CIN: Computers, Informatics, Nursing & Vol. 26, No. 4, 234–241 & Copyright B 2008 Wolters Kluwer Health | Lippincott Williams & Wilkins FEATURE ARTICLE Assessing the Work sampling measured nurse documentation Impact of an time before and after the implementation of an electronic medical record on a medical-surgical Electronic Medical nursing unit. Documentation was separated into subprocesses of admissions, discharges, and Record on Nurse routine/daily documentations. Production rate of documentation time is defined and measured. The results indicate that there is no difference in Documentation Time documentation time between pre-electronic med- ical record and post-electronic medical record for BRIAN HAKES, MBA admissions and routine/daily documentation time. JOHN WHITTINGTON, MD Post-electronic medical record documentation time was longer than that in the pre-electronic medical record for patients discharged to a nursing home. It was demonstrated that the electronic medical record may reduce documentation time after the Healthcare today is facing constant pressure to decrease adoption of computerized physician order entry. waste and reduce cost. Improving efficiency, safety, and quality is the reason for implementing an electronic KEY WORDS medical record (EMR). However, studies have shown EMR Benefit Realization & variable impact.1–7 At OSF HealthCare System in Peoria, Information System Benefit Realization & IL, the EMR vendor, IDX Corporation (GE Healthcare, Nurse Documentation Time & 8 Seattle, WA), issued a press release stating that on Nurse Efficiency and Documentation & average, a nurse using a paper-based documentation Nurse Utilization & system spends about 30% of his/her time—more than 3 Work Sampling of Nurse Documentation Time hours out of each 12-hour shift—on documentation. Using online documentation reduced that proportion by 50%, freeing up an additional 1.5 hours per nurse per commercial vendors provided marketing material sup- 12-hour shift. The premise is that the EMR will reduce porting their performance claims consisting of vendor- the time that nurses spend on documentation and free sponsored studies for benefit analysis. They found that up more time for patient care. independent research studies were limited and difficult Many questions and comments were made about the to replicate, leading to the conclusion that there was impact that the EMR had on a nurse’s time at the OSF little actual evidence to verify the clinical and admin- HealthCare System as it was installing an EMR in its istrative impact of an EMR. six hospitals. As a practical issue, management wanted Bosman et al10 performed work measurement in an to ascertain the net impact of the EMR on the amount 18-bed ICU for postoperative cardiothoracic patients. of time that nurses spent doing documentation. Because At the time of the study, the ICU was doing concurrent of these discussions, the decision was to conduct a charting. Their study segmented the ICU patients into study assessing the impact of the EMR on the amount two groups. One group was charted using an EMR, and of time that nurses use for documentation at the site the other was charted using a paper medical record. Of where the EMR was last to be installed. interest is that they separated the RN time for the Author Affiliations: System Solutions for HealthCare, Peoria, IL BACKGROUND (Mr Hakes); and OSF HealthCare System, Peoria, IL (Dr Whittington). 9 Corresponding author: Brian Hakes, MBA, System Solutions for When Hemby and White researched bedside informa- HealthCare, 7059 N Fox Point Dr, Peoria, IL 61614-2231 (bhakes@ tion systems in 1997, they discovered that many sbcglobal.net). 234 CIN: Computers, Informatics, Nursing & July/August 2008 Copyright @ 2008 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. charting of the admission from the other routine During their work measurement, they delineated charting tasks. Of significance is their temporal analysis between paper documentation and electronic record of the documentation time for the first 24 hours of the keeping but did not report the two separately. Their patient stay in the ICU. results showed that the proportion of nurse time spent Their results showed that the EMR admission on documentation is 15.79%. Documentation on the documentation time, 18.1 minutes, was 1.3 minutes paper chart was 10.55%, and the EMR was 5.24%. longer than the paper chart admission documentation Korst et al11 acknowledged a difference in documenta- time, 16.8 minutes. However, at the 95% confidence tion time on the day and night shifts. During the day, interval (CI), this is not a significant difference. documentation amounted to 19.17% of nurses’ time; Their temporal analysis showed different results. on the night shift, the ratio was 12.41%. They acknowledged that the amount of data generated The Cedars-Sinai authors acknowledged that the by a postoperative patient after cardiothoracic surgery number of patients on the unit influenced the amount is usually greatest after the initial transfer and admis- of documentation and recorded that statistic. However, sion to the ICU. For both their patient groups, the they did not allow for the census variation in their work amount of nursing documentation time decreased the sampling and in the reduced data. They did not report longer the patient stayed in the ICU, that is, more than results of the impact of census variation on documen- the initial 24 hours of patient stay. For this clinical tation time. documentation period, Bosman et al10 reported that the Korst et al11 documented that the median number of EMR reduced the amount of clinical documentation deliveries and nurses varied slightly by shift. The unit time over the paper record. They reported the docu- median number of nurses was nine nurses during the mentation time as proportion of total nursing time, and day and eight nurses on nights for the work sampling the values were 14.4% and 20.5% for the EMR and period. For the corresponding period, during the day paper chart, respectively. shift, the median number of deliveries was three What they did not consider is how census variation deliveries, and at night, the median was one delivery. may have affected the difference. The patient-to-staff The median patient census showed considerable varia- ratio for their ICU is 2:1. The number of patients in the tion between the two shifts. The median day census was study was 145, with 71 patients in the EMR group and 10 patients, and the median night census was four 74 patients in the paper documentation group. The patients. However, the Cedars-Sinai authors stratified median length of stay for all patients in the ICU was the data into periods when the staff were busy or not 22 hours. This means that there were periods when the busy. The busy period was defined as a time when the nurse would have had only one patient and would not patient census was nine patients or more. They reported do as much documentation as when two patients were that the amount of time spent on documentation did available. The authors did not report the patient not appear to vary with the census. volumes, for example, total number of patient hours Bosman et al10 and Korst et al11 recognized several in the ICU, for the two groups. A difference in the general factors that influence the proportion of staff patient load between the two groups would influence time spent on documentation: (1) the mix in the staff, the net proportion difference in the nurse documenta- (2) the type of unit, and (3) the amount of activity on tion time. that unit. They also recognized seasonal variation in the At Cedars-Sinai Medical Center, work measurement documentation time, as the amount of nurse time spent on a single intrapartum unit assessed RN documenta- on patient documentation was sensitive to the time of tion time. The unit had 20 rooms: 12 labor-delivery day (shift). They also recognized and acknowledged rooms and eight labor rooms. Korst et al11 used work census variation influence on the amount of nurse sampling to determine the proportion of time that the documentation time but did not present documentation nursing staff spent on documentation and other patient time as a function of patient activity. The purpose of care and non–patient care activities. They conducted this article was to present a methodology that is more their study during the implementation of an EMR descriptive of quantifying the impact of an EMR on system, in which nurses performed double charting. nurse documentation time and to expand on the work Their sampling plan consisted of recording staff activ- of Bosman et al10 and Korst et al.11 ities at random blocks of time, each spanning about 4 Similar to Bosman et al,10 this article separates hours. Their categories of nursing activities included documentation into discrete activities of admissions documentation, bedside care, bedside supportive care, and discharges. Like Korst et al,11 this article also nonbedside care, and non–patient care. demonstrates how documentation varies by the time of They generated a 95% CI, with the true mean day. The article presents a method of correlating docu- proportion lying within T2% of the sample estimate mentation time to census. The result is a rate of docu- derived from a total sample size of 2100 observations. mentation time per patient per hour. After compiling CIN: Computers, Informatics, Nursing & July/August 2008 235 Copyright @ 2008 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. the relevant pre-EMR and post-EMR statistics, the null Work Measurement hypothesis that there is no difference in documentation time between the paper chart and EMR is tested. The Work sampling is a group timing technique that records aim was to evaluate the impact of an EMR using a side- the discrete activity that each person, that is, the floor by-side comparison of like statistics.