5. Reasons to Develop an
Automated Patient Medical Record

5.1     Project Context: Determining Project Objectives From Organizational Objectives

The reason for doing a project within an organization is to improve the organization. In order to justify spending a lot of money to do a costly project, an organization must identify the reasons for doing the project, comparing these reasons and resulting benefits against the costs. Further, the organization must choose between projects and pick the proper mix of projects to select a mix with the greatest overall benefits to the organization at the lowest overall cost.

One way an organization can evaluate, select and compare projects is to determine the reasons for doing each project, the project objectives, and compare these project objectives against organizational objectives, desired future positions of the organization, as projects must support organizational objectives.  See figure 5.1.

Since all projects, and thus all project objectives, must support organizational objectives, a method for determining project objectives is to look at each organizational objective to determine if the project would positively impact the organizational objective. If so, a project objective can be determined from the organizational objective.

 For example, one “organizational objective” may be to “decrease healthcare organization costs without impacting patient care”. Two projects may each support this organizational objective: An automated patient medical record system will “reduce or eliminate the costs of transporting, finding, copying, storing, filing, and organizing the patient medical record”, while a new membership system will “reduce the costs of collecting capitation fees”. Another organizational objective might be to “improve patient care”. Of the two projects, only the automated patient medical record project supports this organizational objective; it improves patient care by “making the patient medical record always available”.

Once a project is completed, its project objectives can be used to measure its failure or success: a project can be considered successful if it meets its project objectives, enhancing the organizational objectives. Because a project objective is usually not directly measurable, a goal that can be measured should be set up to evaluate whether or not the project objective has been reached. For example, a project objective to “create an always available patient medical record” might have a measurable goal at the end of the project of “observing patient visits with physicians, an analyst must verify that for 99.5% of visits that the patient’s medical record is immediately available on-line to the physician”. This could be done statistically with the analyst visiting various medical locations and observing various physicians giving care, observing whether or not the patient’s medical record is available on-line.

Because the objectives are likely to take a long time to achieve, measurable goals may also be established at many different points during the project that can be used to measure the progress of the project towards the objective—An intermediate goal might be that “40% of HMO patient records will be automated after 3 years from the start of the project”.

Figure 5.8 lists project objectives for our project to automate the patient medical record derived from HMO organizational objectives. In the early stages of a project, project objectives can be used to determine an initial set of business requirements for the project. A business requirement is a required characteristic of the organization at the end of the project. The total set of business requirements for the project, determined as the project progresses, describe the expected project effects upon the organization.

From these initial business requirements, an initial description of products of the project can be determined—In our case the product is an automated patient medical record. These products can be described by a model of them, referred to here as a conceptual view. Section 5.6 presents an initial conceptual view of the automated patient medical record system. More refined conceptual views can later be developed as the project is defined further, as additional business requirements are determined.

But again, before a project is started, it must be selected among many potential projects that might benefit the organization. The next section describes an approach that could be used by an organization to select and evaluate projects.

5.2 Organization and Project Objectives Strategies, and Goals

An approach for selecting and evaluating projects within an organization is presented in reference [1]. It uses the following terminology, which is related to the overall organization rather than to a project:

·        Mission: the business the organization is in.

·        Objectives: desired future positions of an organization.

·        Strategy: the general direction in which the objectives are to be pursued.

·        Goals: specific targets to be sought at specified points in time.

·        Projects: resource-consuming sets of activities through which strategies are implemented and goals are pursued.

Figure 5.2 shows the relationship between these elements together with an example for an HMO.

 

The company has a specific mission. Based upon this mission and the market, the company defines objectives for the future for the company to enable the company to improve and advance. Strategies for achieving these objectives are determined. Goals, ways to measure the organizational objectives, are established.

Appropriate projects to fulfill the goals and objectives are picked. Projects themselves have project objectives. Projects are picked whose project objectives most closely match the organizational objectives taking into consideration constraints on the project, including availability of resources (money, time, people) necessary to do the project, the availability of technology to successfully accomplish the project, and other obstacles that might hinder accomplishment of the project.

After a project is picked, goals to measure project objectives during the course of the project are determined. Strategies for doing the project are later determined as part of the project plan.

5.3     Possible Objectives for an HMO

The following are a set of organizational objectives of our example HMO that enhance the HMO’s organizational mission of “providing quality healthcare at a reasonable cost”:

1.      quality medical care is provided to HMO members

2.      medical care is provided at a reasonable cost to members or to members’ employers paying for the care

3.      the HMO makes money

4.      members are highly satisfied with the HMO and medical care

5.      HMO employees are productive and happy in doing their jobs

6.      the HMO can predict and handle future growth, or non-growth.

The above is a typical mission of an HMO and are typical organizational objectives for an HMO.

5.4     Derived Objectives for a Project

Project objectives can be determined from organizational objectives as follows: (1) Each organizational objective must be evaluated as to how the project would effect it. (2) If the project has a positive effect on the organizational objective, then some aspect of that HMO organizational objective should be a project objective. So let us consider these organizational objectives one by one and how they are positively effected by our project to automate the patient medical record.

(Note that negative effects of the project will be considered later on.)

5.4.1    Quality Medical Care

Automating the patient medical record could have a positive effect on patient care within an HMO.          

5.4.1.1    Reengineering to Improve the Quality of Patient Care

An automated patient medical record system effects the entire clinical workflow. Patient care is a joint effort of many people, including medical personnel and non-medical support personnel. Patient care is the result of interactions between many different departments inside and outside the HMO. The automated patient medical record system would support all these people and all departments that influence patient care.

Changing and improving the way an organization works is called “reengineering”. Reengineering involves looking at and possibly changing the workflows of employees in the HMO to improve the workflows. Creation of an automated patient medical record system provides an excellent opportunity to make patient care better for both patients and employees through reengineering.

In the creation of an automated patient medical record system, the entire clinical workflow of the HMO must be re-evaluated to completely eliminate unnecessary steps and to restructure non-productive steps. At the same time, healthcare organization employees must be insured that the automated patient medical record system improves patient care while not creating additional employee stress. Each step must be evaluated on how it positively or negatively affects both patients and employees, with significant employee and patient input into changes in the care process. Through this reengineering process, the quality of medical care at the HMO would be improved.

5.4.1.2    A Complete, Always Available, Patient Medical Record

With automation of the patient medical record, there is the potential of having a complete patient medical record, containing information from both inside and outside the HMO (a “longitudinal” or “life time” patient medical record) that is available at any time to authorized caregivers both inside and outside the HMO. Whereas, the current (paper) patient chart may be only at one location at a time with transportation time between locations, and might even be temporarily misplaced, the automated patient medical record would be immediately available to all caregivers who needed it.

The fragmentation of the patient medical record described in section 4.7 and shown in figure 5.3 should no longer occur.

Patient care, as a result, could be changed in a positive way in many ways, including the following:

·        There would be fewer gaps in medical information.

·        Critical medical information, such as allergies to medications, advance directives, a complete list of medications the patient is taking, and an up-to-date patient and family history is likely to be immediately available. This information is especially important in the emergency department and the inpatient setting.

·        There would be no delay between closely occurring visits to transport over the chart.

·        A caregiver can telephone a patient when needed instead of having to first wait for the arrival of the chart.

·        A patient can contact a physician by telephone and immediately receive informed advice because the physician will have the (automated) patient medical record in hand.

·        The patient medical record would now be available to advice nurses at the time a patient calls in for advice.

·        Multiple caregivers can view and/or update the patient medical record concurrently, even at two different geographic locations.

With simultaneous caregiver access to the patient medical record, telemedicine becomes more feasible—especially when a nurse is co-located with the patient and the physician is at a remote location, or when physicians at different locations are working in consultation with each other.

Telemedicine [2,3] uses telecommunications, from phones, teleconferencing to teleinstrumentation, together with computing technology to facilitate caregiver/patient and caregiver/caregiver interactions, remote diagnosis and treatment, and the transmission of medical data and images (from photographs to MRI to audiovisual). Telecommunication interactions fall into the following categories :

·        Telementoring: fostering a mentoring relationship between a specialist and primary care physician, between a physician and a nurse practitioner, etc.

·        Teleconsultation: fostering consultations between physicians, usually in very different locations, for time critical care or care requiring a second opinion.

·        Telediagnosis: the provision of diagnostic services to remote locations without the expertise.

Concurrent access to the patient’s medical record would certainly assist the first two situations where caregivers all have the authority to access and update the patient medical record.

Telemedicine with a mutually available patient medical record fosters interactive communication between caregivers located remotely. Remote instrumentation, using medical instruments remotely located with the patient, where results can be immediately included as values, graphics or images in the patient medical record can further foster this communication.

At least the following medical instruments which can be used remotely currently have defined ANSI interface standards [4]: dermscope (for images of skin), ophthalmoscope (for images on the interior of the eye), otoscope (for images of the middle ear), laparoscope (for structures within the abdominal cavity), nasopharyngoscope (for examination of the sinus cavity and nasopharynx), intraoral scope (for inside the month, usually for dentistry), bronchoscope (for examination of the bronchi), sigmoidscope (for examination of the rectum and sigmoid colon with attachments to allow biopsy), and stethoscope (for heart and lung sounds), with the potential of significant results automatically being recorded in the automated patient medical record. Other currently existing remote instrumentation, without ANSI standards also exist (e.g., tele-ultrasound for OB/GYN and real-time interactive digital diagnostic quality transmission of Echocardiograms [5]).

5.4.1.3    Organization So Clinical Information can be Easily Found

If relevant information cannot be found in the patient medical record, then it is essentially equivalent to the information not being there. Patient care can be improved by making it easier to find clinical information in the patient medical record by “organizing” it.

Unlike the paper patient medical record which can be organized in a limited number of ways, the automated patient medical record can be electronically organized in an unlimited number of ways, for example by encounter, significant health problem or type of caregiver.

An “encounter” will be used here to mean a face-to-face interaction between a patient and a healthcare provider or other direct communication that substitutes for a face-to-face interaction, such as some phone calls. Examples of encounters are an inpatient stay, outpatient visit, emergency department visit, advice nurse call, a phone call between a patient and a physician, a home health visit, or a skilled nursing facility (SNF) stay. A “significant health problem” is any significant medical condition or disease.

For example, four methods that can be employed to organize the patient medical record and enable a caregiver to find relevant information are the following:

·        Summaries: Lists can be produced summarizing clinical information for a patient, including past encounters, current medications taken by the patient, significant health problems, immunizations.

·        Drill down: Patient clinical information can be organized in a hierarchical fashion, including by encounter and significant health problem, so drill down is possible to relevant documents containing that information (e.g., drill down from an encounter, perhaps after selection from a list of encounters, to a list of the medical record documents for the encounter, and drill down after selection of the document on the list to the document contents).

·        Organize by type of caregiver: Categorize parts of the patient medical record so the parts relevant to a particular type of caregiver can be found (e.g., an ophthalmologist is interested in all eye problems)

·        Synopses: Display a shortened, summarized, version of what happened during an encounter.

Reference [6] gives examples of possible useful ways to organize a patient medical record in a teaching hospital:

·        The general internist wants a view that will help him or her manage the medical aspects of the case.

·        The subspecialist’s view must contain additional details relevant to his or her special duties in a case.

·        The chief resident needs additional details to support teaching during rounds.

·        The intensive care nurse needs a view that embraces the care and management of the patient to whom he or she is assigned.

·        The pharmacy’s view supports patient medication, including medication history and patient response to drugs.

·        The dietitian requires a view to support diet and nutrition control for the patient.

·        The security department’s view identifies security risks or hazardous patients.

·        The accounting department has a view relating to what should be charged.

·        Multiple reporting views are required to prepare internal or external reports for policymakers as appropriate.

Even when clinical information in an automated patient medical record is not organized (e.g., in free text progress notes), it still can be found by information retrieval (IR). Searches for related information scattered throughout the Internet, produced by different organizations and individuals, use this approach.

5.4.1.4    Continuity and Coordination of Patient Care Among Caregivers

Patient care can be improved by a patient medical record that assists caregivers in the coordination of caregivers who see the patient. There are at least two types of continuity of care: (1) continuity of care across encounters or (2) continuity of care within an encounter.

Continuity of Care Across Encounters

Continuity of care across encounters in an HMO is generally accomplished by having each HMO member pick a primary care physician and by assignment of the patient with a care team. When the patient comes in for care, he will usually be appointed with the primary care physician or a member of the care team.

Independent of whether the patient primarily sees one caregiver, sees a care team, or sees many unaffiliated caregivers, patient care can be improved if there is better communication between the many or even the same caregiver as the caregiver(s) treat the patient across many visits and/or hospital stays. The automated patient medical record is one vehicle for improving this communication.

This requires a way of identifying all encounters that are part of the same treatment plan. Possible ways of doing this are relating encounters in the patient medical record by use of case management techniques or use of clinical pathways.

Case management involves assignment of a case manager for a set of encounters either (1) for a particular medical condition for a patient or (2) for all encounters for a high-risk patient.

A clinical pathway is a structured way to identify care activities and caregiver workflow needed to care for a patient with a particular condition or disease. Paths through a clinical pathway can be adjusted for the particular needs of an individual patient and updated as they change.

Through clinical pathways and cases, there is the potential to coordinate caregivers better within one outpatient visit, within one inpatient stay, or over any combination of inpatient stays and outpatient visits. Coordination of patient care can occur within one facility, a number of facilities, and potentially among care provided in multiple healthcare organizations. Care can be multidisciplinary, across multiple medical departments. This coordination promotes the team approach to providing patient care.

Continuity of Care within an Encounter

Continuity of care within a stay in the hospital is clearly important as many different caregivers are involved. But continuity of care also applies to a single outpatient encounter.

If a patient talks to an advice nurse over the phone to seek advice and this sets up a subsequent outpatient visit with a physician or nurse practitioner, then the physician or nurse practitioner should know everything pertinent that occurred during the advice nurse phone call. This allows the physician or nurse to prepare for the visit and to spend her time more productively with the patient.

A diagnosis for an outpatient visit is often not completed until the results of a physician order come back, for example, the results of a clinical laboratory test, an interpretation of an x-ray, or the results of a procedure. Thus care extends beyond the time of the face-to-face meeting with the patient.

An automated patient medical record that associates all these care activities with the outpatient visit provides a more complete picture of the care given than one that does not.

5.4.1.5    Automated Assistance to the Caregiver in Providing Care

With the automation of the patient medical record, information that is input to the patient medical record can be validated as it is input. Consistent terminology could be checked for and enforced as it is input by the caregiver. Legibility of information is guaranteed.

With automated system ordering, tests can be ordered much more quickly than with a physician order sheet that is on paper. Orders and results would be automatically and immediately recorded in the patient medical record.  The caregiver could be immediately informed of abnormal, STAT or panic results.

“Clinical checking” of medication orders can be done by the automated system as the order is input, including checking for wrong doses, wrong choices, wrong techniques, delays, known allergies, missed doses, wrong drugs, drug-drug interactions, wrong frequencies, and wrong routes. Prescriptions can be printed so mistakes due to poor handwriting and similarly named medications can be avoided. And ordered medications can be checked against the diagnosis—A report from the National Academy of Sciences’ Institute of Medicine estimated that 44,000 to 98,000 people die annually due to mistakes by medical personnel, mainly in improper ordering and administration of medication [7].

The system could inform a caregiver if an “at risk” patient is not complying with a drug regimen (e.g., by not picking up a prescribed medication or by not ordering refills within a specified time period) [8]. The system could inform the caregiver if the patient fails to keep important appointments or does not show up for an important diagnostic test.

The system could inform a caregiver of trends in the patient’s health based upon measurements taken over a period of time (e.g., increase in blood pressure over time).

The automated patient medical record system could record actual outcomes of treatments. Measuring outcomes based upon treatments is a way of evaluating treatments and from the results of these evaluations, determining future clinical practice guidelines for care of HMO patients. Guidelines that are based upon the best current scientific research, that produce the best outcomes, are called best practice guidelines.  Use of the best treatments and practices for diseases that produce the best outcomes for the least cost as determined by the best scientific evidence is called evidence-based medicine.

The Agency for Health Care Policy and Research (AHCPR), a government agency, has established a National Guideline Database of clinical practice guidelines for treatment of a number of medical conditions in association with private and public healthcare organizations based upon the best available scientific evidence [9].  AHCPR guidelines exist for the medical conditions listed in figure 5.4.

 

Figure 5.4—Existing AHCPR Clinical Practice Guidelines

Number

Description

1

acute pain management

2

urinary incontinence in adults

3

pressure ulcers in adults

4

cataract in adults

5

depression in primary care

6

sickle cell disease

7

benign prostatic hyperplasia

8

management of cancer pain

9

unstable angina

10

heart failure

11

otitis media with effusion (earache)

12

quality mammography

13

acute low back problems in adults

14

pressure ulcer treatment

15

post-stroke rehabilitation

 

HMOs might want to do their own determination of best practice guidelines based upon treatments and evaluations of the treatments. These HMO guidelines can provide additional guidelines to be used in the HMO together with the AHCPR guidelines. How treatments can be evaluated, measurements of the results of treatments—their outcomes—is a controversial area. Most encounters involve treatments, but few include outcome measurements.

Some current outcome measures can be considered subjective while others are objective. Some groups consider outcomes to be related to the patient’s perception of his or her health. A good outcome occurs for a disease (e.g., knee replacement or benign prostatic hypertrophy) if the patient’s perception of his quality of life is good. A patient questionnaire--SF-36 or HSQ-12 (Health Status Questionnaire) are existing ones--is used to measure such outcomes [10]. Some outcomes are more directly measurable, such as a decrease in the patient’s (systolic and diastolic) blood pressure, x-rays showing the healing of a broken arm or leg, or the urinary flow rate or percentage of the bladder that is emptied upon urination to measure benign prostatic hypertrophy.

Using such outcome evaluation techniques to determine and evaluate best practice guidelines for the HMO together with automation of the patient medical record can improve quality and effectiveness of medical care for patients. As clinical information on the patient is input, the system can check this information against guidelines that result in practice patterns producing the best outcomes for the diagnosis and patient characteristics, using the best combination of treatments, medications and medication dosages. At the same time, the system can collect research information on these and alternative practice patterns within the HMO that can be used to refine best practice guidelines, taking into account individual differences between patients. “Disease management” information from outside institutions could be incorporated into the automated guidelines (e.g., from specialty care organizations for cancer, asthma, and cardiology and from pharmaceutical companies based upon their research) [11].

As important as guidelines for physicians are guidelines for advice nurses. An advice nurse who receives calls from patients has to distinguish emergency and urgent care situations from others, and insure that a patient comes in quickly for care when necessary; the advice nurse is often the first point of contact for a patient during off-hours. Also, automation can make HMO nursing standards for care consistent with JCAHO requirements, available on-line to inpatient nurses. Standards for care for a particular condition can be found immediately based upon the chief complaint or preliminary diagnoses.  

Various other forms of assistance can be provided to the caregiver in the care of the patient. The system could automatically identify changes in patient health that may help identify future health problems before they become acute (e.g., changes in serum cholesterol and blood pressure beyond normal age related values, clusters of a particular disease, etc.)  The automated system could assist the caregiver in making diagnoses; see section 17.4.7. Clinical checking of clinician orders (checking drug/drug, drug/food, etc. interactions, allergies, etc.) can be done and duplication of tests can be immediately recognized. Body surface area, intravenous, pediatric, parenteral or other types of calculations for medication dosage determination can be provided, allowing these calculations to be documented [12].

On-line medical references, including those on the Internet, an Intranet or an Extranet (i.e., the Internet restricted to inside the HMO or inside the healthcare community), can be accessed during display of the automated patient medical record and during the input of information to the chart. Such on-line medical information may or may not be specific to the organization (e.g., the organizational drug formulary; organizational protocols and practice guidelines; DME formulary; poisons information; pre-travel vaccinations; AHCPR clinical practice guidelines [10]; CDC prevention guidelines for advising patients on how to prevent diseases such sexually transmitted diseases, hepatitis, bicycle injuries [13], National Library of Medicine (NLM) database Medline [14]; or composite systems such as Micromedex).

A caregiver ordering a treatment could be informed of costs, risks (e.g., side effects, interactions) and other factors in selection of treatments. The caregiver could immediately inform the patient, thus enabling the patient and caregiver to make wise joint decisions.

The automated patient medical record system would ideally tie together all existing clinical systems (the ADT system, the ordering system, the pharmacy system, the clinical laboratory system, etc.) and, over time, will require integration of additional non-clinical systems (e.g., inventory, financial) in order to completely automate the patient medical record. Entrance of information in one of the systems would flow to the others and may be recorded in the chart. As a result, the patient would not be asked for the same information over and over again and care of the patient could be completed much faster.

Automation of the patient medical record would allow the physician to communicate with a patient at any time regarding the patient’s medical condition, emphasizing care decisions, and thus promoting greater compliance with these decisions. The physician would no longer have to wait for the patient medical record before communication with the patient over the telephone or via e-mail.

After a patient visit, the physician could complete his or her visit documentation and have orders and instructions for the patient and family printed out for a medical assistant or nurse, who could give it to the patient and review it with the patient or family members. This would also promote patient compliance with physician instructions and orders, including informing family members of how they might assist the patient.  

5.4.1.6    Preventive Care

Automation of the patient medical record enables the HMO to establish preventive care programs (e.g., immunizations, cholesterol screening, blood pressure, weight control, stop smoking programs, flu shots, mammography, pap smear, breast exams, flex sigmoidoscopy, colonoscopy, tetanus shots, and diabetic retinal screening exams). A number of these preventive health measures have been shown to provide better medical care for the patient in that they catch diseases early on so the diseases can more easily be cured; also, many such preventive health measures have been shown to save the HMO money in the long run.

A clinical pathway covering the patient over the patient’s life time (called a “life care path”) can be set up to automatically send out “outreach” letters to a patient to come in to the HMO for a specific preventive care exam, with determination of which exams and how often to send out such letters being based upon the patient’s sex, age, family history, social history and the current health of the patient. Also as part of an outreach program, the automated patient medical record could also identify pregnant women who came in for care, but who have not come back to receive regular prenatal care.

Some preventive care can be supported by what I call trend documents. A trend document is a document that automatically records a value or values that a caregiver wants to track over time as the value(s) are input via other clinician documents. For example, a trend document could capture blood pressures, diagnostic images of a knee with progressive osteoarthritis, height for a child who is shorter than other children in her age group, PSA values to detect prostate cancer, internal pictures of an eye having macular degeneration, etc.

A trend document has the potential of being a valuable tool in preventive health care, diagnosing diseases early on and predicting the progress of diseases. It is also a valuable tool in patient education in demonstrating to the patient the progress of a disease. Dangerous health conditions can be potentially recognized by the automated system and automatically reported to the primary care physician, who could have the patient come in to be checked.

Currently, with a paper patient medical record, much of the information that could be in a trend document is lost for various reasons:

·        older paper charts are archived

·        charts may be fragmented such as shown in figure 4.8

·        in some healthcare organizations, some diagnostic images are destroyed after a short time (e.g., after 7 years).

With an automated patient medical record, information in a trend document can be kept immediately available, no matter how old.

5.4.1.7    Predicting Diseases

According to the Journal of the American Medical Association (JAMA), “Individuals are born with a relatively fixed genetic status that in combination with environmental factors determines the propensities for a variety of disease states” [15].

With an automated patient medical record, there is the potential of having a more complete social, family, environmental, and (with new information on the human genome) genetic history of an HMO member. With this information collected from many people, there is the potential to identify the root causes of diseases. And individually for the HMO member, there is the potential from this collective information to identify diseases for which the member has a propensity to develop; countermeasures could then be started early on to ward off these diseases.

Even if possible, there are factors that make this idea of disease prediction difficult to achieve. Some of these are practical factors, while others are philosophical ones. Practical factors are the inability to get complete and accurate social, family, and environmental information; the possibly prohibitive cost of analyzing the DNA of a particular individual; the real chance that the prediction of a disease, since necessarily probabilistic, may be wrong; and the potential of stereotyping and other misuse of this information, especially by non-medical organizations. Philosophical factors making disease prediction, even if possible, difficult include privacy, ethical concerns, and simply that some people would rather not have any inkling of the future. Because of these fears and difficulties, collection of this information over time should only be done with the specific approval of the HMO member and should be subject to the strictest security rules.

Useful for the process of creating a complete family and genetic history of the patient is the ability to identify the HMO member’s relatives and their relationships, so familial diseases and genetics can be determined directly from the medical history collected for a relative. Two problems are that a family member may not be a member of the HMO and that a family member may not care to participate in this process.

Although much of this social, family, environmental and genetic information may be collected as part of normal examinations, such information may be incomplete or inaccurate. I think that more thorough information would be obtained if the information came from periodic questionnaires filled out by the patient, with the member periodically reviewing and verifying the information’s accuracy, say through interviews with a nurse; however, even information collected in this way is likely to be incomplete or inaccurate.

I will refer to this disease prediction approach—basically a theoretical approach to disease prediction where the total of a person’s genetics and significant environmental factors are recorded over the patients lifetime enabling prediction of future diseases—as the direct cause disease prediction. Such an approach could involve collecting a huge amount of information, so much that it may not be feasible to identify which environmental factors are the true causes of disease. Furthermore, all the significant environmental factors may not have been recorded. One disease may mask another or result in the death of the individual before the other occurs. And such an approach, no matter how thorough, may not be better than any less thorough approach, as the occurrence of a disease may be largely dependent upon chance [16]. This direct cause disease prediction approach then is perhaps like “looking for a needle in a haystack without a guarantee that there is a needle”.

This book identifies three other approaches to disease prediction that may be much more practical than direct cause disease prediction, referred to in this book as analytic disease prediction, probabilistic descriptive prediction, and disease progression analysis.

The first approach, analytic disease prediction, is predicting a disease from what is known to be risk factors for that disease and protective factors for that disease. For example, because a person smokes, he has a greater likelihood of getting lung cancer. Because a person had a severe knee fracture, he has a much greater risk of osteoarthritis and a knee replacement when he grows older. Because a woman has the BRCA1 or BRCA2 gene, she has a much greater risk of developing breast cancer. Once the physician recognizes such a risk factor for disease, the physician could record the factor in the patient medical record, providing easy access to it for later caregivers, who could take countermeasures against the disease. This approach is currently used all the time, but could be supported by an automated patient medical record, which could assure that such risk factors are always recorded and are available to all later caregivers.

Epidemiology is the study of the occurrence, distribution, and causes of disease and the application of this study to the control of health problems. Clinical epidemiology [17] is the use of epidemiology in direct patient care. What this book calls direct cause disease prediction and analytic disease prediction are based upon analytic epidemiology [18], a branch of epidemiology identifying the causations of diseases, also referred to as etiology.

The second more practical approach to disease prediction presented here, descriptive disease prediction, is identifying any pattern of factors that can be associated with a disease. In such an approach, all you need to know is that, over time, if this person has a particular set of factors (say the symptom or condition X followed by Y and also at some time has Z) then the probability that he or she will have disease A or a debilitating form of that disease is increased to R% instead of S%.

Such relationships could be found by a computer program or human doing pattern matching, analyzing many automated medical records of patients who develop a particular disease to determine combinations of factors (symptoms and conditions) that can be used to predict a particular disease. Analyzing all patients with these factors, a probability of a patient having the factors developing the disease can be determined. Another computer program using this information could then identify patients with these factors.

Descriptive disease prediction is based upon descriptive epidemiology [18], a branch of epidemiology identifying patterns or trends in diseases and injuries.

Disease predictions from these two approaches could be used together with preventive care and patient education. If the patient was found to have a higher probability of a disease, he or she could be taught how to detect the disease (e.g., to test for breast cancer) or to avoid the disease (e.g., stop smoking). She or he can be advised to come in for regular preventative care (e.g., mammograms).

A more proactive approach to disease prediction is disease progression analysis. Patients who have a disease or are candidates for a disease could be tracked over time: a medical value or state that is predictive of the disease could be periodically measured and recorded. This progression of medical values or states could be compared against those of patients who have developed the disease, and from this, a prediction of the future progression of a disease or to a disease could be made.

Disease progression analysis might consist of the following:

1.      recording a medical value or state over time as a measurement of a disease:  recording a medical value or state over time for a patient as a measurement of a disease consecutively over a long period of time (e.g., taking a person's blood pressure, recording a PSA value to measure the potential for prostate cancer or onset of BPH, taking x-rays of a knee injury, determining bone density, etc.). This recording could be done, as described earlier in section 5.4.1.6 on a trend document.

2.      prediction: based upon the trend data and past trends of other people, predicting the future changes in medical values or states.

3.      countermeasures:  taking countermeasures to slow or stop the possible progression toward the disease.

4.      prediction of a probable time of when a treatment decision would need to be made, or of a probable time of unset of the disease or of a debilitating condition: based upon trend data and past trends of other people, predicting when the person might have to make a decision on treatment and predicting when the disease or a debilitating condition is likely to occur (e.g., heart disease, the need for a knee replacement)

5.      prediction after countermeasures: doing the evaluation of when the disease or debilitating condition is likely to occur after countermeasures: doing what-if analysis and determining the effects of different countermeasures (e.g., predicting the probability of the onset of debilitation from osteoporosis both before and after the person begins an exercise program).

6.      identification of treatments or potential future treatments that could be applied at a treatment decision point: keeping track of clinical trials for new treatments, so the patient can be told of current treatments in comparison to potential improved treatments for a later decision point.

Some diseases are virtually all environmental (e.g., infectious diseases), while others are genetic (e.g.,down syndrome). But I think that disease prediction is applicable to either category of disease and as well as being applicable to those diseases that fall in between these two extremes, where both genetic and environmental factors have a significant influence. In the future, digital libraries will record the human genome matching genes with referenced research and probable functions [19]; reference [19] predicts that, although epidemiological studies are now mainly done in the laboratory now, “as times goes on, more work in medicine will be done in the digital library rather than at the lab bench.”

For all forms of disease prediction, it is important for the caregiver to be cognizant of the patient’s values. A values questionnaire could collect this values information by asking questions of a patient such as the following: “If medical science was able to predict that you will develop a particular disease, would you like to know in advance? Would this depend upon the type of disease?” Such values questionnaires currently exist for end-of-life decisions; see section 5.4.3.5.

5.4.1.8    Assistance in Clinical Research

Patient care can be improved by more thorough clinical research with larger patient populations. Clinical research can involve an investigation of a new method, medication, or procedure in the treatment of a particular disease or condition in comparison with a current, high-quality, standard method of care for that disease or condition.

With automation of the patient medical record there is likely to be much more medical information available to the researcher. This information could be made available without identifying the patients or caregivers, thus not compromising patient or caregiver privacy. This allows for very large-scale investigations.

Automation of the patient medical record also enables very long-term, longitudinal, studies, spanning from infancy to adulthood. For example, a longitudinal study enables determination of what in infancy or childhood is predictive of what happens in adulthood—For example, do temper tantrums in childhood have any relationship to bi-polar disorders later on in life? Doing such research with paper medical records would not be possible, as the volume of paper records over such a long time would swamp chart rooms, and finding the required information in these paper records would be very difficult.  Large capacity storage devices and automated information searching methods make such longitudinal research feasible with automated patient medical records.

With an automated patient medical record, it is also easier to pick ideal target patient populations for research (e.g., patients with a targeted condition and no others, or patients who do not currently take medications which make it difficult to evaluate a new medication). This may increase the validity of the research.

An automated system also enables quick dissemination of the results of clinical research to caregivers. For example, studies have shown that the incidence of Sudden Infant Death Syndrome (SIDS) is reduced by placing healthy infants on their backs or sides to sleep [20]; the automated medical record system could remind a caregiver to inform new mothers of this research during pediatric visits. 

5.4.1.9    Quality Management, Infection Control and Public Health

Many groups look at overall patient care, not just care for an individual patient. These groups include accreditation and government agencies, public health agencies, and agencies that keep information—called registries—on specific diseases and conditions. The automated patient medical record system could be designed to look at patient care from these many different perspectives, and report on identified problems in patient care that the system detects. The automated patient medical record could (1) automate quality control checks, (2) automate the recognition of infectious outbreaks, both inside and outside the medical organization, and (3) automate the recognition that relevant patient information needs to be sent to registries.

The automated patient medical record system could potentially automatically collect the information required by accreditation and government agencies (e.g., JCAHO, OSHA). It could then alert management when it identifies that care is not up to the standards set by these agencies.

The automated patient medical record system could also be designed to identify infection outbreaks, cases where multiple patients have the same communicable disease. This information could be sent to the HMO department of infection control and epidemiology. Automated notification of incidents of communicable diseases will allow the infection control department to quickly report such incidents to the public health department, to identify nosocomial infections (i.e., infections acquired by patients during their hospital stays), and to insure that proper isolation procedures are followed.

Voluntary and government mandated registries have been established for various disease and health problems, including cancer, AIDS, birth defects, diabetes, implants, organ transplants, measles, trauma and hazardous substances [21]. Some HMOs have their own registries (e.g., Kaiser Permanente has a confidential registry to track patients with the BRCA1 gene that predisposes a patient to breast or ovarian cancer). The automated medical record system could automatically collect registry information or, alternatively, alert a caregiver when registry information should be collected based upon the diagnosis for the patient.  

5.4.1.10  Special Assistance for Radiologists and Other Caregivers Doing Interpretations

Patient care can be improved by digitization of diagnostic images and digitization of test results that require interpretation so that these diagnostic images and test results can be quickly sent across networks to radiologists and other physicians who can interpret them, analyzing them for indications of abnormalities and possible disease. Examples include mammographies, fluorescein angiographies and pulmonary test results. Once completed, interpretations can immediately be sent back over the network directly to the automated patient medical record with the ordering physician being informed. Thus, the time lag between taking a diagnostic image or performing a test and receipt of the interpretation results can be greatly reduced for the ordering physician and for the patient. 

Standards for transmission of diagnostic images are the DICOM-3 standards recognized by the American College of Radiology [22].

Digitization of diagnostic images also allows physicians doing interpretations to be assisted by computerized algorithms that would enable computer recognition of diseases, such as detection of cancers within mammographies.

Because diagnostic images take up a lot of space in chart rooms, they are kept for a shorter period of time than other parts of the patient medical record. Digitization enables efficient storage of diagnostic images. This makes it more cost effective to keep past diagnostic images around, thus promoting the ability to compare past diagnostic images with current ones, which can be useful in visually showing the progress of some diseases (e.g., degeneration of cartilage within the knee or of the spread of cancer within the body).

5.4.2    Cost Control

Automating the patient medical record could help control costs within the HMO.

5.4.2.1    Reengineering to Decrease the Cost of Patient Care

Section 5.4.1.1 that discusses reengineering of workflow to improve patient care with can occur as part of a project to automate the patient medical record. During this analysis of the changed workflow with an automated patient medical record, costly processes within the HMO can be identified, and eliminated or modified.

5.4.2.2    Demand Management

In an HMO, one major way of cutting costs is referred to as demand management [23], a system to provide a patient with the most cost-effective care for the patient’s medical complaint, by instructing a patient in self-care when appropriate, by “triaging” the patient to the most cost-effective caregiver for the medical condition, and by coordinating the care given when there are multiple caregivers. Section 4.6 describes demand management within an HMO in detail.

The following are examples of the ways an automated patient medical record can support demand management:

·        Support for advice nurses: The automated patient medical record would be available to advice nurses, providing summaries of the member’s health problems, medications, past encounters, etc. This would assist the advice nurse in advising the patient on self-care, whether to come in or not, and whether to come in on an emergency basis or urgent basis.

·        Support for caregiver direct communication with patients: The automated patient medical record would be immediately available to physicians, nurse practitioners, physician’s assistants and physicians’ nurses, rather than having to be ordered, supporting direct communication with the patient over the telephone, which might save on the cost of a more costly face to face visit.

·        Coordination of care for acute medical conditions: It could support case management techniques to coordinate care between primary caregivers, specialists and others in care of a patient for an acute medical condition, clearly combining all documentation for the acute condition in one place and clearly identifying the overall treatment plan. This coordinates the care between caregivers and enables evaluation of best, most cost effective, medical practices for a type of acute medical condition.

·        Coordination of care for high risk patients: It could support case management techniques to coordinate care for a high risk patient, for an elderly Medicare patient, for a patient being treated for workman’s compensation claims, for a pregnant woman, or for or a patient with a high impact medical condition (e.g., diabetes), with assignment of the case to a case manager. The case manager would coordinate care and insure that the most cost-effective care was given to the patient.

·        Preventive care: It could support wellness programs through critical pathways that would automatically sent out letters to patients advising them to come in for preventive care testing based upon the patient’s age, sex and medical condition. Examples of such tests are mammographies to test for breast cancer and sigmoidoscopies for test for colon cancer. As a result of these tests, medical conditions could be identified before they become much more costly to treat.

5.4.2.3    Reducing and Simplifying Documentation

Automation of the patient medical record would reduce the costs of medical documentation, both for inpatient stays and outpatient visits. It is estimated that medical documentation accounts for 30% of the costs of an inpatient stay.

Types of medical documents currently existing in a paper patient medical record are described in section 4.4.3 of this book. These would be replaced by automated documentation whose input could be greatly simplified because of automation. Methods to simplify input of information include the following:

·        Templates and abbreviations in progress notes: During the time a caregiver enters a progress note, the problem can be selected from pick lists. A textual template for the identified problem (e.g., migraine headache) can be used, where the physician can fill in the blanks in the textual template or select from pick lists embedded in the template. Alternatively, when the caregiver enters textual information for the progress note, there is system assurance of use of a standardized medical vocabulary, and entrance of caregiver or organization determined abbreviations to expand to the full word (e.g., “AE” for “above elbow”, “t.o.” for “telephone order”, “appy” for “appendectomy”). Caregivers can devise their own templates and abbreviation schemes, or organizations can devise templates and abbreviations for use in all caregiver notes.

·        Entrance “by exception”: Another method of simplifying documentation is to have information entered “by exception” by the system, with the user only entering information when it changes. For example, within a flow sheet entered every hour, an inpatient nurse might record inpatient status information including skin color (e.g., pale), heart rhythm (e.g., normal), pain? (e.g., yes), and nausea? (e.g., no), with the information for the current hour being assumed by the system to be the same as for the previous hour unless there was a change in status.

“Entrance by exception” also applies to progress notes. A template describing a problem in detail might be used in a progress note that covers a large percentage of cases of that problem, where the caregiver would only need to make changes in small portions of the text that do not apply to the current problem.

·        Automated clinical pathways: A clinical pathway for a given medical condition identifies the care to be given for the medical condition based upon the outcomes of treatments. Clinical pathways can include prewritten physicians’ orders. Physician orders would be embedded in the clinical pathway describing normal treatment guidelines based upon outcomes. The clinical pathway would identify outcomes that are routine. Additional narrative documentation would only be required when there were variances from a normal expected outcome. Thus clinical pathways have the potential of significantly reducing the time to input information.

·        Trend tracking: By setting up automatic graphing of a measured variable over time (e.g., blood pressure, weight, height, serum cholesterol), perhaps together with a graph of normal age related values for these variables, any caregiver can quickly identify long term trends in a patient’s health; this is what I have termed a “trend document”. For a caregiver to manually produce such a document would require much work.

·        Automation of nursing diagnoses, interventions and outcomes: At the start of an inpatient stay, nurses identify nursing diagnoses, patient problems identified through assessment findings in comparison with what is considered to be normal. For example, one nursing diagnosis might be “urinary retention”. Nursing interventions that can be performed to alleviate or treat the diagnosis are identified (e.g., patient education on medication or catheterization) and expected outcomes of these interventions during the stay are identified (e.g., absence of urinary infection, recognition of urge to void). These interactions and outcomes could be made part of a nursing care plan for the stay identifying the timing of interventions and expected outcomes. The University of Iowa has identified a comprehensive set of nursing diagnoses, NANDA (North American Nursing Diagnosis Association) diagnoses, NIC (Nursing Intervention Classifications) interventions and NOC (Nursing Outcomes Classifications) outcomes [23]; additionally, they provide information on how nursing diagnoses, interventions and outcomes all interrelate. Together, NANDA, NIC and NOC classifications could be computerized to automate the generation of nursing care plans as part of the automated patient medical record system.

·        Patient tailoring: Further ways to assist the caregiver in the documentation process is to have flow sheets, care plans, clinical pathways and other documents that could be tailored specifically for the patient. Non-applicable information for the patient could be excluded.

·        Interfacing of systems to eliminate redundant entry of information: An automated patient medical record system requires information also needed by other HMO clinical systems. This information includes proper patient identification information such as names, phone numbers and address; the start and end times of encounters; and orders and their results. As part of the automated patient medical record project, HMO clinical systems could be networked to share this information. As a result, caregivers would no longer be required to enter the same information multiple times in multiple HMO clinical systems. Inconsistent entrance of information would be eliminated.

5.4.2.4    Least Costly, and Best, Care Practices

Some care practices are more cost effective than others. Some medications are more cost effective than others.

During the caregiver interview of the patient and establishment of a treatment plan, the system can evaluate caregiver input to the automated patient medical record and compare it against HMO established best practice guidelines, making suggestions to the caregiver about more effective, less costly, treatments that produce the best outcomes [24].

During medication ordering, the automated system could make suggestions on more effective, and more cost-effective, medications. The system can give advice on best dosage and can automatically protect against drug interactions with other drugs, clinical laboratory tests, and foods. Generic medications could be identified which could replace a more costly brand name drug that is prescribed.

The automated system should be set up to recognize special characteristics of a patient that affect care decisions. For example, patient drug allergies could be immediately identified from previous recording of the patient’s allergies, with the system informing the caregiver is he orders such a drug. The caregiver could be warned when one of his care decisions seems to be inconsistent with a pre-existing patient condition.

Through these approaches, the quality and effectiveness of medical care can be improved, resulting in overall care at the lowest cost for the patient and the HMO.

5.4.2.5    Utilizing Healthcare Personnel More Effectively

With automation of the patient medical record, the cost of HMO staff can be reduced in a number of very different ways:

·        Caregivers would spend less time on paper work and more time caring for the patient, thus better utilizing caregiver time.

·        Some caregiver services can now be consolidated and outsourced, thus saving the HMO money.

·        Chart room staff and couriers transporting charts and orders could be decreased, reducing costs.

·        Reports could be automated instead of done by hand.

Better Utilization of Caregiver Time

Automating paper-based systems would result in better use of personnel directly caring for the patient. Thirty percent of the time spent in the care of an inpatient is spent on filling out paperwork for the patient chart. As a result of less redundant entry of information and of use of “charting by exception” (keeping previous information and only requiring recording of changes), there will be significantly less time in such paperwork by nurses and unit assistants. Completely or partially automated coding and grouping of discharge diagnoses, procedures and supplies for inpatients and outpatients (ICD, CPT, DRG codes) is also potentially possible, substituting for the manual process.

Without an automated patient medical record and ordering system, there is significant paperwork involved in the ordering process. The physician writes the order on a paper order form, which is either given to the patient, a unit assistant, courier, or other employee and is transferred to the pharmacy, laboratory or other performing area. The performing area may then re-enter the order in its computers. For the clinical laboratory, radiology department or other performing area, results might later be printed and returned by courier to the physician, with the billing office sent a copy of the test to charge the patient or insurance company. Without an automated patient medical record and ordering system, inpatient medication orders are re-written on a paper Medication Administration Record (MAR) in the unit, with a copy of the MAR later sent to the billing office; the billing office might then re-enter charges in their system.

With an automated patient medical record and ordering system, time and personnel costs of this paper flow would be greatly reduced. The physician’s automated order would be sent in seconds to the pharmacy, laboratory, radiology or other performing area. Prior to sending, the order could be checked for drug interactions, allergic reactions, safe dosages, duplicate tests, etc., and the order could be corrected immediately, instead of later, making it unnecessary for later costly interactions between the performing department and the physician. Once a clinical laboratory test was completed, diagnostic image analyzed and transcribed, etc., results would be immediately sent back to the ordering caregiver, with charges sent automatically to the billing office.  Inpatient medications could be immediately updated to the MAR; when a nurse recorded a medication as taken on the MAR, the system would automatically send the charges to the billing office. If an outpatient did not have a test done, did not pick up a prescription, did not call in for a refill when an expected time period or did not have a procedure done, the ordering caregiver could be alerted of possible non-compliance with the order and call up the patient, possibly saving extra visits, avoiding future health complications and insuring against malpractice suits. With automated ordering, time spent by nurses, physicians, pharmacists, and others in verifying illegible orders would be saved.

Because the patient medical record would always be available, a physician would be able to authoritatively provide advice over the phone, possibly substituting a phone call for a more costly face to face meeting, consequently better using the physician’s time.

The patient can be more quickly discharged, as the attending physician and other caregivers could be alerted that a discharge is pending, so a meeting with the patient for patient instructions can be scheduled and completed, discharge medications received, and follow-on procedures, tests, medications and appointments can be scheduled. Prior to discharge, caregivers can be informed of which discharge activities and documents are outstanding.

Outsourcing and Consolidation of Services Done Independently From the Patient

Outsourcing or consolidating medical services done independently from the patient would be much easier with automation of the patient medical record. For example, a group of radiologists could interpret diagnostic images and transcribe results for multiple health care organization locations or for multiple health care organizations. The automated diagnostic images would remain available to the caregiver caring for the patient during its interpretation. The ordering caregiver could be immediately alerted when results came back.   

The Changing Nature of Chart Rooms

Even though the chart room will be needed until the patient chart is fully automated, there will be less ordering of patient charts because sufficient chart information will often be available on-line. This will result in a decrease in the chart room staff. Also, there will be less time spent in tracking down a misplaced paper chart.

A decrease in chart room staff could also result from off-site storage of chart information. Pacific Bell is trying to establish a service to store diagnostic images of California health care organizations at Pacific Bell; Pacific Bell estimates that such storage of diagnostic image information at Pacific Bell or other outside institution could save a health care organization about 30%-40% of the current dollar costs of storage of diagnostic images [4]. The outside organization, and not the HMO, would have responsibility for storing back-up data off-site, and responsibility for the costs of insurance for possible destruction of the chart information.

Automation of Reporting

Reports based upon patient care can be automatically generated and information can be transferred over networks to outside organizations rather than be compiled by hand. This information may include information for the NCQA (National Committee for Quality Assurance) database [25], a national database for evaluating HMOs, and include information for generation of the HEDIS report, an HMO “report card”. Upon entrance, registry information can be automatically transmitted outside the organization.

5.4.2.6    Direct Savings to the HMO

It is estimated that there are costs of between $7 and $8 for each time a paper chart is retrieved [26]. Storage of diagnostic images digitally could save on the approximate cost of $40 per diagnostic image film. These costs alone could potentially pay for automation of the chart over time.

It is estimated that 11% of tests are reordered because the previous tests cannot be found. Complete automation of the chart will eliminate this. Not only would the HMO save money, but the patient would be spared the discomfort of taking additional unnecessary tests.

Transcription time and associated costs could be saved. Time and costs to do ordering could be cut, as well as costs of transporting paper orders and paper returned results. Communication between caregivers could be speeded, and the process of oversight of nurse practitioners and physician assistants by supervising physicians could be facilitated. All this could also potentially allow more patients to be seen, saving the healthcare organization costs on number of clinicians required.

It is estimated that adverse drug events (ADEs) occurring after hospitalizations cost hospitals $2 billion dollars per year.  Checking during physician order entry could reduce serious medication errors by half [27].  These errors include wrong doses, wrong choices, wrong techniques, delays, known allergies, missed doses, wrong drugs, drug-drug interactions, wrong frequencies, and wrong routes.

Automation of the chart can assist the HMO in identifying medical fraud. "Drug jumping", where a patient may go from clinic to clinic to deliberately get the same drug, can be detected.  Medicare and Medicaid fraud can be detected—It is estimated that fraudulent Medicare and Medicaid claims cost taxpayers as much as $30 billion annually for the many types of fraud [28]. Excessive drug prescriptions and other unusual situations can be automatically identified. Suspicious charges can automatically be flagged by the system, in particular when this involves payments by the HMO for outside medical services.

5.4.2.7    Support for Billing and Electronic Commerce

Money can be saved in the HMO by electronic billing and electronic commerce supported by the automated patient medical record.

An HMO, even one supported primarily by capitation, collects money for patient care from many sources, including from Medicare, Workman’s compensation, and insurance companies. An HMO also pays out money for outside medical services. In these processes, there is a need for billing, collection and payment information, much of which can be captured by the automated patient medical record system, including diseases, procedures, providers, supplies and medications for a patient.

From this same information, medical supplies can be automatically re-ordered to replenish hospital units when necessary.

A standard (1) for payments for services sent and received automatically over electronic networks and (2) for ordering medical supplies and drugs without operator intervention is Electronic Data Interchange (EDI). EDI is a standard for transmitting structured data using agreed upon message standards by “trading partners”, from a sender to a recipient, from the sending computer to the receiving computer [29]. This transmission can occur over private networks, referred to as “value added networks”, or VANs. EDI is increasingly being done over the Internet using industry agreed-upon formats in a language called XML, which is part of the same family of languages as HTML.

EDI can support automated billing and collection of medical payments from the government and outside insurance agencies based upon services recorded by the automated patient chart system. EDI can support automatic ordering of medical supplies based upon consumption of supplies recorded in conjunction with the automated patient chart system.

To support electronic healthcare claims to Medicare, the U.S. government, through the Health Insurance Portability and Accountability Act of 1996 (HIPAA) [30], has mandated standards for electronic claim forms and for provider identifiers, and in the future patient identifiers.

A third financial capability related to the automated patient medical record is caregiver ordering of durable medical equipment for a patient through outside companies (equipment leased or sold to patients for use in their homes, e.g., wheelchairs, walkers, canes) through outside companies. Durable medical equipment can be supported by electronic commerce over the Internet.

5.4.2.8    Decreasing Data Processing Costs Through Standardization

From the discussion so far, it is clear that the automated patient medical record system will be receiving information from many clinical and financial data processing systems in the HMO. These  include

·        encounter systems: As stated in section 5.4.1.3, encounters (inpatient stays, outpatient visits, emergency department visits, advice nurse calls, phone calls between a patient and a physician, home health visits, skilled nursing facility stays, etc.) can be used to organize the automated patient medical record.

·        ordering and results systems: As stated in section 5.4.1.5, these systems must be interfaced with the automated patient medical record system to receive orders from the automated patient medical record system and send results back to the automated patient medical record.

·        billing and financial systems:  As stated in 5.4.2.7, billing and financial systems need to be interfaced with the patient medical record system.

Interfacing these systems with the automated patient medical record system is a huge initial expense to the HMO, but also has the potential for long term savings to the HMO, especially an HMO as large as ours with many different such clinical and financial systems, if this interfacing can at the same time result in standardization of clinical systems and software throughout the HMO.

The potential savings is in future software and hardware maintenance costs. The National Bureau of Standards publication, “Guidance on Software Maintenance” [31] estimates that 60 to 70 percent of total software resources are spent upon maintenance. Reference [32] estimates that as much as 80 percent of a information system’s budget will be allocated to maintenance costs.

If a clinical system of a particular type was replaced by a company standard clinical system of that type at the time it was interfaced with the automated patient medical record system, then the following potential maintenance savings could result:

·        Less cost for installing systems: Installation of a new system for a new user would be less costly, as there would be only one installation procedure instead of many.

·        Less help desk and system support staff: If a computer or software system went down, standardized procedures could be followed to bring the system back up or to provide help for a user.

·        Less system maintenance staff: With fewer total clinical systems to support, less system maintenance staff would be needed for adding new system features or correcting bugs in systems.

·        Less cost for upgrading a type of clinical system: Replacement of a type of clinical system by a more powerful, improved one would be significantly less costly, as systems in the HMO could be all replaced at the same time.

In order to replace clinical systems within the HMO, they may have to first be decoupled. Consider figure 5.5 with shows interfaces that might occur between clinical systems in an HMO.

When the various clinical systems are interfaced with the automated patient medical record system to provide encounter information needed to organize the patient medical record, to send caregiver orders to the clinical systems, and to receive back results of orders, then the clinical systems could be decoupled. Proprietary connections between systems could be removed creating standard interfaces for each type of clinical system in the HMO, removing specialized connections between systems and creating ones where any system needing the information can get it. In such a setup, clinical systems become more independent  (see figure 5.6). One of many possible ways of interfacing these clinical systems is through a computer referred to as an interface engine to which each of the clinical systems is connected.

In creating this independence of systems, care must be taken to not create (1) a bottleneck where all network traffic is concentrated into a single connection with resulting periodic overloads in the system nor to create (2) a single point of failure where a single failure in a software or hardware system could cause the entire network to cease functioning. Later chapters, 10 and 15, discuss these potential problems and ways to avoid them.

An industry standard network protocol for healthcare systems interfaces is HL7 [33]. If this standard is used together with decoupling for each set of clinical systems in the healthcare organization, then there is the potential that any type of clinical system can be more easily replaced by another clinical system of the same type having the same interfacing scheme. Note that most commercial clinical systems have such HL7 interfaces. This concept of interchangeability of systems is referred to in this book as “component adaptability”, and is discussed later in this book.

Over time, proprietary and differing clinical systems could be replaced by common systems throughout the HMO.

All interfaces between clinical systems and with the automated patient chart system must be real-time interfaces unless the information is exceedingly static; otherwise, there is a great chance of information being out of synch. (This is both a cost and quality issue, although more a quality issue.)

A less restrictive approach to having a standard for each type of clinical system in the HMO (clinical lab systems, appointment systems, etc.) is to have different standards for large medical centers versus small ones.

Other types of standardization could also save the HMO money, such as organization wide standards for computers and other hardware. Standardization of hardware in a healthcare organization, whether for clinical systems or the automated patient medical record system, could include standardization of brands and configurations, or alternatively different standards for different types of users, e.g., “power users” versus “non-power users”. Such standardization could also save on help desk and support staff with standardized procedures in the organization for installing and fixing hardware.

5.4.2.9    Protection Against Costly Law Suits

Millions of dollars could be spent on malpractice suits in a large HMO in one year. With automation of the patient medical record, there would be one complete patient medical record available to all caregivers at all times, containing the patient’s complete care history, instead of possibly many paper charts each showing only a small part of the patient’s care. Information could always be found (1) due to the many concurrent ways an automated patient medical record could organized, differently for every category of caregiver (e.g., organized for those dealing with eye problems, those dealing with respiratory problems, for generalists, etc.), (2) due to the ability to use information retrieval techniques to find specific information, and (3) due to the ability to summarize key information. An automated patient medical record would decrease the chance of caregivers making mistakes due to lack of information or not finding information, and make defending against malpractice suits much easier.

When care information is entered into the automated patient medical record by a caregiver, the system could verify that the information entered would be understandable by other caregivers, for example, that a standardized medical vocabulary was used [34] and that abbreviations were either expanded and verified or were part of a standard set of abbreviations that could be understood in facilities in and outside the HMO [35]. This would guarantee that the patient medical record was always readable, and never illegible.

Once a caregiver has completed the information on a patient’s medical record for an encounter, the caregiver would sign off on the information, adding a digital signature [30]. The system would record the date and time and guarantee that the information could not be modified at a later date except through a clearly identified addendum. This removes the uncertainty that occurs in paper charts now that the caregiver could potentially modify the patient medical record after the fact.

Reference [36], the book Legal Aspects of Documenting Patient Care,  lists 25 common provider documentation errors that could be used against the caregiver in a malpractice suit. These errors are presented in figure 5.7 along with an indication of whether automation of the patient medical record could potentially eliminate the error. Most, but not all, errors could be eliminated by automation of the patient medical record.

 

Figure 5.7—Twenty-five Documentation Problems, Errors and Suggestions

Number

Description

Eliminated by Automation of the Patient Chart?

1

Illegible notation

Yes

2

Failure to identify the patient being treated

Yes

3

Failure to identify the date (and time) of treatment

Yes

4

Use of multiple or inconsistent documentation formats by providers in a facility

Yes

5

Failure to use an indelible instrument to record treatment entry

Yes

6

Pen runs out of ink midway through a treatment entry

Yes

7

Provider editing chart removing the originally entered data

Yes

8

Not signing treatment entries

Yes

9

Not properly correcting errors in treatment entries

Yes

10

Unauthorized abbreviations

Yes

11

Improper spelling, grammar, and use of extraneous verbiage not affecting patient care

Yes

12

Physician orders: ambiguous orders, treating patients without written orders

Largely

13

Untimely documentation of patient care

No

14

Identifying the filing of an incident report in the patient record--it should not be mentioned

No

15

Delineating patient care rendered and clinical information supplied by another caregiver

No

16

Blaming or disparaging another provider in the patient treatment record

No

17

Expressing personal feelings about a patient in the treatment record

No

18

Especially the 'O', 'A' and 'P' parts of the SOAP note must be written in objective, unambiguous and where possible quantifiable terms.  Providers should avoid ambiguous conclusions such as "appears within normal limits".

Potentially

19

Not documenting with specificity

Potentially

20

Recording hearsay as fact

No

21

Special caution is not exercised when countersigning another provider's evaluation or treatment

No

22

Failure to document a patient's informed consent to treatment

Yes

23

Failure to thoroughly document discharge, home care, and follow-up instructions issued to patients and/or family members or significant others

To some extent

24

Failure to carefully document a patient's noncompliance with treatment orders

No

25

Failure to carefully document a patient's or family member or significant other's possible contributory negligence

No

 

An automated patient medical record could also promote member satisfaction as the patient would no longer have to wait for the caregiver to receive the chart before the caregiver could take action, such as refilling a prescription. A satisfied member is less likely to bring a lawsuit and also tends to be more compliant to treatment regimes, reducing the chance of a lawsuit [37].

As was seen in figure 4.7 in the last chapter, some HMOs such as ours are protected from lawsuits by requiring members to accept contracts that require arbitration to resolve disputes. If this changed and lawsuits did become a significant problem in the future, for example as a result of changes in the law such as a “Patient Bill of Rights”, then the HMO might potentially have to change its practices to protect against these lawsuits.

One possibility would be practicing defensive medicine by performing more clinical tests, like the fee for service world does. This would increase costs to the HMO and subject the patient to possibly unnecessary tests.

Another possibility would be that physicians are required to follow only “best practice guidelines”. One possibility is using AHCPR whenever possible. If best medical practices are strictly followed, then there is little basis for a lawsuit. 

Use and enforcement of best medical practices could potentially be better tracked and enforced by an automated patient medical record that combines together encounters for the same condition and insures that best practices are being followed across these encounters. These best medical practices could potentially apply both inside and outside the HMO. This could both save money for the HMO and insure consistent medical care throughout the HMO, and outside the HMO when there are “point of service” plans. Through tracking of treatments across encounters together with the outcomes of the treatments, physicians who need retraining could be identified to insure the safety of patients before significant problems occurred. The negatives of using “best medical practices” are that individualized or inventive treatments might become less common.

5.4.3    Satisfying the Needs of Patients         

When the patient is being seen or wants to be seen by a physician or nurse practitioner, the patient should not encounter bottlenecks or roadblocks to receiving care. Once the patient is seen, the patient should be given quality medical care and clear caregiver instructions on follow-up activities and care.

The best quality medical care is useless if the patient does not comply with caregiver’s instructions. This can occur because the patient does not remember or understand the caregiver’s instructions, or because the information is not communicated clearly to the patient.

5.4.3.1    Helping the Patient Live a Normal, Comfortable Life

It is the responsibility of a caregiver to provide care that allows a patient to live, as much as possible, a normal, comfortable life. Two ways this can be done is (1) to insure that the patient comprehends and is able to follow care instructions from the caregiver, and (2) to identify where care can be changed to improve the patient’s well being, especially where a change in medication can improve a patient’s well being.

Caregivers’ instructions should be incorporated into the every day life of the patient to make the patient’s life more comfortable. This could be accomplished by easy-to-read and understandable written instructions given to the patient, which the patient would keep with her or him or reference when necessary. Where the patient and his family does not understand English, it may be appropriate for these instructions to be written in the language of the patient (e.g., Spanish, Chinese).

When a caregiver is caring for a patient, the automated patient medical record system, where possible, should recognize possible medication reactions. Where the patient is experiencing a symptom that might indicate a possible reaction to a current medication, the caregiver should be informed. Where there is such a symptom, or where the required dosage of the medication has increased significantly and has made the taking of a medication possibly unmanageable, and there might be a more appropriate medication without these effects, the caregiver should be informed. For example, a dosage requiring that the patient takes pills three times a day in regularly spaced intervals is less desirable to a situation where the patient gets the same results by taking another medication only one time a day, and a replacement of a medication by a more effective—though more costly—one might have dramatic positive effects upon the patient’s life (e.g., the new medication might no longer result in the patient being drowsy all of the time). A primary purpose of medicine is restoring or improving the well being of the patient—costs are a secondary consideration. 

5.4.3.2    Greater Patient Compliance with Doctor’s Orders

If the patient does not follow his doctor’s orders then he might as well have not gone to the seen the doctor. Again easy-to-read and understandable written instructions given to the patient that the patient would keep with him or her would help with compliance. Especially for patients take multiple medications, help with the scheduling the taking of medications could be provided.

5.4.3.3    Improving Patient Access to Care

Patient access to care involves the following:

·        The HMO enables the patient to quickly make an appointment with an appropriate physician, nurse practitioner or other care provider--this contact is either on the telephone or just after an outpatient visit

·        The HMO provides quality medical care for the patient or his family in a pleasant medical center.

·        The HMO eliminates roadblocks to the member and his family in receiving prompt care.

In creating and implementing the automated patient medical record, medical centers may have to be redesigned to accommodate automation and caregiver workflows may have to be redefined. In this process, the above aspects of patient access to care could be improved.

5.4.3.4    Personalized Care

“Personalized care” means treating the patient as an individual different than any other individual. In order to provide personalized care, the caregiver must be able to identify what the patient considers to be most important. What does the patient consider to be his or her most significant ongoing medical problem? Is she a mother and what are the names of her children? What is his or her wife or husband’s name? What else is important to him or her? What ongoing specialty care is the patient being seen for?

What would be very useful during all phases of care of the patient would be a quick summary of important attributes of the patient. This summary could both come from the patient, and from information known by the automated patient medical record system, such as ongoing specialty care currently being provided to the patient. This book will call this personalized information a personal profile.

Through a personal profile, a caregiver may be able to quickly identify what the HMO member considers most important, including what the member considers to be his or her most significant health problems, and perhaps including information such as a list of the member’s spouse and children. This would enable a caregiver to determine the medical conditions that concern the patient the most, and would enable care to be switched from the parent to his or her children or spouse.

With the list of specialty care currently being given, a caregiver would be able to quickly schedule a patient for this ongoing specialty care.

Personal information in the proposed personal profile would only be available with the approval of the patient. This personal profile would only be used in the local region of the HMO and would not be available outside the HMO. 

5.4.3.5    Greater Participation of the Patient and His Family in the Care Process

Automation of the patient medical record enables the patient to participate more fully in his or her own care.

Immediately after a patient visit, the patient could be given a printed document which could be reviewed with a nurse, which would explain the diagnosis, what the patient needs to do to get or keep healthy, and what medications have been prescribed or tests have been ordered. This is a document that is in the language the patient understands, matching the medical sophistication and interest level of the patient.

Additionally, the patient could check for the accuracy of information in the patient record and correct any inaccuracies. And when the patient is verbally told information during a visit, a patient can insure that this is recorded in the chart. Accuracy of information in the patient medical record would facilitate later evaluation of clinical outcomes.

Full disclosure also opens up another possibility: that a patient could become significantly more involved in his own care and treatment. Patients may want to get a much better knowledge about the details of their medical conditions and treatments, as stated in their medical records, and have a stronger influence on how their care is handled.

This process is already occurring in many different ways. The gay community has had a big impact on new treatments for AIDS/HIV. Societies have sprung up for diseases as diverse as psoriasis and breast cancer; often, members of these societies know much more about research in the area of their medical interest than all but a few physicians. Lots of medical information, though not always accurate, is available on the Internet.  Andy Grove, the CEO of Intel, has used his scientific research background to do an analysis of the treatments for prostate cancer, a disease he contracted [38]. Ideally, such societies, patients, laypersons, physicians and researchers can all support each other.

Younger patients require the support of their family members in the care process. Other patients may also want the support of their family members. Information produced for the patient could be in a form useable by a patient’s family, as well as the patient.

An advance directive enables a patient to participate in his or her care when the patient is unconscious or otherwise incapacitated. Again, advance directives are written instructions a patient has prepared for medical personnel to inform them of the patient’s wishes for treatments and care when the patient is incapacitated.

(Seriously ill patients and their families should be encouraged to discuss end of life decisions and the idea of assigning a surrogate or proxy decision maker for the patient via a healthcare proxy. A values history questionnaire illiciting a patient’s own personal view of care at the end of life—such as appears in the book Handbook for Mortals: Guidance for People Facing Serious Illness [39] and at the website in reference [40]—can used to assist family members in these discussions.)

5.4.3.6    Patient-Centered Care for High Risk Patients: “Guardian Angel” and Monitoring Systems

An automated patient medical record system could provide support for “patient-centered care”. A high risk patient could be given a computer system, perhaps a small computer using a pen for input (a personal digital assistant or PDA), for use outside the healthcare organization, at home or away from home, that can be used to monitor the patient’s health either via patient input or instrumentation input. The PDA based system could give advice, health education and therapy plans to the patient, inform the patient of appointments or inform the patient to schedule appointments. The system could interface with the automated patient medical record to alarm physicians or other caregivers of critical situations. This approach has been used in the “Guardian Angel” systems at MIT [41] for patients with the following medical conditions: insulin-dependent diabetes, hypertension, angina, chronic anticoagulation, chronic renal disease, and pulmonary disease (COPD).

Similar patient-centered systems in the hospital are computerized equipment connected to patients in the coronary and respiratory ICU.  These systems, like “Guardian Angel” systems, are monitoring systems.  Such monitoring systems, whether an ICU or “Guardian Angel” system, may collect a large amount of information from the patient. To feasibly store this information in the automated patient medical record, information, such as waveforms or input directly from patients, must be filtered, analyzed and selected for inclusion or not. As a result of the analysis, an alarm could be sent to a caregiver when a critical situation occurs. Enough additional information, besides the critical data, must be preserved to put the reasons for an alarm into context.

A less complicated version of a “Guardian Angel” system is a PDA that could be given to an elderly HMO member taking lots of medications, to inform the patient when a medication is to be taken. The member would need to respond, perhaps by a simple message box, when the medication was taken, with information on this sent to the automated patient medical record. If the member does not respond for a critical medication, he or she could be reminded by telephone. (It is estimated that 7-9% of admissions to hospitals are due to complications from medications.)

5.4.3.7    Patient Education

Through an automated system, up-to-date medical references of all kinds would be available. These could potentially include medical references for the patient to increase the patient’s understanding of his conditions.

For example, during a hospital stay or after an outpatient visit, the patient could be given an up-to-date FAQs sheet on each condition she or he is being treated for (e.g., pneumonia, pressure ulcers, etc.) printed through the automated system  (FAQ stands for “frequently asked questions”.)  The sheet should be compatible with any AHCPR guideline used for the condition; see section 5.4.1.5.

On-line multi-media presentations to demonstrate medical conditions and procedures could potentially be available. These multi-media presentations could be made available at the patient’s home through the Internet.

5.4.4    HMO Employee Satisfaction  

HMO employee satisfaction is a product of

·        the employee receiving a decent salary

·        the employee having a good working environment*

·        the employee having an opportunity of making a difference in the healthcare organization*

·        the employee’s work activities matching the most productive and least stressful methods of providing care*.

The *’s identify items that could be positively effected by automation of the patient medical record. Reengineering of patient care with strong caregiver input, removing roadblocks and creating a less stressful atmosphere for caregivers, could greatly improve caregiver satisfaction. 

With an automated patient medical record, strong support for provider workflow management could be provided. For example, it is known that providing a list of patients of interest to each caregiver which provides quick access to a patient’s medical record provides enormous benefits to a caregiver in managing his workflow.

5.4.5    Providing Information for Future Planning

According to reference [42], one problem of most managed care organizations is that they cannot accurately determine the true costs of patient care and they cannot predict future growth, or non-growth, of the HMO. As a result, information for future planning of the HMO is not available, e.g., for calculation of optimal capitation fees. The automated patient medical record facilitates capture of this cost and growth prediction information for future planning.

The automated patient medical record system and other systems can accurately record (1) services, tests and procedures given to patients, (2) costs of supplies and medications dispensed, (3) provider time caring for patients, (4) numbers and types of hospitalizations and visits, and (5) trends in membership growth and patient utilization. As a result, the actual costs of patient care can be better determined and trends in utilization can be determined.

This information facilitates future planning of personnel requirements and of budgets. It facilitates determination of optimal capitation fees, for example to control or increase growth.

5.4.6    Summary: Project Objectives Fulfilling Organizational Objectives

Our project furthers almost all the organizational objectives listed in section 5.3.

The purpose of the discussion in this section was to determine project objectives for our automated patient medical record project from HMO organizational objectives. Figure 5.8 shows the results--It shows HMO organizational objectives and related project objectives for the automated patient medical record.

 

Figure 5.8—Project Objectives Derived from Organizational Objectives

Number

Organizational Objective

Number

Project Objective

Metrics for Goals Measuring the Objective

1

Quality medical care is provided to HMO members (section 5.4.1)