A practical framework for data management processes and their evaluation in population-based medical registries.

INTRODUCTION: We present a framework for data management processes in population-based medical registries. Existing guidelines lack the concreteness we deem necessary for them to be of practical use, especially concerning the establishment of new registries. Therefore, we propose adjustments and concretisations with regard to data quality, data privacy, data security and registry purposes. MATERIALS AND METHODS: First, we separately elaborate on the issues to be included into the framework and present proposals for their improvements. Thereafter, we provide a framework for medical registries based on quasi-standard-operation procedures. RESULTS: The main result is a concise and scientifically based framework that tries to be both broad and concrete. Within that framework, we distinguish between data acquisition, data storage and data presentation as sub-headings. We use the framework to categorise and evaluate the data management processes of a German cancer registry. DISCUSSION: The standardisation of data management processes in medical registries is important to guarantee high quality of the registered data, to enhance the realisation of purposes, to increase efficiency and to enable comparisons between registries. Our framework is destined to show how one central impediment for such standardisations - lack of practicality - can be addressed on scientific grounds.

Active learning strategies for the deduplication of electronic patient data using classification trees.

INTRODUCTION: Supervised record linkage methods often require a clerical review to gain informative training data. Active learning means to actively prompt the user to label data with special characteristics in order to minimise the review costs. We conducted an empirical evaluation to investigate whether a simple active learning strategy using binary comparison patterns is sufficient or if string metrics together with a more sophisticated algorithm are necessary to achieve high accuracies with a small training set. MATERIAL AND METHODS: Based on medical registry data with different numbers of attributes, we used active learning to acquire training sets for classification trees, which were then used to classify the remaining data. Active learning for binary patterns means that every distinct comparison pattern represents a stratum from which one item is sampled. Active learning for patterns consisting of the Levenshtein string metric values uses an iterative process where the most informative and representative examples are added to the training set. In this context, we extended the active learning strategy by Sarawagi and Bhamidipaty (2002). RESULTS: On the original data set, active learning based on binary comparison patterns leads to the best results. When dropping four or six attributes, using string metrics leads to better results. In both cases, not more than 200 manually reviewed training examples are necessary. CONCLUSIONS: In record linkage applications where only forename, name and birthday are available as attributes, we suggest the sophisticated active learning strategy based on string metrics in order to achieve highly accurate results. We recommend the simple strategy if more attributes are available, as in our study. In both cases, active learning significantly reduces the amount of manual involvement in training data selection compared to usual record linkage settings.

Missing values in deduplication of electronic patient data.

INTRODUCTION: Systematic approaches to dealing with missing values in record linkage are still lacking. This article compares the ad-hoc treatment of unknown comparison values as 'unequal' with other and more sophisticated approaches. An empirical evaluation was conducted of the methods on real-world data as well as on simulated data based on them. MATERIAL AND METHODS: Cancer registry data and artificial data with increased numbers of missing values in a relevant variable are used for empirical comparisons. As a classification method, classification and regression trees were used. On the resulting binary comparison patterns, the following strategies for dealing with missingness are considered: imputation with unique values, sample-based imputation, reduced-model classification and complete-case induction. These approaches are evaluated according to the number of training data needed for induction and the F-scores achieved. RESULTS: The evaluations reveal that unique value imputation leads to the best results. Imputation with zero is preferred to imputation with 0.5, although the latter shows the highest median F-scores. Imputation with zero needs considerably less training data, it shows only slightly worse results and simplifies the computation by maintaining the binary structure of the data. CONCLUSIONS: The results support the ad-hoc solution for missing values 'replace NA by the value of inequality'. This conclusion is based on a limited amount of data and on a specific deduplication method. Nevertheless, the authors are confident that their results should be confirmed by other empirical analyses and applications.

A data protection scheme for medical research networks. Review after five years of operation.

BACKGROUND: The data protection requirements matured in parallel to new clinical tests generating more personal data since the 1960s. About ten years ago it was recognized that a generic data protection scheme for medical research networks is required, which reinforces patient rights but also allows economically feasible medical research compared to "hand-carved" individual solutions. OBJECTIVES: To give recommendations for more efficient IT infrastructures for medical research networks in compliance with data protection requirements. METHODS: The IT infrastructures of three medical research networks were reviewed with respect to the relevant data management modules. Recommendations are derived to increase cost efficiency in research networks assessing the consequences of a service provider approach without lowering the data protection level. RESULTS: The existing data protection schemes are very complex. Smaller research networks cannot afford the implementation of such schemes. Larger networks struggle to keep them sustainable. Due to a modular redesign in the medical research network community, a new approach offers opportunities for an efficient sustainable IT infrastructure involving a service provider concept. For standard components 70-80% of the costs could be cut down, for open source components about 37% over a three-year period. CONCLUSIONS: Future research networks should switch to a service-oriented approach to achieve a sustainable, cost-efficient IT infrastructure.

Evaluation of record linkage methods for iterative insertions.

OBJECTIVES: There have been many developments and applications of mathematical methods in the context of record linkage as one area of interdisciplinary research efforts. However, comparative evaluations of record linkage methods are still underrepresented. In this paper improvements of the Fellegi-Sunter model are compared with other elaborated classification methods in order to direct further research endeavors to the most promising methodologies. METHODS: The task of linking records can be viewed as a special form of object identification. We consider several non-stochastic methods and procedures for the record linkage task in addition to the Fellegi-Sunter model and perform an empirical evaluation on artificial and real data in the context of iterative insertions. This evaluation provides a deeper insight into empirical similarities and differences between different modelling frames of the record linkage problem. In addition, the effects of using string comparators on the performance of different matching algorithms are evaluated. RESULTS: Our central results show that stochastic record linkage based on the principle of the EM algorithm exhibits best classification results when calibrating data are structurally different to validation data. Bagging, boosting together with support vector machines are best classification methods when calibrating and validation data have no major structural differences. CONCLUSIONS: The most promising methodologies for record linkage in environments similar to the one considered in this paper seem to be stochastic ones.

[Registries for rare diseases. Compliance and data protection]. / Register zu seltenen Krankheiten : Patientencompliance und Datenschutz.

For rare diseases, clinical and epidemiological research suffers from very small numbers of cases. A comprehensive collection of data and information in registries is an essential precondition to improve this situation. To this end, a number of disease specific networks have started collecting data with support from the German Ministry of Research. The past experiences of the German Childhood Cancer Registry show that voluntary participation, based on informed consent, can result in a satisfactory completeness of data collection and, thus, enable successful medical research. There are several ways to build registries and research networks conforming to the data protection rules.

Integrated information systems for translational medicine.

OBJECTIVES: Translational medicine research needs a two-way information highway between 'bedside' and 'bench'. Unfortunately there are still weak links between successfully integrated information roads for bench, i.e. research networks, and bedside, i.e. regional or national health information systems. The question arises, what measures have to be taken to overcome the deficiencies. METHODS: It is examined how patient care-related costs of clinical research can be separated and shared by health insurances, whether quality of patient care data is sufficient for research, how patient identity can be maintained without conflict to privacy, how care and research records can be archived, and how information systems for care and research can be integrated. RESULTS: Since clinical trials improve quality of care, insurers share parts of the costs. Quality of care data has to be improved by introducing minimum basic data sets. Pseudonymization solves the conflict between needs for patient identity and privacy. Archiving patient care records and research records is similar and XML and CDISC can be used. Principles of networking infrastructures for care and research still differ. They have to be bridged first and harmonized later. CONCLUSIONS: To link information systems for care (bed) and for research (bench) needs technical infrastructures as well as economic and organizational regulations.

[Clinical data protection and the internet]. / Klinischer Datenschutz und Internet.

The benefit of using the internet in health care is growing: Information retrieval, communication, and telemedical cooperation support the physician and serve the patient. On the other hand connecting a hospital to the internet leads to considerable security problems, exposes sensible data to danger, and corrupts the professional discretion, unless the connection is installed and managed with extreme care. The state of the art for an internet connection is a firewall system. Moreover the security in a networked environment depends on a cryptographic infrastructure. This will soon be provided in Germany by Health Professional Cards.

[IT safety in medical networks--current problems and approach to solutions]. / IT-Sicherheit in medizinischen Netzen--aktuelle Probleme und Lösungsansätze.

Designers and users of medical networks have to face strong requirements for data protection and security. Professional discretion and data protection laws allow the transfer of or access to patient data only in a therapeutic context. These data should also be protected from the network provider. Patients should be safe from any harm by faulty data or buggy procedures. On the other hand the security of the most used software products gets worse and worse. The use of the internet endangers more and more the integrity of the user's computer. The security requirements can be met only through strict care in planning, building, and configuring the infrastructure. Some concrete recommendations and guiding principles can immediately be realized. If these recommendations are followed, the internet can be of immense value for health care.

Retention of hyaluronic acid in alginate beads: aspects for in vitro cartilage engineering.

Alginate has been used successfully for three-dimensional chondrocyte cultures and may be important for cartilage transplant formation. However, alginate is not a natural component of the cartilage matrix. The aim of this study was (a) to supplement alginate with the extracellular matrix component hyaluronic acid; and (b) to analyze the hyaluronic acid retention in different alginate gels. Hyaluronan is assumed to improve proteoglycan retention and may be important for in vitro matrix formation, tissue turgor, and biomechanical quality. Alginate and hyaluronan were mixed with chondrocytes and polymerized as were alginate, hyaluronan, and fibrinogen. [3H]hyaluronan was used to quantitate the leakage of hyaluronan from the gel beads. After 28 days in culture, 1.2% alginate beads supplemented with 0.26% hyaluronan contained only 9% of the initial amount of hyaluronan whereas 2.4% alginate beads still contained about 55% of the initial 0.22% hyaluronan. Release of hyaluronan from the beads was significantly lower if the beads additionally contained fibrin. Alginate beads supplemented with hyaluronan or fibrin showed increased chondrocyte proliferation compared to controls. Supplemented hyaluronan greatly diffuses out of alginate gels of lower densities. It must be assumed also that most of the hyaluronan newly synthesized by chondrocytes in these cells diffuses into the surrounding culture medium. The in vitro development of a sufficiently hygroscopic cartilage ground substance therefore may be very limited. Sufficient hyaluronic acid retention can be achieved in alginate gels with concentrations above 1.2% or by addition of fibrin.

THEMPO: a knowledge-based system for therapy planning in pediatric oncology.

This article describes the knowledge-based system THEMPO (Therapy Management in Pediatric Oncology), which supports protocol-directed therapy planning and configuration in pediatric oncology. THEMPO provides a semantic network controlled by graph grammars to cover the different types of knowledge relevant in the domain, and offers a suite of acquisition tools for knowledge base authoring. Medical problem solvers, operating on the oncological network, reason about adequate therapeutic and diagnostic timetables for a patient. Furthermore, a corresponding patient record, also based on semantic networks and graph grammars, has been implemented to represent the course of therapy of an oncological patient.

[Data protection measures in medical information systems]. / Datenschutzmassnahmen in medizinischen Informationssystemen.

Information technology and networks should improve health care in the next few years. But the information systems in use are open systems; processing and transporting data by them is incompatible with the high privacy requirements of patient information. We have the technical means to achieve better security, but we cannot build them into existing systems by way of addition--the suppliers of hard- and software must integrate them into the systems, beginning with the system design, in a way that doesn't detract users from their proper tasks; users shouldn't have to do more than log into the IT system with their Health Professional Card. On the other hand the expense of organisational means and security personnel in health care environment is not negligible. But a thorough realization of security features will substantially improve the quality of information and the compliance with the law. The costs will reduce the profit of information technology, but not completely.

Pseudonyms for cancer registries.

In order to conform to the rigid German legislation on data privacy and security we developed a new concept of data flow and data storage for population-based cancer registries. A special trusted office generates a pseudonym for each case by a cryptographic procedure. This office also handles the notification of cases and communicates with the reporting physicians. It passes pseudonymous records to the registration office for permanent storage. The registration office links the records according to the pseudonyms. Starting from a requirements analysis we show how to construct the pseudonyms; we then show that they meet the requirements. We discuss how the pseudonyms have to be protected by cryptographic and organizational means. A pilot study showed that the proposed procedure gives acceptable synonym and homonym error rates. The methods described are not restricted to cancer registration and may serve as a model for comparable applications in medical informatics.

A graph-grammar approach to represent causal, temporal and other contexts in an oncological patient record.

The data of a patient undergoing complex diagnostic and therapeutic procedures do not only form a simple chronology of events, but are closely related in many ways. Such data contexts include causal or temporal relationships, they express inconsistencies and revision processes, or describe patient-specific heuristics. The knowledge of data contexts supports the retrospective understanding of the medical decision-making process and is a valuable base for further treatment. Conventional data models usually neglect the problem of context knowledge, or simply use free text which is not processed by the program. In connection with the development of the knowledge-based system THEMPO (Therapy Management in Pediatric Oncology), which supports therapy and monitoring in pediatric oncology, a graph-grammar approach has been used to design and implement a graph-oriented patient model which allows the representation of non-trivial (causal, temporal, etc.) clinical contexts. For context acquisition a mouse-based tool has been developed allowing the physician to specify contexts in a comfortable graphical manner. Furthermore, the retrieval of contexts is realized with graphical tools as well.

A new concept to ensure data privacy and data security in cancer registries.

In order to conform to the rigid German legislation on data privacy and security, a new concept of data flow and data storage for population-based cancer registries has been developed. This paper describes the basic principles of this approach and presents results from a pilot study in Rheinland-Pfalz. The cancer registry consists of two offices. The first office handles the notification of cases and communicates with the reporting physicians. It performs encryption of identifying data and passes encrypted records to the second office for permanent storage. Record linkage is performed with the encrypted data in the second office. This paper describes in detail how different encryption techniques are used for different purposes in order to ensure optimal security and the usability of the data. Our record linkage study shows that it is possible to keep both synonym and homonym error rates below 1%--acceptable both for routine performance of a registry and for scientific analyses. The methods described are not restricted to cancer registration and may therefore serve as a model for comparable MI applications.

Bead cellulose derivatives as supports for immobilization and chromatographic purification of proteins.

Characteristic data are presented for Divicell, a macroporous bead cellulose with excellent flow parameters. The preparation of Divicell derivatives and their properties are described with respect to their application as chromatographic supports. The ion exchangers Divicell DEAE and Divicell CM were manufactured in two types with different exclusion limits and an available capacity for proteins of up to 100 mg/ml gel. Divicell Blue is a bead cellulose with covalently bound Cibacron Blue F3G-A and was found to be a very suitable adsorbent for the selective separation and purification of human serum albumin. Activation of Divicell with sodium periodate, epichlorohydrin and 5-norbornene-2,3-dicarboximido carbonochloridate provided activated supports used for immobilization of ligands in organic solvents and in aqueous solutions. Coupling of amines, diamines, amino acids, carbohydrates and proteins is described. The immobilized ligands retained their biological activity as determined by their specific adsorption of proteins. Divicell alkyl derivatives were tested in hydrophobic interaction chromatography with bovine serum albumin as a model. Examples are presented of the application of Divicell derivatives to the purification of biomacromolecules such as immunoglobulins and lectins by affinity chromatography. The results were comparable to those obtained using the corresponding Sepharose-derived absorbents.

Modification of carboxyl groups on NADPH-cytochrome P-450 reductase involved in binding of cytochromes c and P-450 LM2.

Carboxyl groups of NADPH-cytochrome P-450 reductase have been modified with the water-soluble carbodiimide EDC. Although there is no significant loss in DCPIP reduction the activity with cytochrome c and cytochrome P-450 LM2 as electron acceptors was inhibited by about 60 and 85%, respectively (1 h incubation time, 20 mM EDC). The inactivation by EDC was nearly completely prevented in the presence of cytochrome P-450 LM2, but not by bovine serum albumin. These results and crosslinking studies suggest that carboxyl groups of NADPH-cytochrome P-450 reductase are involved in charge-pair interactions to cytochrome c and to at least two amino groups of cytochrome P-450 LM2.

p-Nitrophenetole deethylase activity of rat liver microsomes entrapped in polyelectrolyte capsules.

Liver microsomes from phenobarbital induced rats are entrapped in capsules prepared from polyelectrolytes. A comparative analysis of the deethylase activity against p-nitrophenetole by encapsulated and freely suspended microsomes is carried out. The pH optimum occurs at about 7.2 for encapsulated as well as free microsomes. The pH activity profile of encapsulated microsomes, however, is strongly flattened. The maximal velocity of entrapped microsomes is about a quarter of that of free microsomes. The Michaelis-Menten constants are virtually equal. Despite of the lowered activity of encapsulated microsomes this kind of immobilization of enzymes may be useful in biotechnology and medicine because mild immobilization conditions like room temperatures and aqueous solutions are realized.

Spin state control of cytochrome P-450 reduction and catalytic activity in a reconstituted P-450 LM2 system as induced by a series of benzphetamine analogues.

Reconstituted liposomal cytochrome P-450 LM2 was reacted with a series of benzphetamine analogues as substrates. Based on the thermodynamical model of Ristau et al. (Biochim. Biophys. Acta, 536 (1978) 226-234) the free enthalpy of substrate binding to the high spin form of the enzyme was shown to correlate with the total high spin content of the respective enzyme substrate complex. Reduction and substrate N-demethylation rates as well have been evidenced to linearly correlate with the substrate-induced spin shift delta alpha and moreover with the spin content alpha. The data obtained provide further experimental support for the spin state regulation of the reduction and conversion rate of cytochrome P-450 LM2.

[Activity of urease in microcapsules]. / Aktivitätseigenschaften von Urease in Mikrokapseln.

The activity of urease in microcapsules formed by a symplex membrane from cellulose sulphate and polydimethyldiallyl ammonium chloride is studied and compared with that of free urease in solution. Both free and encapsulated urease exhibits a similar pH optimum. The interaction of urease with cellulose sulphate causes a decrease of the activity down to 35 to 38%. In spite of its lower absolute activity the encapsulated urease shows a higher long time stability than the free urease. The presented immobilization method for urease could be of importance for an active detoxification in uremia.