PDF(Pubmed)
Abstract：
BACKGROUND: The COVID-19 pandemic highlighted the importance of telemedicine in health care. However, video telemedicine requires adequate broadband internet speeds. As video-based telemedicine grows, variations in broadband access must be accurately measured and characterized.
OBJECTIVE: This study aims to compare the Federal Communications Commission (FCC) and Microsoft US broadband use data sources to measure county-level broadband access among veterans receiving mental health care from the Veterans Health Administration (VHA).
METHODS: Retrospective observational cohort study using administrative data to identify mental health visits from January 1, 2019, to December 31, 2020, among 1161 VHA mental health clinics. The exposure is county-level broadband percentages calculated as the percentage of the county population with access to adequate broadband speeds (ie, download >25 megabits per second) as measured by the FCC and Microsoft. All veterans receiving VHA mental health services during the study period were included and categorized based on their use of video mental health visits. Broadband access was compared between and within data sources, stratified by video versus no video telemedicine use.
RESULTS: Over the 2-year study period, 1,474,024 veterans with VHA mental health visits were identified. Average broadband percentages varied by source (FCC mean 91.3%, SD 12.5% vs Microsoft mean 48.2%, SD 18.1%; P<.001). Within each data source, broadband percentages generally increased from 2019 to 2020. Adjusted regression analyses estimated the change after pandemic onset versus before the pandemic in quarterly county-based mental health visit counts at prespecified broadband percentages. Using FCC model estimates, given all other covariates are constant and assuming an FCC percentage set at 70%, the incidence rate ratio (IRR) of county-level quarterly mental video visits during the COVID-19 pandemic was 6.81 times (95% CI 6.49-7.13) the rate before the pandemic. In comparison, the model using Microsoft data exhibited a stronger association (IRR 7.28; 95% CI 6.78-7.81). This relationship held across all broadband access levels assessed.
CONCLUSIONS: This study found FCC broadband data estimated higher and less variable county-level broadband percentages compared to those estimated using Microsoft data. Regardless of the data source, veterans without mental health video visits lived in counties with lower broadband access, highlighting the need for accurate broadband speeds to prioritize infrastructure and intervention development based on the greatest community-level impacts. Future work should link broadband access to differences in clinical outcomes.
OBJECTIVE: This study aims to compare the Federal Communications Commission (FCC) and Microsoft US broadband use data sources to measure county-level broadband access among veterans receiving mental health care from the Veterans Health Administration (VHA).
METHODS: Retrospective observational cohort study using administrative data to identify mental health visits from January 1, 2019, to December 31, 2020, among 1161 VHA mental health clinics. The exposure is county-level broadband percentages calculated as the percentage of the county population with access to adequate broadband speeds (ie, download >25 megabits per second) as measured by the FCC and Microsoft. All veterans receiving VHA mental health services during the study period were included and categorized based on their use of video mental health visits. Broadband access was compared between and within data sources, stratified by video versus no video telemedicine use.
RESULTS: Over the 2-year study period, 1,474,024 veterans with VHA mental health visits were identified. Average broadband percentages varied by source (FCC mean 91.3%, SD 12.5% vs Microsoft mean 48.2%, SD 18.1%; P<.001). Within each data source, broadband percentages generally increased from 2019 to 2020. Adjusted regression analyses estimated the change after pandemic onset versus before the pandemic in quarterly county-based mental health visit counts at prespecified broadband percentages. Using FCC model estimates, given all other covariates are constant and assuming an FCC percentage set at 70%, the incidence rate ratio (IRR) of county-level quarterly mental video visits during the COVID-19 pandemic was 6.81 times (95% CI 6.49-7.13) the rate before the pandemic. In comparison, the model using Microsoft data exhibited a stronger association (IRR 7.28; 95% CI 6.78-7.81). This relationship held across all broadband access levels assessed.
CONCLUSIONS: This study found FCC broadband data estimated higher and less variable county-level broadband percentages compared to those estimated using Microsoft data. Regardless of the data source, veterans without mental health video visits lived in counties with lower broadband access, highlighting the need for accurate broadband speeds to prioritize infrastructure and intervention development based on the greatest community-level impacts. Future work should link broadband access to differences in clinical outcomes.
摘要：
背景：COVID-19大流行凸显了远程医疗在医疗保健中的重要性。然而,视频远程医疗需要足够的宽带互联网速度。随着基于视频的远程医疗的发展，必须准确测量和表征宽带接入的变化。
目的：本研究旨在比较联邦通信委员会（FCC）和微软美国宽带使用数据源，以衡量接受退伍军人健康管理局（VHA）精神保健的退伍军人的县级宽带接入。
方法：使用行政数据进行回顾性观察性队列研究，以确定1161个VHA心理健康诊所中从2019年1月1日至2020年12月31日的心理健康访视。暴露量是县级宽带百分比，以获得足够宽带速度的县级人口的百分比计算（即，下载>25兆比特每秒)，由FCC和微软测量。在研究期间接受VHA心理健康服务的所有退伍军人都被包括在内，并根据他们对视频心理健康访问的使用进行分类。在数据源之间和内部比较了宽带接入，按视频和没有视频远程医疗使用进行分层。
结果：在2年的研究期间，确定了1,474,024名接受VHA心理健康访问的退伍军人。平均宽带百分比因来源而异(FCC平均为91.3%，SD12.5%与微软平均48.2%，SD18.1%；P<.001）。在每个数据源中，从2019年到2020年，宽带百分比普遍上升。调整后的回归分析估计了大流行发作后与大流行前相比的变化，季度基于县的心理健康访视计数按预定的宽带百分比进行。使用FCC模型估计，假设所有其他协变量都是恒定的，并假设FCC百分比设置为70%，COVID-19大流行期间县级季度心理视频访视的发生率比(IRR)是大流行前的6.81倍(95%CI6.49~7.13).相比之下,使用微软数据的模型表现出更强的相关性(IRR7.28；95%CI6.78-7.81)。这种关系在评估的所有宽带接入级别中都保持。
结论：这项研究发现，与使用Microsoft数据估算的数据相比，FCC宽带数据估算的县级宽带百分比更高，变化更小。无论数据源如何，没有心理健康视频访问的退伍军人生活在宽带接入较低的县，强调需要准确的宽带速度，以便根据社区层面的最大影响优先考虑基础设施和干预发展。未来的工作应该将宽带接入与临床结果的差异联系起来。
目的：本研究旨在比较联邦通信委员会（FCC）和微软美国宽带使用数据源，以衡量接受退伍军人健康管理局（VHA）精神保健的退伍军人的县级宽带接入。
方法：使用行政数据进行回顾性观察性队列研究，以确定1161个VHA心理健康诊所中从2019年1月1日至2020年12月31日的心理健康访视。暴露量是县级宽带百分比，以获得足够宽带速度的县级人口的百分比计算（即，下载>25兆比特每秒)，由FCC和微软测量。在研究期间接受VHA心理健康服务的所有退伍军人都被包括在内，并根据他们对视频心理健康访问的使用进行分类。在数据源之间和内部比较了宽带接入，按视频和没有视频远程医疗使用进行分层。
结果：在2年的研究期间，确定了1,474,024名接受VHA心理健康访问的退伍军人。平均宽带百分比因来源而异(FCC平均为91.3%，SD12.5%与微软平均48.2%，SD18.1%；P<.001）。在每个数据源中，从2019年到2020年，宽带百分比普遍上升。调整后的回归分析估计了大流行发作后与大流行前相比的变化，季度基于县的心理健康访视计数按预定的宽带百分比进行。使用FCC模型估计，假设所有其他协变量都是恒定的，并假设FCC百分比设置为70%，COVID-19大流行期间县级季度心理视频访视的发生率比(IRR)是大流行前的6.81倍(95%CI6.49~7.13).相比之下,使用微软数据的模型表现出更强的相关性(IRR7.28；95%CI6.78-7.81)。这种关系在评估的所有宽带接入级别中都保持。
结论：这项研究发现，与使用Microsoft数据估算的数据相比，FCC宽带数据估算的县级宽带百分比更高，变化更小。无论数据源如何，没有心理健康视频访问的退伍军人生活在宽带接入较低的县，强调需要准确的宽带速度，以便根据社区层面的最大影响优先考虑基础设施和干预发展。未来的工作应该将宽带接入与临床结果的差异联系起来。
