Bose, Eliezer
(2015)
TIME SERIES ANALYSIS AND CLUSTERING TO CHARACTERIZE CARDIORESPIRATORY INSTABILITY PATTERNS IN STEP-DOWN UNIT PATIENTS.
Doctoral Dissertation, University of Pittsburgh.
(Unpublished)
Abstract
Background: Cardiorespiratory instability (CRI) in noninvasively monitored step-down unit (SDU) patients has a variety of etiologies, and therefore likely manifests in different patterns of vital signs (VS) changes. Objective: We sought to describe differences in admission characteristics and outcomes between patients with and without CRI. We explored use of clustering techniques to identify VS patterns within initial CRI epoch (CRI1) and assessed inter-cluster differences in admission characteristics, outcomes and medications.
Methods: Admission characteristics and continuous monitoring data (frequency 1/20 Hz) were recorded in 307 patients. Vital sign (VS) deviations beyond local instability trigger criteria for 3 consecutive minutes or for 4 out of a 5 minute moving window were classified as CRI events. We identified CRI1 in 133 patients, derived statistical features of CRI1 epoch and employed hierarchical and k-means clustering techniques. We tested several clustering solutions and used 10-fold cross validation and ANOVA to establish best solution. Inter-cluster differences in admission characteristics, outcomes and medications were assessed.
Main Results: Patients transferred to the SDU from units with higher monitoring capability were more likely to develop CRI (n=133, CRI 44% vs no CRI n=174, 31%, p=.042). Patients with at least one event of CRI had longer hospital length of stay (CRI 11.3 + 10.2 days vs no CRI 7.8 + 9.2, p=.001) and SDU unit stay (CRI 6.1 + 4.9 days vs no CRI 3.5 + 2.9, p< .001). Four main clusters(C) were derived. Clusters were significantly different based on age (p=0.001; younger patients in C1 and older in C2), number of comorbidities (p<0.01; more C2 patients had ≥2), and admission source (p=0.008; more C1 and C4 patients transferred in from a higher intensity monitoring unit). Patients with CRI differed significantly (p<.05) from those without CRI based on medication categories.
Conclusions: CRI1 was associated with prolonged hospital and SDU length of stay. Patients transferred from a higher level of care were more likely to develop CRI, suggesting that they are sicker. Future study will be needed to determine if there are common physiologic underpinnings of VS clusters which might inform monitoring practices and clinical decision-making when CRI first manifests.
Share
| Citation/Export: |
|
| Social Networking: |
|
Details
| Item Type: |
University of Pittsburgh ETD
|
| Status: |
Unpublished |
| Creators/Authors: |
|
| ETD Committee: |
|
| Date: |
27 July 2015 |
| Date Type: |
Publication |
| Defense Date: |
21 July 2015 |
| Approval Date: |
27 July 2015 |
| Submission Date: |
27 July 2015 |
| Access Restriction: |
2 year -- Restrict access to University of Pittsburgh for a period of 2 years. |
| Number of Pages: |
280 |
| Institution: |
University of Pittsburgh |
| Schools and Programs: |
School of Nursing > Nursing |
| Degree: |
PhD - Doctor of Philosophy |
| Thesis Type: |
Doctoral Dissertation |
| Refereed: |
Yes |
| Uncontrolled Keywords: |
time-series analysis, clustering, cardiorespiratory instability, step-down unit |
| Date Deposited: |
27 Jul 2015 21:40 |
| Last Modified: |
27 Jul 2017 05:15 |
| URI: |
http://d-scholarship.pitt.edu/id/eprint/25818 |
Metrics
Monthly Views for the past 3 years
Plum Analytics
Actions (login required)
 |
View Item |
![[img]](http://d-scholarship.pitt.edu/style/images/fileicons/application_pdf.png)
![[feed]](/images/feed-icon-32x32.png){kind=icon}