The value of electronic patient records for determining the prevalence and prognosis of cognitive and physical frailty in a large hospital-based cohort
Bibliographic record
Abstract
Older people with frailty account for an increasing proportion of hospital admissions but access to specialised geriatric care varies. Data on the burden of cognitive and physical frailty hospital-wide and by specialty are currently limited but such data are necessary for informing clinical guidance, service-planning and policy. Routine clinical data captured in hospital electronic patient records (EPRs) offers an alternative to existing prospective and administrative coding data study designs but the value of these data for measuring frailty has not yet been studied. Therefore, my thesis aimed to evaluate ‘the value of hospital EPRs for determining the prevalence and prognosis of cognitive and physical frailty’ primarily using data from a large hospital-based cohort, i.e., the Oxford Cognitive Comorbidity, Ageing and Frailty Research Database-Electronic Patient Records (ORCHARD-EPR) study. First, in a systematic review/meta-analysis of 45 hospital-wide and general medicine cohorts, I showed that global frailty, measured using validated tools, was prevalent in older people with unplanned hospital admissions although heterogeneity was high and not explained by frailty tool, setting, or risk of bias. Despite variation in prevalence, frailty was consistently associated with mortality, length of stay (LoS) and discharge destination including after adjustment for confounders, although findings on readmissions were mixed. Notably, cognitive impairment, and in particular delirium, was poorly ascertained by all the frailty measures used in included studies and therefore prevalence of cognitive frailty and the degree of overlap with physical frailty was uncertain. In addition, all studies were either small prospective studies or large administrative datasets based on ICD-10 diagnostic coding except for three which used EPR data but were not validated. Second, I assembled, cleaned and validated data from the Oxford and Reading Cognitive Comorbidity, Frailty and Ageing Research Database-Electronic Patient Record (ORCHARD-EPR) 2017-2019 dataset which contains pseudo-anonymised Oxfordshire University Hospital NHS Foundation Trust EPR data for >=100,000 unplanned hospital admissions. Importantly, ORCHARD-EPR includes the results of mandatory cognitive screening in those >=70 years (existing dementia diagnosis, delirium diagnosis informed by the Confusion Assessment Methods-CAM and recorded as “certain” or “uncertain”, Abbreviated Mental Test-AMT). I validated cognitive frailty data from the cognitive screen for general medicine admissions in ORCHARD-EPR against a reference cohort. The prevalence of cognitive frailty (certain delirium, dementia, AMT<8) in general medicine admissions was 35% in ORCHARD-EPR (increasing to 41% when uncertain delirium diagnosis was included) compared to 50% in the reference data with the difference in prevalence largely explained by lower rates of delirium diagnosis in ORCHARD-EPR. Third, I determined that cognitive frailty (certain+uncertain delirium, dementia, AMT<8) was present in 35% of all ORCHARD-EPR admissions (n=51,202) across 29 specialties, with delirium the most common diagnosis. I also showed that any cognitive frailty predicted survival over up to four years of follow-up as well as LoS, delayed discharge, and discharge destination over and above age, sex, comorbidity and illness severity, with associations strongest for delirium. Additionally, only delirium in non-care home residents predicted readmission. Fourth, using a modified version of the Hospital Frailty Risk Score to measure physical frailty from ICD-10 codes, I found that moderate/severe physical frailty was present in 73% of individuals with cognitive frailty, but only 46% with physical frailty were cognitively frail. Despite being independently associated with survival, physical frailty added little to mortality risk in those with cognitive frailty but added markedly to LoS and risks of delayed discharge and discharge to a destination other than home. In conclusion, routinely acquired clinical frailty data is a valuable tool for research, capable of combining large, inclusive, hospital-wide sampling with accurate ascertainment as shown using ORCHARD-EPR. Similar approaches could be adopted at other centres but will likely only be useful if routine screening is embedded into the EPR and shown to reliably identify frail patients. Moreover, findings of high frailty prevalence across multiple specialties and impact on outcomes supports more widespread implementation of routine cognitive and physical frailty screening in older people with unplanned hospital admission in line with current guidance. Additionally, findings support increased emphasis on delirium in policy and research.
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 | 0.039 | 0.119 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.002 | 0.004 |
| Bibliometrics | 0.004 | 0.005 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.003 | 0.002 |
| Open science | 0.001 | 0.002 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 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".