Theme
Imagining an Electronic Medical Record for Turning Cancer Screening Knowledge into Practice

https://doi.org/10.1016/j.amepre.2009.09.037Get rights and content

Introduction

Policy and healthcare leaders expect electronic medical records (EMRs) to close the gap between what is known and practiced in medicine—the now well-recognized quality chasm.1, 2, 3, 4 This chasm includes problems with the translation of guidelines into practice (the content of care) and problems with how patients and providers communicate throughout the trajectory of care (the processes of care). The EMR is the software and hardware within a medical practice that is expected to replace the paper medical record and help with the content and process of care. A fully functional EMR has been defined as one that can (1) record health information and data on individuals, (2) assist in the management of test results, (3) provide direct entry of pharmacy orders, and (4) provide information to assist physicians with medical decisions.5 At present, however, only about 4% of physicians use a fully functional EMR in their practice.6 The failure to adopt EMRs is partly due to their costs and also their inability to meet practitioners' needs.7, 8, 9 To better express what is needed to make an EMR useful for practice, this article documents the process and results of an effort to identify the EMRs' functional requirements for the Health Resources Services Administration's (HRSA) Cancer Collaborative.10 The purpose of this article is to provide an example of an iterative process of EMR development and implementation that many have recognized is needed to realize a record that will be adopted and used by practitioners.11

The HRSA Cancer Collaborative involved practice improvement teams (e.g., receptionist, medical records clerk, nurses, nurse practitioners, and physicians) who set measurable clinical goals based on existing guidelines and evidence from the medical literature.12, 13, 14, 15 These teams conduct regular assessments and change their practices to improve measures relevant to the clinical goals.16, 17, 18, 19 The overall method of considering these changes as a collaborating team is called a planned approach, and it has been shown to improve both the content and process of care for diabetes, cardiovascular disease, and cancer.10, 14, 20 Because the work depends on measuring and actively managing individual and group data, planned approaches depend on having an automated source of data management like spreadsheets, clinical management software, or an EMR.

Planned approaches to care go beyond improving the content of care and also focus on how to make the process of care more patient centered.1, 21 Patient-centered care includes four fundamental concepts: (1) maintaining dignity and respect, (2) openly sharing information, (3) encouraging participation of the patient, and (4) encouraging collaboration among providers and patients.22 An appropriately designed EMR offers a tool for sharing information while encouraging participation and collaboration in care.23

The cancer screening process is a good prototype for conceiving a patient-centered EMR because it presents the challenge of tracking the care of individuals across the cancer care continuum from prevention through screening, diagnosis, treatment, and end-of-life care (Figure 1).24 Like many other clinical processes, cancer screening involves multiple discrete steps and transitions in care that must be managed for the care to be successful (Figure 1, Table 1).24 Furthermore, many individuals in a practice are eligible for cancer screening and they are traversing the steps and transitions at different points in time. The entire population eligible for cancer screening may reach several hundred patients in a practice or thousands of patients in a group of practices.

The care of such a large group of individuals across the steps and transitions shown in Figure 1 offers many opportunities for an EMR to be an effective tool, including helping the practitioner identify patients due for screening, facilitate the discussion of the screening options, mail reminders for those due, track the results of screening and symptomatic evaluations, and stimulate the physician and patient to take the appropriate actions, while providing educational material for the patient and decision support for the providers. Because many of these steps and transitions in cancer care represent processes that are common to other conditions (e.g., recruitment, follow-up, tracking of individual progress), developing a tool that assists with patient-centered cancer care implementation could have implications for the care of other conditions as well.

A patient case provides a good example of the trajectory of care from a patient's perspective and demonstrates how things can go right and wrong. Although details and names have been changed, the basic facts reflect an actual experience in the cervical cancer screening process.

At age 22, Janis J. moved to a busy urban setting to begin a new job as an advertising executive. She left a small town, where her long-time physician had performed her previous Pap 3 years before. At the time of Janis's first visit to her new physician, Dr. Jones, she was alarmed when Dr. Jones's face looked concerned during the Pap.

Janis felt reassured when Dr. Jones told her she would be called if there were any abnormalities, and she forgot about the visit as she started moving to a new apartment. Seven weeks later, she noted blood spotting her underwear. She decided to go back to see Dr. Jones and noted that she expressed confusion that Janis had not been notified that her Pap result had been abnormal. Dr. Jones noted that Janis might have “carcinoma in situ” of the cervix. Janis reported that that sounded like the carcinoma of the lung that killed her father, so she didn't remember much after that point in the visit. She thought her physician referred her to a gynecologist but she wasn't sure and she felt so healthy she decided to just go on with her life. After several weeks, Dr. Jones called her and insisted she see the gynecologist, who took her history all over again, established a diagnosis of invasive cervical cancer, and recommended a hysterectomy.

Janis considered her options for several more weeks, but finally she decided to have a hysterectomy and receive radiation. Once Janis decided to begin treatment, she insisted that her primary care physician and gynecologist talk to each other because they were giving conflicting recommendations about activity levels and diet. Throughout the experience, Janis noted that no one mentioned human papilloma virus or explained that it was a likely cause of her cervical cancer. It was 6 months from her abnormal Pap to the beginning of her treatment. She has now lived 17 years without evidence of recurrence.

This story highlights the steps in the screening process for the patient and the opportunities for improvement for healthcare providers. Janis reported that she was unaware of the etiology of cervical cancer and the things she could have done to avoid this preventable disease. She was nearly lost to follow-up because of the failure to receive notification of abnormal results. Her cancer was treated because she paid attention to her vaginal spotting and returned for care. Her physicians made efforts to communicate with her, but the system of care did not ensure coordinated follow-up ad did not provide educational opportunities or notify her physician that a key follow-up in care had not occurred. A patient-centered system committed to planned care and prevention could have given her the information she needed at each step along her diagnostic and treatment path. Even better, it could have helped her avoid the disease by providing her and her physicians with screening reminders to assist in achieving the identification and treatment of precursor conditions before invasive cervical cancer developed. To begin building such a system, the major steps indicated in each section below were undertaken.

Janis's case highlights the hit-or-miss nature of tracking, measurement, and support of patient progress through the screening process in conventional healthcare systems. If an EMR had existed among the providers involved in Janis's care, then a better care trajectory could be imagined. A simple postcard reminder from her original physician, initiated by an EMR tracking the dates of her last Pap, might have triggered a visit to her new physician.25 There is also evidence26 that tailored reminders could take her specific characteristics into account and be more effective. The treatment and the steps involved could have been explained in language specific to her educational level and cultural background. Improving Janis's knowledge and improving the communications among her physicians could have simplified the instructions regarding diet and activity and provided one consistent set of recommendations. Patient-oriented, condition-specific content links embedded in the EMR could have provided printed or reliable online information about her diagnosis and potential treatments. The EMR could have provided health history and other summaries to relieve the gynecologist's burden of independently obtaining the details.

Although the above functions would have improved care, there are other ways that current Internet and computer technology could create a platform that enables a variety of health behaviors and creates an electronic health record, rather than a medical record. She eventually obtained appropriate screening and treatment, but she did not learn about the true cause of her condition or its prevention until after she had the disease. With the development of an EHR and a portal to tailored information about health, she could have learned how to reduce her risk of HPV infection through monogamy and condoms. Vaccination was not available to her then, but it is now and could also be explained through a web portal accessing health information. The need for screening and appropriate primary and secondary prevention activities for women of her age could also have been accessed through a web portal to health information. A secure location for her unique record could have sent reminders regarding her needed care independent of her provider. Advice and reminders could have been tailored to her specific needs per evidence-based recommendations. She could have accessed information about her care at each step in the process, even if that information was not provided by her physician. In this way, the electronic health record would provide ways for her to gather information relevant to her care and more actively participate and collaborate in the care, achieving at least three goals of patient-centeredness.

Although improvements in care through an EMR that is shared among all the practicing providers can be imagined, such a system does not currently exist. This paper therefore describes the process and results of providers considering and articulating their functional needs for an EMR.

To identify and articulate the needed functionality for an electronic health record, a team was developed to consider care, and the needed functions of an electronic medical record that would serve its needs were described. This team included a physician and nurse practitioner, a physician assistant from a rural setting, an information technology specialist, and a physician researcher (practitioners). They met by telephone once a week for 6 months. During this time, they considered each step and transition in the screening process diagrammed in Figure 1 and discussed how the EMR could support their clinical work to achieve the goals described below each step and transition in care. Knowing that the requirements also needed to apply to other conditions besides cancer, they established a general template to summarize the characteristics. They also agreed that they needed to describe requirements that were grounded in delivering and measuring care. They wanted requirements that the type of data be absolutely necessary for care and its measurement, not just nice to collect.

The system they described was meant to be a tool for achieving changes in practice. They therefore wanted the EMR to help consciously organize the major steps in the screening process for a defined population. For example, the first step was to identify who was in their practice (the denominator of care). They defined that population as all people who met the age criteria for screening and who had been seen in a community health center at least once in the previous 3 years (the denominator of care), and the EMR would need to be able to do that for them. They knew that within the denominator of care, practitioners needed to identify who was due for screening according to the guidelines. EMRs needed to provide reminders that occurred in the course of care (in-reach reminders) and create letters or lists for phone calls to people who had not been seen in the office (outreach reminders). Both types of reminders have a strong evidence base,27, 28 and Janis's case exemplified how outreach might have made a difference in her care. The two requirements for care meant that practitioners needed individual-level data showing screening status as the patient was seen for care and population-based measures that showed the screening status of all the age-eligible people in the practice. From a population perspective, it became clear that identifying those who were not in compliance with the guidelines for care was more important than identifying those who were. The practitioners needed tools to facilitate outreach mailing or phone call reminders to those due for cancer screening discussions.

The discussions in the development team were informed by experience with a prototype clinical management system (CMS) that had some of the rudimentary needs. This multicondition CMS operated independently of billing and clinical systems and was modified by users to capture measures for the Cancer Collaborative. This system included variables for patient demographics and information on dates and results of particular tests, such as mammography, colonoscopy, and colposcopy, performed at centers outside the federally qualified health centers (FQHCs). However, each of these variables had to be entered directly onto a data-entry screen or populated by entering data from a form completed at the time of a visit. The CMS also documented shared decision making about screening that would not be documented in a billing system because it was not a billable activity. Although the CMS was useful to many improvement teams, it became clear that practices needed a fully functional EMR that could exchange patient information with other electronic data systems because the burden of data entry and the inability to transfer information from system to system created a highly unwieldy system. Although the usefulness of the CMS as a tool for planned screening was limited, it did suggest generic processes that would be valuable in an EMR.

The broad functional characteristics established for the EMR for planned cancer screening included collecting information about additional medical history on individuals to appropriately prompt action. For example, a woman without a cervix and no important cancer risk factors does not need a Pap. An individual with a normal colonoscopy does not need a fecal occult blood test (FOBT) the following year. Practitioners therefore needed the information system to exclude some individuals, and for others, to record both a result from a test and the resulting action taken. To do the latter, practitioners also needed to report three result values and identify when action was needed: normal, no action necessary; abnormal, needs further immediate evaluation; or abnormal, needs action in the future. For example, a positive FOBT should trigger a provider prompt that colonoscopy is necessary now. A “probably benign finding” on mammography should trigger a repeat mammogram in 6 months.

The practitioners saw this process of developing a tool to manage individual care needs as fundamental to patient-centered care. Knowing the unique needs of individuals means that providers can tailor care to exclude individuals who were no longer eligible for screening, identify how a patient wants to be notified about results, or offer a specific kind of colorectal cancer screening based on an expressed preference.

To ensure that the EMR would be as clear and responsive to provider needs as possible, they consciously deconstructed each of the three major steps in the screening process. Table 1 shows the template for describing the functional requirements identified for each of the three steps and substeps in the screening process for breast, cervical, and colorectal cancer. The data requirements linked each step in the screening process to an explicit measure (Table 1). For example, optimal care starts with a two-way discussion between the care team and patient about screening recommendations (shared decision making).29, 30 EMR software would examine data for the denominator of care and document whether the discussion had occurred. Although no national benchmark for the occurrence of shared decision making exists, if screening rates are not increasing, the first question is whether shared decision making is being offered. Having a measure of whether the discussion is occurring is critical to understanding the options for improving screening rates.

The requirements included in Table 2 show functionality at both the individual and population levels. The practitioners also recognized potential difference between what practitioners would like to do and what is actually needed for care. Table 2 therefore also specifies required and desirable functionality, and their respective data sets reflect the two levels of need. An example of the complete data specification for cervical cancer screening is included as Table 3, and specifications for all three cancers may be found on the HRSA Knowledge Gateway website in the section on cancer (www.healthdisparities.net). Another key EMR requirement is that customized care management reports must be easily created with pull-down menus by someone with few computer skills.

Follow-up care is frequently a complicated process that involves multiple steps with referral to multiple providers of care. For example, Janis's Pap required referral for colposcopy. When colposcopy demonstrated cancer, then surgery needed to be scheduled, completed, and the results reviewed. Healthcare providers recognize the limitations of the paper record and are concerned about the failure to follow-up on abnormal tests. That may be why Dr. Jones was agitated, and the scope of the problem has raised the question of whether screening can achieve its potential to reduce mortality.31 Therefore, the practitioners envisioned that the EMR would need to monitor the steps in the follow-up of abnormalities. An example of follow-up specifications is included for cervical cancer (Table 4, Table 5).

For any given practice, the proportion of people who actually need cancer treatment is very small compared to the entire population. An EMR can flag these individuals and provide prompts to providers to ensure that patients are referred appropriately (Table 6). This requires sophisticated software that anticipates the need to prompt the provider if an appointment is not kept, a procedure or pathology report is not received, or further evaluation and surveillance is required.

In summary, the development team determined that an EMR must perform three major tasks in order to facilitate providers' work to provide screening services: (1) provide patient-level data; (2) measure the screening process for the population of individuals in a practice, health center, or healthcare organization; and (3) provide population management data and functions that would give practitioners the tools to contact individuals or subgroups regarding specific care needs. The EMR must have software that can provide succinct, individualized cancer screening prompts at the time of care based on age, gender, screening history, and factors that might exclude individuals for screening. It also must have flexible data import and export capabilities and nonproprietary output in multiple formats so that operators can reuse data without technical assistance. Results of prior testing need to be couched in simple, searchable results such as “positive,” “negative,” “needs further evaluation,” and “excluded from screening” so that subgroups needing further care can be identified at a population level for follow-up or referral at an individual level.

Section snippets

Discussion and Implications

This article summarizes the results of a 6-month process to describe the functionality of an EMR that might meet the needs of a planned approach to screening implementation. This is recognizably one process, but it included some important characteristics that appear to be missing from the development process for EMRs: clear clinical goals, discussion among providers, and experience with clinical management systems to help providers imagine a better EMR tool. It is now well recognized that EMRs

First page preview

First page preview
Click to open first page preview

References (33)

  • R.L. Wears et al.

    Computer technology and clinical work: still waiting for Godot

    JAMA

    (2005)
  • R.D. Kush et al.

    Electronic health records, medical research, and the Tower of Babel

    N Engl J Med

    (2008)
  • P. Hartzband

    Off the record—avoiding the pitfalls of going electronic

    N Engl J Med

    (2008)
  • S.H. Taplin

    Implementing colorectal cancer screening in community health centers: addressing cancer health disparities through a regional cancer collaborative

    Med Care

    (2008)
  • P.P. Reid et al.

    Building a better delivery system: a new engineering/health care partnership

    (2005)
  • T. Bodenheimer et al.

    Improving primary care for patients with chronic illness: the chronic care model, Part 2

    JAMA

    (2002)
  • Cited by (9)

    • Early Detection in the Age of Information Technology

      2016, Oncology Informatics: Using Health Information Technology to Improve Processes and Outcomes in Cancer
    • Health care reform and equity: Promise, pitfalls, and prescriptions

      2011, Annals of Family Medicine
      Citation Excerpt :

      Federal certification of EHR vendors for meaningful use should spur improvements in EHRs.41 Features, if sufficiently user-friendly, such as patient registries, reminders, decision support, computerized order entry, and electronic prescribing, offer the potential for improving equity through improved tracking, population management, standardization of care, and possibly reduced decision-making bias.42 Establishment of regional Health Information Technology Extension programs and Beacon Community Cooperative Agreement programs may further facilitate technology diffusion, infrastructure, and exchange capabilities within regions across the country.

    • Systems strategies to support cancer screening in U.S. primary care practice

      2011, Cancer Epidemiology Biomarkers and Prevention
    View all citing articles on Scopus
    View full text