Can administrative healthcare data be used to predict post-discharge emergency room visits in seniors with colon cancer?
Bibliographic record
Abstract
Background: The comprehensive geriatric assessment (CGA) is a multidimensional in-depth evaluation that can be used to assess and estimate life expectancy, risk of morbidity and the physiological age of older cancer patients. Conducting a CGA, however, is resource-intensive. The CGA is also not specifically targeted towards assessing cancer patients and fails to take into account the impact of past medical events. Objectives: We sought to determine if age-specific risk factors comprising the CGA as well as patterns of healthcare use could be assessed in patients 65 years and older undergoing colon cancer surgery between 2000-2006 using administrative healthcare data. We also aimed to determine whether any associations exist between these risk factors and the occurrence of post-discharge emergency room visits (PERVs). Methods: We first conducted a systematic review (PROSPERO registration number CRD42012002476) using MEDLINE, CINAHL, EMBASE and CANCERLIT databases. to identify CGA domains most predictive of adverse cancer-related outcomes, including treatment-related toxicity, mortality and postoperative complications. Studies published in English or French between May 1997 and May 2012, in which a CGA was conducted in patients over the age of 65 initiating cancer treatment, were assessed for eligibility, of which 9 studies were selected for this review. We subsequently conducted a historical cohort study using administrative healthcare data. This involved using administrative claims provided by Quebec's healthcare insurance program (RAMQ) and hospitalization data to identify patients 65 years and older receiving colon cancer surgery between January 1, 2000 and December 31, 2006. Using a one-year look-back period, ICD-9 and generic drug codes were used to characterize patients' history of relevant comorbidities. Service claims, hospitalization and prescription data were also used to characterize past patterns of healthcare use. Following bivariate analyses, a multivariate logistic regression was used to quantify risk factors for ER visits occurring within 30 days of discharge. Results: Our systematic review indicated that, in predicting mortality, in at least one study or another, all CGA domains were found to be significant. Most frequently, the following domains were reported for predicting mortality: nutritional status, the presence of geriatric syndromes such as depression, and functional status. With regards to chemotherapy-related toxicity, similar findings were obtained where functional status and the presence of geriatric syndromes, such as impaired hearing, had the most significant predictive value. Only one study reported on the incidence of post-operative complications for which severe comorbidity was found to be highly associated with experiencing severe complications, while functional status was found to be significantly associated with experiencing any complication. 3789 patients were included in our historical cohort, of which 17.18% made a PERV. The results of the multivariate logistic regression indicated that certain CGA domains were predictive of PERVs. Specifically, individuals that had recently received care for either diabetes or cardiovascular disease had a greater odds of experiencing a PERV. In addition, individuals with increased medication use, as measured by the number of unique medications dispensed within 6 months preceding surgery were more likely to experience a PERV. A number of patterns of past healthcare use also demonstrated predictive utility for the PERV outcome, including a history of visiting the ER, and whether a patient had visited the ER within 30 days preceding surgery for colon cancer-related symptoms. Conclusions: Certain age-specific risk factors and past patterns of healthcare use may predict PERVs. This has important implications in the development of age-sensitive electronic risk-profiling tools.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.026 | 0.169 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.004 | 0.007 |
| Bibliometrics | 0.015 | 0.018 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.003 | 0.003 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.002 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".