NIELSEN, PETER E. MD; THOMSON, BROOK A. MD; JACKSON, RICHARD B. MD; KOSMAN, KAREN; KILEY, KEVIN C. MD
For 10 years, many national organizations have called for standardized documentation of prenatal care to facilitate communication between care providers and outcome analyses between different facilities.1,2 A multidisciplinary panel convened by the March of Dimes in 1993 recommended that a formal process be developed and supported to implement a surveillance system for collection of population-based and clinical perinatal data.2
In 1996, Frederic Frigoletto, Jr, MD, ACOG's 47th president, declared that the key to ensuring the highest quality patient care was to improve data access, retrieval, and management. ACOG president Frank Miller, MD reiterated that declaration in his presidential address in 1999. Promoting and developing a computer-literate membership was listed as goal number 6 for ACOG over the next 10 years.3 In 1996, Landstuhl Regional Medical Center began a computerized obstetric patient record to meet the needs of the clinical and the administrative staff. It is the only military medical facility that provides tertiary care for obstetric patients in the European Regional Medical Command. Military downsizing and budgetary restrictions have challenged the medical command. Limited resources require optimal use of personnel and facilities without adversely affecting maternal-fetal or neonatal outcomes. With multiple satellite clinics, an area responsible for 5000 deliveries a year, and high-risk patients in 30% of deliveries, the need for timely access to records and the ability to allocate resources and track outcomes led to the development of the Standard Obstetric Record Charting System (STORC). The purpose of this study was to develop, implement, and evaluate this new electronic inpatient and outpatient obstetric care record.
Materials and Methods
STORC is an automated information system for standardized collection and reporting of clinical and administrative data and maternal-fetal and neonatal outcomes for evaluation of obstetric resources. It tracks antepartum, intrapartum, and postpartum care records, and seamlessly integrates outpatient and inpatient data. Continuity and reliability of pregnancy data is a major goal of the system, providing standard forms for charts and risk management, monthly and quarterly quality assessment, and delivery projection reports. It does not simply place forms online, but provides screens with data entry fields unique for different encounters and allows providers to enter only data required for each encounter. The screens were based on typical inpatient and outpatient obstetric encounters and patient flow. A subcontractor developed STORC by using off-the-shelf software with direct input, review, and oversight by obstetrics health care providers, to meet clinical and administrative needs. A group comprising physicians, nurses, and administrative staff met frequently to review numerous screen designs and reports, create valid value pick lists and standard notes, review charts for completeness and accuracy, and solve problems of users. The group ensured a flexible, yet standard system. The cost of development was approximately $500,000, spread over 3 years, for the contractor, database software, and hardware expenses.
Data entry design for STORC was encounter-based, so data are easily tracked and reported. Ambulatory data include evaluation and management codes, patient-waiting times, appointment lengths, visit acuity, standard International Classification of Diseases, 9th revision, and current procedural terminology codes. Data entry in STORC is done when patients are seen. For outpatient visits, screening technicians enter the basic visit information, such as blood pressure (BP) and weight, when patients first arrive. Providers enter all other information. For routine and complicated patients, the system provides several pick lists to ensure consistent data across providers and several defaults to facilitate data entry, including those for routine visits that provide note completion with one mouse click. Providers can add and modify information as needed, and after it has been entered, the system displays a formatted progress note for the user to review and sign. Users can also select other providers to review and cosign notes. For patients with chronic conditions, their diagnoses are updated automatically as part of the assessment for each encounter, whether outpatient or inpatient. Diagnosis or gestational-age–specific treatment recommendations are automatically added to plans, with reminders generated for clinicians. That feature is invaluable for reducing deficiencies with gestational-age–sensitive tests, such as maternal serum multiple marker analyte screening for birth defects and screening for gestational diabetes.
For inpatient admissions, STORC automatically pulls relevant patient information entered at prior visits for current or past pregnancies into the history section. Nurses and other providers enter current data for admission, as needed. If delivery occurs during that admission, nurses enter detailed labor and delivery data for patients and neonates. Obstetricians review and modify data as necessary before electronic signature. Pediatricians, neonatologists, or nurses can enter progress notes into mothers' charts online, provided they have access. Attending physicians are required to review and sign each admission. At discharge, pertinent clinical information is pulled into discharge notes, which have replaced narrative summary dictation. When charts are printed at discharge, all applicable forms are in them.
STORC includes several standard progress notes, including spontaneous and operative vaginal delivery, cesarean delivery, and outpatient procedures such as amniocentesis. Additional standard and free-text notes can be written at any time. Standard progress and procedure note templates can be completed rapidly using pick lists for each field. Data within notes can be queried at a later date.
STORC also selects additional providers for counter-signature and review of notes or procedures, which allows nurse practitioners, residents, and technicians to input data or notes, which automatically appear on the designated provider's screen for review. Notes that require countersignature by a specific provider can be queried and accessed, which provides rapid chart accountability and completion. A security measure also was added to the system so that important patient information would be available at the local level, in the event of server failure, until the system is back online.
A MEDLINE search between January 1, 1980 and April 30, 2000, using the search words “electronic,” “automated,” or “computerized obstetric record” identified seven publications that specifically addressed an encounter-based, automated obstetric record.4–10 An encounter-based system is critical for easy administrative and clinical data entry and access to those data. STORC provides patient data, administrative documentation of record deficiencies, and the ability to query for information vital to resource allocation, provider trends, and process improvement. A powerful, easy-to-use query tool gives users any desired information from the database (Figures 1 and 2).
Currently, data are stored on a central server at Landstuhl Regional Medical Center, and the hospital network is connected to about 60 personal computers throughout the hospital and outlying clinics. A user name and password are required for access into STORC, and the password security validation can be done on the STORC server or within the hospital's domain. Besides requiring passwords for access, there are 28 different security roles that can be assigned. Security roles prevent or grant access to specific areas and tasks within STORC. A separate security module allows authorized users to add or deactivate users and to assign appropriate security roles to each user. The departmental chairman oversees security role assignments and access to the data query tool.
To evaluate this electronic obstetric record, we compared STORC data to several quality assurance measures, including our monthly departmental review and ORYX performance measures. ORYX is the Joint Commission's mandated performance measurement initiative for commission-accredited organizations and its data (usually gathered retrospectively) is to be used for comparing hospitals. ORYX data are generated from a retrospective review of 10% of available inpatient medical records, so discrepancies can result because of selection of small, nonrandom samples of records.
The architecture of STORC is client-server, developed using Borland Delphi client-server development software, version 3.0, and Borland ReportSmith reporting package, version 3.02 (Inprise Corp., Scotts Valley, CA). The server must have Windows NT version 4.0 (Microsoft, Inc., Redmond, WA) as the operating system software and Informix version 7.30 (Informix, Menlo Park, CA) as the database engine software. Client personal computers must have Informix Connect software, Borland Database Engine software, ReportSmith Run-time Viewer software, and application executable and report files. The database uses continuous transaction logging to tape, and level 0 backups are done nightly. Minimum specifications for the production server are a 200-MHz Pentium processor (Intel, Inc., Santa Clara, CA), 128-megabyte random access memory, 12 gigabytes of disk space in a redundant array of inexpensive disks-5 configuration, and two digital audiotape drives, and it must be network ready. For client personal computers, the minimum configuration is a 200-MHz Pentium processor, 64-megabyte random access memory, and at least 100 megabytes of disk space available, and they must be network ready.
STORC was officially implemented at Landstuhl Regional Medical Center as the complete obstetrics record in March 1998. Since then, over 35,000 patient encounters have been entered. More than 175 users currently manage 1159 pregnancies and 2333 deliveries in the STORC system. Before STORC, records were kept in written files at various satellite clinics, and delivery information was handwritten in a log. Quality assurance was done manually by reviewing the delivery log at monthly intervals to collect departmental statistics. Charts were reviewed for deficiencies or sentinel indicators, and tracking recurrent events and trends in practice was difficult. Patients frequently presented to labor and delivery without records because charts were very difficult to get from outlying clinics after office hours. This situation was particularly dangerous when patients with complicated pregnancies were not compliant with care or histories were poor. STORC has virtually eliminated that problem.
As STORC was implemented, quality assurance was done both manually and by using the automated query tool to generate the same report. The manually generated report for one quarter took approximately 5 hours to prepare, whereas the identical report using STORC took 14 minutes. Statistics for last quarter of 1998 showed a greater than 99% correlation between the hand-prepared and STORC-generated reports. All errors investigated in the STORC report were in the paper record with one exception. Use of STORC provided an additional benefit because all records were available immediately if a deficiency existed or required peer review.
Data quality control also is done as the user enters information. Quality control measures include required fields and various error and warning messages that depend on the user's security role and the specific data entered. For example, the number of fields required when a provider enters and saves data from a visit is greater than the number required when a technician enters basic information for that visit. Likewise, the number of fields required for an outpatient visit increases once a disposition has been entered for that visit.
Retrospectively gathered ORYX outcome measures from the first quarter of 1999 at Landstuhl Regional Medical Center were compared with the prospectively gathered STORC data. Unplanned cesarean deliveries were identified in 0% (0 of 18) of the charts reviewed for the retrospective ORYX initiative during that period; however, STORC immediately determined the actual rate of unplanned cesarean deliveries to be 8.3% (23 of 276). Table 1 compares the complete ORYX performance measure data elements collected retrospectively with the same data elements collected prospectively and identified by STORC. Until a complete and prospective data collection tool such as STORC is universally available and used to extract performance outcome measures from all deliveries, such differences in reported outcomes might persist.
Currently STORC does not interface with the hospital's primary health care database to collect patient demographics and laboratory results, but that has not been a major problem because most information is captured during obstetrics orientation classes and is entered by clinic personnel before patients' first visits. Such an interference would further improve STORC's capabilities and might be available later.
Initially there was varying compliance with data input among providers, which made generation of reports and forms difficult. However, since January 1999 compliance has improved dramatically because of several changes made to the software to facilitate data entry and screen navigation in response to user comments, increased use of the system over time, implementation of the provider review “to do” lists that automatically display at key times so providers know what they still need to do, and implementation of the ad hoc query tool that decreases time spent on administrative tasks and improves the user's ability to quickly identify delinquent charts and make clinical data queries for case-list collection. Users also had concerns that the system might cause delays because of patient data entry, especially during busy clinic hours. That is not unusual and has been reported by others. Kainz et al11 studied electronic obstetrics databases throughout Austria and found that 25% of the medical staff did not believe that computerization saved time. However, they noted an increase in chart completeness and accuracy. Some solutions to prevent delays are to establish a feedback group comprising users from various locations and positions to address issues and concerns of users, to provide increased memory for the personal computers to increase data entry speed, to provide adequate training to existing and new users by demonstrating the most efficient ways to accomplish certain tasks, to temporarily increase appointment time slots by 10–15 minutes until users become more comfortable with the system, and to create additional standard notes so users only have to enter patient-specific information and the rest of the text is merged to produce formatted notes.
The nonproprietary nature of STORC is one of its greatest strengths. Local control allows us to rapidly process changes and improvements, making the system meet clinician needs. For the past 2 years, the system has been working with minimal problems and minimal need for support. In the event of problems, Landstuhl Regional Medical Center has a half-time system administrator who monitors the system, does back-ups, installs software onto new personal computers as needed, and provides basic support to users.
The most important aspect of this comprehensive obstetric record is seamless integration of the inpatient and outpatient clinical data and standardization of the data, providing clinicians immediate access to clinical data and easy retrieval for case-lists, procedure documentation, complications, and charting delinquencies.
Most electronic records focus on outpatient or inpatient settings, which makes clinical use and outcome tracking difficult. Obstetrics is a unique specialty that results in inpatient hospital stays after a series of outpatient visits. Integrating those two data sets provides comprehensive data, flexibility, and easily extracted clinical outcomes that are unavailable to physicians who use most electronic medical records. Giving clinicians access to development and update processes increases acceptance of electronic medical record systems.
By using electronic records such as STORC, obstetric data collection and reporting can be standardized to facilitate communication between providers and health care systems and to facilitate outcome analysis between facilities. Further work is needed to enhance the development of this concept of seamless electronic integration of outpatient and inpatient data.