Methods for identifying health state transitions from administrative data: the case of metastasis in prostate cancer
Introduction Health administrative data are a rich source of population-based information, useful for building state transition models for medical decision making. These models require identification of health state transitions and associated times. Indirect methods are needed to predict this inform...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Swansea University
2018-08-01
|
| Series: | International Journal of Population Data Science |
| Online Access: | https://ijpds.org/article/view/708 |
| _version_ | 1827615025354768384 |
|---|---|
| author | Nicholas Mitsakakis Karen Bremner Ruth Croxford Lusine Abrahamyan Welson Ryan Steven Carcone Murray Krahn |
| author_facet | Nicholas Mitsakakis Karen Bremner Ruth Croxford Lusine Abrahamyan Welson Ryan Steven Carcone Murray Krahn |
| author_sort | Nicholas Mitsakakis |
| collection | DOAJ |
| description | Introduction
Health administrative data are a rich source of population-based information, useful for building state transition models for medical decision making. These models require identification of health state transitions and associated times. Indirect methods are needed to predict this information, as it is rarely available in administrative data.
Objectives and Approach
We considered a set of criteria to identify transitions to metastasis for prostate cancer patients in administrative data, utilizing dates of diagnostic and medical billing codes for secondary malignancy, palliative radiation therapy, chemotherapy and bone disorders or procedures. We evaluated the criteria using the true date of metastasis from medical charts of 195 patients linked to health care administrative data in Ontario, Canada. We also built a recursive partitioning tree to optimally combine these criteria and construct rules for identifying metastatic patients. For the evaluation, both misclassification and discrepancy between true and predicted dates for the true positives were considered.
Results
Criteria involving chemotherapy drugs or hospital visits with secondary malignancy ICD10 diagnosis gave the best results, with high sensitivity and specificity. Criteria involving bone related problems, radiation therapy or diagnosis of metastatic cancer in physician billing data were very specific but not sensitive. The criterion involving prescriptions for narcotics was sensitive but not specific. The fitted tree was parsimonious involving only two of the criteria, while improving the accuracy over individual criteria. Most criteria gave a “delayed” prediction, with criterion based on chemotherapy giving on average the smallest delay, as well as exhibiting the least variability. Criteria involving narcotics and bone related problems predicted metastasis date very prematurely, probably triggered by conditions other than prostate cancer.
Conclusion/Implications
Several criteria from administrative databases satisfactorily classified prostate cancer patients with metastasis. A classification tree was built and improved the results over single criteria, demonstrating the added benefits in using advanced statistical learning methods for this task. However, “transition to metastasis” dates were predicted inaccurately, often with significant delay. |
| first_indexed | 2024-03-09T09:05:15Z |
| format | Article |
| id | doaj.art-90151cd35be7438dbabd339d148d1819 |
| institution | Directory Open Access Journal |
| issn | 2399-4908 |
| language | English |
| last_indexed | 2024-03-09T09:05:15Z |
| publishDate | 2018-08-01 |
| publisher | Swansea University |
| record_format | Article |
| series | International Journal of Population Data Science |
| spelling | doaj.art-90151cd35be7438dbabd339d148d18192023-12-02T10:41:45ZengSwansea UniversityInternational Journal of Population Data Science2399-49082018-08-013410.23889/ijpds.v3i4.708Methods for identifying health state transitions from administrative data: the case of metastasis in prostate cancerNicholas Mitsakakis0Karen Bremner1Ruth Croxford2Lusine Abrahamyan3Welson Ryan4Steven Carcone5Murray Krahn6University of TorontoUniversity Health NetworkInstitute of Clinical Evaluative SciencesUniversity of TorontoTHETA CollaborativeTHETA CollaborativeTHETA CollaborativeIntroduction Health administrative data are a rich source of population-based information, useful for building state transition models for medical decision making. These models require identification of health state transitions and associated times. Indirect methods are needed to predict this information, as it is rarely available in administrative data. Objectives and Approach We considered a set of criteria to identify transitions to metastasis for prostate cancer patients in administrative data, utilizing dates of diagnostic and medical billing codes for secondary malignancy, palliative radiation therapy, chemotherapy and bone disorders or procedures. We evaluated the criteria using the true date of metastasis from medical charts of 195 patients linked to health care administrative data in Ontario, Canada. We also built a recursive partitioning tree to optimally combine these criteria and construct rules for identifying metastatic patients. For the evaluation, both misclassification and discrepancy between true and predicted dates for the true positives were considered. Results Criteria involving chemotherapy drugs or hospital visits with secondary malignancy ICD10 diagnosis gave the best results, with high sensitivity and specificity. Criteria involving bone related problems, radiation therapy or diagnosis of metastatic cancer in physician billing data were very specific but not sensitive. The criterion involving prescriptions for narcotics was sensitive but not specific. The fitted tree was parsimonious involving only two of the criteria, while improving the accuracy over individual criteria. Most criteria gave a “delayed” prediction, with criterion based on chemotherapy giving on average the smallest delay, as well as exhibiting the least variability. Criteria involving narcotics and bone related problems predicted metastasis date very prematurely, probably triggered by conditions other than prostate cancer. Conclusion/Implications Several criteria from administrative databases satisfactorily classified prostate cancer patients with metastasis. A classification tree was built and improved the results over single criteria, demonstrating the added benefits in using advanced statistical learning methods for this task. However, “transition to metastasis” dates were predicted inaccurately, often with significant delay.https://ijpds.org/article/view/708 |
| spellingShingle | Nicholas Mitsakakis Karen Bremner Ruth Croxford Lusine Abrahamyan Welson Ryan Steven Carcone Murray Krahn Methods for identifying health state transitions from administrative data: the case of metastasis in prostate cancer International Journal of Population Data Science |
| title | Methods for identifying health state transitions from administrative data: the case of metastasis in prostate cancer |
| title_full | Methods for identifying health state transitions from administrative data: the case of metastasis in prostate cancer |
| title_fullStr | Methods for identifying health state transitions from administrative data: the case of metastasis in prostate cancer |
| title_full_unstemmed | Methods for identifying health state transitions from administrative data: the case of metastasis in prostate cancer |
| title_short | Methods for identifying health state transitions from administrative data: the case of metastasis in prostate cancer |
| title_sort | methods for identifying health state transitions from administrative data the case of metastasis in prostate cancer |
| url | https://ijpds.org/article/view/708 |
| work_keys_str_mv | AT nicholasmitsakakis methodsforidentifyinghealthstatetransitionsfromadministrativedatathecaseofmetastasisinprostatecancer AT karenbremner methodsforidentifyinghealthstatetransitionsfromadministrativedatathecaseofmetastasisinprostatecancer AT ruthcroxford methodsforidentifyinghealthstatetransitionsfromadministrativedatathecaseofmetastasisinprostatecancer AT lusineabrahamyan methodsforidentifyinghealthstatetransitionsfromadministrativedatathecaseofmetastasisinprostatecancer AT welsonryan methodsforidentifyinghealthstatetransitionsfromadministrativedatathecaseofmetastasisinprostatecancer AT stevencarcone methodsforidentifyinghealthstatetransitionsfromadministrativedatathecaseofmetastasisinprostatecancer AT murraykrahn methodsforidentifyinghealthstatetransitionsfromadministrativedatathecaseofmetastasisinprostatecancer |