Proceeding of Toward An Electronic Patient Record'97 (Vol. 2), pp139- 143, 1997.

Dynamic Template Driven Data Entry System for an EPR System

Hiroshi Takeda, Yasushi Matsumura, Takeo Okada, Shigeki Kuwata (Department of Medical Information Science, Osaka University Hospital) Norhiro Hazumi, Hiroaki Nakazawa (NEC corporation)

1. Introduction

As physicians' order entry and report retrieval system have been well operated in a totally integrated hospital information system in the Osaka University Hospital since 1995 [1][2] , the next goal of the hospital information system (HUMANE, human-oriented universal medical assessment system under network environment) will be a total electronic patient record system (EPROU, electronic patient record system at Osaka University hospital) [3] . The project has been in cooperation with Nippon Electric Company (NEC) . Fundamental structure of EPROU is illustrated in Figure 1.

A computerized system enabling faster input of necessary clinical data compared to the current handwritten procedure is essential to the realization of the electronic patient record. Furthermore, the entered data should provide the basis for synthetic and analytic purposes in clinical research, education, hospital management as well as clinical care. To merge the operability and the utility of contents of an EPR system, we have developed a system (DTDES,"dynamic template" driven data entry system) which allows users to select the desired data from a list of items by means of a graphical user interface.

2. Methods

2.1 System Environment

A personal computer (PC9801 series, NEC ) with windows 95 operating system (Microsoft) and Visual Basic 4.0 (Microsoft) and Excel (Microsoft) were used for the development of the system.

2.2 System Outline

The dynamic template is introduced for two reasons. One is to reduce the physicians' work load for inputing clinical data, and the other is to improve or keep the quality of the EPR contents. In clinical settings, most patient cases are requested for detail description regarding selected key words such as chief complaints and physical findings. The needs to guide a physician for medical recording in accordance with decision making process in clinical care. For example, if the chief complaint of a patient is pain,a template is shown in display and the physician will choose second key words among pain-related word sets such as chest, abdomen and so on. After taking the history of the present illness, the physician may proceed to taking physical findings, laboratory test and so on. Some data may be simple or others may be complicated. If each entry item were to be displayed in the same manner, the user would have to follow the same procedure as inputing detailed information even for simplified data. In order to input simplified and detailed data simultaneously, we have concluded that entry items must be displayed in a template in which the size will be adjusted due to the quantity of the information, and that the system will need to be driven by a kind of navigation system which displays entry items based on the previously entered value or category. We named this entry system, the dynamic template driven data entry system (DTDES).

DTDES is equipped with the following functions:
(1) Useless items for clinical data entry are not displayed.
(2) Detailed and simplified information input mode are simultaneously available.
(3) Reference of previously entered contents and input of altered points are allowed.
(4) Repeated input can be reduced.
(5) Input of frequent data patterns can be simplified.
(6) Choices can be displayed clearly and selected easily.
(7) Registered template contents are generated and changed by end- users.

This system has made physicians' data entry simple and easy by allowing users to refer to previous patient records. Physicians can copy necessary parts from the previous data and add revisions simply by selecting a displayed item. This easy entry method encourages users to use EPROU.

Furthermore, it would be useful to simplify data entry of regular patterns by revising default values. Selective branches should be displayed clearly and in a comprehensive matter to all users. Therefore, users may make selections through a diagram as well. For example, it is easier to select areas of a body on a diagram (Figure 2).

2.3 Structure of Registered Data

Data entered through the dynamic template are registered in the following structure.

Individual symptoms (eg. headache, chest pain, abdominal pain) and physical findings (eg. blood pressure, pulse rate, heart sound) are handled as one unit and the title name of the descriptive unit is called as "item". The item is set up so the detailed description of the item rarely has to be handled separately. For example, since systolic and diastolic blood pressure rarely have separate description, blood pressure is one item.

Entry of detailed description of items are made through an item name and value which are called atom name and atom value respectively. If these data are registered in the relational database, its field parameters are as follows; patient ID, registration date, registration time, branch number, item code, item name, parent atom serial number within the item, daughter atom serial number within the parent atom, daughter atom name code, atom name, atom value code and atom value. The parentage between atoms within each item is indicated by the atom serial number.

For example, a patient's chest pain item would be represented in the following way. Parent atom serial number within the item, daughter atom serial number within the family, atom name and atom value are indicated within the parenthesis.

Item : Chest pain; (1, 0, date of onset, June 1996), (2, 0, painful region, left anterior chest ), (3, 0, degree of pain, medium degree), (4, 0, character of pain, oppressive), (5, 0, attack style, paroxysmal), (6, 5, inducer, effort), (7, 6, degree of effort, stair climbing), (8, 5, time frequency of occurrence, irregular), (9, 5, duration of attack, 2 to 3 minutes), (10, 5, frequency of attack, 1 to 2 times/week), (11, 5, up to date progress, becoming frequent).

2.4 Display of Registered Data

When later referring to registered data, data entered from the template are translated into a form of natural language and described on the progress note of EPROU. The above example (Figure 3) is displayed as "medium degree left anterior chest pain since June 1996. The attack is paroxysmal and occurs during effort (climbing stairs). Frequency is indefinite. Lasts a few minutes. Degree of occurrence is 1-2 times/week. Frequency tends to increase after the onset."

2.5 Template Master

The contents and navigation method of templates and translation method into natural language need to be created through repeated trial and error. The template master must have two parts in which controlling data for the template navigation and contents of items are registered. The master file may be generated and edited by end-users. Rules for translation into natural language are registered in the another master file. This file must be maintained by the system manager.

3. Results & Discussion

We tested entry templates for cardiovascular diseases. Physicians evaluated that DTDES has enabled swift data entry, and that the entered clinical data are organized and translated into natural language. The design of the template is a crucial element which determines the operability of the system. Therefore, allowing end- users to freely design the structure of the template was deemed essential.

This system is considered as a reliable entry method of electronic patient records.

4. Conclusion

The EPR system of our university hospital (EPROU) is now under development and will operate in a few years. Dynamic template driven entry system (DTDES) is one of the optimal solution in Japan where physicians' direct data entry is very common. However, templates is a kind of tools to assist the data input and there need more efforts to obtain template master files which closely reflect a thesaurus medical knowledge. In another words, generation of template master is to reconstruct knowledge-base in medicine in terms of medical informatics.

References

[1] H.Takeda, Y.Matsumura, H.Kondo, H.Imai and M. Inoue. Development of a totally integrated hospital information system:An intelligent hospital in Osaka University. Proc. MEDINFO92 (eds,K.Lun et al.), 241- 246,1992.

[2] Y.Matsumura, H.Takeda and M.Inoue. Implementation of the totally integrated hospital information system (HUMANE) in Osaka University Hospital. Proc. MEDINFO95 (eds,R.Greenes et al. ), 590-593, 1995.

[3] H.Takeda, Y.Matsumura, T.Okada, S.Kuwata, M.Inoue, N.Hazumi and J.Aoki. Development of a cardiovascular disease-oriented electronic patient record model in a Japanese university hospital. Proc.Toward an Electronic Patient Record96, 520-523, 1996.