Health Information Technology's Potential to Improve Care

What Is the Reality?  

Texas Medicine Logo(1)

Symposium on Health Information Technology - July 2006

By Mark G. Weiner, MD; David A. Dorr, MD, MS; Jennifer Hornung Garvin, PhD; Abel N. Kho, MD; Eric Pifer, MD; and David W. Bates, MD  

"Potential just means you ain't done it yet."  - Darrell Royal  

With a growing body of evidence supporting the benefits of clinical decision support and clinical physician order entry (CPOE), and even a presidential executive order (1) calling for the nationwide implementation of electronic medical records (EMRs) within 10 years, the potential of health information technology (HIT) to improve the practice of medicine seems closer than ever.

This optimistic view is tempered by many factors, not the least of which is that the promise of information technology becoming commonplace "in the not-too-distant future" has existed for at least 35 years (2). The duration of this as-yet-unmet potential is a source of understandable cynicism regarding HIT and likely plays a role in decisions to delay implementation of EMR systems, especially in light of highly publicized failures of implementation (3-5) and the impact on smaller clinical practices for whom the financial risks and other uncertainties are too great to bear. 

Adding to the hesitation in adoption, the literature on the value of HIT is not uniformly positive, and indeed, some reports have found negative unintended consequences of HIT (6,7).    

We believe that for various reasons, the HIT wave is about to break. Clearly, however, the current level of HIT adoption is low in Texas as well as the rest of the nation.

This paper will provide a context for the conflicting assessments of HIT's value in reducing medical errors and improving physician performance and quality of care. It also will describe keys to realizing HIT benefit, since implementation alone does not ensure that potential benefits will be realized.

Clinical controversies abound within medicine, and the decision to implement an EMR system can be construed as any other therapeutic intervention whose impact in patient care must be studied. Is thrombolytic therapy, angioplasty, or bypass surgery more appropriate in the care of patients with myocardial infarction? What threshold low-density lipoprotein (LDL) value represents optimal cholesterol management? What is the appropriate mammogram screening interval, and when should screening begin? Each of these clinical areas, and many others, has been examined extensively in studies totaling millions of dollars and enrolling tens of thousands of patients. While we have achieved some consensus on these issues, controversy remains and new studies continue, as well they should, as the clinical context in which these studies were initially conducted continues to change and the results obtained 10 years ago may no longer apply.

Despite clinical controversies about their value and role in health care, technological developments within the pharmaceutical and medical instrumentation industries seem to enjoy a more rapid adoption, sometimes with beneficial effects, though at times adoption is criticized for being too rapid and too widespread (8-11).

There are numerous reasons why information technologies have not been adopted as rapidly as other similarly controversial technologies. Some relate to patient demand, sometimes ahead of the evidence (12); some are adopted because cost-effectiveness models are developed to convince insurance companies to cover the treatments and services and some because the issue is well-studied through adequate funding. None of these circumstances apply to HIT. Patients have not demanded it, the financial value of HIT is split across too many stakeholders for anyone to have the courage to make the investment, and the funding for studying the benefit of information technology has lagged behind other high-cost interventions that have a much narrower spectrum of potential benefit than information technology, which potentially could effect every person every time he or she comes into contact (in person or remotely) with a health care professional.

In this context, it is not surprising that the numerous published reviews of the value of information technology have been lukewarm at best. A recent review of computerized clinical decision support (13) reported that improvements in process and outcomes measures related to the use of HIT occurred in only 64 percent of studies that were specially selected for their methodological rigor. Other reviews with different selection criteria have reported similar results (14,15). The authors of these reviews wisely do not suggest a threshold level of success that would warrant adoption, as waiting for a target threshold overlooks the value of the successful components of these clinical decision support systems.

We believe two major changes will result in key benefits of HIT: implementation of EMRs linked with clinical decision support and development of broad clinical data exchange.  In this paper, we first discuss issues relating to EMRs and clinical decision support, and then the steps toward interoperability of clinical data.  

EMRs and Decision Support Systems  

The definition of an EMR and its components varies widely, though usually it consists of computerized interfaces that manage one or more of the following components of a clinical encounter:

  • Documentation (notes, allergies, problem lists),
  • Physician order entry,
  • Pharmacy management,
  • Results reporting, and
  • Communication with other members of the care team.

Computerized implementation of each of these components has a counterpart process in the paper world and, at a minimum, offers intuitive and important advantages over the paper processes, including improved legibility and information availability. Unlike the paper chart, an EMR is never unavailable because it is checked out to another physician or health care professional or, worse yet, lost. 

In one study (16), missing clinical information was associated with at least 10 percent of all reported errors in nine family medicine practices. Another study (17) noted the requirement of spending precious time to track down unavailable data, which detracts from time better spent on direct patient care. Other studies have shown the cost savings related to the reduction in tests reordered simply because the prior results were not available in the paper medical record (18,19). EMRs do not eliminate these problems in information access, but they do make the occurrence far less likely (17).

One of HIT's most significant benefits arises from its ability to provide clinical decision support through intelligently integrating data from different components. This is too cumbersome, if not impossible, using paper. Clinical decision support improves the practice of medicine through proactive and reactive guidance that makes it harder for clinicians to make mistakes and easier for them to manage patients appropriately. Computerized clinical decision support systems (CDSS) can be categorized in the following areas:

  •  Therapeutic alerts - reactive alarms triggered in response to a clinician's action
    •  Therapeutic duplication 
    •  Overall drug dose limits 
    •  Dose limits based on diagnosis, age, laboratory data 
    •  Drug allergy warnings 
    •  Drug-to-drug interactions 
  •  Corollary orders - recommendations for testing or additional therapy in response to another order made by a clinician 
  •  Drug-dosing suggestions - proactive recommendations regarding appropriate dosing of medication chosen by a physician 
  •  Ancillary test alerts - proactive alarms triggered in response to a test abnormality that a doctor may otherwise overlook or not recognize in a timely fashion 
  •  Clinical reminders - proactive suggestions by the system that recommend an action the physician may otherwise fail to perform, triggered in response to patient characteristics and time since the action was last taken 
  •  Clinical guideline implementation -- an ordered series of clinical management suggestions triggered in response to the presence of disease  

Historically, the most consistently successful forms of CDSS have been therapeutic alerts. Numerous studies exist showing the benefit of CPOEs that incorporate alerts for drug dosing, drug allergy, and drug-to-drug interactions. A clinical trial at the LDS Hospital in Salt Lake City (20) showed reduced inappropriate antibiotic dosing - prescribing antibiotics to which the patient had an expressed allergy - and overall adverse drug events (ADEs). 

At the Brigham and Women's Hospital in Boston, using a therapeutic alert system reduced rates of serious medication errors (preventable ADEs + potential ADEs) by 55 percent (21). After subsequent improvements (22), the rate of non-intercepted serious medication errors fell further, to a rate that was 86 percent below the baseline. 

At the Wishard Memorial Hospital in Indianapolis, a randomized study of inpatient medical teams using CPOE generated 12.7 percent lower charges and a 0.9-day shorter length of stay than teams using handwritten orders (23,24).

While therapeutic alerts have reduced overall medical error rates, evidence suggests that these alerts do not always provide useful information and are frequently overridden (25). The distraction of unnecessary alerts can cause a physician to overlook an important warning or incorrectly assume the alerts are meaningful. However, these findings also point to mechanisms to improve the appropriateness and acceptability of alerts by tailoring them to those most clinically significant and most frequently accepted by clinicians. 

In particular, alerts regarding duplicate drug class categories were accepted 77 percent of the time. Alerts for absolute drug-drug contraindications occurred 13 times with a forced 100-percent compliance. Relative drug-to-drug interactions that could be overridden after providing a rationale were accepted 41 percent of the time (26).

Corollary orders are related to therapeutic alerts in that they occur in response to another medication. However, rather than suggesting a physician not proceed with the order, a corollary order is intended to remind the clinician that an additional order may be required to complement the intended order. Examples include recommendations to order drug level, pregnancy, or kidney or liver function testing when ordering certain medications. As with therapeutic alerts, studies of corollary orders have consistently shown improvement in practitioner performance, although recommendations for corollary orders are not consistently followed (27).

Because these studies were conducted at institutions where the culture of information technology has existed for many years and typically in the inpatient setting, their findings may not be generalized to other institutions and to the ambulatory setting. However, the consistent success of drug allergy and drug dose alerts and corollary orders in reducing errors, the availability of information and plausible mechanisms needed to trigger these alerts accurately, and the importance of the information toward appropriate clinical care make them an essential function of any EMR system (28).

Even in a computerized health care environment in which dosing errors are minimized and reminders about important corollary orders are provided, errors are possible when the physician fails to recognize in a timely fashion the critical results from the laboratory and other ancillary tests ordered. Rather than passively waiting for a clinician to log into the system, EMRs can alert the physician to critical lab values and other test results.  In a controlled study of an automated pager system linked to an inpatient lab results monitoring system, the time to address an abnormal lab result was decreased more than 30 minutes in the group that received the pages. 

A trend toward improving the time to treatment was noted in the subgroup with critical labs in which the non-pager group received the usual hospital protocol of a phone call to the nursing unit.

Despite the well-recognized benefit of vaccinations, the proportion of people older than 65 who receive a flu shot has not improved from 66 percent between 1997 and 2002 (29). Only 64.1 percent reported that they received a vaccination in 2003 (30). 

Computerized reminders for vaccinations (31) and other health maintenance activities such as mammography (32) increase adherence with these modalities. Cervical cancer screening with a Pap smear is an important exception (33). Many factors may contribute to the lack of improvement in Pap smear rates using computerized reminder systems. One may have to do with the clinical office's workflow and resources. The Pap smear, unlike the mammography, is an in-office procedure, and unlike the vaccination, may require extra physician time, space, and special equipment. Such a hypothesis is supported by the view that a computerized decision support system cannot overcome and in fact may highlight inadequate office workflow. 

Drug-dosing recommendations differ from drug-dose alerts in that a recommended dose and strategies for changing it are provided proactively to the physician, rather than reacting to a drug order the clinician provides. Improved administration of drugs having narrow therapeutic windows such as warfarin (34), heparin (35), theophylline (36), and lidocaine (37) have been shown, but results are not consistent (38,39).  

While, overall, these results are promising, systems that focus only on making recommendations regarding drug dosing address only half of the issue. The clinician must still make the correct decisions regarding the need for the medication in the first place. Therefore, decision support for drug dosing will be more successful in the context of an electronic clinical guideline system that can address the appropriateness of initiating the medication and considering alternative and additional strategies for management of the particular disease.

There are multiple challenges with electronic implementation of clinical guidelines that are associated with their likelihood of success. At the most fundamental level is the strength of evidence that underlies the clinical guideline itself (40). An electronic version of a guideline cannot be expected to improve outcomes if the therapeutic modalities recommended by the guideline are not effective.

Second, the language of clinical guidelines often includes a great deal of subjective language, making it difficult to apply electronic data in an automated fashion to the decision making regarding the suitability of a patient expected to benefit from the guideline.

Last, the guideline implementation must be intimately linked with the EMR system. The mere availability of online guideline flow sheets is not conducive to their application in a busy medical practice where accessing the guideline distracts from the workflow of the patient encounter.

Addressing all three challenges is difficult. Even where there is evidence to support the value of an intervention, such as blood pressure reduction, there is disagreement over the most effective pharmaceutical modalities to achieve the desired outcome. Therefore, using computerized guidelines to direct medical management of hypertension to lower blood pressure has not been successful (41,42). Similarly, studies of computerized guidelines in managing heart disease (43) and asthma (44) show neither physician adherence to guidelines nor better patient outcomes when physicians were given recommendations according to computerized guidelines.

Interestingly, a randomized controlled trial of a different computerized guideline system that e-mailed physicians outside the context of a clinical visit regarding the potential need for medical management of cholesterol was successful at effecting change to the medical regimen earlier and with a trend toward improved LDL control (45). 

Managing Chronic and Complex Illnesses  

Recently, there has been an increased focus on the ability of information technology to address patients with complex management needs (46). Examples include patients with social, economic, and cultural barriers to care and health, those with multiple chronic illnesses, and the frail elderly (who often also have barriers to care and multiple chronic illnesses) (47). Patients with multiple conditions, especially, are at the heart of the issue, as they represent the largest concentration of health care utilization and costs (67 percent of the Medicare population have two or more chronic illnesses and represent 96 percent of the costs). They are twice as likely to have adverse events (48).

Most HIT, as implemented, is not used to generate information about the quality of care for populations of patients, a key function in care and disease management. A Commonwealth Fund report found EMR practices could generate this data only 21 percent of the time (49). 

Care management models, such as the Chronic Care Model, highlight the need for specific information structures and functions to manage a population (e.g., a disease registry with reports), but most implemented EMR systems do not have this functionality, even when they have the data points needed. They cannot adequately summarize results for individuals or populations. For small practices, where the EMR cost must be recouped relatively quickly, implementing such functionality is unreimbursed under most payment systems (50). In high HIT use systems, standalone registries and care management databases are frequently kept separately from EMRs.       

Finally, when implemented, most population management systems focus only on the generation of quality measures, not patient barriers, exclusions, or other social needs.  In a large study of a broad care management system, these barriers represented 20 percent of patients' needs (51). These barriers can be overcome by actively reaching out to these populations (52,53), but these issues must be addressed by systems and measurements to avoid penalizing these populations further.

Most care and disease management successes have been based on single diseases. While many of the key functions (education, self-management support, team-based care) may extend beyond single diseases, few programs have been able to achieve strong results with patients who have multiple diseases. 

One such program, at Intermountain Healthcare, saw strong improvements in diabetes and depression outcomes, as well as improvements that were not as strong for heart disease and other conditions (54). HIT may clearly help address each of these issues, but the system into which it is placed requires significant change - in workflow, in leadership, in teamwork, and in reimbursement - to extend successes to these vulnerable patients (55).

Finally, implementation of guidelines is difficult for these patients. Studies frequently exclude patients with multiple conditions, and guidelines are developed for single diseases, making the applicability of recommendations questionable (48).

In a study of applying standard clinical guidelines to a complex, elderly patient with a plausible set of diagnoses, the investigators found that by following medication recommendations from guidelines that covered different diseases, drug-to-drug interactions would occur, while in other cases, therapeutic recommendations were mutually exclusive (47). Only 18 percent of studies have shown changes in outcomes related to decision support in these diseases without broader changes to care delivery and population management (13). Further research is needed to understand how to overcome these problems. 

Detecting Adverse Drug Events  

In the preceding sections, we discussed the challenges and models for success of EMR systems toward improving the quality of care.  The common feature of these systems is they can only address known errors the system is designed to look for.

Another benefit of using EMRs is the ability to integrate and present information that can be analyzed to identify unexpected new adverse events and allow a more thorough accounting of the true incidence of known ADEs and medical errors than currently possible.

An ADE is an injury due to a medication that may or may not be the result of a medication error.  For example, cough is an idiosyncratic reaction to an angiotensin-converting enzyme (ACE) inhibitor. Therefore, a patient who coughs while taking an ACE inhibitor for the first time is experiencing an ADE. However, this does not represent a medical error unless the patient was known to have the symptom before, which warranted the cessation of the drug (56).

Several studies have evaluated ADEs and medication errors in outpatient settings. One study determined the incidence of drug complications in outpatients using prescriptions to be 18 percent by patient survey and only 3 percent by manual chart review (57). Still another prospective cohort study using survey and chart review methods found the incidence of ADEs to be 27 per 100 outpatients. Again the patient survey provided a greater number of adverse events using survey methods over chart review (58). Although these studies suggest that patient surveys will be important to collect ongoing information on ADEs, the EMR can be a common repository for this information and can provide a mechanism to inform a physician about symptoms he or she may not otherwise learn about in the typical clinical encounter.

EMR technology has been studied for its ability to capture ADEs and medical errors to eliminate the laborious chart reviews or surveys typically used in research. In particular, triggers - such as the presence of pharmacy data suggesting an antidote; lab findings such as an elevated drug level or a drug-specific change in blood count or kidney or liver function; and diagnostic codes - are retrievable with discrete searching of the structured components of the EMR database.

One area of particular promise is natural language processing that analyzes the unstructured text that comprises a great deal of clinical documentation, and mapping the content to databases of clinical concepts (59). By using these types of technology, the occurrence of ADEs can be detected and tracked automatically and, in a percentage of cases, can be potentially prevented.

Morimoto et al demonstrated specific detection and classification methods to facilitate ADE detection with either manual review processes or automated methods described above (56). The investigators found that both automated methods and manual chart review discovered distinct events, and they suggested that both methods need to be used for thorough data capture.

Gurwitz et al applied computer-generated signals in conjunction with manual methods and found ADE rates of 50 per 1,000 persons, of which 28 percent were preventable, i.e., true errors (60).  

In another study, Honigman et al found that screening EMRs by using a combination of lab and pharmacy data and applying automated natural language techniques to progress notes found an incidence of 5.5 ADEs per 100 patients (61).

These combined ADE detection methods were moderately sensitive and specific and did pick up a number of ADEs that would otherwise have been missed by manual methods alone. However, because ADE prevalence is low, the positive predictive of the putative ADE detected by the combined methods is quite low.

The Institute of Medicine publication, To Err is Human: Building a Safer Health System , found that between 44,000 and 98,000 deaths per year are attributable to medical errors in the inpatient setting, with presumably many more related to errors in the ambulatory setting (62). While the accuracy of the actual number of deaths has been debated (63,64), it is generally held that whatever the number of deaths, it is too high. The report galvanized the discussion of medication errors, their causes, and mechanisms to mitigate them. 

What is not frequently discussed is that the estimations of death in that report were based on a series of studies in 1991 that used data collected from one state in 1984. Regardless of the accuracy of the analysis, the fact that the best available data for this landmark report are 20 years old strongly suggests that good data on the incidence of medical errors are hard to find. Therefore, in addition to the role of EMRs in improving patient safety, automated analysis of EMRs can provide the data that can show when the IOM report no longer applies.

Health Information Exchange Potential  

The full benefits of EMR implementation will be realized when EMRs have the capability to readily exchange information. EMR systems have developed in provincial shops, both commercial and academic. However, health care delivery in this country is not so well delineated. Rather, patients can and do receive health care through multiple venues (65). Regional health information organizations (RHIOs) can span systems to provide clinicians with up-to-date information for emergency care of itinerant patients. 

Working RHIOs, such as the Indiana Health Information Exchange (66) and the Massachusetts Health Data Consortium (67) serve as showcase models for how sharing information benefits patient care across multiple institutions. By sharing information on prior treatments and tests, practitioners in emergency departments avoid redundant costs and efforts (68). Public health departments capture chief complaint data for biosurveillance (69). And electronic laboratory reporting speeds detection and improves completeness of reporting of notifiable infectious diseases (70).  

To accomplish these goals, RHIOs depend on integrating commonly recognized standards and policies. Standard vocabularies such as LOINC (71) and SNOMED (Systematized Nomenclature of Medicine; College of American Pathologists) serve as a common language that institutions use, and the HL7 messaging standard (available at .) sets rules for how institutions can use this common tongue to communicate. Creating and maintaining a global patient index ensures that patients can be correctly identified at any participating institution (72). Adopting standardized vocabularies remains free through the National Library of Medicine, and commercial companies provide increasingly sophisticated tools to implement HL7 and a global patient index across an enterprise application.  

Anecdotally, and by common sense, there is little doubt that sharing health information across institutions can positively impact patient care. Having the latest electrocardiogram to compare against an urgent new study for chest pain is invaluable. Quantifying these benefits, however, requires near-complete capture of communitywide information on test usage, medications, patient outcomes, and payer costs (68). For the near future, such comprehensive data capture is unlikely, and RHIOs will prove their mettle clinically to improve emergency care and public health.

More than 100 communities are creating health information exchanges (73). To ensure that these separate efforts can ultimately share information with each other, a decentralized network that nevertheless shares common standards and policies is proposed (74). Common policies governing the use and security of stored information create an environment of trust and assurance that data use remains in the control of the contributing institution and ultimately the patients they serve. Ironically, this push for interoperability between institutions underscores the need for institutional EMRs to adhere to recognized standards. 


Implementing EMR systems with clinical decision support and clinical data exchange will have greater benefits when including both together. The biggest hurdle to EMR implementation is financial, especially in small offices. More subtle, but perhaps equally important, is the quality of the decision support. Clinical data exchange represents a political and societal challenge, as many of the issues relate to the tensions between societal benefit and confidentiality.

For various reasons, we believe HIT is finally about to achieve its potential. For this to occur, however, many actions are needed, and what is done at the organizational and state levels will be important, especially with respect to financing and interoperability. 

HIT and EMR systems have many potentially beneficial uses, but we still have much to learn about their use, correct implementation, and underlying components for success. A number of projects are under way to help understand what contextual and structural elements create success in EMR systems, but our ability to generate success with technology depends fundamentally on the system into which you apply it. Appropriately deploying CPOE, decision support, ADE detection and mitigation, and care management requires both a sense of optimism at its promise and caution at the unexplored and unforeseen consequences from use. 

Although not every EMR innovation has been successful, some of the simplest capabilities have proven to be effective - from error reduction and cost savings related to improved chart availability, to avoidance of unnecessary morbidity from drug interactions and allergies, to the enhanced ability to adhere to recognized standards of good clinical practice. 

For these reasons alone, HIT has demonstrated its value. The informatics and broader research community will continue to innovate and conduct careful research to fully optimize the benefit of EMRs. 


  1. Bush GW. Executive Order: Incentives for the Use of Health Information Technology and Establishing the Position of the National Health Information Technology Coordinator. April 27, 2004. Available at Accessed April 13, 2006. 
  2. Schwartz WB. Medicine and the computer. The promise and problems of change. N Engl J Med . 1970;283:1257-1264. 
  3. Massaro TA. Introducing physician order entry at a major academic medical center: I. Impact on organizational culture and behavior. Acad Med . 1993;68:20-25.  
  4. Ornstein C. Hospital heeds doctors, suspends use of software; Cedars-Sinai physicians entered prescriptions and other orders in it, but called it unsafe. Los Angeles Times . January 22, 2003; B1. 
  5. Scott JT, Rundall TG, Vogt TM, Hsu J. Kaiser Permanente's experience of implementing an electronic medical record: a qualitative study. BMJ . 2005; 331(7528):1313-1316.
  6. Ash JS, Berg M, Coiera E. Some unintended consequences of information technology in health care: the nature of patient care information system-related errors. J Am Med Inform Assoc. 2004;11(2):104-112. 
  7. Koppel R, Metlay JP, Cohen A, et al. Role of computerized physician order entry systems in facilitating medication errors. JAMA . 2005;293:1197-1203. 
  8. Halm EA, Chassin MR, Tuhrim S, et al. Revisiting the appropriateness of carotid endarterectomy. Stroke . 2003;34:1464-1471. 
  9.  Holmes JH, Metlay J, Holmes WC, Mikanatha N. Poster: Developing a patient intervention to reduce antibiotic overuse. Proc AMIA Symp . 2003:864. 
  10.  Lousuebsakul V, Knutsen SM, Singh PN, Gram IT. Is colposcopic biopsy overused among women with a cytological diagnosis of atypical squamous cells of undetermined significance (ASCUS)? J Womens Health (Larchmt). 2003;12:553-559. 
  11. Schneider EC, Leape LL. Weissman JS. Piana RN. Gatsonis C. Epstein AM. Racial differences in cardiac revascularization rates: does "overuse" explain higher rates among white patients? Ann Intern Med . 2001;135:328-337. 
  12.  Farquhar C, Marjoribanks J, Basser R, Hetrick S, Lethaby A. High dose chemotherapy and autologous bone marrow or stem cell transplantation versus conventional chemotherapy for women with metastatic breast cancer. Cochrane Database Syst Rev . 2005 July 20;(3):CD003142. 
  13.  Garg AX, Adhikari NK, McDonald H, et al. Effects of computerized clinical decision support systems on practitioner performance and patient outcomes: a systematic review. JAMA . 2005;293:1223-1238. 
  14. Johnston ME, Langton KB, Haynes RB, Mathieu A. Effects of computer-based clinical decision support systems on clinician performance and patient outcome: a critical appraisal of research. Ann Intern Med . 1994;120:135-142.
  15. Hunt DL, Haynes RB, Hanna SE, Smith K. Effects of computer-based clinical decision support systems on physician performance and patient outcomes: a systematic review. JAMA . 1998;280:1339-1346.
  16. Elder NC, Vonder Meulen M, Cassedy A. The identification of medical errors by family physicians during outpatient visits. Ann Fam Med . 2004;2:125-129.
  17. Smith PC, Araya-Guerra R, Bublitz C, et al. Missing clinical information during primary care visits. JAMA. 2005;293:565-571.
  18. Tierney  WM, McDonald CJ, Martin DK, Rogers MP. Computerized display of past test results: effects on outpatient testing. Ann Intern Med. 1987;107:569-574.
  19. Stair TO. Reduction of redundant laboratory orders by access to computerized patient records . J Emerg Med . 1998;16:895-897.  
  20. Evans RS, Pestotnik SL, Classen DC, et al. A computer-assisted management program for antibiotics and other antiinfective agents. N Engl J Med . 1998;338:232-238.
  21. Bates DW, Leape LL, Cullen DJ, et al. Effect of computerized physician order entry and a team intervention on prevention of serious medication errors. JAMA . 1998;280:1311-1316.
  22. Bates DW, Teich JM, Lee J, et al. The impact of computerized physician order entry on medication error prevention. J Am Med Inform Assoc . 1999;6:313-321.
  23. McDonald CJ, Overhage JM, Tierney WM, et al. The Regenstrief Medical Record System: a quarter century experience. Int J Med Inform . 1999;54:225-253.
  24. Tierney WM, Miller ME, Overhage JM, McDonald CJ. Physician inpatient order writing on microcomputer workstations: effects on resource utilization. JAMA . 1993;269:379-383.
  25. Weingart SN, Toth M, Sands DZ, Aronson MD, Davis RB, Phillips RS. Physicians' decisions to override computerized drug alerts in primary care. Arch Intern Med . 2003;163:2625-2631.
  26. Shah NR, Seger AC, Seger DL, et al. Improving acceptance of computerized prescribing alert s in ambulatory care . J Am Med Inform Assoc . 2006;13:5-11.  
  27. Overhage JM, Tierney WM, Zhou XH, McDonald CJ. A randomized trial of "corollary orders" to prevent errors of omission. J Am Med Inform Assoc . 1997;4:364-375.
  28. Bell DS, Cretin S, Marken RS, Landman AB. A conceptual framework for evaluating outpatient electronic prescribing systems based on their functional capabilities. J Am Med Inform Assoc . 2004;11(1):60-70.
  29. Centers for Disease Control and Prevention. Influenza vaccination levels among persons aged ≥65 years and among persons aged 18-64 years with high-risk conditions - United States, 2003. MMWR Morb Mortal Wkly Rep . 2005;54:1045-1049.
  30. Centers for Disease Control and Prevention. Influenza and pneumococcal vaccination coverage among persons aged ≥65 years and persons aged 18-64 years with diabetes or asthma - United States, 2003. MMWR Morb Mortal Wkly Rep . 2004;53;1007-1012.
  31. McDowell I, Newell C, Rosser W.  Comparison of three methods of recalling patients for influenza vaccination. CMAJ . 1986;135;991-997.
  32. Burack RC, Gimotty PA, George J, Stengle W, Warbasse L, Moncrease A.  Promoting screening mammography in inner-city settings: a randomized controlled trial of computerized reminders as a component of a program to facilitate mammography. Med Care . 1994;32:609-624.
  33. Shea S, DuMouchel W, Bahamonde L. A meta-analysis of 16 randomized controlled trials to evaluate computer-based clinical reminder systems for preventive care in the ambulatory setting. J Am Med Inform Assoc . 1996;3:399-409.
  34. Abbrecht PH, O'Leary TJ, Behrendt DM.  Evaluation of a computer-assisted method for individualized anticoagulation: retrospective and prospective studies with a pharmacodynamic model. Clin Pharmacol Ther . 1982;32:129-136.
  35. Mungall DR, Anbe D, Forrester PL, et al. A prospective randomized comparison of the accuracy of computer-assisted versus GUSTO nomogram-directed heparin therapy . Clin Pharmacol Ther . 1994;55:591-596.
  36. Verner D, Seligmann H, Platt S, et al. Computer assisted design of a theophylline dosing regimen in acute bronchospasm: serum concentrations and clinical outcome.  Eur J Clin Pharmacol . 1992;43:29-33.
  37. Rodman JH, Jelliffe RW, Kolb E, et. al. Clinical studies with computer-assisted initial lidocaine therapy. Arch Intern Med . 1984;144:703-709.
  38. Poller L, Wright D, Rowlands M.  Prospective comparative study of computer programs used for management of warfarin. J Clin Pathol . 1993;46:299-303.
  39. Casner PR, Reilly R, Ho H. A randomized controlled trial of computerized pharmakokinetic theophylline dosing versus empiric physician dosing. Clin Pharmacol Ther . 1993;53:684-690.
  40. van Tulder MW, Tuut M, Pennick V, Bombardier C, Assendelft WJ. Quality of primary care guidelines for acute low back pain. Spine . 2004;29(17):E357-E362.
  41. Coe FL, Norton E, Oparil S, Tatar A, Pullman TN. Treatment of hypertension by computer and physician: a prospective controlled study. J Chronic Dis . 1977;30:81-92.
  42. McAlister NH, Covvy HD, Tong C, Lee A, Wigle ED. Randomized controlled trial of computer assisted management of hypertension in primary care. BMJ . 1986;293:670-674.
  43. Tierney WM, Overhage JM, Murray MD, et al. Effects of computerized guidelines for managing heart disease in primary care. J Gen Intern Med . 2003;18:967-976.
  44. Tierney WM, Overhage JM, Murray MD, et al. Can computer-generated evidence-based care suggestions enhance evidence-based management of asthma and chronic obstructive pulmonary disease? A randomized, controlled trial. Health Serv Res . 2005;40(2):477-497.
  45. Lester WT, Grant R, Barnett GO, Chueh H. Randomized controlled trial of an informatics-based intervention to increase statin prescription for secondary prevention of coronary disease. J Gen Intern Med . 2006;21:22-29.
  46. Hillestad R, Bigelow J, Bower A, et al. Can electronic medical record systems transform health care? Potential health benefits, savings, and costs. The adoption of interoperable EMR systems could produce efficiency and safety savings of $142-$371 billion. Health Aff (Millwood). 2005;24(5):1103-1117.
  47. Boyd CM, Darer J, Boult C, Fried LP, Boult L, Wu AW. Clinical practice guidelines and quality of care for older patients with multiple comorbid diseases: implications for pay for performance. JAMA . 2005;294:716-724.
  48. Anderson GF. Medicare and chronic conditions. N Engl J Med . 2005;353(3):305-309.
  49. Audet AM, Doty MM, Shamasdin J, Schoenbaum SC. Measure, learn, and improve: physicians' involvement in quality improvement. Health Aff (Millwood). 2005;24(3):843-853.
  50. Miller RH, West C, Brown TM, Sim I, Ganchoff C. The value of electronic medical records in solo or small group practices. Physicians' EMR adoption is slowed by a reimbursement system that rewards the volume of services more than it does their quality. Health Aff (Millwood). 2005;24(5):1127-1137.
  51. Dorr DA, Wilcox A, Burns L, Brunker CP, Narus SP, Clayton PD. Implementing a multidisease chronic care model in primary care using people and technology. Dis Manag . 2006;9(1):1-15.
  52. Kreps GL. Disseminating relevant health information to underserved audiences: implications of the Digital Divide Pilot Projects. J Med Libr Assoc . 2005;93(4 suppl):S68-S73.
  53. Shea S, Weinstock RS, Starren J, et al. A randomized trial comparing telemedicine case management with usual care in older, ethnically diverse, medically underserved patients with diabetes mellitus. J Am Med Inform Assoc . 2006;13(1):40-51.
  54. Dorr DA, Wilcox A, Donnelly SM, Burns L, Clayton PD. Impact of generalist care managers on patients with diabetes. Health Serv Res . 2005;40(5 pt 1):1400-1421.
  55. Scott JT, Rundall TG, Vogt TM, Hsu J. Kaiser Permanente's experience of implementing an electronic medical record: a qualitative study. BMJ. 2005;331(7528):1313-1316.
  56. Morimoto T, Gandhi TK, Seger AC, Hsieh TC, Bates DW. Adverse drug events and medication errors: detection and classification methods. Qual Saf Health Care. 2004;13:306-314.
  57. Gandhi TK, Burstin HR, Cook EF, et al.Drug complications in outpatients. J Gen Intern Med . 2000;15(3):149-154.
  58. Gandhi TK, Weingart SN, Borus J, et al. Patient safety: adverse drug events in ambulatory care. N Engl J Med . 2003;348:1556-1564.
  59. Bates DW, Evans RS, Murff H, Stetson PD, Pizziferri L, Hripcsak G.  Detecting adverse events using information technology. J Am Med Inform Assoc . 2003;10:115-128.
  60. Gurwitz JH, Field TS, Harrold LR, et al. Incidence and preventability of adverse drug events among older persons in ambulatory care. JAMA . 2003;289:1107-1116.
  61. Honigman B, Lee J, Rothschild J. et al. Using computerized data to identify adverse drug events in outpatients. J Am Med Inform Assoc . 2001;8:245-266.
  62. Kohn KT, Corrigan JM, Donaldson MS. To Err Is Human: Building a Safer Health System. Washington, DC: National Academy Press; 1999.
  63. McDonald CJ, Weiner M, Hui SL. Deaths due to medical errors are exaggerated in Institute of Medicine report. JAMA. 2000;284:93-95.
  64. Leape LL. Institute of Medicine medical error figures are not exaggerated. JAMA. 2000;284:95-97.
  65. Finnell JT, Overhage JM, Dexter PR, Perkins SM, Lane KA, McDonald CJ. Community clinical data exchange for emergency medicine patients. Proc AMIA Symp . 2003:235-238.
  66. McDonald CJ, Overhage JM, Barnes M, et al. The Indiana Network for Patient Care: a working local health information infrastructure. An example of a working infrastructure collaboration that links data from five health systems and hundreds of millions of entries. Health Aff (Millwood) . 2005;24(5):1214-1220.
  67. Halamka J, Aranow M, Ascenzo C, et al. Health care IT collaboration in Massachusetts: the experience of creating regional connectivity. J Am Med Inform Assoc . 2005;12:596-601.
  68. Overhage JM, Dexter PR, Perkins SM, et al. A randomized, controlled trial of clinical information shared from another institution. Ann Emerg Med . 2002;39(1):14-23.
  69. Tsui FC, Espino JU, Dato VM, Gesteland PH, Hutman J, Wagner MM. Technical description of RODS: a real-time public health surveillance system. J Am Med Inform Assoc . 2003;10(5):399-408.
  70. Panackal AA, M'ikanatha NM, Tsui FC, et al. Automatic electronic laboratory-based reporting of notifiable infectious diseases at a large health system. Emerg Infect Dis . 2002;8(7):685-691.
  71. Forrey AW, McDonald CJ, DeMoor G, et al. Logical observation identifier names and codes (LOINC) database: a public use set of codes and names for electronic reporting of clinical laboratory test results. Clin Chem . 1996;42(1):81-90.
  72. Gudea S. Deterministic, probabilistic, or fuzzy? A primer on the search algorithms that drive MPI quality. J AHIMA . 2005;76(8):50-54.
  73. Overhage JM, Evans L, Marchibroda J. Communities' readiness for health information exchange: the National Landscape in 2004. J Am Med Inform Assoc . 2005;12(2):107-112.
  74. Halamka J, Overhage JM, Ricciardi L, Rishel W, Shirky C, Diamond C. Exchanging health information: local distribution, national coordination. As more communities develop information-sharing networks, a coordinated approach is essential for linking these networks. Health Aff (Millwood). 2005;24:1170-1179. 

Mark G. Weiner, MD, is the associate director for patient informatics within the Office of Human Research, co-chief of the Biostatistics and Informatics Core of the VA Center for Health Equity Research and Promotion, and committee chair of the University of Pennsylvania Institutional Review Board.  

David A. Dorr, MD, MS, is an assistant professor of medical informatics at Oregon Health & Science University.  

Jennifer Hornung Garvin, PhD, is a medical informatics postdoctoral fellow at the Center for Health Equity Research and Promotion sponsored by the Philadelphia VA Medical Center in association with the University of Pennsylvania School of Medicine.  

Abel N. Kho, MD, is a National Library of Medicine medical informatics fellow with the Regenstrief Institute, Inc., in Indianapolis.  

Eric Pifer, MD, is chief medical informatics officer in the Penn Health System and assistant professor of medicine in the University of Pennsylvania School of Medicine.  

David W. Bates, MD, is the chief of the Division of General Medicine at Brigham and Women's Hospital in Boston and a practicing general internist. He also serves as medical director of clinical and quality analysis for Partner's Healthcare Systems.  

July 2006 Texas Medicine Contents
Texas Medicine Main Page