Views of diagnosis distribution in primary care in 2.5 million encounters in Stockholm: a comparison between ICD-10 and SNOMED CT.

BACKGROUND: Primary care (PC) in Sweden provides ambulatory and home health care outside hospitals. Within the County Council of Stockholm, coding of diagnoses in PC is mandatory and is done by general practitioners (GPs) using a Swedish primary care version of the International Statistical Classification of Diseases, version 10 (ICD-10). ICD-10 has a mono-hierarchical structure. SNOMED CT is poly-hierarchical and belongs to a new generation of terminology systems with attributes (characteristics) that connect concepts in SNOMED CT and build relationships. Mapping terminologies and classifications has been pointed out as a way to attain additional advantages in describing and documenting healthcare data. A poly-hierarchical system supports the representation and aggregation of healthcare data on the basis of specific medical aspects and various levels of clinical detail. OBJECTIVE: To describe and compare diagnoses and health problems in KSH97-P/ICD-10 and SNOMED CT using primary care diagnostic data, and to explore and exemplify complementary aggregations of diagnoses and health problems generated from a mapping to SNOMED CT. METHODS: We used diagnostic data collected throughout 2006 and coded in electronic patient records (EPRs), and a mapping from KSH97-P/ICD-10 to SNOMED CT, to aggregate the diagnostic data with SNOMED CT defining hierarchical relationship Is a and selected attribute relationships. RESULTS: The chapter level comparison between ICD-10 and SNOMED CT showed minor differences except for infectious and digestive system disorders. The relationships chosen aggregated the diagnostic data to 2861 concepts, showing a multidimensional view on different medical and specific levels and also including clinically relevant characteristics through attribute relationships. CONCLUSIONS: SNOMED CT provides a different view of diagnoses and health problems on a chapter level, and adds significant new views of the clinical data with aggregations generated from SNOMED CT Is a and attribute relationships. A broader use of SNOMED CT is therefore of importance when describing and developing primary care.

Visualization of disease distribution with SNOMED CT and ICD-10.

Methods for presentation of disease and health problem distribution in a health care environment rely among other things on the inherent structure of the controlled terminology used for coding. In the present study, this aspect is explored with a focus on ICD-10 and SNOMED CT. The distribution of 2,5 million diagnostic codes from primary health care in the Stockholm region is presented and analyzed through the "lenses" of ICD-10 and SNOMED CT. The patient encounters, originally coded with a reduced set of ICD-10 codes used in primary health care in Sweden, were mapped to SNOMED CT concepts through a mapping table. The method used for utilizing the richer structure of SNOMED CT as compared to ICD-10 is presented, together with examples of produced disease distributions. Implications of the proposed method for enriching a traditional classification such as ICD-10 through mappings to SNOMED CT are discussed.

Archetype-based conversion of EHR content models: pilot experience with a regional EHR system.

BACKGROUND: Exchange of Electronic Health Record (EHR) data between systems from different suppliers is a major challenge. EHR communication based on archetype methodology has been developed by openEHR and CEN/ISO. The experience of using archetypes in deployed EHR systems is quite limited today. Currently deployed EHR systems with large user bases have their own proprietary way of representing clinical content using various models. This study was designed to investigate the feasibility of representing EHR content models from a regional EHR system as openEHR archetypes and inversely to convert archetypes to the proprietary format. METHODS: The openEHR EHR Reference Model (RM) and Archetype Model (AM) specifications were used. The template model of the Cambio COSMIC, a regional EHR product from Sweden, was analyzed and compared to the openEHR RM and AM. This study was focused on the convertibility of the EHR semantic models. A semantic mapping between the openEHR RM/AM and the COSMIC template model was produced and used as the basis for developing prototype software that performs automated bi-directional conversion between openEHR archetypes and COSMIC templates. RESULTS: Automated bi-directional conversion between openEHR archetype format and COSMIC template format has been achieved. Several archetypes from the openEHR Clinical Knowledge Repository have been imported into COSMIC, preserving most of the structural and terminology related constraints. COSMIC templates from a large regional installation were successfully converted into the openEHR archetype format. The conversion from the COSMIC templates into archetype format preserves nearly all structural and semantic definitions of the original content models. A strategy of gradually adding archetype support to legacy EHR systems was formulated in order to allow sharing of clinical content models defined using different formats. CONCLUSION: The openEHR RM and AM are expressive enough to represent the existing clinical content models from the template based EHR system tested and legacy content models can automatically be converted to archetype format for sharing of knowledge. With some limitations, internationally available archetypes could be converted to the legacy EHR models. Archetype support can be added to legacy EHR systems in an incremental way allowing a migration path to interoperability based on standards.

Representing a chemotherapy guideline using openEHR and rules.

Computerized guidelines can provide decision support and facilitate the use of clinical guidelines. Several computerized guideline representation models (GRMs) exist but the poor interoperability between the guideline systems and the Electronic Health Record (EHR) systems limits their clinical usefulness. In this study we analyzed the clinical use of a published lymphoma chemotherapy guideline. We found that existing GRMs have limitations that can make it difficult to meet the clinical requirements. We hypothesized that guidelines could be represented as data and logic using openEHR archetypes, templates and rules. The design was tested by implementing the lymphoma guideline. We conclude that using the openEHR models and rules to represent chemotherapy guidelines is feasible and confers several advantages both from a clinical and from an informatics perspective.

Non-compliance with a postmastectomy radiotherapy guideline: decision tree and cause analysis.

BACKGROUND: The guideline for postmastectomy radiotherapy (PMRT), which is prescribed to reduce recurrence of breast cancer in the chest wall and improve overall survival, is not always followed. Identifying and extracting important patterns of non-compliance are crucial in maintaining the quality of care in Oncology. METHODS: Analysis of 759 patients with malignant breast cancer using decision tree induction (DTI) found patterns of non-compliance with the guideline. The PMRT guideline was used to separate cases according to the recommendation to receive or not receive PMRT. The two groups of patients were analyzed separately. Resulting patterns were transformed into rules that were then compared with the reasons that were extracted by manual inspection of records for the non-compliant cases. RESULTS: Analyzing patients in the group who should receive PMRT according to the guideline did not result in a robust decision tree. However, classification of the other group, patients who should not receive PMRT treatment according to the guideline, resulted in a tree with nine leaves and three of them were representing non-compliance with the guideline. In a comparison between rules resulting from these three non-compliant patterns and manual inspection of patient records, the following was found: In the decision tree, presence of perigland growth is the most important variable followed by number of malignantly invaded lymph nodes and level of Progesterone receptor. DNA index, age, size of the tumor and level of Estrogen receptor are also involved but with less importance. From manual inspection of the cases, the most frequent pattern for non-compliance is age above the threshold followed by near cut-off values for risk factors and unknown reasons. CONCLUSION: Comparison of patterns of non-compliance acquired from data mining and manual inspection of patient records demonstrates that not all of the non-compliances are repetitive or important. There are some overlaps between important variables acquired from manual inspection of patient records and data mining but they are not identical. Data mining can highlight non-compliance patterns valuable for guideline authors and for medical audit. Improving guidelines by using feedback from data mining can improve the quality of care in oncology.

Integration of tools for binding archetypes to SNOMED CT.

BACKGROUND: The Archetype formalism and the associated Archetype Definition Language have been proposed as an ISO standard for specifying models of components of electronic healthcare records as a means of achieving interoperability between clinical systems. This paper presents an archetype editor with support for manual or semi-automatic creation of bindings between archetypes and terminology systems. METHODS: Lexical and semantic methods are applied in order to obtain automatic mapping suggestions. Information visualisation methods are also used to assist the user in exploration and selection of mappings. RESULTS: An integrated tool for archetype authoring, semi-automatic SNOMED CT terminology binding assistance and terminology visualization was created and released as open source. CONCLUSION: Finding the right terms to bind is a difficult task but the effort to achieve terminology bindings may be reduced with the help of the described approach. The methods and tools presented are general, but here only bindings between SNOMED CT and archetypes based on the openEHR reference model are presented in detail.

An archetype-based testing framework.

With the introduction of EHR two-level modelling and archetype methodologies pioneered by openEHR and standardized by CEN/ISO, we are one step closer to semantic interoperability and future-proof adaptive healthcare information systems. Along with the opportunities, there are also challenges. Archetypes provide the full semantics of EHR data explicitly to surrounding systems in a platform-independent way, yet it is up to the receiving system to interpret the semantics and process the data accordingly. In this paper we propose a design of an archetype-based platform-independent testing framework for validating implementations of the openEHR archetype formalism as a means of improving quality and interoperability of EHRs.

Creating a medical dictionary using word alignment: the influence of sources and resources.

BACKGROUND: Automatic word alignment of parallel texts with the same content in different languages is among other things used to generate dictionaries for new translations. The quality of the generated word alignment depends on the quality of the input resources. In this paper we report on automatic word alignment of the English and Swedish versions of the medical terminology systems ICD-10, ICF, NCSP, KSH97-P and parts of MeSH and how the terminology systems and type of resources influence the quality. METHODS: We automatically word aligned the terminology systems using static resources, like dictionaries, statistical resources, like statistically derived dictionaries, and training resources, which were generated from manual word alignment. We varied which part of the terminology systems that we used to generate the resources, which parts that we word aligned and which types of resources we used in the alignment process to explore the influence the different terminology systems and resources have on the recall and precision. After the analysis, we used the best configuration of the automatic word alignment for generation of candidate term pairs. We then manually verified the candidate term pairs and included the correct pairs in an English-Swedish dictionary. RESULTS: The results indicate that more resources and resource types give better results but the size of the parts used to generate the resources only partly affects the quality. The most generally useful resources were generated from ICD-10 and resources generated from MeSH were not as general as other resources. Systematic inter-language differences in the structure of the terminology system rubrics make the rubrics harder to align. Manually created training resources give nearly as good results as a union of static resources, statistical resources and training resources and noticeably better results than a union of static resources and statistical resources. The verified English-Swedish dictionary contains 24,000 term pairs in base forms. CONCLUSION: More resources give better results in the automatic word alignment, but some resources only give small improvements. The most important type of resource is training and the most general resources were generated from ICD-10.

A data mining approach to analyze non-compliance with a guideline for the treatment of breast cancer.

Postmastectomy radiotherapy (PMRT) is prescribed in order to reduce the local recurrence of breast cancer and improve overall survival. A guideline supports the trade-off between benefits and adverse effects of PMRT. However, this guideline is not always followed in practice. This study tries to find a method for revealing patterns of non-compliance between the actual treatment and the PMRT guideline. Data from breast cancer patients admitted to Linköping University Hospital between 1990 and 2000 were analyzed in this study. Cases that were not treated in accordance with the guideline were selected and analyzed by decision tree induction (DTI). Thereafter, four resulting rules, as representations for groups of patients, were compared to the guideline. Finding patterns of non-compliance with guidelines by means of rules can be an appropriate alternative to manual methods, i.e. a case-by-case comparison when studying very large datasets. The resulting rules can be used in a knowledge base of a guideline-based decision support system to alert when inconsistencies with the guidelines may appear.

Graphical overview and navigation of electronic health records in a prototyping environment using Google Earth and openEHR archetypes.

This paper describes selected earlier approaches to graphically relating events to each other and to time; some new combinations are also suggested. These are then combined into a unified prototyping environment for visualization and navigation of electronic health records. Google Earth (GE) is used for handling display and interaction of clinical information stored using openEHR data structures and 'archetypes'. The strength of the approach comes from GE's sophisticated handling of detail levels, from coarse overviews to fine-grained details that has been combined with linear, polar and region-based views of clinical events related to time. The system should be easy to learn since all the visualization styles can use the same navigation. The structured and multifaceted approach to handling time that is possible with archetyped openEHR data lends itself well to visualizing and integration with openEHR components is provided in the environment.

Creating a medical English-Swedish dictionary using interactive word alignment.

BACKGROUND: This paper reports on a parallel collection of rubrics from the medical terminology systems ICD-10, ICF, MeSH, NCSP and KSH97-P and its use for semi-automatic creation of an English-Swedish dictionary of medical terminology. The methods presented are relevant for many other West European language pairs than English-Swedish. METHODS: The medical terminology systems were collected in electronic format in both English and Swedish and the rubrics were extracted in parallel language pairs. Initially, interactive word alignment was used to create training data from a sample. Then the training data were utilised in automatic word alignment in order to generate candidate term pairs. The last step was manual verification of the term pair candidates. RESULTS: A dictionary of 31,000 verified entries has been created in less than three man weeks, thus with considerably less time and effort needed compared to a manual approach, and without compromising quality. As a side effect of our work we found 40 different translation problems in the terminology systems and these results indicate the power of the method for finding inconsistencies in terminology translations. We also report on some factors that may contribute to making the process of dictionary creation with similar tools even more expedient. Finally, the contribution is discussed in relation to other ongoing efforts in constructing medical lexicons for non-English languages. CONCLUSION: In three man weeks we were able to produce a medical English-Swedish dictionary consisting of 31,000 entries and also found hidden translation errors in the utilized medical terminology systems.

Interactive visualization and navigation of complex terminology systems, exemplified by SNOMED CT.

Free-text queries are natural entries into the exploration of complex terminology systems. The way search results are presented has impact on the user's ability to grasp the overall structure of the system. Complex hierarchies like the one used in SNOMED CT, where nodes have multiple parents (IS-A) and several other relationship types, makes visualization challenging. This paper presents a prototype, TermViz, applying well known methods like "focus+context" and self-organizing layouts from the fields of Information Visualization and Graph Drawing to terminologies like SNOMED CT and ICD-10. The user can simultaneously focus on several nodes in the terminologies and then use interactive animated graph navigation and semantic zooming to further explore the terminology systems without loosing context. The prototype, based on Open Source Java components, demonstrates how a number of Information Visualisation methods can aid the exploration of medical terminologies with millions of elements and can serve as a base for further development.

An approach for generating fuzzy rules from decision trees.

Identifying high-risk breast cancer patients is vital both for clinicians and for patients. Some variables for identifying these patients such as tumor size are good candidates for fuzzification. In this study, Decision Tree Induction (DTI) has been applied to 3949 female breast cancer patients and crisp If-Then rules has been acquired from the resulting tree. After assigning membership functions for each variable in the crisp rules, they were converted into fuzzy rules and a mathematical model was constructed. One hundred randomly selected cases were examined by this model and compared with crisp rules predictions. The outcomes were examined by the area under the ROC curve (AUC). No significant difference was noticed between these two approaches for prediction of recurrence of breast cancer. By soft discretization of variables according to resulting rules from DTI, a predictive model, which is both more robust to noise and more comprehensible for clinicians, can be built.

Exploring cancer register data to find risk factors for recurrence of breast cancer--application of Canonical Correlation Analysis.

BACKGROUND: A common approach in exploring register data is to find relationships between outcomes and predictors by using multiple regression analysis (MRA). If there is more than one outcome variable, the analysis must then be repeated, and the results combined in some arbitrary fashion. In contrast, Canonical Correlation Analysis (CCA) has the ability to analyze multiple outcomes at the same time. One essential outcome after breast cancer treatment is recurrence of the disease. It is important to understand the relationship between different predictors and recurrence, including the time interval until recurrence. This study describes the application of CCA to find important predictors for two different outcomes for breast cancer patients, loco-regional recurrence and occurrence of distant metastasis and to decrease the number of variables in the sets of predictors and outcomes without decreasing the predictive strength of the model. METHODS: Data for 637 malignant breast cancer patients admitted in the south-east region of Sweden were analyzed. By using CCA and looking at the structure coefficients (loadings), relationships between tumor specifications and the two outcomes during different time intervals were analyzed and a correlation model was built. RESULTS: The analysis successfully detected known predictors for breast cancer recurrence during the first two years and distant metastasis 2-4 years after diagnosis. Nottingham Histologic Grading (NHG) was the most important predictor, while age of the patient at the time of diagnosis was not an important predictor. CONCLUSION: In cancer registers with high dimensionality, CCA can be used for identifying the importance of risk factors for breast cancer recurrence. This technique can result in a model ready for further processing by data mining methods through reducing the number of variables to important ones.

Computerisation, coding, data retrieval and related attitudes among Swedish general practitioners-a survey of necessary conditions for a database of diseases and health problems.

OBJECTIVE: To investigate necessary conditions for the establishment of a database of diseases and health problems for research and health care planning, based on electronic patient records in everyday clinical use among general practitioners (GPs). DESIGN: Postal questionnaire study. SETTING: Primary health care in Sweden. SUBJECTS: Three hundred randomly selected GPs. MAIN OUTCOME MEASURES: Degree of computerisation of patient records. User frequency and characteristics of diagnosis classification systems and coding tools. Frequency of coding activities and retrieval of codes, and related attitudes. Opinions on a primary health care version of ICD-10. RESULTS: A total of 184 GPs (61% of the 300 GPs) were included in the study. About 92% used an electronic record system, some type of diagnostic classification was used by 93%, and ICD based classifications by 88%. The classification in use was computerised for 74%. Mainly simple tools were used to retrieve diagnostic codes. About 76% of GPs reported classifying at least one symptom or disease per encounter. The codes were retrieved 'once a month' or more by 19%. Classification of diseases was considered important for follow-up by 83%, and for the care of the patient by 75% of the GPs. The primary health care version of ICD-10 with a total of 972 codes was considered too limited in size by 31%. CONCLUSION: Electronic patient records in everyday clinical use in Swedish general practice provide several fundamentals for a database of diagnostic data. However, there are several barriers to the establishment of such a database that is both valid and reliable.

Requirements and prototyping of a home health care application based on emerging JAVA technology.

IT support for home health care is an expanding area within health care IT development. Home health care differs from other in- or outpatient care delivery forms in a number of ways, and thus, the introduction of home health care applications must be based on a rigorous analysis of necessary requirements to secure safe and reliable health care. This article reports early experiences from the development of a home health care application based on emerging JAVA technologies. A prototype application for the follow-up of diabetes patients is presented and discussed in relation to a list of general requirements on home health care applications.

Using the MEDLINE database to study the concept of urinary tract infections in different domains of medicine.

As a way of exploring differences between medical domains regarding management of urinary tract infections, we investigated the MEDLINE database for differences in indexing patterns. Further, our intention was to assess the MEDLINE database as a source for studying medical domains. We examined the use of main headings, subheadings and the level of main headings in six medical domains that manage urinary tract infections. Many intuitive but also some counterintuitive results were found indicating that the MEDLINE database is difficult to use for studying medical domains mainly due to unclear semantics both in the headings and the indexing process, which results in variability in indexing. This variability probably hides significant results. We also conclude that the differences found indicate that in addition to differences between domains, there are also large variations within domains.

Enriching a primary health care version of ICD-10 using SNOMED CT mapping.

BACKGROUND: In order to satisfy different needs, medical terminology systems must have richer structures. This study examines whether a Swedish primary health care version of the mono-hierarchical ICD-10 (KSH97-P) may obtain a richer structure using category and chapter mappings from KSH97-P to SNOMED CT and SNOMED CT's structure. Manually-built mappings from KSH97-P's categories and chapters to SNOMED CT's concepts are used as a starting point. RESULTS: The mappings are manually evaluated using computer-produced information and a small number of mappings are updated. A new and poly-hierarchical chapter division of KSH97-P's categories has been created using the category and chapter mappings and SNOMED CT's generic structure. In the new chapter division, most categories are included in their original chapters. A considerable number of concepts are included in other chapters than their original chapters. Most of these inclusions can be explained by ICD-10's design. KSH97-P's categories are also extended with attributes using the category mappings and SNOMED CT's defining attribute relationships. About three-fourths of all concepts receive an attribute of type Finding site and about half of all concepts receive an attribute of type Associated morphology. Other types of attributes are less common. CONCLUSIONS: It is possible to use mappings from KSH97-P to SNOMED CT and SNOMED CT's structure to enrich KSH97-P's mono-hierarchical structure with a poly-hierarchical chapter division and attributes of type Finding site and Associated morphology. The final mappings are available as additional files for this paper.

Predicting metastasis in breast cancer: comparing a decision tree with domain experts.

Breast malignancy is the second most common cause of cancer death among women in Western countries. Identifying high-risk patients is vital in order to provide them with specialized treatment. In some situations, such as when access to experienced oncologists is not possible, decision support methods can be helpful in predicting the recurrence of cancer. Three thousand six hundred ninety-nine breast cancer patients admitted in south-east Sweden from 1986 to 1995 were studied. A decision tree was trained with all patients except for 100 cases and tested with those 100 cases. Two domain experts were asked for their opinions about the probability of recurrence of a certain outcome for these 100 patients. ROC curves, area under the ROC curves, and calibration for predictions were computed and compared. After comparing the predictions from a model built by data mining with predictions made by two domain experts, no significant differences were noted. In situations where experienced oncologists are not available, predictive models created with data mining techniques can be used to support physicians in decision making with acceptable accuracy.