 8)	3.5 (1.1)	23.9 (11.1)	–39.1 (25.7)	1,502 (92)	66.5 (39.1)
	LPN	310 (14, 23)	5.1 (1.8)	35.1 (14.3)	–43.5 (24.9)	305 (98)	83.5 (45.7)
Medical-4	All staff	2,215 (100, 9)	3.5 (3.2)	22.7 (25.0)	–23.8 (28.5)	1,624 (73)	61.4 (47.6)
	RN	1,518 (69, 7)	2.1 (1.0)	14.2 (9.4)	–17.5 (28.4)	1,047 (69)	46.2 (33.2)
	LPN	390 (18, 29)	9.8 (2.2)	70.2 (19.8)	–44.6 (14.6)	380 (97)	124.5 (29.5)
Surgical-1	All staff	2,215 (100, 9)	2.8 (1.0)	16.2 (9.1)	–31.7 (23.3)	1,963 (89)	56.3 (36.1)
	RN	1,989 (90, 9)	2.9 (1.0)	17.5 (8.4)	–33.3 (21.0)	1,864 (94)	58.0 (35.6)
	LPN	27 (1, 2)	3.9 (1.5)	22.9 (10.6)	–23.4 (32.6)	23 (85)	69.4 (27.4)
Surgical-2	All staff	2,053 (100, 8)	2.9 (1.3)	16.6 (10.2)	–36.7 (21.2)	1,792 (87)	52.7 (39.5)
	RN	1,705 (83, 8)	2.8 (0.9)	16.6 (7.7)	–39.2 (18.0)	1,586 (93)	52.3 (39.4)
	LPN	154 (8, 11)	5.6 (1.7)	34.0 (13.1)	–38.2 (18.5)	148 (96)	71.9 (34.9)
Step Down-1	All staff	3,427 (100, 13)	3.1 (1.0)	19.4 (10.8)	–40.6 (25.0)	3,130 (91)	61.1 (38.0)
	RN	2,742 (80, 13)	3.2 (0.9)	22.9 (9.2)	–42.1 (23.6)	2,611 (95)	67.4 (36.6)
	LPN	35 (1, 3)	1.2 (1.2)	11.3 (7.2)	–16.3 (25.5)	[a]	[a]
Step Down-2	All staff	2,385 (100, 9)	2.9 (1.2)	16.4 (8.3)	–41.2 (19.5)	2,239 (94)	50.3 (37.4)
	RN	2,034 (85, 9)	2.7 (0.7)	18.0 (7.7)	–41.8 (18.5)	1,931 (95)	53.6 (38.1)
Critical Care-1	All staff	3,064 (100, 12)	1.6 (0.6)	7.4 (4.4)	–19.0 (27.8)	1,714 (56)	38.0 (35.9)
	RN	2,659 (87, 12)	1.6 (0.5)	7.9 (4.4)	–18.1 (26.2)	1,494 (56)	39.7 (36.1)
Critical Care-2	All staff	2,103 (100, 8)	1.6 (0.7)	7.5 (4.5)	–27.1 (29.0)	1,040 (49)	45.4 (38.7)
	RN	1,741 (83, 8)	1.5 (0.5)	8.3 (4.5)	–24.9 (28.7)	815 (47)	51.3 (39.4)
Critical Care-3	All staff	1,622 (100, 6)	1.5 (0.5)	9.3 (6.1)	–26.2 (29.9)	782 (48)	44.7 (40.1)
	RN	1,406 (87, 7)	1.5 (0.5)	10.4 (5.9)	–25.3 (29.6)	695 (49)	48.9 (40.1)

Abbreviations: LPN, licensed practical nurse; RN, registered nurse.

Note: Duration is calculated only for medication pass instances with two or more patients. LPN data not reported for units with no LPN data or LPN data for less than 11 d. Other staff data reported only at the hospital level.

^a Data for less than 11 medication pass instances.

Examination of data over time shows the range of variation within units for number of patients and patients per staff ([Figs. 2] [3] [4]). Unit Medical-3 ([Fig. 5]) appears to have shifted their staffing model for medication administration midway through FY15, increasing the number of patients receiving medication from LPNs and increasing the number of patients per LPN for the peak-time medication pass. This suggests that unit-level shifts in staffing mix or distribution of medication workload may be detectable in BCMA data.

Fig. 2 Patients per staff for RNs during the peak-time medication pass each day for each unit-by-unit-type during FY15. RN, registered nurse

Fig. 3 Patients per staff for LPNs during the peak-time medication pass each day for each unit (with >150 days with LPNs) by unit-type during FY15. LPN, licensed practical nurse.

Fig. 4 Number of patients receiving peak-time medications each day on each unit-by-unit-type during FY15.

Fig. 5 A mid-year shift in the medication administration staffing pattern on unit Medical-3: number of patients, number of staff, and patients per staff by role and over time. (A and B) Daily number of patients receiving peak-time medications from RNs (A) and LPNs (B). (C and D) Daily number of RNs (C) and LPNs (D) administering peak-time medications each day. (E and F) Daily number of patients receiving peak-time medications per RN (E) and per LPN (F). LPN, licensed practical nurse; RN, registered nurse.

Across all RNs, a median of 16 peak-time medications was scanned daily ([Table 2]). Hospital wide, LPNs scanned a median 33 peak-time medications each. On unit Medical-4, the median number of medications per LPN was 72, a very high medication workload for LPNs. For other staff hospital wide, the median number of peak-time medications was five per staff member.

Medication pass duration was calculated for medication pass instances with two or more patients (n = 20,901; 82% of all medication pass instances). On critical care units RN peak-time medication pass duration ranged from median 27 to 33 minutes, and 36 to 61 minutes across all other unit types. Median medication pass duration for other staff hospital wide was 26 minutes. Medication pass duration ranged from median 52 to 70 minutes for LPNs on units with >150 days with LPN activity, except for unit Medical-4 with median 127 minutes per LPN. This duration indicates that LPNs on unit Medical-4 were frequently unable to administer all assigned peak-time medications on time (± 60 minutes from scheduled time). It is likely that this is due to high LPN workload on unit Medical-4 (median ten patients, 72 medications per peak-time medication pass). Relationships between workload (number of patients and number of medications) and duration for RNs and LPNs were explored via linear mixed effects models. Due to movement across units, other staff were excluded from the models.

[Tables 3] and [4] present parameter estimates for the linear mixed effects models for duration and start time, respectively.

Table 3 Parameter estimates for linear mixed effects model of duration

	Duration (RN)[a]			Duration (LPN)[a]
Predictors[b]	Estimates	CI	p-Value	Estimates	CI	p-Value
Intercept	29.64	23.72 to 35.56	<0.001	49.51	27.62 to 71.39	<0.001
Medical-1	1.25	−1.97 to 4.46	0.447	−2.95	−14.42 to 8.51	0.613
Medical-2	4.46	1.33 to 7.59	0.005	3.04	−9.70 to 15.79	0.640
Medical-3	−3.04	−6.68 to 0.59	0.101	5.20	−4.77 to 15.17	0.306
Medical-4	−4.84	−8.57 to −1.11	0.011	−4.84	−15.50 to 5.83	0.374
Surgical-1	−1.22	−4.49 to 2.04	0.462
Surgical-2	−1.36	−4.60 to 1.89	0.413
Step Down-1	3.55	0.32 to 6.79	0.031
Step Down-2	−0.69	−3.63 to 2.26	0.648
Critical Care-2	3.74	−0.37 to 7.84	0.074
Critical Care-3	0.38	−4.37 to 5.12	0.876
Sunday	−2.36	−3.56 to −1.16	<0.001	0.10	−4.49 to 4.69	0.965
Monday	−0.45	−1.67 to 0.76	0.465	−1.34	−5.36 to 2.69	0.514
Tuesday	0.18	−0.97 to 1.33	0.760	−1.33	−5.69 to 3.04	0.552
Wednesday	0.22	−0.94 to 1.38	0.711	−2.96	−6.61 to 0.69	0.112
Thursday	0.65	−0.51 to 1.80	0.273	1.37	−2.89 to 5.63	0.528
Friday	2.41	1.25 to 3.57	<0.001	5.71	1.58 to 9.84	0.007
January	−1.30	−2.86 to 0.26	0.101	−2.96	−8.52 to 2.61	0.297
February	0.09	−1.49 to 1.68	0.908	−0.37	−6.21 to 5.47	0.901
March	−1.17	−2.72 to 0.38	0.138	−1.38	−6.77 to 4.01	0.616
April	−0.80	−2.44 to 0.85	0.342	9.30	3.95 to 14.65	0.001
May	−0.79	−2.45 to 0.88	0.354	3.41	−1.97 to 8.78	0.214
June	0.76	−0.87 to 2.39	0.362	4.26	−1.73 to 10.24	0.163
July	0.62	−1.05 to 2.29	0.468	−0.43	−6.47 to 5.61	0.888
August	1.76	0.06 to 3.45	0.042	1.39	−5.07 to 7.85	0.673
September	0.25	−1.38 to 1.89	0.763	−2.17	−8.39 to 4.05	0.495
October	−1.34	−2.92 to 0.24	0.095	0.12	−4.94 to 5.19	0.962
November	0.97	−0.64 to 2.57	0.237	−3.95	−9.67 to 1.77	0.175
Number of medication[c]	0.61	0.52 to 0.70	<0.001	0.37	0.15 to 0.59	0.001
Mean number medications[d]	0.12	−0.08 to 0.32	0.240	−0.05	−0.51 to 0.41	0.824
Number of patients[c]	13.36	12.34 to 14.39	<0.001	7.94	5.91 to 9.97	<0.001
Mean number of patients[d]	0.04	−1.76 to 1.83	0.968	0.64	−3.02 to 4.30	0.732
Number of patients[e]	−0.57	−1.02 to −0.12	0.014	−0.95	−2.27 to 0.38	0.162
Number of nurses[e]	0.09	−1.04 to 1.22	0.881	−2.57	−7.21 to 2.06	0.277
Number of patients × number of nurses[e]	0.03	−0.03 to 0.10	0.301	0.15	−0.11 to 0.41	0.267
Random effects
σ 2	1,010.21			868.82
τ00	49.91 staff_id			138.78staff_id
ICC	0.05			0.14
N	265 staff_id			13staff_id
Observations	17,446			1,255
Marginal R 2/conditional R 2	0.199/0.237			0.474/0.546
Deviance	170,480.457			12,055.405
AIC	170,524.133			12,035.620
AICc	170,524.295			12,037.242

Note: Bolded values denote p < 0.05

Abbreviations: LPN, licensed practical nurse; RN, registered nurse.

^a In minutes.

^b Categorical variables (units, days, and months) were coded using sum contrasts. Each level in the table reflects deviation from the grand mean. For the RN model the expected duration for unit Critical Care-1 is 2.23 min less than the grand mean. For the LPN model the expected duration for unit Surgical-2 is 0.45 min less than the grand mean. All predictors were centered at 2.

^c Staff-daily level.

^d Unit-daily mean across staff on a given day.

^e Unit-daily.

Table 4 Parameter estimates for linear mixed effects model of start time

	Start time (RN)[a]			Start time (LPN)[a]
Predictors[b]	Estimates	CI	p-Value	Estimates	CI	p-Value
Intercept	490.79[c]	486.22–495.36	<0.001	513.86[c]	492.37–535.36	<0.001
Medical-1	2.03	−1.95 to 6.00	0.317	−0.15	−7.86 to 7.56	0.970
Medical-2	−9.28	−13.57 to −5.00	<0.001	−10.30	−20.98 to 0.38	0.059
Medical-3	−9.24	−15.32 to −3.16	0.003	−2.99	−12.02 to 6.03	0.515
Medical-4	10.30	5.42–15.18	<0.001	8.93	1.16–16.70	0.024
Surgical-1	−4.78	−9.06 to −0.50	0.029
Surgical-2	−9.39	−13.90 to −4.88	<0.001
Step Down-1	−3.66	−7.88 to 0.56	0.089
Step Down-2	−4.88	−8.54 to −1.22	0.009
Critical Care-2	12.03	7.15–16.91	<0.001
Critical Care-3	3.89	−2.51 to 10.29	0.234
Sunday	−3.03	−3.74 to −2.31	<0.001	0.03	−2.59 to 2.64	0.984
Monday	0.21	−0.50 to 0.93	0.559	0.06	−2.27 to 2.40	0.957
Tuesday	−1.06	−1.75 to −0.37	0.002	−1.32	−3.83 to 1.20	0.306
Wednesday	2.55	1.86–3.24	<0.001	−0.02	−2.12 to 2.07	0.984
Thursday	1.22	0.53–1.91	0.001	−1.93	−4.40 to 0.54	0.125
Friday	0.75	0.06–1.43	0.033	0.81	−1.57 to 3.19	0.503
January	0.45	−0.50 to 1.40	0.351	−1.59	−4.81 to 1.62	0.331
February	1.22	0.25–2.18	0.013	−1.25	−4.62 to 2.11	0.465
March	0.91	−0.03 to 1.85	0.058	−2.14	−5.23 to 0.95	0.175
April	0.11	−0.86 to 1.08	0.820	−1.50	−4.57 to 1.58	0.340
May	1.16	0.18–2.14	0.020	−1.17	−4.28 to 1.94	0.462
June	0.27	−0.70 to 1.24	0.582	−0.45	−3.91 to 3.01	0.798
July	−1.71	−2.68 to −0.74	0.001	−0.43	−3.85 to 2.99	0.806
August	−0.87	−1.85 to 0.11	0.082	1.91	−1.77 to 5.58	0.308
September	1.21	0.24–2.18	0.015	5.17	1.67–8.68	0.004
October	0.36	−0.59 to 1.30	0.456	1.62	−1.31 to 4.56	0.278
November	−1.57	−2.52 to −0.61	0.001	0.06	−3.29 to 3.40	0.973
Number of medications[d]	−0.25	−0.31 to −0.20	<0.001	−0.05	−0.18 to 0.07	0.420
Mean number of Medications[e]	−0.06	−0.21 to 0.09	0.439	0.07	−0.30 to 0.44	0.716
Number of patients[d]	−4.20	−4.76 to −3.64	<0.001	−1.44	−2.54 to −0.34	0.010
Mean number of patients[e]	5.30	4.04–6.56	<0.001	−0.41	−3.11 to 2.29	0.766
Number of patients[f]	0.53	0.17–0.88	0.004	0.10	−0.90 to 1.10	0.843
Number of nurses[f]	0.93	0.32–1.53	0.003	−0.51	−3.50 to 2.49	0.740
Number of patients × number of nurses[f]	−0.02	−0.06 to 0.01	0.234	0.08	−0.07 to 0.23	0.313
Random effects
σ2	436.09			292.58
τ00	706.90 staff_id			1,373.18 staff_id
ICC	0.62			0.82
N	366 staff_id			27 staff_id
Observations	21,509			1,292
Marginal R 2/conditional R 2	0.085/0.651			0.017/0.827
Deviance	192,983.053			11,070.905
AIC	193,044.013			11,076.279
AICc	193,044.144			11,077.853

Abbreviations: AIC, Akaike information criterion; AICc, small-sample corrected Akaike information criterion; LPN, licensed practical nurse; RN, registered nurse.

^a Minutes from scheduled time.

^b Categorical variables (units, days, and months) were coded using sum contrasts. Each level in the table reflects deviation from the grand-mean. For the RN model the expected start time for unit Critical Care-1 is 12.98 min later than the grand mean. For the LPN model the expected start time for unit Surgical-2 is 4.51 min later than the grand mean. Number of patients was centered at one patient. All other continuous predictors were centered at 2.

^c Minutes since midnight.

^d Staff daily.

^e Unit-daily mean across staff on a given day.

^f Unit daily.

RN Medication Pass Duration: the model explained significant variability in duration (conditional R²  = 0.237). Variance explained due to fixed effects was 0.199. Intraclass correlation for the model was 0.05, suggesting that much of the variability in duration was due to daily unit-level factors. The average medication pass duration for an RN with two patients receiving one medication each was 29.64 minutes (95% CI: [23.76, 35.56], p <0.001). There was a significant variability across RNs (SD = 7.05 minutes). For every additional patient, an RN spent 13.36 additional minutes (95% CI: [12.34, 14.39], p <0.001) and for every additional medication an additional 0.61 minutes was spent (95% CI: [0.52, 0.70], p <0.001). At the unit level, every additional patient on a given day resulted in shortening of duration by 0.57 minutes (95% CI: [−1.02, −0.12], p <0.014). Changes in unit-level number of staff did not have a significant effect on individual RN medication pass duration. On average, an RN with four patients, each with six medications, would be expected to take 70 minutes to complete the peak-time medication pass.

RN Start Time: the model explained a significant proportion of variance (conditional R²  = 0.651) in medication pass start time for RNs. Fixed predictors explained 8.5% of variability in start time. Conditional intraclass correlation was 0.62 suggesting that a significant portion of variability in start time was due to individual RNs. Expected start time for an RN with one patient receiving two medications was 8:11 (SD 26.59 minutes). There was a significant variability across units (−9.4, 12.0) with seven out of 11 units deviating from the average. This suggests that RN practices vary according to local unit conditions. For each additional patient, RNs started earlier by 4.20 minutes (95% CI: [−4.76, −3.64], p <0.001). For every additional medication, start time was 0.25 minutes earlier (95% CI: [−0.31, −0.20], p <0.001). At the unit level, as the average number of patients per staff increased on a given day, the expected start time was delayed by 5.30 minutes (95% CI: [4.04, 6.56], p <0.001). An increase in the number of patients and staff in the unit delayed the medication pass start time by 0.53 minutes (95% CI: [0.17, 0.88], p <0.01) and 0.93 (95% CI: [0.32, 1.53], p <0.01), respectively.

LPN Duration: the model explained a substantial proportion of variability (marginal R²  = 0.474, conditional R²  = 0.546) in LPN duration. Expected duration for an LPN with two patients, each with one medication, was 49.5 minutes. There was a significant variability across LPNs (SD = 11.8 minutes). LPN duration did not vary across units suggesting that the nature of work for LPNs is similar in these units. At the individual level, each additional patient and medication lengthened LPN duration by 7.94 (95% CI: [5.91, 9.97]) and 0.37 (95% CI: [0.15, 0.59]) minutes, respectively. Other predictors had no significant effect. On average, an LPN with four patients, each with six medications, would be expected to take 74 minutes to complete the peak-time medication pass.

LPN Start Time: for LPN medication pass start time, fixed predictors did not explain significant variability (Marginal R ² = 0.02). Expected start time for an LPN with a single patient receiving two medications was 8:34 (SD 37.06 minutes), and LPNs started the medication pass on average 1.4 minutes earlier for each additional patient assigned to them (95% CI: [−2.54, −0.34]). Intraclass correlation was 0.82, suggesting that a significant proportion of variability in start time was attributable to individual LPNs.

Discussion

To our knowledge, this is the first study showing that BCMA data can be used to quantitatively examine the individual staff medication pass for a given scheduled medication time allowing for description of point-in-time variation in individual workload (as measured by number of patients and number of medications administered) and process of care across different types of staff and acute inpatient units. An initial set of descriptors for the individual medication pass derived from BCMA data includes: scheduled medication time, staff member, unit(s), number of patients, number of medications, start time, and duration. Staff role and unit name required linking with data outside of BCMA. Descriptive summaries of the individual medication pass showed clear differences in staffing patterns and workload by unit and by role. Differences across units even of the same unit type were considerable, suggesting that aggregating at the level of unit type would obscure important unit-level differences. Linear mixed effects models of medication pass duration and start time offered preliminary evidence for how workload relates to nursing practice. Higher workload in the form of number of patients and number of medications was a key driver in lengthening the duration of an individual nurse's peak-time medication pass, and while individual variation was present, factors other than the individual staff member were the greater source of variation. It is likely that this variation is related to local conditions or patient-related factors not measured in this study. Individual variation predominated in models of start time, though generally staff began their medication pass earlier when assigned more patients and more medications.

Our findings are consistent with an observational study prior to the adoption of BCMA, where nurses took more than 15 minutes to administer medications to a single patient.[14] In addition, a systematic review identified three studies reporting that the amount of time spent by nurses on medication administration decreased after adoption of BCMA.[13] The contribution of time due to medications in the duration models is small relative to the time attributed to patients. This aligns with the notion that there are fixed costs (in time) for gathering the medications for, traveling to, ensuring the identity of, assessing, engaging with, and otherwise attending to the needs of patients.[39] [40] [41] [42] [43] [44] [45] [46] [47] [48]

Our findings suggest that higher numbers of patients assigned to nurses may be incompatible with timely medication administration. Timing errors in medication administration are frequent but often discounted in reports of medication errors.[29] [49] [50] [51] [52] [53] [54] [55] The persistent presence of wrong-time errors indicates that nurses are often unable to administer all medications assigned to them in the allotted time. This has implications for non-medication patient care as well. As the number and complexity of patients assigned to a given nurse increases, more time may be required to administer medications, and less time is available to attend to other patient needs, potentially resulting in missed care.[7] [8] [56] [57]

Our findings also suggest that variation in medication pass start time is greatest at the individual staff level and that nurses modify the start time of their medication pass in response to the workload assigned to them, adapting their behavior in response to an assessment of the work required for a given patient assignment. These data are consistent with nurses using critical thinking to organize their temporal schedule.[58] With more patients and medications on a given day, they tend to begin the medication pass earlier, perhaps to offset the greater time required for more patients and an awareness of the need to meet the medication timeliness standard regardless of their patient load. This is in tension with unit-level factors (i.e., average patients per staff, number of patients, and number of staff) that are associated with delayed start time. Reasons for these unit-level delays (e.g., competition for resources, geographic layout of larger units) could be further explored.

This work is similar to audit log studies in that we utilized timestamped data to analyze clinician workflow.[59] As recommended by Rule et al, we have thoroughly described the sample in terms of patients, staff, and clinically relevant counts of activities, we have described in detail how we developed the dataset of activities, and we have validated our results against prior observational assessments of medication administration.

There are some limitations to be acknowledged in this work. Duration calculated from first to last BCMA scan for medications scheduled at a given time will leave out preparation time before the first medication is scanned and medication activity after the last medication is scanned, and therefore will be an underestimate of total time required to administer all medications. In addition, medication characteristics such as route of administration may affect duration but are not accounted for here. Non-medication patient needs related to the complexity of their condition or education needs are also not included in the current models (except through differences across units) and are likely to influence medication pass duration. Likewise, due to the small number of units, we were unable to attribute differences across units to specific unit-level characteristics, such as patient characteristics of those hospitalized on the unit. In addition, consideration should be made for assessing the effect of workarounds during medication administration, which are typically aimed at reducing time spent to accomplish a task, on medication pass characteristics such as duration.[22] [60] [61]

Of course, peak-time gives a point-in-time snapshot for only one time of day, and 24-hour nursing care requires two or three shifts of staff each day. However, examining BCMA data for the peak time provides a daily snapshot view of patient census, staffing, staffing mix, workload, and nursing practice when the medication pass is the primary nursing process on the unit. Nearly all (98.3%) patients receiving medications received a peak-time scheduled medication. The number of patients per staff for the peak-time medication pass may be lower than actual patients per staff, if patients do not have peak-time scheduled medications or if anticipated patients have yet to be admitted; patient census and staffing records could be used to assess the degree of difference that may exist between medication pass and actual patients per staff. Staffing practices differ in the VA when compared with non-VA health care systems, which could result in different staffing levels when compared with national data. Utilizing data from FY15 when both RNs and LPNs administered acute inpatient medications supports generalizability of these findings to settings with a variety of staffing approaches. Lastly, data presented here are for one VA hospital, however, the process for summarizing medication pass characteristics, including computing medication pass duration, should be similar for any hospital with a BCMA system that records a patient, staff member, medication, location, and time for each medication administered.

Conclusion

This paper presents a proof-of-concept for using the individual staff medication pass derived from BCMA data to examine variation in the dimensions of a nursing process of care. Medication pass analysis can provide important information about point-in-time staffing and workload and, as the most immediate process indicator influenced by staff workload, medication pass duration can support evaluation of conceptual frameworks aiming to relate staffing and workload to patient care and outcomes. The findings here provide supporting evidence for the relationship between staffing and nursing practice, establishing a direct link between basic measures of workload and medication pass duration. Future research should consider how types of medication, routes of administration, complexity of patient care needs, nurse experience, and other important workload factors beyond the number of patients and number of medications contribute to variation in medication pass duration.

Clinical Relevance Statement

Understanding how staffing levels and workload affect the way individual nurses allocate the fixed amount of time they have available during a shift is of particular importance for those seeking to provide excellent care to patients and ensure staff well-being in a setting of limited resources. Health system leaders currently limited by inconsistent, infrequent, and/or paper-based reporting practices can partner with informaticians to incorporate medication pass data into reports of real-time nurse staffing and workload which could be used to identify and correct cases of potential workload imbalance that may lead to burnout.

Multiple-Choice Questions

1. Bar code medication administration data can be used to identify:

   • Peak time for medication administration.

   • Start time for the individual staff medication pass.

   • Number of patients receiving medications scheduled for a given time.

   • All of the above.

   Correct Answer: The correct answer is option d. BCMA data can provide insight into local policy (e.g., peak time for medication administration), approximate patient census (e.g., number of patients receiving medications scheduled for a given time), and the individual staff medication pass process (e.g., start time for the individual staff medication pass).

2. Bar code medication administration data are collected by what method:

   • Surveys of nurses after their shifts.

   • During the routine provision of nursing care.

   • Daily reporting by unit nurse managers during standardized times of day.

   • Payroll processing.

   Correct Answer: The correct answer is option b. BCMA data are generated in real-time during the routine provision of nursing care (i.e., during the course of medication administration).

Conflict of Interest

None declared.

Protection of Human and Animal Subjects

This research was approved by the Baylor College of Medicine Institutional Review Board.

• References

• 1 Cohen CC, Barnes H, Buerhaus PI. et al. Top priorities for the next decade of nursing health services research. Nurs Outlook 2021; 69 (03) 265-275
• 2 Winter SG, Bartel AP, de Cordova PB. et al. The effect of data aggregation on estimations of nurse staffing and patient outcomes. Health Serv Res 2021; 56 (06) 1262-1270
• 3 Fore A, Islim F, Shever L. Data collected by the electronic health record is insufficient for estimating nursing costs: an observational study on acute care inpatient nursing units. Int J Nurs Stud 2019; 91: 101-107
• 4 Caruso R, Arrigoni C, Conte G. et al. The byzantine role of big data application in nursing science: a systematic review. Comput Inform Nurs 2020; 39 (04) 178-186
• 5 Welton JM, Jenkins P, Perraillon MC. A micro-costing or ‘bottom-up’ approach to measuring nursing costs using data from electronic health records. Nurs Econ 2018;36(01):
• 6 Hackl WO, Rauchegger F, Ammenwerth E. A nursing intelligence system to support secondary use of nursing routine data. Appl Clin Inform 2015; 6 (02) 418-428
• 7 Womack DM, Hribar MR, Steege LM, Vuckovic NH, Eldredge DH, Gorman PN. Registered nurse strain detection using ambient data: an exploratory study of underutilized operational data streams in the hospital workplace. Appl Clin Inform 2020; 11 (04) 598-605
• 8 Westbrook JI, Duffield C, Li L, Creswick NJ. How much time do nurses have for patients? A longitudinal study quantifying hospital nurses' patterns of task time distribution and interactions with health professionals. BMC Health Serv Res 2011; 11: 319
• 9 Hendrich A, Chow MP, Skierczynski BA, Lu Z. A 36-hospital time and motion study: how do medical-surgical nurses spend their time?. Perm J 2008; 12 (03) 25-34
• 10 Michel O, Garcia Manjon AJ, Pasquier J, Ortoleva Bucher C. How do nurses spend their time? A time and motion analysis of nursing activities in an internal medicine unit. J Adv Nurs 2021; 77 (11) 4459-4470
• 11 Congdon J, Craft J, Christensen M. Are we measuring nursing workflow correctly? A literature review. Br J Nurs 2020; 29 (21) 1252-1259
• 12 Sun C, Cato K. How much time do nurses spend using electronic devices at work?. Nurs Manage 2020; 51 (03) 22-29
• 13 Moore EC, Tolley CL, Bates DW, Slight SP. A systematic review of the impact of health information technology on nurses' time. J Am Med Inform Assoc 2020; 27 (05) 798-807
• 14 Elganzouri ES, Standish CA, Androwich I. Medication Administration Time Study (MATS): nursing staff performance of medication administration. J Nurs Adm 2009; 39 (05) 204-210
• 15 Keohane CA, Bane AD, Featherstone E. et al. Quantifying nursing workflow in medication administration. J Nurs Adm 2008; 38 (01) 19-26
• 16 Jennings BM, Sandelowski M, Mark B. The nurse's medication day. Qual Health Res 2011; 21 (10) 1441-1451
• 17 Barker KN, Flynn EA, Pepper GA, Bates DW, Mikeal RL. Medication errors observed in 36 health care facilities. Arch Intern Med 2002; 162 (16) 1897-1903
• 18 Donaldson N, Aydin C, Fridman M, Foley M. Improving medication administration safety: using naïve observation to assess practice and guide improvements in process and outcomes. J Healthc Qual 2014; 36 (06) 58-68
• 19 Womack DM, Vuckovic NN, Steege LM, Eldredge DH, Hribar MR, Gorman PN. Subtle cues: qualitative elicitation of signs of capacity strain in the hospital workplace. Appl Ergon 2019; 81: 102893
• 20 Zayas-Cabán T, Chaney KJ, Rucker DW. National health information technology priorities for research: a policy and development agenda. J Am Med Inform Assoc 2020; 27 (04) 652-657
• 21 Johnson CL, Carlson RA, Tucker CL, Willette C. Using BCMA software to improve patient safety in Veterans Administration Medical Centers. J Healthc Inf Manag 2002; 16 (01) 46-51
• 22 Staggers N, Iribarren S, Guo JW, Weir C. Evaluation of a BCMA's electronic medication administration record. West J Nurs Res 2015; 37 (07) 899-921
• 23 Coyle GA, Heinen M. Evolution of BCMA within the Department of Veterans Affairs. Nurs Adm Q 2005; 29 (01) 32-38
• 24 Spetz J, Burgess JF, Phibbs CS. The effect of health information technology implementation in Veterans Health Administration hospitals on patient outcomes. Healthc (Amst) 2014; 2 (01) 40-47
• 25 New report exposes problems with hospital medication errors. The Leapfrog Group. March 28, 2019. Accessed November 19, 2021 at: https://www.leapfroggroup.org/news-events/new-report-exposes-problems-hospital-medication-errors
• 26 Shah K, Lo C, Babich M, Tsao NW, Bansback NJ. Bar code medication administration technology: a systematic review of impact on patient safety when used with computerized prescriber order entry and automated dispensing devices. Can J Hosp Pharm 2016; 69 (05) 394-402
• 27 Poon EG, Keohane CA, Yoon CS. et al. Effect of bar-code technology on the safety of medication administration. N Engl J Med 2010; 362 (18) 1698-1707
• 28 Paoletti RD, Suess TM, Lesko MG. et al. Using bar-code technology and medication observation methodology for safer medication administration. Am J Health Syst Pharm 2007; 64 (05) 536-543
• 29 Strudwick G, Reisdorfer E, Warnock C. et al. Factors associated with barcode medication administration technology that contribute to patient safety: an integrative review. J Nurs Care Qual 2018; 33 (01) 79-85
• 30 Barakat S, Franklin BD. An evaluation of the impact of barcode patient and medication scanning on nursing workflow at a UK teaching hospital. Pharmacy (Basel) 2020; 8 (03) 148
• 31 Khammarnia M, Kassani A, Eslahi M. The efficacy of patients' wristband bar-code on prevention of medical errors: a meta-analysis study. Appl Clin Inform 2015; 6 (04) 716-727
• 32 Welton JM, Kleiner C, Valdez C, Richardson S, Boyle K, Lucas E. Using time-referenced data to assess medication administration performance and quality. J Nurs Adm 2018; 48 (02) 100-106
• 33 Loresto Jr FL, Welton J, Grim S, Valdez C, Eron K. Exploring inpatient medication patterns: a big data and multilevel approach. J Nurs Adm 2019; 49 (06) 336-342
• 34 Patterson ES, Rogers ML, Render ML. Fifteen best practice recommendations for bar-code medication administration in the Veterans Health Administration. Jt Comm J Qual Saf 2004; 30 (07) 355-365
• 35 ISMP Acute care guidelines for timely administration of scheduled medications. Institute for Safe Medication Practices. 2011. Accessed April 15, 2021 at: https://www.ismp.org/sites/default/files/attachments/2018-02/tasm.pdf
• 36 State operations manual appendix A – Survey protocol, regulations and interpretive guidelines for hospitals. Centers for Medicare & Medicaid Services. 2014. Accessed April 15, 2021 at: https://www.cms.gov/Regulations-and-Guidance/Guidance/Transmittals/Downloads/R116SOMA.pdf
• 37 Yang C, Kuebeler MK, Jiang R. et al. Adapting clinical data from electronic health records for reporting at the nursing unit level. Poster abstract accepted to American Medical Informatics Association 2022 Clinical Informatics Conference; 24–26 May 2022; Houston, TX
• 38 Bates D, Machler M, Bolker B, Walker S. Fitting linear mixed-effects models using lme4. J Stat Softw 2015; 67 (01) 1-48
• 39 Eisenhauer LA, Hurley AC, Dolan N. Nurses' reported thinking during medication administration. J Nurs Scholarsh 2007; 39 (01) 82-87
• 40 Potter P, Wolf L, Boxerman S. et al. Understanding the cognitive work of nursing in the acute care environment. J Nurs Adm 2005; 35 (7-8): 327-335
• 41 Myers RA, Parikh PJ. Nurses' work with interruptions: an objective model for testing interventions. Health Care Manage Sci 2019; 22 (01) 1-15
• 42 Hopkinson SG, Jennings BM. Interruptions during nurses' work: a state-of-the-science review. Res Nurs Health 2013; 36 (01) 38-53
• 43 Sitterding MC, Ebright P, Broome M, Patterson ES, Wuchner S. Situation awareness and interruption handling during medication administration. West J Nurs Res 2014; 36 (07) 891-916
• 44 Thomas L, Donohue-Porter P, Stein Fishbein J. Impact of interruptions, distractions, and cognitive load on procedure failures and medication administration errors. J Nurs Care Qual 2017; 32 (04) 309-317
• 45 Hayes C, Jackson D, Davidson PM, Power T. Medication errors in hospitals: a literature review of disruptions to nursing practice during medication administration. J Clin Nurs 2015; 24 (21-22): 3063-3076
• 46 Rafferty AM, Franklin BD. Interruptions in medication administration: are we asking the right questions?. BMJ Qual Saf 2017; 26 (09) 701-703
• 47 Hopkinson SG, Wiegand DL. The culture contributing to interruptions in the nursing work environment: an ethnography. J Clin Nurs 2017; 26 (23-24): 5093-5102
• 48 Sloss EA, Jones TL. Alert types and frequencies during bar code-assisted medication administration: a systematic review. J Nurs Care Qual 2020; 35 (03) 265-269
• 49 Keers RN, Williams SD, Cooke J, Ashcroft DM. Prevalence and nature of medication administration errors in health care settings: a systematic review of direct observational evidence. Ann Pharmacother 2013; 47 (02) 237-256
• 50 Westbrook JI, Sunderland NS, Woods A, Raban MZ, Gates P, Li L. Changes in medication administration error rates associated with the introduction of electronic medication systems in hospitals: a multisite controlled before and after study. BMJ Health Care Inform 2020; 27 (03) e100170
• 51 Krähenbühl-Melcher A, Schlienger R, Lampert M, Haschke M, Drewe J, Krähenbühl S. Drug-related problems in hospitals: a review of the recent literature. Drug Saf 2007; 30 (05) 379-407
• 52 Furnish C, Wagner S, Dangler A. et al. Evaluation of medication administration timing—are we meeting our goals?. J Pharm Pract 2021; 34 (05) 750-754
• 53 Zimmerman S, Love K, Sloane PD, Cohen LW, Reed D, Carder PC. Center for Excellence in Assisted Living-University of North Carolina Collaborative. Medication administration errors in assisted living: scope, characteristics, and the importance of staff training. J Am Geriatr Soc 2011; 59 (06) 1060-1068
• 54 McLeod MC, Barber N, Franklin BD. Methodological variations and their effects on reported medication administration error rates. BMJ Qual Saf 2013; 22 (04) 278-289
• 55 Berdot S, Gillaizeau F, Caruba T, Prognon P, Durieux P, Sabatier B. Drug administration errors in hospital inpatients: a systematic review. PLoS One 2013; 8 (06) e68856
• 56 Bridges J, Griffiths P, Oliver E, Pickering RM. Hospital nurse staffing and staff-patient interactions: an observational study. BMJ Qual Saf 2019; 28 (09) 706-713
• 57 Dwibedi N, Sansgiry SS, Frost CP. et al. Effect of bar-code-assisted medication administration on nurses' activities in an intensive care unit: a time-motion study. Am J Health Syst Pharm 2011; 68 (11) 1026-1031
• 58 Hong JY, Ivory CH, VanHouten CB, Simpson CL, Novak LL. Disappearing expertise in clinical automation: Barcode medication administration and nurse autonomy. J Am Med Inform Assoc 2021; 28 (02) 232-238
• 59 Rule A, Chiang MF, Hribar MR. Using electronic health record audit logs to study clinical activity: a systematic review of aims, measures, and methods. J Am Med Inform Assoc 2020; 27 (03) 480-490
• 60 van der Veen W, Taxis K, Wouters H, Vermeulen H, Bates DW, van den Bemt PMLA. BCMA Study Group. Factors associated with workarounds in barcode-assisted medication administration in hospitals. J Clin Nurs 2020; 29 (13-14): 2239-2250
• 61 Nguyen OT, Shah S, Gartland AJ. et al. Factors associated with nurse well-being in relation to electronic health record use: a systematic review. J Am Med Inform Assoc 2021; 28 (06) 1288-1297

Publication History

Received: 14 September 2022

Accepted: 02 December 2022

Accepted Manuscript online:
06 December 2022

Article published online:
01 February 2023