Information quality: a conceptual framework and empirical validation

Quality of information is a critical issue in today's interconnected society. Yet information is still poorly defined and theoretical problems with current information quality models exist, such as circularities between levels, external attributes or factors, and ambiguous or undefined terms. Few such models have been empirically validated. This study developed an information quality model that addressed conceptual flaws in existing models and empirically validated the resulting model in the health care claims processing context. Two parallel data sets, regarding individuals' perceptions of the quality of claims information submitted and received from claims insurers, were gathered simultaneously from 198 United States health care professionals responsible for processing claims information. A structural equation model based on the information quality model was developed and the model fit tested with each data set by partial least squares. The model significantly predicted 37% of the variance in perceived overall Information Quality in the supplier context, and 59% of the variance in perceived overall Information Quality in the consumer context. For each context, perceived Interpretability and Integrity of information were shown to be significant predictors of perceived Information Quality. In each context, perceived Accuracy, Completeness, Consistency and Existence of information were also shown to be significant determinants of perceived Integrity of information. Two attributes of information well-accepted as determinants of information quality—Accessibility and Relevance—did not significantly contribute to its prediction in either the supplier or consumer context, though their contributions in the consumer context approached significance. Perceived Currency of information was nonetheless shown to be a significant predictor of perceived information Relevance. Results supported the model developed and treatment of Information Quality as a formative construct. They also suggested that: (1) importance ratings do not adequately identify attributes of information that explain perceptions of information quality; (2) conventional theory and wisdom about the core information attributes that contribute to information quality perceptions may not apply to certain contexts, such as health care claims processing; and, (3) evaluations and implementations of information quality in these or other domains that rely on assessed perceptions of Accessibility and Relevance may be misguided.