Original Contribution

Effect of introduction of electronic patient reporting on the duration of Ambulance calls^?

Markku Kuisma MD, PhD^a,?, Taneli Vayrynen MD^a, Tuomas Hiltunen MD^a,

Kari Porthan^b, Janne Aaltonen MD^c

^aHelsinki EMS, Helsinki University Central Hospital, PO Box 112, FIN-00099 Helsingin kaupunki, Finland

^bHelsinki City Rescue Department, PO Box 112, FIN-00099 Helsingin kaupunki, Finland

^cHospital District of Helsinki and Uusimaa, Headquarters, PO Box 100, FIN-00029 HUS, Finland

Received 22 May 2008; revised 9 July 2008; accepted 29 July 2008

Abstract

Objectives: We examined the effect of the change from paper records to the electronic patient records (EPRs) on ambulance call duration.

Methods: We retrieved call duration times 6 months before (group 1) and 6 months after (group 2) the introduction of EPR. Subgroup analysis of group 2 was fulfilled depending whether the calls were made during the first or last 3 months after EPR introduction.

Results: We analyzed 37 599 ambulance calls (17 950 were in group 1 and 19 649 were in group 2). The median call duration in group 1 was 48 minutes and in group 2 was 49 minutes (P = .008). In group 2, call duration was longer during the first 3 months after EPR introduction. In multiple Linear regression analysis, urgency category (P b .0001), unit level (P b .0001), and transportation decision (P b .0001) influenced the call duration. The documentation method was not a significant factor.

Conclusions: Electronic patient record system can be implemented in an urban ambulance service in such a way that documentation method does not become a significant factor in determining call duration in the long run. Temporary performance drop during the first 3 months after introduction was noticed, reflecting adaptation process to a new way of working.

(C) 2009

Introduction

Electronic patient records (EPRs) are widely used in health care services. There have been successful implementations [1,2] and, on the other hand, serious flops [3]. However, only few experiences have been published, and therefore, one can expect publishing bias. Published reports have been focused

^? Preliminary results have been presented in the International Congress of Emergency Medicine, San Francisco, Calif, April 3 to 6, 2008.

* Corresponding author.

E-mail address: [email protected] (M. Kuisma).

mainly on internally developed systems, minority of papers address commercially developed systems [4].

Many emergency medical services (EMSs) are currently introducing or at least planning to implement EPR systems. The effect EPR systems have on time spent on ambulance call has not been published. The length of an ambulance call is critical both medically (eg, rapid transportation in case of major trauma) and on Resource management basis because EMS is a time-critical service. In busy urban ambulance services, permanent increase in call duration may delay the readiness for the next mission and lead in the need of extra units. Changes in EMS processes that increase the call

0735-6757/$ - see front matter (C) 2009 doi:10.1016/j.ajem.2008.07.033

duration may be justified, but benefits have to be carefully weighed against risks and costs.

The purpose of the study was to report the effect that change from paper reporting to EPR has on ambulance call duration. Our hypothesis was that the total duration of the ambulance calls is not dependent on documentation method.

Methods

Description of EMS system

The Helsinki Dispatching Center serves the capital city of Helsinki (population, 567 000). Dispatching center prioritize emergency medical calls into 4 urgency categories from A to D (Table 1). Medical prioritizing criteria are listed in the dispatcher’s guidebook. Prioritizing is based on the serious- ness of the patient’s chief complaint (eg, chest pain, abdominal pain, fall, hemorrhage) and on the patient’s current condition (eg, awake or not, breathing normally or not, being able to talk).

Helsinki EMS system is 3 tiered, and the services for A, B, and C calls are provided by the Rescue Department. Nonemergency calls (category D) are carried out by private ambulance companies. In category A calls, the likelihood of being immediately life threatening is high, in B calls, the risk is moderate or unknown, C calls are emergency calls with a low risk of being life threatening in the next 20 to 30 minutes, and D calls are nonemergency calls. The first tier consists of 8 basic life support (BLS) ambulances and 8 fire engines used as first responding units (FRUs) and are manned by emergency medical technicians . In the second tier, there are 3 Advanced life support ambulances manned with paramedics and one medical supervisor unit. The

Urgency Dispatched units Maximum Blue lights response time and siren (min)

Nearest ambulance + 7? Yes MICU or medical supervisor + FRU if necessary Nearest ambulance + 7? Yes FRU if necessary Most appropriate 20 No ambulance Private ambulance 90 No

* Ninety percent of calls should have a maximum of 7 minutes and 95.4% of calls a maximum of 10 minutes of response time for the first unit. The FRUs are used to meet the response time target when necessary.

physician staffed mobile intensive care unit (MICU) makes up the third tier. Ambulances are located at 8 fire and ambulance stations. Emergency medical service does not carry out interhospital patient transfers.

Implementation of EPR system

Helsinki EMS introduced EPR (Merlot Medi commercial software 1.0; Logica, Helsinki, Finland) in clinical use in 2007 after 4 years of planning. It was made operational between January 15 to 21, 2007. Before introduction, 314 EMTs and paramedics, 8 medical supervisors and senior officers, and 10 physicians received 17 hours of training. At no time, there was double documentation (ie, filling both paper charts and electronic reports) because it was not considered feasible in out-of-hospital setting. This was also a partial reason for intensive training. Traditional paper forms were left in ambulances to be used only if EPR would fall down, and they were needed approximately once a day in the whole EMS system. During the first 4 weeks, there was a telephone help desk service 24/7 run by the implementing personnel of the system. After that, the operational 24/7 support services were integrated to the work of the on duty medical supervisor. In case the call duration had increased substantially and led into the shortage of ambulances, there was a possibility to increase ambulance capacity temporarily by one BLS unit.

Description of EPR system

The EPR used has been tailored for the out-of-hospital use. It is composed currently of 4 different components: (1) electronic medical record module (which replaced paper chart), (2) control/management module for physician and medical supervisor on duty, (3) disaster and major incident module (with reduced user interface), and (4) administration module including reporting and invoicing features. In this article, we focus on electronic medical record module of EPR. Ambulances and FRUs are equipped with touch screen laptops (Xplore Technologies iX104C3 Tablet Computers, Austin, TX) and small printers (Pentax PocketJet 3 Plus, Golden, CO). Laptops run EPR as client-server basis. Connection to the remote server is created by either wireless local area network or general packet radio service. Ambulance printers are used to provide a patient with paper copy of the report, including discharge instructions if a patient is not transported to the emergency department (ED). The alarm, including necessary dispatching information, is sent from the dispatching center to the ambulance laptop. The EPR includes some compulsory fields that have to be filled out in every case, for example, the cause of the call, the time of symptom or injury onset, information on who asked for help, current medication, and relevant chronic diseases. After these questions, the personnel can choose between answering predefined symptom or injury-specific questions

Table 1 Dispatched units, maximum Response times, and the use of blue lights and siren according the urgency category of the ambulance call

(preferred) or, alternatively, typing free text. Data from patient’s vital signs (heart rate, blood pressure, oxygen saturation, end-tidal CO₂, and 12-lead electrocardiogram) are transmitted via Bluetooth from monitors (Lifepak 12; Physiocontrol, Minneapolis, Minn) to the laptop and then to the EPR server. Electronic patient record includes all EMS

protocols and patient referral instructions. If the criteria to contact online medical direction are fulfilled, the system will light a red alert on the screen to remind EMTs and/or paramedics. The ambulance personnel send the patient report electronically to a printer at the ED during transportation. It means that the ambulance report is available at the ED

Fig. 1 The process of data transmission and the function of EPR system. General packet radio service is a packet-oriented mobile data service that can be used for services such as SMS and MMS messaging, and for Internet communication services. LAN indicates local area network, a computer network covering a small local area; TETRA, TErrestrial Trunked RAdio, a professional mobile radio designed for use by government agencies and emergency services.

approximately 5 to 12 minutes before patient arrival. In case the laptop would loose connection to the server (eg, in certain underground premises), the personnel can continue to use the system locally. The process of data transmission and the function of EPR are presented in Fig. 1.

Before the introduction of EPR, EMTs and paramedics filled out a patient report in paper form. It was handed over to the ED staff if the patient was transported or to the patient if the patient was left at the scene. A copy of the report was retained for EMS archive. In cases that MICU participated the call, 2 paper reports were made, one by an EMT or a paramedics and the other by a physician. After the introduction of EPR, no paper was used on the scene, and the MICU report was integrated into the general electronic patient report.

Study design

We conducted a retrospective cohort study in Helsinki EMS using an electronic database in which Prehospital time intervals of ambulance calls are automatically recorded. We retrieved the call duration times 6 months before (group 1, paper documentation) and 6 months after (group 2, electronic documentation) the introduction of EPR. The exact dates for group 1 were July 1 to December 31, 2006, and for group 2 were February 1 to July 31, 2007. Call duration was defined as the Total prehospital time from dispatching of the unit until the ambulance call had ended. The end of call was the time when the vehicle was back in service. After the call had ended, no changes were possible in the EPR report. On purpose, we did not use the time of “arrival at hospital” because it may underestimate the time needed to fill out a complete patient report especially in cases necessitating rapid transportation. All patients with adequate date were included in the study whether they were transported or not. In group 2 subgroup analysis, the call duration during the first 3 months of EPR use (group 2a) were compared with call duration during the last 3 months of study period (group 2b). The study plan did not include category D calls because

companies taking care of D calls were not equipped with EPR. Calls with incomplete time interval data were excluded (n = 2689). The data were defined as incomplete if the time of dispatching and/or the time of the end of the call was missing. In the study setting, the following parameters were registered: total duration of the ambulance call (minutes), urgency category of the call (A, B, or C), transportation (transported to the ED or left at the scene), and unit level (BLS or ALS ambulance).

Statistics

The study plan was approved by the institutional review board as a substudy of a research project monitoring the effects of the dispatching center reform. Statistical analysis was done using SPSS version 13.0 statistical software package (Chicago, IL). Mann-Whitney U and ?2 tests were used for statistical comparison of groups where appropriate. Multiple linear regression was used to find out the independent effect of the following 4 parameters to the duration of the ambulance call: urgency (A and B vs C), unit level (BLS vs ALS), transportation (transported vs left at the scene), and documentation method (paper vs EPR). Data are presented as medians and interquartile ranges (IQRs). Median values were used because the data were not normally distributed. P values less than .05 were considered statisti- cally significant. The primary end point of the study was the total duration of the ambulance call, and the secondary end points were the total duration of the ambulance call in predefined subgroups (according to the urgency category, unit level, and transportation decision and, in group 2, according to the period: first 3 months vs last 3 months).

Results

During the study period, there were 40 288 ambulance calls of which 2689 calls (6.7%) had incomplete electronic

Fig. 2 The median duration of ambulance calls (minutes) in different subgroups before introduction of EPR system (group 1), 1 to 3 months after introduction (group 2a) and 4 to 6 months after introduction (group 2b).

data on call duration and were excluded. The exclusion rate was 4.7% during the paper phase and 7.2% during the electronic phase of the trial. The study material consisted of 37 599 consecutive ambulance calls of which 17 950 took place before the introduction of EPR during paper documentation (group 1) and 19 649 after the introduction of electronic documentation (group 2). The groups differed in terms of urgency distribution of calls, that is, the proportion of C calls was lower and the proportions of A and B calls were higher in group 2. Therefore, multiple linear regression analysis was done to supplement uni- variate analysis.

Univariate analyses

The median (IQR) duration of the call was 48 minutes (31-66 minutes) in group 1 and 49 minutes (32-67 minutes) in group 2 (P = .008). Fig. 2 presents call duration in various subgroups.

The call duration increased 2 minutes in category B calls (P = .003) and 1 minute in C calls (P = .005) but decreased

4 minutes in category A calls (P b .0001) after the introduction of EPR. Subgroup analysis within group 2 showed that call duration first extended but then returned to baseline level during months 4 to 6, except in category A calls in which call duration remained decreased (Table 2).

The call duration of ALS units increased with 1 minute (P = .045) and BLS units with 2 minutes (P = .064) after the introduction of EPR. Subgroup analysis within group 2 showed that call duration decreased 2 minutes (P b .0001) in both ALS and BLS unit calls after 3 months of EPR system use (Table 3).

Among the transported patients, the call duration remained unchanged (P = .15); among the nontransported

Table 2 The duration of the ambulance call according to the urgency category of the call

Comparison of duration (min) between groups
Unit level Group 1 Group 2			P
ALS 49 (33-66) 50 (34-68)			.045
BLS 47 (30-66) 49 (31-67)			.064
Comparison of duration (min) within group 2
Unit level	Group 2a	Group 2b	P
ALS	51 (35-69)	49 (32-67)	b.0001
BLS	50 (33-67)	48 (30-66)	b.0001
The data are presented as medians with IQRs. Group 1 includes 17 950 ambulance calls before introduction and group 2 includes 19 649 ambulances calls after introduction of electronic patient reporting. Group 2a includes cases from months 1 to 3 and group 2b cases from months 4 to 6 after the implementation of EPR.

patients, it increased 3 minutes (P b .0001) after the introduction of EPR. In the subgroup analysis within group 2, the call duration decreased 3 minutes in the transported patients (P b .0001) and 2 minutes in the nontransported patients (P b .0001) between groups 2a and 2b (Table 4). Compared with the baseline level, the call duration remained 1 minute reduced (P b .0001) in the transported patients and increased 2 minutes (P b .0001) in the nontransported patients (group 1 vs group 2b).

Multiple linear regression analysis

Table 3 The duration of the ambulance call according to the unit level

Factors that were found to influence the call duration in multiple linear regression analysis were urgency category (in C calls, the duration was 3 minutes shorter, P b .0001), unit level (call duration was 2 minutes shorter in calls taken care of by BLS ambulances, P b .0001), and transportation decision (call duration was 19 minutes shorter if patients

Table 4 The duration (minutes) of the ambulance call according to the transportation decision

Comparison of duration (min) between groups
Urgency	Group 1	Group 2	P
A 54 (31-73) 50 (22-69) b.0001
B 50 (33-69) 52 (35-71) .003
C 46 (30-63) 47 (32-64) .005
Comparison of duration (min) within group 2
Urgency	Group 2a	Group 2b	P
A B C	51 (25-70) 53 (36-72) 48 (34-65)	49 (20-67) 51 (33-70) 46 (31-64)	.053 b.0001 b.0001
The data are presented as medians and IQRs. Group 1 includes 17 915 ambulance calls before introduction and group 2 includes 19 614 ambulance calls after introduction of electronic patient reporting. Group 2a includes cases from months 1 to 3 and group 2b cases from months 4 to 6 after the implementation of EPR. Cases with incomplete data on urgency category (n = 70) have been excluded.

Comparison of duration (min) between groups
Transportation	Group 1	Group 2	P
Transported 57 (41-73) 57 (40-73) .15 Not transported 33 (24-42) 36(27-47) b.0001 Comparison of duration (min) within group 2
Transportation	Group 2a	Group 2b	P
Transported	59 (42-74)	56 (38-73)	b.0001
Not transported	37 (28-47)	35 (26-46)	b.0001
The data are presented as medians with IQRs. Group 1 includes 17 950 ambulance calls before introduction and group 2 includes 19 649 ambulances calls after introduction of electronic patient reporting. Group 2a includes cases from months 1 to 3 and group 2b cases from months 4 to 6 after the implementation of EPR.

Parameter Influence 95% CI P (min) (min)

Urgency category C vs A and B -3 (-4 to -3) b.0001 Group 2 vs 1 0 (0 to 1) .566 Unit level BLS vs ALS -2 (-3 to -2) b.0001 Transportation Left at the scene vs -19 (-20 to -19) b.0001 transported

CI indicates confidence interval. The model includes 4 parameters. Urgency categories A and B were combined and compared with category C calls.

were not transported, P b.0001). The documentation method (paper report vs electronic medical record) was not a significant factor (Table 5).

Table 5 Multiple linear regression analysis of factors influencing the duration of the ambulance call (n = 37 529)

Discussion

The main finding of this study was that documentation method itself was not a significant factor in determining the total duration of an ambulance call. Therefore, the fear that implementing EPR would lead into the considerable call time prolongation in busy ambulances services seems not to be justified. The temporary increase in call duration immediately after the change from paper reporting to EPR reflects adaptation to a new system and to a new way of working.

In subgroup analyses, the call duration remained perma- nently decreased in category A calls after the implementation of EPR. In most of category A calls, an MICU or a medical supervisor unit is dispatched with the nearest BLS or ALS unit. So category A calls are usually attended by 2 units, which enables simultaneous use of 2 laptops to speed up the documentation process. The decrease in call duration can also be explained by omitting double documentation. The MICU physician did not fill out a separate report any more. Some of the medical supervisors had an instructor status in imple- menting phase of EPR. This may also have had a positive effect on time use in A calls.

The call duration remained 2 minutes elevated in nontransported patients when EPR was used. This can be explained by insufficient documentation level when paper charts were used. Several local audits have shown that most documentation deficiencies are found in the charts of nontransported patients. A special electronic nontransporta- tion sheet was included already in the planning phase to increase the quality of documentation. Another reason for increased call duration may be in the use of ambulance printers. During the first months, there were frequent problems connecting the ambulance printer and the laptop

and in the use of a special printing paper required by the printer. The statistical differences registered were highly significant in most subgroup analyses, but the clinical relevance of a 1- to 3-minute time difference during a 50- minute ambulance call may well be debated. However, from the system management point of view, a 1- to 3-minute time difference becomes relevant in urban systems taking care of tens or hundreds of thousands of calls annually.

A systematic review of Health information technology demonstrated 3 major benefits on quality: increased adherence to guideline-based care, enhanced surveillance and monitoring, and decreased Medication errors [4]. The major efficiency benefit found was decreased use rate of health care services, but data on time use were mixed. None of the studies concerned ambulance services. Another systematic review of randomized controlled trials found that decision support systems significantly improved clinical practice in 68% of trials [5]. These findings have to be confirmed also in ambulance services.

Krall [6] showed that physicians took 30 days to return their baseline level of productivity after the implementation of electronic health records, and the patient visit time increased on average by 2 minutes per encounter. Another outpatient study showed minimal (0.5 minutes) reduction in visit time [7]. Studies done in Intensive care units have shown an absolute 5.4% to 10.9% decrease in documentation time [1,8-10]. Two of the studies also showed increased time devoted to patient care [1,10]. In a Finnish ICU study, after the implementation of a clinical information system, both documentation time (not significant [NS]) and time spent on Patient Care Activities increased, and time used for supporting activities decreased [11]. Plans to reduce ICU staff with the aid of clinical information systems were judged to be unrealistic. In outpatient care, after the introduction of electronic health records, the age-adjusted rates of patient visits fell by 9% [2]. It is most unlikely to be able to reduce the number of ambulance calls with the aid of EPR. Instead, EPR might help in reducing the number of patients needing transportation and an immediate ED visit because of standardized methodology and timely updated guidelines aiding at analysis of patient’s need of care and transportation. The major benefits of EPR in an ambulance service in our opinion can be found in the areas of documentation compliance and uniformity, accuracy of vital signs registra- tion, appropriate use of online medical direction, and integration with the hospital medical record. Early notifica- tion of ED personnel with accurate patient data is likely to improve immediate diagnostics and patient care at the ED. Administration costs are considerably lower because no secretarial assistance is any more needed in handling 40 000 paper reports. These benefits have to be balanced against disadvantages of which failures in electronic transmission of data and costs are of particular importance. A total failure of data transmission is a rare event, but delays in transmission have occurred frequently. The initial investment costs were high, and maintenance costs have to be considered, too. The

cost for tailored software was 400 000 EUR, but with this cost, software can be used in the whole hospital district area (population more than 1.4 million inhabitants). Addi- tional investment costs came from laptops, printers, and training of personnel. The cost for maintenance is approxi- mately 0.15 EUR/inhabitant a year.

The change from paper reporting to electronic reporting was a major challenge. Electronic patient record is a tool, but it is also a complex system requiring change of earlier working methods in a standardized way. Any tool may be used better or worse. In the analysis of complex systems and process changes, more relevant question than what tool is used is how the tool is used. If the change had not been managed properly, the call duration might have been prolonged substantially, and other major problems might have occurred. It is important to realize that introducing EPR includes much more than just changing a paper chart to a computer touch screen. Reorganization of working processes is an important factor in the successful implementation of EPR. Reluctance to change working processes when EPR is introduced may lead to loose many potential benefits of electronic systems. We recommend other EMS systems to implement EPR systems with careful planning, vulnerability analysis, multidisciplinary cooperation, keeping the person- nel informed, proper training of personnel, and help desk service after go-live. Our results can be generalized to other EMS systems only if these vital infrastructural elements are considered in proper manner.

It is expected that the time used for documentation will decrease in the future. The software will be developed according to the user needs and feedback, and network problems between ambulance laptops and the server will reach minimal level. The current software version is 2.0, and the one used in the study was 1.0. The filling out of invoicing information by personnel into the EPR will be minimized and made automatic. New EMTs and paramedics will receive the user training already during their basic studies. Although we did not study it, our perception is that the younger generation of employees is more rapid in learning and using electronic documentation compared with personnel older than 35 to 40 years. This is supported by experiences in ICU environment [11]. At the ED side, the process will be made fully electronic in 2009. Printing at the ED will not be needed any more when both prenotification of the coming patient and the patient report itself will be received into the electronic information systems of hospitals.

Our study was limited by excluding approximately 7% of all calls from analysis because of incomplete time interval data. However, this exclusion is not likely to bias the results because the exclusion proportion was relatively low compared with sample size. We did not directly measure how much time personnel used for documentation in each call. We chose to measure the duration of an ambulance call as a surrogate marker. In ambulance care, the extra time used in documentation is likely to turn into the increase in the duration of a call, whereas, for example,

in the hospital ICU, the length of stay is not likely to be prolonged. Instead in the ICU environment, the extra time used for documentation is merely away from patient care or other activities.

Ambulance call duration when EPR is used is only one area of interest. Relevant future research areas include, for example, effect on quality and cost-effectiveness of care, protocol adherence, medication error reduction, decision support, user perceptions, data security, and the impact of EPR on other organizations, that is, hospital EDs, primary care centers, and coroner’s office.

Conclusions

We showed that the documentation method was not a significant factor determining the call duration in an urban ambulance service if EPR system implementation is planned and fulfilled carefully. Call duration was dependent on the urgency category of the call, the unit type, and the transportation decision. Call duration was longer during the first 3 months after the introduction of EPR, reflecting adaptation process to a new system and to a new way of working. Good planning, training of personnel, and reorganization of working processes are mandatory require- ments for the successful implementation of EPR.

Acknowledgments

We thank Jarmo Laukkanen, Dispatching Advisor, Helsinki Dispatching Center, for help in data management and Kari Velling, Information Technology Advisor, Helsinki City Rescue Department, for his constructive comments in the final preparation of the manuscript.

References

1. Bosman R, Rood E, Oudemans-van Straaten H, et al. Intensive care information system reduces documentation time of nurses after cardiothoracic surgery. Intensive Care Med 2003;29:83-90.
2. Garrido T, Jamieson L, Zhou Y, et al. Effect of electronic health records in ambulatory care: retrospective, serial, Cross sectional study. BMJ 2005;330:581-5.
3. Scott J, Rundall T, Vogt T, Hsu J. Kaiser Permanente’s experience of implementing an electronic medical record: a qualitative study. BMJ 2005;331:1313-6.
4. Chaudhry B, Wang J, Wu S, et al. Systematic review: impact of health information technology on quality, efficiency, and costs of medical care. Ann Intern Med 2006;144:E12-E22.
5. Kawamoto K, Houlihan C, Balas E, Lobach D. Improving clinical practice using clinical support systems: a systematic review of trials to identify features critical to success. BMJ 2005;330:765.
6. Krall M. Acceptance and performance by clinicians using an ambulatory medical record in an HMO. Proc Annu Symp Comput Appl Med Care 1995:708-11.
7. Pizziferri L, Kittler A, Volk L, et al. Primary care physician time utilization before and after implementation of an electronic health record: a Time-motion study. J Biomed Inform 2005;38:176-88.
8. Pierpont G, Thilgen D. Effect computerized charting on nursing activity in intensive care. Crit Care Med 1995;23:1067-73.
9. Korst L, Eusebio-Angeja A, Chamorro T, et al. Nursing documentation time during implementation of an electronic medical record. JONA 2003;33:24-30.
10. Wong D, Gallegos Y, Weinger M, et al. Changes in intensive care unit nurse task activity after installation of a third-generation intensive care unit information system. Crit Care Med 2003;31: 2488-94.
11. Saarinen K, Aho M. Does the implementation of a clinical information system decrease the time intensive care nurses spend on documentation of care. Acta Anaesthesiol Scand 2005;49: 62-5.