An exploratory data quality analysis of time series physiologic signals using a large-scale intensive care unit database

Abstract

Physiological data, such as heart rate and blood pressure, are critical to clinical decision-making in the intensive care unit (ICU). Vital signs data, which are available from electronic health records, can be used to diagnose and predict important clinical outcomes; While there have been some reports on the data quality of nurse-verified vital sign data, little has been reported on the data quality of higher frequency time-series vital signs acquired in ICUs, that would enable such predictive modeling. In this study, we assessed the data quality issues, defined as the completeness, accuracy, and timeliness, of minute-by-minute time series vital signs data within the MIMIC-III data set, captured from 16009 patient-ICU stays and corresponding to 9410 unique adult patients. We measured data quality of four time-series vital signs data streams in the MIMIC-III data set: heart rate (HR), respiratory rate (RR), blood oxygen saturation (SpO2), and arterial blood pressure (ABP). Approximately, 30% of patient-ICU stays did not have at least 1 min of data during the time-frame of the ICU stay for HR, RR, and SpO2. The percentage of patient-ICU stays that did not have at least 1 min of ABP data was ∼56%. We observed ∼80% coverage of the total duration of the ICU stay for HR, RR, and SpO2. Finally, only 12.5%%, 9.9%, 7.5%, and 4.4% of ICU lengths of stay had ≥ 99% data available for HR, RR, SpO2, and ABP, respectively, that would meet the three data quality requirements we looked into in this study. Our findings on data completeness, accuracy, and timeliness have important implications for data scientists and informatics researchers who use time series vital signs data to develop predictive models of ICU outcomes.

Digital Health Data Quality Issues: Systematic Review

“…Studies often referred to accuracy as the “ correctness” of data, which is the degree to which data correctly communicate the parameter being represented [ 32 ]. By contrast, other studies focused on plausibility , which is the extent to which data points are believable [ 35 ].…”

“…The currency dimension (18/227, 7.9%) is composed of a single subtheme: timeliness . Currency, or timeliness, is defined by Afshar et al [ 32 ] and Makeleni and Cilliers [ 31 ] as the degree to which data represent reality from the required point in time. From an EHR perspective, data should be up to date, available, and reflect the profile of the patient at the time when the data are accessed [ 32 , 50 ].…”

“…Currency, or timeliness, is defined by Afshar et al [ 32 ] and Makeleni and Cilliers [ 31 ] as the degree to which data represent reality from the required point in time. From an EHR perspective, data should be up to date, available, and reflect the profile of the patient at the time when the data are accessed [ 32 , 50 ]. Lee et al [ 35 ] extended this to include the recording of an event at the time when it occurs such that a value is deemed current if it is representative of the clinically relevant time of the event.…”

“…Validity, correctness, integrity, conformance, plausibility, veracity, and accurate diagnostic data "The degree to which data reveal the truth about the event being described" [31] Accuracy Inconsistent data capturing, standardization, concordance, uniqueness, data variability, temporal variability, system differences, semantic consistency, structuredness, and representational consistency "Absence of differences between data items representing the same objects based on specific information requirements. Consistent data contain the same data values when compared between different databases" [31] Consistency Missing data, level of completeness, representativeness, fragmentation, and breadth of documentation "The absence of data at a single moment over time or when measured at multiple moments over time" [79] Completeness Contextual DQ, fitness for use, granularity, and relevancy Assessment of DQ is "dependent on the task at hand" [18] Contextual validity Accessible DQ and availability How "feasible it is for users to extract the data of interest" [18] Accessibility Timeliness "The degree to which data represent reality from the required point in time" [32] Currency…”

Development of an Open-Source Annotated Glaucoma Medication Dataset From Clinical Notes in the Electronic Health Record

“…Several of these aforementioned datasets have been most notably used for developed de-identification algorithms, particularly for PHI. 27 , 33 , 42 – 44 Additional use cases for these datasets include big data retrospective analyses for risk factors, 45 , 46 analysis of documentation similarity, 47 and text extraction for fungal endophthalmitis, 48 among others. Our dataset complements existing de-identified datasets and offers a new avenue of exploration with annotated ophthalmic medication data.…”