The Limits of Planning

What are the special characteristics of business processes in hospitals?

Krahmer: At first glance, a hospital does not differ at all from a production company. If you look at the process chain, you see various stations at which specific actions are performed in both cases. Depending on how well or poorly the process is organized, the throughput times can differ.

The special characteristic of hospitals is that the “goods” are human beings. Everything that we do focuses on the patient. That’s why our challenge is to deal with events that can be planned for and those that cannot be planned.

What events can confuse processes?

Krahmer: For example, no one can forecast how many patients will come to a hospital on any given day or know what problems the patients will have. Accordingly, clinics must reserve capacity, even if doing so causes additional costs. In normal operations, a patient who has already been admitted might be unable to meet a scheduled appointment for an in-house examination because the patient has suddenly taken a turn for the worse.

Of course, similar problems occur in industry when a machine unexpectedly fails or when production must be reorganized. But the comparison with goods proves inadequate when human beings are involved.

What does IT contribute to the management of hospitals?

Krahmer: IT primarily supports efficient process management in hospitals: in logistics, for example. But you cannot overestimate the role of IT: the most important challenges are not technological. External practitioners and hospitals should better coordinate their work with each other. That would avoid performing complex diagnoses twice, and patients would not block the emergency room because of a lack of previous examinations. Healthcare must adhere to all sorts of standards in regard to patient protection and liability issues, especially during surgeries and diagnostic procedures.

Finally, it’s always a matter of budgets and who can and must take responsibility for payment. IT in healthcare depends on these factors.

Does IT in hospitals enable real process innovations?

Krahmer: In our industry, medical advances usually drive innovative processes. But aside from purely medical innovations, costs and legal requirements are the basic challenges that force every clinic to change its processes.

We find ourselves between a rock and a hard place. On one side, hospitals should operate like economic units and companies. On the other side, they should care for each patient individually.

The question here is which IT solution brings which benefits. Many areas in a hospital can use standard software, especially in logistics and controlling. But there are limits, especially in material consumption accounting. It’s impossible to define every possible parameter as you would on an assembly line. The software has to take care of that.

How can applications meet the requirements of clinics?

Krahmer: I believe that the software currently in use basically gives us a solid platform. Nevertheless, the solutions for practical work must be adjusted to the specifics of the industry and individual hospitals. The bandwidth stretches from small specialty clinics that want complete electronic integration of as many processes as possible to individual, paper-based practices.

Only minor details are often missing to use the software optimally. To close the gap between software developers and users, the University Clinic of Mannheim participates in the Enterprise Services Community program of SAP for the healthcare industry. We have worked with colleagues and experts from SAP to determine which supplemental or modified enterprise services hospitals need for their business processes. SAP has included the results of our working group in enhancement package 3 for SAP ERP, which provides the corresponding functionality.

What are those functions, for example?

Krahmer: With the functionality of enhancement package 3 from SAP, hospitals are better able to trigger downstream processes without complete master data. That is an important problem we have addressed. When patients are admitted to the hospital, many of them can provide only incomplete information on some topics, like their history.

Their condition makes it impossible for many patients to provide any information at all. In such situations, the administration of the hospital is confronted with incomplete data records. But it needs complete master data to complete the admitting transaction, assign the patient a number in the core enterprise resource planning system, and begin downstream processes.

If the system does not contain a record for an emergency patient, the sticker for the test tube used in a blood test cannot be printed because the subsystem in the lab needs the patient number from the core system. Complete identification of the test tube, however, is a precondition for automatic analysis of the blood sample and assignment of the results to the patient.

Redirection of master data is also important for admitted patients – for meals, for example. Another subsystem is responsible for purchasing, logistics, and preparation in this case. The system needs data that states that a specific patient exists and the location of the patient. It must also determine which provisions a patient needs – a special diet, for example. When a patient is admitted in the morning, the kitchen is already preparing lunch for that day. In some cases, however, the patient number that the subsystem needs to trigger the related order is not yet available.

Of course, the master data of a patient can be captured and stored as early as possible before the actual admittance. Many hospitals try to do just that. But that approach creates numerous superfluous data records – when a patient cancels on short notice before an operation, for example.

How extensively can patient-care processes be organized and mapped electronically?

Krahmer: Hospitals can analyze the approaches to care that have proven themselves in the past. Today’s IT systems feature special points to capture data for analysis. For example, this approach can be used to determine how much time is needed for a standard operation – from admittance to procedure – and which lab tests are required. An efficient process can be derived from this knowledge – but only in the context of a limited ability to plan and legal requirements for traceability.

Within a short window after admittance, hospitals can plan in great detail and attempt to keep to their plans. The number of unknowns increases thereafter because a patient’s condition can develop in any number of ways. In this phase, targets are still helpful – for mobilization of patients, for example. But only milestones are defined here. At those times, the personnel checks the health of the patient to derive additional measures. An IT system can map this structure quite well.

Large hospitals use management software and specialized individual solutions for various specialty departments. How can anyone get a handle on this heterogeneous IT landscape?

Krahmer: Service-oriented architectures are an interesting approach here. Standardized protocols have certainly already improved data exchange. But with an integrated architecture, hospitals would find it much easier to define the software for individual disciplines as services and to group them around a stable core system.

Prenatal diagnostics are a good example. Diagnostic findings are based on the results of ultrasound measurements that are compared with benchmarks from a central database in the clinic’s core system. The particular feature here is that both the measurement parameters and the benchmark data differ from those of other ultrasound examinations. The special query for prenatal diagnostics can be defined as a service which communicates with the core system. This architecture is, of course, much more flexible and cost-efficient than developing an additional interface for every diagnostic area.

Nevertheless, one must define exactly how a service is encapsulated and who is responsible up to a given point in the process. If diagnostic data is transmitted incorrectly, one must determine who is at fault. Who made a mistake? Was it the manufacturer of the diagnostic device, of the service, or of the core system? Or was it the clinic itself, which is responsible for operations of the entire system?

Any discussion must remember that a human organism pays no attention to well-defined IT systems.