Defining Key Deprescribing Measures from Electronic Health Data: A Multisite Data Harmonization Project
Abstract Background Deprescribing, or systematically stopping or reducing risky or unneeded medications, could improve older adults’ health. Electronic health records (EHR) hold promise for supporting deprescribing studies, but there are currently no standardized measures for key variables. With benzodiazepines and other sedative-hypnotics (Z-drugs) as a case study, we developed and examined EHR-based definitions for chronic medication use and discontinuation.Methods We conducted a retrospective cohort study set within 5 U.S. healthcare systems. The study population was adults age 65+ from 2017-2019 with chronic benzodiazepine or Z-drug use, without dementia or serious mental illness. We used EHR data for medication orders and dispensings to define key variables, including chronic benzodiazepine/Z-drug use and discontinuation. We explored definitions for discontinuation based on 1) gaps in medication availability during follow-up (no active order/dispensing) or 2) not having medication available at a fixed time point. We examined the impact of varying gap length from 30 to 180 days, accounting for stockpiling, and requiring a 30-day period without orders/dispensings (“halo”) around the fixed time point. We also compared results from medication orders vs. dispensings for the same population.Results 1.6-2.6% of older adults had chronic use of a benzodiazepine or Z-drug. Depending on the definition and site, the proportion discontinuing use over 12 months ranged from 6% to 49%. Requiring a longer gap in orders/dispensings or a 30-day “halo” around a fixed time point resulted in lower estimates. Orders data were less likely to identify discontinuation than dispensing data.Conclusions Requiring a medication gap of ≥90 days or a 30-day period with no orders/dispensings around a fixed time point may improve the likelihood that an outcome represents true discontinuation. Orders data appear to underestimate discontinuation compared to dispensing data. More work is needed to adapt and test the proposed definitions for other drug classes and care settings.Impact Statement We certify that this work is novel. Prior papers have identified a need for greater standardization of definitions for medication exposure and discontinuation in deprescribing studies. To our knowledge, no prior paper has systematically examined the construction of deprescribing variables from electronic health data. This is the first paper to present standardized definitions for variables needed for deprescribing studies based on electronic health data, to implement these definitions in multiple healthcare systems and data types, and to examine their performance. This is also the first paper to examine the impact of using medication orders vs. dispensing data to define key deprescribing variables.Key points <jats:list list-type="bullet">Using benzodiazepines and Z-drugs as a case study, we developed and implemented standardized definitions for key variables needed for deprescribing studies using electronic health records data from 5 diverse U.S. healthcare systems.Requiring a gap of ≥90 days without an active order/dispensing or, for a fixed time point, requiring a period of ≥30 days surrounding it with no order/dispensing will likely increase the accuracy of identifying true medication discontinuation.Applying definitions to medication orders data generated higher estimates of chronic use and lower estimates of medication discontinuation than dispensing data.Why does this paper matter? The use of standardized variable definitions in deprescribing studies will improve the ability to synthesize data and compare results between studies, advancing knowledge and supporting more evidence-based guidelines for clinical care..
| Medienart: |
|
|---|
Preprint| Erscheinungsjahr: |
2023 |
|---|---|
| Erschienen: |
2023 |
| Enthalten in: |
bioRxiv.org - (2023) vom: 10. Nov. Zur Gesamtaufnahme - year:2023 |
|---|
| Sprache: |
Englisch |
|---|
| Beteiligte Personen: |
Dublin, Sascha [VerfasserIn] |
|---|
| Links: |
Volltext [kostenfrei] |
|---|
| Themen: |
|---|
| doi: |
10.1101/2023.11.06.23298060 |
|---|
| funding: |
|
|---|---|
| Förderinstitution / Projekttitel: |
|
| PPN (Katalog-ID): |
XBI041462963 |
|---|
| LEADER | 01000caa a22002652 4500 | ||
|---|---|---|---|
| 001 | XBI041462963 | ||
| 003 | DE-627 | ||
| 005 | 20231205143742.0 | ||
| 007 | cr uuu---uuuuu | ||
| 008 | 231108s2023 xx |||||o 00| ||eng c | ||
| 024 | 7 | |a 10.1101/2023.11.06.23298060 |2 doi | |
| 035 | |a (DE-627)XBI041462963 | ||
| 035 | |a (biorXiv)10.1101/2023.11.06.23298060 | ||
| 040 | |a DE-627 |b ger |c DE-627 |e rakwb | ||
| 041 | |a eng | ||
| 100 | 1 | |a Dublin, Sascha |e verfasserin |4 aut | |
| 245 | 1 | 0 | |a Defining Key Deprescribing Measures from Electronic Health Data: A Multisite Data Harmonization Project |
| 264 | 1 | |c 2023 | |
| 336 | |a Text |b txt |2 rdacontent | ||
| 337 | |a Computermedien |b c |2 rdamedia | ||
| 338 | |a Online-Ressource |b cr |2 rdacarrier | ||
| 520 | |a Abstract Background Deprescribing, or systematically stopping or reducing risky or unneeded medications, could improve older adults’ health. Electronic health records (EHR) hold promise for supporting deprescribing studies, but there are currently no standardized measures for key variables. With benzodiazepines and other sedative-hypnotics (Z-drugs) as a case study, we developed and examined EHR-based definitions for chronic medication use and discontinuation.Methods We conducted a retrospective cohort study set within 5 U.S. healthcare systems. The study population was adults age 65+ from 2017-2019 with chronic benzodiazepine or Z-drug use, without dementia or serious mental illness. We used EHR data for medication orders and dispensings to define key variables, including chronic benzodiazepine/Z-drug use and discontinuation. We explored definitions for discontinuation based on 1) gaps in medication availability during follow-up (no active order/dispensing) or 2) not having medication available at a fixed time point. We examined the impact of varying gap length from 30 to 180 days, accounting for stockpiling, and requiring a 30-day period without orders/dispensings (“halo”) around the fixed time point. We also compared results from medication orders vs. dispensings for the same population.Results 1.6-2.6% of older adults had chronic use of a benzodiazepine or Z-drug. Depending on the definition and site, the proportion discontinuing use over 12 months ranged from 6% to 49%. Requiring a longer gap in orders/dispensings or a 30-day “halo” around a fixed time point resulted in lower estimates. Orders data were less likely to identify discontinuation than dispensing data.Conclusions Requiring a medication gap of ≥90 days or a 30-day period with no orders/dispensings around a fixed time point may improve the likelihood that an outcome represents true discontinuation. Orders data appear to underestimate discontinuation compared to dispensing data. More work is needed to adapt and test the proposed definitions for other drug classes and care settings.Impact Statement We certify that this work is novel. Prior papers have identified a need for greater standardization of definitions for medication exposure and discontinuation in deprescribing studies. To our knowledge, no prior paper has systematically examined the construction of deprescribing variables from electronic health data. This is the first paper to present standardized definitions for variables needed for deprescribing studies based on electronic health data, to implement these definitions in multiple healthcare systems and data types, and to examine their performance. This is also the first paper to examine the impact of using medication orders vs. dispensing data to define key deprescribing variables.Key points <jats:list list-type="bullet">Using benzodiazepines and Z-drugs as a case study, we developed and implemented standardized definitions for key variables needed for deprescribing studies using electronic health records data from 5 diverse U.S. healthcare systems.Requiring a gap of ≥90 days without an active order/dispensing or, for a fixed time point, requiring a period of ≥30 days surrounding it with no order/dispensing will likely increase the accuracy of identifying true medication discontinuation.Applying definitions to medication orders data generated higher estimates of chronic use and lower estimates of medication discontinuation than dispensing data.Why does this paper matter? The use of standardized variable definitions in deprescribing studies will improve the ability to synthesize data and compare results between studies, advancing knowledge and supporting more evidence-based guidelines for clinical care. | ||
| 650 | 4 | |a Biology |7 (dpeaa)DE-84 | |
| 650 | 4 | |a 570 |7 (dpeaa)DE-84 | |
| 700 | 1 | |a Albertson-Junkans, Ladia |4 aut | |
| 700 | 1 | |a Pham Nguyen, Thanh Phuong |0 (orcid)0000-0003-3589-764X |4 aut | |
| 700 | 1 | |a Pavon, Juliessa M. |4 aut | |
| 700 | 1 | |a Hastings, Susan N. |4 aut | |
| 700 | 1 | |a Maciejewski, Matthew L. |4 aut | |
| 700 | 1 | |a Willis, Allison |4 aut | |
| 700 | 1 | |a Zepel, Lindsay |4 aut | |
| 700 | 1 | |a Hennessy, Sean |4 aut | |
| 700 | 1 | |a Albers, Kathleen B. |4 aut | |
| 700 | 1 | |a Mowery, Danielle |0 (orcid)0000-0003-3802-4457 |4 aut | |
| 700 | 1 | |a Clark, Amy G. |4 aut | |
| 700 | 1 | |a Thomas, Sunil |4 aut | |
| 700 | 1 | |a Steinman, Michael A. |4 aut | |
| 700 | 1 | |a Boyd, Cynthia M. |4 aut | |
| 700 | 1 | |a Bayliss, Elizabeth A. |4 aut | |
| 773 | 0 | 8 | |i Enthalten in |t bioRxiv.org |g (2023) vom: 10. Nov. |
| 773 | 1 | 8 | |g year:2023 |g day:10 |g month:11 |
| 856 | 4 | 0 | |u http://dx.doi.org/10.1101/2023.11.06.23298060 |z kostenfrei |3 Volltext |
| 912 | |a GBV_XBI | ||
| 951 | |a AR | ||
| 952 | |j 2023 |b 10 |c 11 | ||