A search of literature discussing protocols, practice guidelines, standard orders, and critical paths was conducted to compile elements critical to the design and successful functioning and implementation of both manual and automated protocol systems.

Research confirmed that timeliness, flexibility and analysis of variance were critical factors for success. Rules, guidelines, or care plans had to be available to the provider at the time decisions were being made. Plans had to be adaptable to complicating circumstances, such as multiple problems, each with its own critical path, in a way that was helpful rather than burdensome to staff members. Their output needed to become the basis of charting, rather than an added record keeping burden. Variances from critical paths needed to be monitored, analyzed, and used as the basis for future modifications. Where the system was not flexible enough to be used in most cases for which it was intended, it would become a headache and fail to be utilized.

Definitions

It is possible for a single software application to deal effectively with clinical or treatment protocols, clinical guidelines, and standard orders, since these are differing approaches in the attempt to apply standards and rules to health care, and analyze results and outcomes with the goal of continually refining and improving the rules and their implementation. However they do have some significant differences and design requirements which must be accommodated in any application that hopes to deal with them effectively.

1. Critical Paths refer to activities, (answering the questions WHO, WHAT, WHEN). Their primary focus is on efficiency. They assume appropriate diagnosis has been made and appropriate procedures selected. Time requirements and dependencies are critical factors in their layouts (Yandell, B., 1995).
2. Clinical Practice Guidelines are most concerned with decision making (WHY?). Their primary focus is on appropriateness, and the decision points where the activities that will be performed are selected (Yandell, B., 1995).
3. Clinical / Treatment Protocols are most concerned with methods (HOW?). They describe a sequence of steps, often with branches based on results of previous steps. They generally address one discipline, and do not explicitly deal with the dimension of time. Flexibility in branching is a key element of their design requirements (Yandell, B., 1995).
4. Standard Order Sets can combine elements of the three other approaches, but are somewhat less formal, and are standing orders including different disciplines, and they may or may not include a time dimension and contingency statements. Their main advantage is one of flexibility, since they can easily be modified. (Yandell, B., 1995).

All approaches have a basic common structure and implementation. All have a triggering condition or set of conditions, all describe actions which need to be implemented in a sequential order, generally with some dependency that one of the previous actions has been completed, or that some specific event has occurred or outcome has been achieved. All require some method of recording and analyzing variances, in order to measure performance quality, or to refine the standards to make them more appropriate. And most importantly because of the volume of data tracking involved, all are dependent on information technology as a critical component needed for their successful implementation (Lumsdon, K. & Hagland, M., 1993). Such successful implementations depends on finding ways to lighten rather than increase the work load of staff members. Protocols implemented on a computer in real time can actually reduce unnecessary usage of tests, and have some effect on improving quality. Preliminary orders would be printed out by computer. Clinician would then put in the order. (Nightingale, P.G., Peters, M., et. al., 1994)

Abbot lists the principal causes of problems in the early attempts at automating clinical pathways:

• layering pathway on top of the Nursing Care plan (forcing extra or duplicate entry);
• inability to accommodate a multidisciplinary orientation;
• rigid adherence to the day of stay;
• task orientation;
• inability to merge multiple pathways;
• variance reporting not well integrated with the pathway.

He concludes that a successful automated implementation must have the following features:

• it should represent multidisciplinary contributions and responsibilities;
• integrate documentation and variance;
• provide flexibility in handling complex, changing patient needs;
• allow merging of multiple paths;
• shift from task oriented to outcome oriented care (acceleration or regression based on outcomes);
• reduce manual order entry.

He concludes that the consistent identification and analysis of variances is critical to the success of any automated implementation. (Abbott, J., Hronek, C,. and Mirecki, J.K., 1995)

Repeatedly discussions regarding how to make an implementation of protocols or other quality improvement standards succeed focused on the need to insure that protocols not only improved care quality, but did not increase workload. If the quality tool actually decreased staff workload, it stood a much better chance of being successfully implemented. There was extensive discussion about both avoiding the nightmare of maintaining duplicate documentation and development of parallel information systems, and finding ways to automatically put critical paths or protocols directly into the patient record (Coffee, R.J., Othman, J.E., Walters, J.I., 1995). In a situation where Standard Order Sets are produced and suggested by computer, physicians have the opportunity to modify them. Then once signed they become the orders. Physicians have found this method very convenient, and it has completely replaced their use of critical paths. (Yandell, B., 1995). Another term for this is "Charting by Exception." It introduces tremendous efficiencies as well as insuring that providers see what is recommended. (Abbott, J., Hronek, C,. and Mirecki, J.K., 1995)

Heymann discusses the other advantages that can be gained by incorporating standards of care directly into patients' case notes. Their use facilitates shared care, is an aid to staff training, and gives greater legal protection. Networked electronic patient records which incorporate protocols will allow the enormous value of protocols as quality tools and organizational drivers to be realized (Heymann, T., 1994).

In addition Critical Path Management can be used to improve the efficiency of clinical processes by focusing on time analysis of activities, and necessity of finding problem spots which delay everything else (80% of problems occurences are caused by 20% of possible causes). The critical path is the component of the pathway to the outcome that requires the most time to complete. (Luttman, R.J., Laffel, G.L., Pearson, S.D., 1995).

DESIGN ELEMENTS

Most discussions of protocol software design elements were concerned with actions, events and outcomes, variances, or the integration of the protocol software with other healthcare systems. Spath gave a more general description of the processes or functions that needed to be monitored including: Acuity, Time since diagnosis or admission, Consultation, Tests, Treatments, Medications, IV, Diet/Nutritional Support, Patient Activity, Teaching, DC planning (Spath, P.L., 1993).

Events

Events or milestones, such as outcomes, which have no time duration, need to be distinguished from activities which require time and resources to complete (Coffee, R.J., Othman, J.E., Walters, J,I., 1995). Events, like activities, can have dependencies. Or conversely, activity start times and durations can be tracked and then analyzed. Since events and outcomes can involve patients and family members, they have to be included in the critical path process. (Schriefer, J., 1995).

Actions / Activities

As already mentioned activities require time and resources to complete, and can be broken down in several different ways. How much of the provider's time is required, vs. how much total time is required for completion, vs. the amount of time that can be made available without affecting other activities might be determined. Dependencies must be tracked as well as the time limiting paths of activities, most importantly the initiation, duration, and completion time. Activities could be listed on an activity precedent table which numbered all activities, and listed all chronological dependencies, which Coffee, et. al. called precedents. A precedent is an activity that must be completed before another activity can begin. Activity categories such as consultations, tests, patient activities, treatments, medications, diet, and patient and family education seem to be fairly generally accepted. (Coffee, R.J., Othman, J.E., Walters, J,I., 1995)

Making Actions the basic building blocks is the key to developing a modular system. Actions can be linked together to form Routines. Protocols are assemblies of actions and routines organized in a branching logic algorithm. To facilitate this there are Questions to help determine how to branch. To help with descriptive parts of protocols, ways of recording comments are needed. (Michelson, S., Ben-Sassat, A., Weil, M.H., 1982)

All treatment and care activities should be handled as objects, each of which carries costing, coding, and resource implications as well as access to help. Single activities may be selected to build up individual patient-specific protocols or added to personalize existing protocols. Users need to be able to view the activity set by time, profession and activity status. They should be required to tick, cross or comment on all activities and be requested to file an exception report where necessary. At the touch of a button, all exceptions can be reviewed. (Heymann, T., 1994)

To analyze and manage Critical Pathways it is necessary to break time estimates down into categories such as most optimistic, normal, and most pessimistic time, using standard deviations. It is also necessary to break activity time into sub-components. Ordering/scheduling, processing, and reporting might be activity sub-components for providers. Treatment and recovery are sub-components for patients. The critical path is the component of the pathway to the outcome that requires the most time to complete. (Luttman, R.J., Laffel, G.L., Pearson, S.D., 1995).

In describing activity schedules related to administration of drugs or tests we need to be able to employ references to days of week, i.e.: M,W,F, and days since treatment began, i.e.: 1,8,12,15 (McColligan, E.E., 1987)

Integration with Other Data

Pryor stresses the importance of going after data from source definitions. By having the capability of monitoring lab data as it is placed in a patient record, alerts could be sent out immediately. Notification should be expanded from printed reports and terminal messages to use beepers as well. Some computerized protocols need to be data driven when particular data elements are stored in patients record. Others should be executed on demand by the nurse (Pryor, T.A., 1994).

An interface between a data base with patient status and the lab computers to automatically get results from tests is critical to allow both static and dynamic protocols to be enforced. It is important to provide preliminary printouts or reminders of what protocol describes prior to Physician actually putting in final orders. (Nightingale, P.G., Peters, M., et. al., 1994).

Variances

Pryor discusses the importance of following up on caregivers choosing not to follow an alert (variance), and to find out why in order to refine their model later. Medication critiquing alerts are also very important (Pryor, T.A., 1994).

Schriefer feels that variances must be linked directly into the case management process, and data entry of variances needs to be streamlined by having a field in which anticipated variances could be listed for any activity, with each of these variances having data definitions. Then once the user checked off that an outcome was unmet, the reason could be coded almost immediately. Variances can be categorized into the following groups: Patient/Family, Caregiver/Clinician, Hospital/System, Community (Schriefer, J., 1995).

Variances need to be tracked for outcomes such as: insurance refusals to pay for services, and re-admissions of patients. Certain variances sometimes need to become part of the path. When they occur, they will be addressed by a new set of activities (Schriefer, J., 1995).

Farley repeatedly stresses how critical it is to keep the documentation load down (Farley, K., 1995). Without streamlining and flexible support by a computer system, variance tracking is going to be immensely unpopular. Categories for variances were: Patient physical complication, patient social factors, physician non-compliance, nurse non-compliance, ancillary department causes, placement or discharge issues, or undetermined (Yandell, B., 1995).

From our review of the status and needs of protocol software design, we can conclude that the most critical considerations for the effective implementation and success of a computerized protocol tool are:

• Protocols cannot succeed if their use substantially increases staff burdens, workloads or documentation requirements. To insure staff acceptance protocols need to reduce documentation requirements.
• Activities are the most complex part of protocols. In order for a system to meet the likely needs of a wide variety of users, users must be given the ability to describe a wide range of activities.
• Variances, their description, follow-up and analysis plays a critical role in the ultimate effectiveness of protocol implementation. Their documentation cannot be too much of a burden for users, must be an integral part of care documentation (charting by exception), and must be used to refine protocols.