HIS-Based Support of Follow-Up Documentation - Concept and Implementation for Clinical Studies

Abstract

Objective: Follow-up data must be collected according to the protocol of each clinical study, i.e. at certain time points. Missing follow-up information is a critical problem and may impede or bias the analysis of study data and result in delays. Moreover, additional patient recruitment may be necessary due to incomplete follow-up data. Current electronic data capture (EDC) systems in clinical studies are usually separated from hospital information systems (HIS) and therefore can provide limited functionality to support clinical workflow. In two case studies, we assessed the feasibility of HIS-based support of follow-up documentation.

Methods: We have developed a data model and a HIS-based workflow to provide follow-up forms according to clinical study protocols. If a follow-up form was due, a database procedure created a follow-up event which was translated by a communication server into an HL7 message and transferred to the import interface of the clinical information system (CIS). This procedure generated the required follow-up form and enqueued a link to it in a work list of the relating study nurses and study physicians, respectively.

Results: A HIS-based follow-up system automatically generated follow-up forms as defined by a clinical study protocol. These forms were scheduled into work lists of study nurses and study physicians. This system was integrated into the clinical workflow of two clinical studies. In a study from nuclear medicine, each scenario from the test concept according to the protocol of the single photon emission computer tomography/computer tomography (SPECT/CT) study was simulated and each scenario passed the test. For a study in psychiatry, 128 follow-up forms were automatically generated within 27 weeks, on average five forms per week (maximum 12, minimum 1 form per week).

Conclusion: HIS-based support of follow-up documentation in clinical studies is technically feasible and can support compliance with study protocols.

Keywords: Follow-up; clinical studies; completeness; hospital information system; single source information system.

Fig. 1

Architecture of dual source information systems: HIS and research databases are separated systems. Case report forms (CRFs) are entered into the research database and are not available in the hospital information system (HIS) [5]; CIS: clinical information system; LIMS: laboratory information systems; RIS/PACS: radiological information system/picture archiving and communication system.

Fig. 2

Architecture of single source information systems: Routine and research data are collected within HIS. Research data are exported into the research database. Monitoring takes care of incomplete or incorrect data elements [5].

Fig. 3

Architecture of automatic generation of follow-up forms: A scheduled database procedure executes defined queries on existing documentation. For related case IDs, a follow-up event is created and transferred by the communication server to the import interface of the CIS, RIS/PACS, LIMS, data warehouse or a departmental system respectively generating the follow-up form.

Fig. 4

Flow chart to generate follow-up events: Periodic queries identify due follow-up forms for n studies where n ∈ {1, …, n}. The system was implemented using procedural language/structured query language (PL/SQL).

Fig. 5

Flow chart of the implementation to generate follow-up events for the SPECT/CT study: Study-specific periodic queries identify due follow-up forms.

Fig. 6

Screenshot of the nurse’s/physician’s work list: Tasks from the clinical routine are presented in these lists; therefore, nurses and physicians regularly review the list items. Automatically generated follow-up protocols are enqueued in this list.

Fig. 7

Screenshot of a follow-up form in nuclear medicine: Radio buttons and check boxes are preferred data elements in order to speed up documentation. Data entry efforts are minimized by the use of many conditional items.

Fig. 8

Screenshot of a follow-up form in nuclear medicine with regard to loss of follow-up status: The follow-up form allows documentation of loss of follow-up patients.

Fig. 9

Flow chart of the implementation to generate follow-up events for the depression study: Periodic queries identify due follow-up forms.

Fig. 10

Bar diagram of automatically generated follow-up forms per calendar week: Bar diagram of automatically generated follow-up forms per calendar week. Within 27 weeks, 128 forms were generated, on average five per week (WK: week).