Diabetes Patient Tracker, a Personal Digital Assistant-Based Diabetes Management System for Primary Care Practices in Oklahoma

Transcription

1 DIABETES TECHNOLOGY & THERAPEUTICS Volume 5, Number 6, 2003 Mary Ann Liebert, Inc. Diabetes Patient Tracker, a Personal Digital Assistant-Based Diabetes Management System for Primary Care Practices in Oklahoma ZSOLT NAGYKALDI, Ph.D. and JAMES W. MOLD, M.D., M.P.H. ABSTRACT It has been demonstrated that electronic patient registries combined with a clinical decision support system have a significant positive impact on the documentation and delivery of services provided by health care professionals. While implementation of available commercial systems has not always been proven effective in a number of primary care practices, development and implementation of such a system in a practice-based research network might enhance successful implementation. Physicians in our practice-based research network (Oklahoma Physicians Resource/Research Network) initiated a project that aimed at designing, testing, and implementing a personal digital assistant-based diabetes management system. We utilized the best practice approach to determine the principles on which the application must operate. System development and beta testing were also accomplished based on the direct feedback of user clinicians. Practice Enhancement Assistants (PEAs) were available in the practices for assistance with implementation. Implementation of the Diabetes Patient Tracker (DPT) resulted in a significant improvement (p, 0.05) in nine of 10 diabetic quality of care measures compared with pre-intervention levels in 20 primary care practices. Regular PEA visits similarly increased the number of foot exams and retinal exams performed in the last year (p and 0.02, respectively). DPT is a low-cost, feasible, easily implementable, and very effective paper-less tool that significantly improves patient care and documentation in primary care practices. INTRODUCTION THE OKLAHOMA CENTER for Family Medicine Research (OCFMR) and the Oklahoma Physicians Resource/Research Network (OKPRN), a primary care practice-based research network, have completed a number of projects that involve participatory research and development. This approach begins with a request from participating clinicians, who remain involved in all phases of the process through testing and implementation. One particularly effective method, called best practice research has recently been described by Mold and Gregory. 1 Solutions to clinician-identified challenges are systematically sought from within the network itself. Exemplars are identified, and their methods are studied and, in some cases, enhanced with technology. Methodological assistance and information technology expertise are provided by the OCFMR. The OCFMR also provides Practice Department of Family and Preventive Medicine, College of Medicine, University of Oklahoma Health Sciences Center, Oklahoma City, Oklahoma. 997

2 998 Enhancement Assistants (PEAs), trained masters-level professionals, who work with small groups of practices over sustained periods of time to help them to implement the new methods. A variety of studies, including the current study, were done in collaboration and with the support of the Oklahoma Foundation for Medical Quality (OFMQ). The effectiveness of diabetes electronic management systems has been demonstrated and described in the literature. 2 4 However, these studies implemented PC-based, or web-based, systems, 5 while a diabetes registry and patient management system incorporating mobile solutions, such as personal digital assistants (PDAs), has not been reported. In addition, reported systems did not designate nurses as users and did not implement the best practice research method. IMPETUS FOR DEVELOPMENT OF THE SYSTEM The Diabetes Patient Tracker (DPT) was developed as a part of an OKPRN diabetes best practice research and development project aimed at improving documentation and delivery of medical services for diabetic patients. Based upon chart audits of 40 OKPRN clinicians, we identified five exemplary physicians, who achieved high scores on established quality of care guidelines (Diabetes Quality Improvement Project). From interviews conducted by phone with these exemplars, we identified six principles that seemed to characterize exemplary performance. These principles were then shared with other network practices, and customized technology was used to assist them with implementation. The technology specifically addressed three of the principles: use of a systematic protocol for diabetic office visits, use of a diabetic registry, and inclusion of diabetes flow charts in medical records. DESCRIPTION OF THE SYSTEM NAGYKALDI AND MOLD The DPT application was designed in Pendragon Internet Forms (PIF), a leading database enterprise solution for handheld computers. We utilized Palm m505 and Palm m515 handhelds in our test clinics to collect data and to build our database. The PDAs were connected to a Pendragon Enterprise Server located in the University of Oklahoma Health Sciences Center, Department of Family and Preventive Medicine via the Internet. The PDAs used existing office desktop computers running Windows 98, 2000, or XP operating system and an S-link conduit to synchronize with the Server. Office printers were enhanced, if necessary, with infrared ports. Data transfer was 56-bit encrypted throughout the entire communication process. We applied rigorous administrative and information technology measures to comply with applicable HIPAA guidelines for confidential patient data handling and communications. We established a multiple-level security system and granted different access rights to the database for participants of the study and information technology personnel. We made sure that database access and communication are properly tracked by the system according to built-in features. Full (administrative) access to the system was provided for only two system administrators. Clinics could review and edit only their own data and had no access to another clinic s database. Because DPT was considered an extension of the clinic s own medical record that was never shared as identifiable patient information with third parties, we did not obtain individual informed consent from patients. However, participating patients were informed about the electronic version of their medical records, and opting-out was offered as an option. During our studies no opting-out was requested, and no patient complaint occurred at any investigation site. Participating physicians and clinic personnel signed a consent form for the Diabetes Best Practice Study approved by the University of Oklahoma Health Sciences Center Institutional Review Board. Clinic sites could also use a browser-based web interface to the Server to review and edit their patients data. This interface was in compliance with the security and administrative measures mentioned above. The Server stored incoming data in an SQL-database. Database administrators were able to review, edit, im-

3 DIABETES PATIENT TRACKER 999 port, and export database components into flat files (CSV, TXT, and XLS) to produce complex reports and perform data analysis for intervention sites and for the study. It was very easy to query the unified database regarding intervention rates, the physician s compliance rates with quality of care guidelines, comparative performance, and patterns of patient population served by the clinics and to determine what services were due for the patients and when. We also have the capability to fully automate these queries and reports in the future and with the proper encryption technology pass these reports, or patient reminders, to the clinics on a regular basis. The PDAs ran the PIF s PDA-side application iforms featuring multiple tables, pageby-page (single record within one patient encounter entry), and review mode (all records of one patient encounter) in an easily searchable and filterable database. The PDA application captures the patient demographics, a current medication list, vital signs, fingertip sugar, and hemoglobin A1c (HbA1c) values by patient visit, diet, exercise, smoking information, urine protein, urine microalbumine complete lipid panel, retinal, foot, and dental exams, flu, Pneumovax (Merck Frosst) and adult diphtheria tetanus vaccinations, purified protein derivative status, and diabetes education. The patient report printed from the PDA includes this information for the last four visits. In addition, we developed an electronic chart auditing application in Visual Basic that reads and analyzes the patient database on the handheld and sends the audit results to an infrared printer for quality improvement purposes at the clinic site. From this summary the clinic staff can easily determine what services are due for their patients. The intervention protocol calls for immediate ( point-of-service ) data entry into the PDA by the nurse when the patient comes in for his or her regular diabetes checkup. Generally, the nurse prints an up-to-date patient report via an infrared-enabled printer from the PDA to review or update medical history and to identify interventions required during that visit. According to individual clinic procedures, the nurse can automatically administer immunizations. The physician then reviews the report and decides how to proceed with other interventions, such as change in medications, ordering labs, and counseling the patient (smoking, diet, diabetes education, etc). Interventions and new lab test results are recorded by the nurse in the PDA to keep the record up-to-date. The Web interface to the Server is an alternative option for data entry. Optionally, a 1-page chart report can be printed and attached to the patient s paper chart. These reports and the database in the PDA provide a comprehensive diabetes visit history, and the staff can quickly and effectively review what happened with the patient in the past regarding virtually every intervention related to the patient s diabetes. More complex audit reports can be generated for the physicians for intramural and for siteto-site comparative analysis from the Server. EFFECTIVENESS OF THE SYSTEM Patient charts of 30 OKPRN family physicians (residents and faculty) were audited by the OFMQ to determine the baseline rate of preventive and interventive services for established patients with diagnosis codes of , , and seen in the last 3 months. These services included: HbA1c measurement within 1 year; urine for protein within 1 year; lipid panel within 1 year; retinal examination within 1 year (both ordered and documented); foot examination within 1 year; use of angiotensin converting enzyme inhibitor for blood pressure; use of angiotensin converting enzyme inhibitor for proteinuria; flu shot within 1 year and Pneumovax ever. We introduced the principles learned from exemplar physicians and supplied a Diabetes Toolkit that contained our findings and resources for implementation. Following the chart audit, we consulted with the physicians and the office staff and taught them how to use the PDA and the DPT handheld application. After an average of 4 months, patient charts in participating physician offices were audited by the OFMQ to determine the post-intervention rates of delivering and documenting services. We put a particular emphasis on physician and nurse satisfaction. We spent a considerable amount of time developing and optimalizing

4 1000 the PDA application based on the feedback of our OKPRN beta-testers, so that practices could take full advantage of a mobile solution without compromising speed and without spending a significant time with data management. Menus and data entry fields were designed to provide a fast, one-tap entry method, and Graffiti use on the handheld was minimalized. Multiple ways of data management (PDA, Web-browser, and conventional print-outs) also proved to be an effective solution for providing sufficient flexibility in the data management protocol and customization of the tool to the provider s needs and particular workflow. Technical and clinical support was provided by PEAs who regularly visited participating sites and helped solve their problems related to this and other projects, taught the clinic personnel how to use the PDAs, monitored the study and protocols, and assisted in every step of implementation. This form of support proved to be very effective and resulted in high compliance rates and user satisfaction. We also queried the physicians about their experience with the DPT and the best practice method via a faxed questionnaire and personal contact. We received a very favorable response. RESULTS Eighty percent (80%; 20 of 25) of the physicians implemented the PDA-based electronic registry in their practices, and 76% (19 of 25) had regular PEA visits to help them with the implementation of guidelines and the handheld system. The total number of implemented NAGYKALDI AND MOLD principles was between three and six per physician. Statistical analysis by patient demonstrated a significant improvement (p, 0.05) in nine of 10 measures included in the audit (see Table 1). The absolute increase in these values was between 3% and 25% within a 4-month period. The study was conducted in the late winter months, and this factor had an impact on flushot rates. The short study period (9 months total) probably made documentation of eye exams more difficult. Implementation of the DPT electronic registry also showed a strong correlation with the quality indicator improvement of the measurement and documentation of HbA1c values measured in the last 3 months when pre- and post-intervention values were compared by physician (p, 0.005), while PEA visits similarly increased the number of foot exams and retinal exams performed in the last year (p and 0.02, respectively) in Pearson s correlation analysis. These improvements correspond with previous reports on the effectiveness of electronic systems to improve disease management. 6 SUMMARY From our pilot study we conclude that the DPT, a PDA-based patient management system that is focusing on diabetes care, is a low-cost, feasible, easily implementable, and very effective paper-less tool that significantly improves patient care and documentation in primary care practices. A practice-based research network provides an ideal environment to test and implement this tool, and methodological or TABLE 1. QUALITY IMPROVEMENT RATES BY PATIENT ACEI, Retinal Urinanalysis Retinal if Foot exam HbA1c in ACEI, if exam Flu Pneumovax Coverage HTN exam (ordered) measurement proteinuria proteinuria (documented) shot shot Pre-intervention 69% 72% 41% 89% 46% 74% 32% 53% 38% Post-intervention 91% 82% 62% 97% 68% 93% 35% 69% 63% Improvement (%) 22% 10.5% 21% 8% 21% 19% 3% 16% 25% There were 595 patients pre-intervention and 582 patients post-intervention (p, for all changes). ACEI, angiotensin converting enzyme inhibitor; HTN, hypertensive.

5 DIABETES PATIENT TRACKER 1001 technical support of PEAs ensures clinic staff compliance and satisfaction. ACKNOWLEDGMENTS This work was funded in part by the Oklahoma Foundation for Medical Quality. REFERENCES 1. Mold JW, Gregory ME: Best practice research. Fam Med 2003;35: O Connor PJ: Electronic medical records and diabetes care improvement: are we waiting for Godot? Diabetes Care 2003;26: Montori VM, Dinneen SF, Gorman CA, Zimmerman BR, Rizza RA, Bjornsen SS, Green EM, Bryant SC, Smith SA: The impact of planned care and a diabetes electronic management system on community-based diabetes care: the Mayo Health System Diabetes Translation Project. Diabetes Care 2002;25: Gorman CA, Zimmerman BR, Smith SA, Dinneen SF, Knudsen JB, Holm D, Jorgensen B, Bjornsen S, Plant K, Hanson P, Rizza RA: DEMS a second generation on diabetes electronic management system. Comput Methods Programs Biomed 2000;62: Baker AM, Lafata JE, Ward RE, Whitehouse F, Divine G: A web-based diabetes care management support system. Jt Comm J Qual Improv 2001;27: Sidorov J, Gabbay R, Harris R, Shull RD, Girolami S, Tomcavage J, Starkey R, Hughes R: Disease management for diabetes mellitus: impact on hemoglobin A1c. Am J Manag Care 2000;6: Address reprint requests to: Zsolt Nagykaldi, Ph.D. Department of Family and Preventive Medicine College of Medicine University of Oklahoma Health Sciences Center 900 NE 10th Street Oklahoma City, OK zsolt-nagykaldi@ouhsc.edu