Predicting the risk of falls in patients with an indication for antihypertensive treatment: development and validation of the stratify-falls prediction model
OBJECTIVE: Falls are one of the most common side effects associated with antihypertensive medication. In patients at high risk of falls, the additional risk of harm from medication may outweigh the potential benefits in terms of cardiovascular risk reduction. However, it is currently unclear which p...
| Prif Awduron: | , , , , , , , |
|---|---|
| Fformat: | Journal article |
| Iaith: | English |
| Cyhoeddwyd: |
Lippincott, Williams & Wilkins
2022
|
| _version_ | 1826310083199369216 |
|---|---|
| author | Archer, L Koshiaris, C Snell Ie, K Riley, RD Stevens, R Hobbs, FR McManus, RJ Sheppard, JP |
| author_facet | Archer, L Koshiaris, C Snell Ie, K Riley, RD Stevens, R Hobbs, FR McManus, RJ Sheppard, JP |
| author_sort | Archer, L |
| collection | OXFORD |
| description | OBJECTIVE: Falls are one of the most common side effects associated with antihypertensive medication. In patients at high risk of falls, the additional risk of harm from medication may outweigh the potential benefits in terms of cardiovascular risk reduction. However, it is currently unclear which patients are at high risk of falls. This study aimed to develop and validate a clinical prediction model for risk of hospitalisation or death from falls in adults eligible for antihypertensive treatment. DESIGN AND METHOD: Eligible patients were aged 40 and above, with at least one blood pressure measurement between 130-179 mm Hg. The outcome was a fall resulting in hospitalisation or death within 10 years of entering the study. Model development was conducted in data from CPRD GOLD using a Fine-Gray approach, accounting for the competing risk of death from other causes, with subsequent recalibration at specific time-points using pseudo-values. External validation was conducted using data from CPRD Aurum, with performance assessed through calibration curves, the Observed/Expected (O/E) ratio, C-statistic, and D-statistic, each pooled across GP practices. RESULTS: Analysis included 1,773,224 patients (62,691 events) for model development, and 3,805,322 (206,956 events) for external validation. Conditional on other variables, increasing age, being female, of white ethnicity, being a heavy drinker and living in an area of high social deprivation were associated with an increased risk of falls. Upon external validation, the re-calibrated model showed good discrimination, with pooled C-statistics of 0.843 (95%CI: 0.841 to 0.844) and 0.833 (95%CI: 0.831 to 0.835) at 5 and 10 years respectively. Original model calibration was poor on visual inspection and whilst this improved with re-calibration, under-prediction of risk remained (Observed/Expected ratio 1.839 at 10 years, 95%CI: 1.811 to 1.865) (Figure 1). CONCLUSIONS: STRATIFY-Falls uses commonly recorded clinical characteristics and shows good discrimination on external validation, accurately distinguishing between patients at high and low risk of falls in the next 5-10 years. This model may be used in routine clinical practice to help identify those at high risk of falls, for whom the benefits of antihypertensive medication may be outweighed by the harms. |
| first_indexed | 2024-03-07T07:45:27Z |
| format | Journal article |
| id | oxford-uuid:69c6e6d0-48d4-4b37-961a-7ab36766b362 |
| institution | University of Oxford |
| language | English |
| last_indexed | 2024-03-07T07:45:27Z |
| publishDate | 2022 |
| publisher | Lippincott, Williams & Wilkins |
| record_format | dspace |
| spelling | oxford-uuid:69c6e6d0-48d4-4b37-961a-7ab36766b3622023-06-05T10:48:05ZPredicting the risk of falls in patients with an indication for antihypertensive treatment: development and validation of the stratify-falls prediction modelJournal articlehttp://purl.org/coar/resource_type/c_dcae04bcuuid:69c6e6d0-48d4-4b37-961a-7ab36766b362EnglishSymplectic ElementsLippincott, Williams & Wilkins 2022Archer, LKoshiaris, CSnell Ie, KRiley, RDStevens, RHobbs, FRMcManus, RJSheppard, JPOBJECTIVE: Falls are one of the most common side effects associated with antihypertensive medication. In patients at high risk of falls, the additional risk of harm from medication may outweigh the potential benefits in terms of cardiovascular risk reduction. However, it is currently unclear which patients are at high risk of falls. This study aimed to develop and validate a clinical prediction model for risk of hospitalisation or death from falls in adults eligible for antihypertensive treatment. DESIGN AND METHOD: Eligible patients were aged 40 and above, with at least one blood pressure measurement between 130-179 mm Hg. The outcome was a fall resulting in hospitalisation or death within 10 years of entering the study. Model development was conducted in data from CPRD GOLD using a Fine-Gray approach, accounting for the competing risk of death from other causes, with subsequent recalibration at specific time-points using pseudo-values. External validation was conducted using data from CPRD Aurum, with performance assessed through calibration curves, the Observed/Expected (O/E) ratio, C-statistic, and D-statistic, each pooled across GP practices. RESULTS: Analysis included 1,773,224 patients (62,691 events) for model development, and 3,805,322 (206,956 events) for external validation. Conditional on other variables, increasing age, being female, of white ethnicity, being a heavy drinker and living in an area of high social deprivation were associated with an increased risk of falls. Upon external validation, the re-calibrated model showed good discrimination, with pooled C-statistics of 0.843 (95%CI: 0.841 to 0.844) and 0.833 (95%CI: 0.831 to 0.835) at 5 and 10 years respectively. Original model calibration was poor on visual inspection and whilst this improved with re-calibration, under-prediction of risk remained (Observed/Expected ratio 1.839 at 10 years, 95%CI: 1.811 to 1.865) (Figure 1). CONCLUSIONS: STRATIFY-Falls uses commonly recorded clinical characteristics and shows good discrimination on external validation, accurately distinguishing between patients at high and low risk of falls in the next 5-10 years. This model may be used in routine clinical practice to help identify those at high risk of falls, for whom the benefits of antihypertensive medication may be outweighed by the harms. |
| spellingShingle | Archer, L Koshiaris, C Snell Ie, K Riley, RD Stevens, R Hobbs, FR McManus, RJ Sheppard, JP Predicting the risk of falls in patients with an indication for antihypertensive treatment: development and validation of the stratify-falls prediction model |
| title | Predicting the risk of falls in patients with an indication for antihypertensive treatment: development and validation of the stratify-falls prediction model |
| title_full | Predicting the risk of falls in patients with an indication for antihypertensive treatment: development and validation of the stratify-falls prediction model |
| title_fullStr | Predicting the risk of falls in patients with an indication for antihypertensive treatment: development and validation of the stratify-falls prediction model |
| title_full_unstemmed | Predicting the risk of falls in patients with an indication for antihypertensive treatment: development and validation of the stratify-falls prediction model |
| title_short | Predicting the risk of falls in patients with an indication for antihypertensive treatment: development and validation of the stratify-falls prediction model |
| title_sort | predicting the risk of falls in patients with an indication for antihypertensive treatment development and validation of the stratify falls prediction model |
| work_keys_str_mv | AT archerl predictingtheriskoffallsinpatientswithanindicationforantihypertensivetreatmentdevelopmentandvalidationofthestratifyfallspredictionmodel AT koshiarisc predictingtheriskoffallsinpatientswithanindicationforantihypertensivetreatmentdevelopmentandvalidationofthestratifyfallspredictionmodel AT snelliek predictingtheriskoffallsinpatientswithanindicationforantihypertensivetreatmentdevelopmentandvalidationofthestratifyfallspredictionmodel AT rileyrd predictingtheriskoffallsinpatientswithanindicationforantihypertensivetreatmentdevelopmentandvalidationofthestratifyfallspredictionmodel AT stevensr predictingtheriskoffallsinpatientswithanindicationforantihypertensivetreatmentdevelopmentandvalidationofthestratifyfallspredictionmodel AT hobbsfr predictingtheriskoffallsinpatientswithanindicationforantihypertensivetreatmentdevelopmentandvalidationofthestratifyfallspredictionmodel AT mcmanusrj predictingtheriskoffallsinpatientswithanindicationforantihypertensivetreatmentdevelopmentandvalidationofthestratifyfallspredictionmodel AT sheppardjp predictingtheriskoffallsinpatientswithanindicationforantihypertensivetreatmentdevelopmentandvalidationofthestratifyfallspredictionmodel |