Gunnar Nilsson1, Hans Ahlfeldt, Lars-Erik Strender. 1. Department of Clinical Science, Family Medicine Stockholm, Karolinska Institutet, Stockholm, Sweden. gunnar.nilsson@nlpo.sll.se
Abstract
OBJECTIVE: To investigate textual content, health problems and diagnostic codes in everyday electronic patient records. DESIGN: Retrospective and observational database study. SETTING: Primary health care in Stockholm. SUBJECTS: Twenty randomly selected general practitioners with 20 records each. MAIN OUTCOME MEASURES: The frequency of use of problem-oriented medical records. The number of words, problems and diagnostic codes. The completeness and correctness of the diagnostic codes. RESULTS: About 14.5% of 400 studied records were problem-oriented. The mean number of words per record was 99.4, and the mean number of problems managed per record was 1.2. On average, there were 1.1 diagnostic codes per record and this differed widely among GPs and also among the electronic patient record systems. The mean number of codes per problem was 0.9, and the proportion of correct codes was 97.4%. CONCLUSIONS: The electronic patient records in general practice in Stockholm have an extensive textual content. A vast majority of the problems are coded and the completeness and correctness of diagnostic codes are high. It seems that problem-oriented electronic patient record systems enforce coding activities. It is feasible to establish a database of diagnostic data for research and health care planning based on electronic patient records.
OBJECTIVE: To investigate textual content, health problems and diagnostic codes in everyday electronic patient records. DESIGN: Retrospective and observational database study. SETTING: Primary health care in Stockholm. SUBJECTS: Twenty randomly selected general practitioners with 20 records each. MAIN OUTCOME MEASURES: The frequency of use of problem-oriented medical records. The number of words, problems and diagnostic codes. The completeness and correctness of the diagnostic codes. RESULTS: About 14.5% of 400 studied records were problem-oriented. The mean number of words per record was 99.4, and the mean number of problems managed per record was 1.2. On average, there were 1.1 diagnostic codes per record and this differed widely among GPs and also among the electronic patient record systems. The mean number of codes per problem was 0.9, and the proportion of correct codes was 97.4%. CONCLUSIONS: The electronic patient records in general practice in Stockholm have an extensive textual content. A vast majority of the problems are coded and the completeness and correctness of diagnostic codes are high. It seems that problem-oriented electronic patient record systems enforce coding activities. It is feasible to establish a database of diagnostic data for research and health care planning based on electronic patient records.
Authors: Anna Vikström; Maria Hägglund; Mikael Nyström; Lars-Erik Strender; Sabine Koch; Per Hjerpe; Ulf Lindblad; Gunnar H Nilsson Journal: BMC Fam Pract Date: 2012-01-09 Impact factor: 2.497
Authors: Lennart Carlsson; Roland Morgell; Lars-Erik Strender; Britt Arrelöv; Gunnar H Nilsson Journal: BMC Public Health Date: 2009-05-27 Impact factor: 3.295
Authors: Sven G Engström; Lennart Carlsson; Carl-Johan Ostgren; Gunnar H Nilsson; Lars A Borgquist Journal: BMC Public Health Date: 2006-02-16 Impact factor: 3.295
Authors: Axel C Carlsson; Per Wändell; Urban Ösby; Ramin Zarrinkoub; Björn Wettermark; Gunnar Ljunggren Journal: BMC Public Health Date: 2013-07-18 Impact factor: 3.295