Atchison Christina J, Gilby Nicholas, Pantelidou Galini, Clemens Sam, Pickering Kevin, Chadeau-Hyam Marc, Ashby Deborah, Barclay Wendy S, Cooke Graham S, Darzi Ara, Riley Steven, Donnelly Christl A, Ward Helen, Elliott Paul
School of Public Health, Imperial College London, London, United Kingdom.
Ipsos, London, United Kingdom.
JMIR Public Health Surveill. 2025 Jan 9;11:e60022. doi: 10.2196/60022.
High response rates are needed in population-based studies, as nonresponse reduces effective sample size and bias affects accuracy and decreases the generalizability of the study findings.
We tested different strategies to improve response rate and reduce nonresponse bias in a national population-based COVID-19 surveillance program in England, United Kingdom.
Over 19 rounds, a random sample of individuals aged 5 years and older from the general population in England were invited by mail to complete a web-based questionnaire and return a swab for SARS-CoV-2 testing. We carried out several nested randomized controlled experiments to measure the impact on response rates of different interventions, including (1) variations in invitation and reminder letters and SMS text messages and (2) the offer of a conditional monetary incentive to return a swab, reporting absolute changes in response and relative response rate (95% CIs).
Monetary incentives increased the response rate (completed swabs returned as a proportion of the number of individuals invited) across all age groups, sex at birth, and area deprivation with the biggest increase among the lowest responders, namely teenagers and young adults and those living in more deprived areas. With no monetary incentive, the response rate was 3.4% in participants aged 18-22 years, increasing to 8.1% with a £10 (US $12.5) incentive, 11.9% with £20 (US $25.0), and 18.2% with £30 (US $37.5) (relative response rate 2.4 [95% CI 2.0-2.9], 3.5 [95% CI 3.0-4.2], and 5.4 [95% CI 4.4-6.7], respectively). Nonmonetary strategies had a modest, if any, impact on response rate. The largest effect was observed for sending an additional swab reminder (SMS text message or email). For example, those receiving an additional SMS text message were more likely to return a completed swab compared to those receiving the standard email-SMS approach, 73.3% versus 70.2%: percentage difference 3.1% (95% CI 2.2%-4.0%).
Conditional monetary incentives improved response rates to a web-based survey, which required the return of a swab test, particularly for younger age groups. Used in a selective way, incentives may be an effective strategy for improving sample response and representativeness in population-based studies.
在基于人群的研究中需要高应答率，因为无应答会减少有效样本量，而偏倚会影响准确性并降低研究结果的普遍性。
我们在英国英格兰一项基于全国人群的新冠病毒监测项目中测试了不同策略，以提高应答率并减少无应答偏倚。
在19轮研究中，通过邮件邀请英格兰普通人群中5岁及以上的个体随机样本，完成一份基于网络的问卷并返还用于新冠病毒检测的拭子。我们进行了多项嵌套随机对照试验，以测量不同干预措施对应答率的影响，包括（1）邀请信、提醒信和短信的变化，以及（2）提供返还拭子的有条件金钱激励，报告应答的绝对变化和相对应答率（95%置信区间）。
金钱激励提高了所有年龄组、出生时性别和地区贫困程度人群的应答率（返还的已完成拭子占受邀个体数量的比例），应答率最低的人群（即青少年、年轻人以及生活在更贫困地区的人）提高幅度最大。在没有金钱激励的情况下，18 - 22岁参与者的应答率为3.4%，提供10英镑（12.5美元）激励时应答率升至8.1%，提供20英镑（25.0美元）时为11.9%，提供30英镑（37.5美元）时为18.2%（相对应答率分别为2.4 [95%置信区间2.0 - 2.9]、3.5 [95%置信区间3.0 - 4.2]和5.4 [95%置信区间4.4 - 6.7]）。非金钱策略对应答率的影响即使有也很有限。观察到的最大影响是发送额外的拭子提醒（短信或电子邮件）。例如，与接受标准电子邮件 - 短信方式的人相比，收到额外短信的人更有可能返还已完成的拭子，分别为73.3%和70.2%：百分比差异为3.1%（95%置信区间2.2% - 4.0%）。
有条件的金钱激励提高了基于网络的调查的应答率，该调查需要返还拭子检测，特别是对于较年轻的年龄组。有选择地使用激励措施可能是提高基于人群研究中样本应答率和代表性的有效策略。
