Cori Campbell, Tingyan Wang, Eleanor Barnes, Philippa C. Matthews, the UK Health Informatics Collaborative for Viral Hepatitis and Liver Disease
<p>Thank you for the opportunity to respond to the letter from Katkuri et al. [<span>1</span>]. We thank this group for providing positive feedback on our report of treatment eligibility for adults living with Hepatitis B virus (HBV) infection in the UK [<span>2</span>], which was undertaken by the Health Informatics Collaborative (HIC) for viral hepatitis and liver disease [<span>3</span>].</p><p>First, we address the feedback about the HIC approach of using a single positive HBV test as a marker of chronic infection. We recognise that this approach deviates from the gold standard of two tests over a period of at least 6 months, and may therefore (in theory) misassign some cases of acute infection as chronic. However, the decision to adopt this relaxed approach to case definition was taken and ratified by an expert board with clinical and data experts from all the HIC sites and has been previously justified, published and peer reviewed [<span>4</span>]. In England, the incidence of acute HBV infection has been declining over time, and is low (< 300 cases/year since 2020 [<span>5</span>]). The potential for acute infection to influence any analytical conclusions based on analysis of the population under chronic care clinics is therefore negligible. Based on the longitudinal follow-up in this dataset, the chances of including acute cases is further reduced (those with acute infection will not have data points for inclusion beyond 6 months under observation). By removing those with only one diagnostic test, we would risk excluding valuable data from people with true chronic infection.</p><p>Second, in considering our statistical methods and presentation of results, we agree there is a potential problem of reverse causality. On these grounds we analysed data and presented conclusions carefully, deliberately avoiding characterisation of any associations as causal. On these grounds, we simply present ‘factors associated with treatment’ without any statements on either temporality or causality, which we are indeed unable to assign in this analysis. Note that figure 1 in our paper [<span>2</span>] presents an odds ratio for ‘ever receiving’ treatment which does not imply any direction of association.</p><p>Many factors associated with treatment status (such as sex, chronic co-infection, ethnicity, and IMD quintile) are not subject to temporality. However, we recognise that the lack of significant association between viral load and treatment is likely explained by a temporality issue (viral load is lower in those receiving treatment and therefore—despite being a major part of eligibility assessment—does not emerge as a significant signal predicting treatment. This point is addressed in the paper's discussion [<span>2</span>]). We agree and acknowledge that there are limitations associated with gaps in recording of treatment start date; this is explicitly included in the original paper (section 4.5 in the discussion expands on this challenge in some de
感谢您给我机会回复Katkuri et al. b[1]的来信。我们感谢该小组对我们的报告提供了积极的反馈,该报告是由健康信息学协作组织(HIC)对病毒性肝炎和肝脏疾病[3]进行的,针对英国[3]的成人乙型肝炎病毒(HBV)感染的治疗资格。首先,我们解决了关于HIC方法的反馈,即使用单一阳性HBV检测作为慢性感染的标记。我们认识到,这种方法偏离了在至少6个月的时间内进行两次检测的黄金标准,因此(理论上)可能会将一些急性感染病例误认为是慢性感染。然而,采用这种宽松的病例定义方法的决定是由一个由来自所有卫生保健中心的临床和数据专家组成的专家委员会做出并批准的,并且此前已得到证明、发表和同行评议。在英国,随着时间的推移,急性HBV感染的发病率一直在下降,并且很低(自2020年以来每年300例)。因此,急性感染对基于慢性护理诊所人口分析的任何分析结论的潜在影响可以忽略不计。根据该数据集中的纵向随访,纳入急性病例的机会进一步降低(急性感染患者在观察6个月后将没有纳入的数据点)。通过排除那些只有一种诊断测试的人,我们可能会排除来自真正慢性感染患者的有价值数据。其次,在考虑我们的统计方法和结果的呈现时,我们同意存在反向因果关系的潜在问题。基于这些理由,我们仔细分析了数据并提出了结论,故意避免将任何关联描述为因果关系。基于这些理由,我们只是简单地提出“与治疗有关的因素”,而不作任何关于时间性或因果性的说明,因为在这种分析中,我们确实不能确定因果性。请注意,我们的论文[2]中的图1给出了“曾经接受”治疗的比值比,这并不意味着任何关联方向。许多与治疗状况相关的因素(如性别、慢性合并感染、种族和IMD五分位数)不受时间性的影响。然而,我们认识到病毒载量与治疗之间缺乏显著相关性可能是由于暂时性问题(接受治疗的患者的病毒载量较低,因此尽管是资格评估的主要部分,但并不作为预测治疗的重要信号)。这一点在本文的讨论中得到了解决。我们同意并承认在治疗开始日期的记录中存在与空白相关的限制;这在原始论文中明确包含(讨论的第4.5节对这一挑战进行了一些详细的扩展)。第三,在基于HIC数据评估治疗资格方面,确实采取了简化的方法。我们在论文中解释了原因:这一国家合作过程捕获了人口统计、成像和实验室数据,因为使用电子系统[2]稳健地捕获了定量变量。其他变量,如家族史、合并症或个人偏好,在现实世界的临床实践中通常应用于决策,但不幸的是,电子记录尚未始终如一地捕获。我们题为“注意事项和限制”的一节描述了这个问题,我们继续详细说明这些缺失参数的含义,这可能导致对个性化评估后接受治疗的比例的低估或高估。Katkuri和合著者提出了一种敏感性分析方法来调查这些缺失数据的影响。然而,当这些参数没有在整个数据集中收集时,这样的工作就不能应用(只有当队列的一个子集缺少数据时,才能进行敏感性分析,这些数据可以被定义和删除,以便进行更细粒度的重新分析)。我们在论文的未来愿望部分中进一步探索了缺失参数的方法,其中描述了对增强数据的需求。其中一些参数,如用于诊断的ICD-10代码,将允许将来扩展。其他数据,如个人偏好,在人口水平上仍然不太容易记录,因为这是患者与其护理提供者之间的微妙讨论,而现有的电子平台无法捕获。我们清楚地认识到这些注意事项,因此将我们的工作作为“初步”描述性分析。 我们同意Katkuri的观点,即这些差异在数字上很小,不会改变治疗结果的幅度或趋势,也不会影响定性结论(基于这些理由,期刊编辑团队建议不需要发表正式的勘误)。总的来说,我们非常感谢对我们的出版物的周到和详细的反馈。这次对话是一个有趣的机会,可以反思、讨论并从使用现实世界临床数据的机遇和挑战中学习。尽管存在这些警告,但我们报告的分析是建立与在人口水平上应对HBV的规划政策和服务相关的模式的垫脚石。作为一个临床和研究界,我们无疑必须继续推进和改进常规卫生服务数据的保密整理和查询方式,以了解人口水平的特征和需求。这些愿望被强调为英国国家卫生服务战略的一部分,作为推动从模拟到数字医疗的一部分。随着国际社会致力于更广泛地推广HBV治疗,将其作为推动我们到2030年消除作为公共卫生威胁的病毒性肝炎战略的组成部分,在获取高质量人口数据方面进行持续投资是一项基本愿望。弗朗西斯克里克研究所得到了英国癌症研究中心、英国医学研究理事会和威康信托基金会的核心资金支持(参考CC2223)。她还获得了伦敦大学学院医院NIHR生物医学研究中心(BRC)的资助。获得葛兰素史克公司博士奖学金资助,由P.C.M.和E.B.指导(2019-2023)。英国病毒性肝炎和肝脏疾病健康信息学合作组织承认来自NIHR和GSK的资助。P.C.M.是英国病毒性肝炎国家战略小组(NSGVH)的联合主席,从牛津大学出版社获得图书出版的版税,并从强生公司获得一次教育活动的演讲费(2025年)。E.B.拥有HBV疫苗专利,并已获得GSK和Barinthus在HBV领域的研究资金。关于“乙型肝炎病毒(HBV)在英国的治疗资格:回顾性纵向队列数据探讨临床指南变化的影响”https://doi.org/10.1111/jvh.70132.Data的评论,分享不适用于本文,因为在目前的研究中没有生成或分析数据集。
{"title":"Response to Comments on ‘Hepatitis B Virus (HBV) Treatment Eligibility in the UK: Retrospective Longitudinal Cohort Data to Explore the Impact of Changes in Clinical Guidelines’","authors":"Cori Campbell, Tingyan Wang, Eleanor Barnes, Philippa C. Matthews, the UK Health Informatics Collaborative for Viral Hepatitis and Liver Disease","doi":"10.1111/jvh.70133","DOIUrl":"10.1111/jvh.70133","url":null,"abstract":"<p>Thank you for the opportunity to respond to the letter from Katkuri et al. [<span>1</span>]. We thank this group for providing positive feedback on our report of treatment eligibility for adults living with Hepatitis B virus (HBV) infection in the UK [<span>2</span>], which was undertaken by the Health Informatics Collaborative (HIC) for viral hepatitis and liver disease [<span>3</span>].</p><p>First, we address the feedback about the HIC approach of using a single positive HBV test as a marker of chronic infection. We recognise that this approach deviates from the gold standard of two tests over a period of at least 6 months, and may therefore (in theory) misassign some cases of acute infection as chronic. However, the decision to adopt this relaxed approach to case definition was taken and ratified by an expert board with clinical and data experts from all the HIC sites and has been previously justified, published and peer reviewed [<span>4</span>]. In England, the incidence of acute HBV infection has been declining over time, and is low (< 300 cases/year since 2020 [<span>5</span>]). The potential for acute infection to influence any analytical conclusions based on analysis of the population under chronic care clinics is therefore negligible. Based on the longitudinal follow-up in this dataset, the chances of including acute cases is further reduced (those with acute infection will not have data points for inclusion beyond 6 months under observation). By removing those with only one diagnostic test, we would risk excluding valuable data from people with true chronic infection.</p><p>Second, in considering our statistical methods and presentation of results, we agree there is a potential problem of reverse causality. On these grounds we analysed data and presented conclusions carefully, deliberately avoiding characterisation of any associations as causal. On these grounds, we simply present ‘factors associated with treatment’ without any statements on either temporality or causality, which we are indeed unable to assign in this analysis. Note that figure 1 in our paper [<span>2</span>] presents an odds ratio for ‘ever receiving’ treatment which does not imply any direction of association.</p><p>Many factors associated with treatment status (such as sex, chronic co-infection, ethnicity, and IMD quintile) are not subject to temporality. However, we recognise that the lack of significant association between viral load and treatment is likely explained by a temporality issue (viral load is lower in those receiving treatment and therefore—despite being a major part of eligibility assessment—does not emerge as a significant signal predicting treatment. This point is addressed in the paper's discussion [<span>2</span>]). We agree and acknowledge that there are limitations associated with gaps in recording of treatment start date; this is explicitly included in the original paper (section 4.5 in the discussion expands on this challenge in some de","PeriodicalId":17762,"journal":{"name":"Journal of Viral Hepatitis","volume":"33 2","pages":""},"PeriodicalIF":2.3,"publicationDate":"2026-01-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12828635/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146030241","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}