地磁校准 椭球校准_校准行为感染风险

地磁校准 椭球校准

Recent national COVID-19 statistics and reports have highlighted the total number of new cases over the last 14 days per 100,000 of population as a normalised estimate of the prevalence of the virus. It allows for a like-for-like comparison between different counties and regions but in recent days I have been wondering how useful such a metric is from a public communication viewpoint. Do people have adequate intuitive sense of what a particular value of this metric means? It might help us to compare counties but does it help us to calibrate our behaviour where we live?

是最近几个国家COVID-19的数据和报告都强调新发病例总数在过去14天每10万人口 ，因为病毒的流行的标准化估计。 它允许在不同县和地区之间进行类似的比较，但是最近几天，我一直想知道从公共传播的角度来看，这种度量有多有用。 人们是否对该指标的特定值有足够的直观认识？ 这可能有助于我们比较县，但是否可以帮助我们校准居住地的行为？

At the risk of courting controversy, I’m going to claim that even though this metric is epidemiologically informative, it is not so easy for the general public to understand and, as a result, knowing how to calibrate our behaviour can be difficult. This shortcoming is important because adjusting our behaviour in the right way is involves a delicate balance between starving the virus of transmission opportunities while getting on with our lives. Underestimating the level of risk will help the virus to grow but overestimating it brings a different set of problems, including having a deleterious effect on our mental health and increasing the chances of irreparably damaging the economy.

冒着引起争议的风险，我要声明，即使该指标在流行病学上具有参考意义，但对于一般公众而言，理解起来并不容易，因此，知道如何校准我们的行为可能会很困难。 这个缺点很重要，因为以正确的方式调整我们的行为需要在饥饿的传播机会的病毒与我们的生活继续之间保持微妙的平衡。 低估风险水平将有助于病毒的增长，但高估它会带来一系列不同的问题，包括对我们的心理健康产生有害影响以及增加对经济造成无法弥补的损害的机会。

In this article I describe a simple alternative for captuing the current status of the virus, one that is more intuitive and more actionable. The basic idea is to triangulate current levels of the virus with respect to the highs of lockdown, when the virus was out of control, and the lows of summer, when the virus was largely contained. The point is that we can all remember how it felt during lockdown and after restrictions were eased in June and July — we can recall the perception of risk and how we modified our behaviour during these months — and relating current levels to these periods provides us with a useful reference point against which to evaluate current risk and modify our behaviour accordingly.

在本文中，我描述了一种用于捕获病毒当前状态的简单替代方法，该替代方法更直观，更可行。 基本思想是根据病毒失控时的锁定高点和主要包含该病毒的夏季低点，对病毒的当前水平进行三角测量。 关键是，我们都可以记住在锁定期间以及6月和7月限制解除后的感觉-我们可以回想起对风险的感知以及在这几个月中我们如何改变行为-将当前水平与这些时期相关联可以为我们提供一个有用的参考点，可用来评估当前风险并相应地改变我们的行为。

As an example, Dublin’s current prevalence is approximately 28 cases per 100,000 of population, but what does this mean for our behaviour? Knowing that this is almost 10-times lower than Dublin’s average lockdown level (relative prevalence with respect to lockdown), but 3-times higher than Dublin’s Summer lows (relative prevalence with respect to Summer) is more useful, because it suggests that Dubliner’s should be more alert than they were during June and July without being so alarmed as to avoid getting on with their lives by self-imposed lockdown.

例如，都柏林目前的患病率约为每10万人中28例，但这对我们的行为意味着什么？ 知道这几乎比都柏林的平均锁定水平低10倍 ( 相对于锁定的相对流行率 )，但是比都柏林的夏季低点(相对于夏天的相对流行)高3倍 ，因为它表明都柏林人应该比6月和7月的时候更加警惕，不要太警惕，以免因自我施加的封锁而过着自己的生活。

The graph below plots the current virus prevalence (14-day case totals per 100,000 population) for counties in the Republic of Ireland using this approach. The x-axis (log scale) is the relative prevalence with respect to the Summer months (June and July) while the y-axis (also log scale) is the relative prevalence with respect to lockdown (March 13–May 18). The diameter of each county marker is proportional to the current prevalence level (per 100,000 of population) and this value is indicated along with the county name. Informally, three clusters can be distinguished and are highlighted in this graph, although it is important to point out that this is based on observation alone.

吨他图下方地块使用这种方法在爱尔兰共和国的县当前病毒流行率(每10万人14天的情况下，总计)。 x轴(对数标度)是相对于夏季月份(6月和7月)的相对患病率，而y轴(也对数标度)是相对于锁定(3月13日至5月18日)的相对患病率。 每个县标记的直径与当前患病率(每100,000人口)成正比，并且该值与县名一起显示。 非正式地，可以区分三个聚类并在此图中突出显示，尽管必须指出这仅基于观察。

The red cluster includes counties where virus prevalence is at or near lockdown levels and significantly higher (5–25x higher) than Summer levels. Not surprisingly this cluster includes Laois, Offaly, and Kildare but also Kilkenny, Carlow, Limerick and others. In terms of behaviour, it seems reasonable that people in these counties would be on their highest level of alert for the virus. They should be behaving much more cautiously than they were in Summer and, arguably, in a manner that is close to the way they behaved during lockdown.

红色集群包括病毒流行率处于或接近锁定水平且明显高于夏季水平(高5至25倍)的县。 毫不奇怪，这个集群包括Laois，Offaly和Kildare，还包括Kilkenny，Carlow，Limerick等。 在行为方面，这些县的人们对病毒保持最高戒备似乎是合理的。 他们应该比夏季更加谨慎，并且可以说，其行为方式与锁定期间的行为方式相近。

The orange cluster includes counties (e.g. Dublin, Kerry, Meath, Longford etc.) where current prevalence is significantly higher (2–10x higher) than Summer lows but still considerably lower (typically 2–5x lower) than lockdown highs. People in these counties should be on a higher level of alert, compared with June and July, while recognising that levels still remain far from the highs of lockdown. Arguably, Donegal might be a better fit for the red cluster, because its high current prevalence relative to the Summer sets it apart from others in this cluster.

橙色集群包括县(例如都柏林，克里，米斯，朗福德等)，这些县的当前流行率比夏季低点高得多(高2-10倍)，但比锁定高点低得多(通常低2-5倍)。 与6月和7月相比，这些县的人们应该提高警惕水平，同时要认识到，这一水平与锁定的高点仍然相去甚远。 可以说，多尼戈尔可能更适合红色星团，因为相对于夏季而言，它的高电流流行使其与该星团中的其他星区分开。

Finally, the yellow cluster includes counties (e.g. Cork, Wiklow, Galway, Cavan etc.) where levels continue to be at or below Summer levels, and much lower (10–100x lower) than the levels during lockdown. People in these counties should continue to behave as they have been — since it is obviously working — and take pride in their achievement to date.

最后， 黄色集群包括县(例如科克，威克洛，戈尔韦，卡文等)，这些县的水平继续保持或低于夏季水平，并且比锁定期间的水平低得多(低10–100倍)。 这些县的人们应该继续保持其原有的举止-因为这显然正在奏效-并为迄今取得的成就感到自豪。

There are a couple of caveats to note about this approach. The main one is that it implies that all counties were similarly hard hit during lockdown and did similarly well during the Summer months. While this is typically true, it is not always so, which means some (atypical) counties may be misplaced, if for example, they appear to be doing well now but only compared to a poor Summer, or vice versa.

Ť这里有几个关于这种方法的注意事项要注意的。 最主要的是，这意味着在封锁期间所有县遭受了类似的重创，而在夏季月份中表现良好。 尽管这通常是正确的，但并非总是如此，这意味着某些(非典型)县可能放错了位置，例如，如果它们现在看起来表现良好，但仅与夏季不佳相比，反之亦然。

For example, Monaghan experienced high case numbers in early June resulting in a high average prevalence in June and July, so that now Monaghan appears to be doing well, on a par with its Summer prevalence (14 cases per 100k), when in comparison with others in the orange cluster this current prevalence appears to be high; it’s x-axis position is underestimated. In contrast, Waterford had a relatively low average prevalence during lockdown but its current prevalence, relative to lockdown, appears to be high because it is close to this low lockdown level; its y-axis position is overestimated.

例如，莫纳汉(Monaghan)在6月初的病例数很高，导致6月和7月的平均患病率很高，因此现在与现在的夏季患病率(每100k 14例)相比，莫纳汉(Monaghan)表现良好。橙色集群中的其他国家，当前流行率似乎很高； 它的x轴位置被低估了。 相反，沃特福德在锁定期间的平均患病率相对较低，但相对于锁定而言，当前的患病率似乎很高，因为它接近此较低的锁定水平。 其y轴位置被高估了。

One way to address this is by making actual lockdown and Summer prevalence levels explicit in the visualisation; this doesn't adjust the position of the counties but it makes their lockdown and Summer levels transparent. In the version below we show one way to achieve this by annotating each county with two concentric rings, in addition to its filled, main marker. The larger outer ring, outlined with the same colour as the main marker, indicates the mean prevalence during lockdown. Most counties present with a lockdown ring that is significantly larger than the main marker, indicating lockdown levels were much higher than current levels; Carlow and Kilkenny are exceptions as they are currently experiencing prevalence levels that are similar to those in lockdown. The second ring, outlined in black, shows average Summer prevalence levels, and the size of this ring usually indicates Summer levels that were much lower than lockdown, and usually lower than current levels too; Cavan, Sligo and Westmeath are the exceptions; they have current prevalence levels that are even less than their Summer levels.

解决此问题的一种方法是在可视化中明确显示实际的锁定和夏季患病率。 这不会调整县的位置，但是会使他们的锁定和夏季级别透明。 在下面的版本中，我们展示了一种实现此目标的方法，即在每个郡县之外，还用两个同心圆环标注主要的标记。 外圈较大，轮廓与主要标记的颜色相同，表示锁定期间的平均患病率。 大多数县的锁定环比主要标记大得多，这表明锁定水平远高于当前水平。 卡洛(Carlow)和基尔肯尼(Kilkenny)是个例外，因为他们目前的患病率与锁定水平相似。 第二个圆环用黑色勾勒出，表示夏季的平均患病水平，该圆环的大小通常表示夏季水平远低于锁定水平，通常也低于当前水平。 Cavan，Sligo和Westmeath除外； 他们目前的患病率甚至低于夏季水平。

In this new version of the graph the solid vertical and horizontal lines correspond to the lockdown and Summer prevalence levels (relative prevalence of 1) while the dashed lines show the median of the current prevalence relative to the lockdown and Summer periods.

在此图表的新版本中，实心垂直线和水平线对应于锁定和夏季流行程度(相对流行度1)，而虚线表示相对于锁定和夏季时期的当前流行程度的中位数。

Of course by adding additional features to the graph in this way, we run the risk of complicating the core message. Nevertheless it does provide a path to enriching the visualisation if such additional information is necessary.

当然，通过以这种方式向图表添加其他功能，我们冒着使核心信息复杂化的风险。 但是，如果需要这样的附加信息，它的确提供了丰富可视化的途径。

The purpose of this analysis is not to provide yet another COVID-19 metric, we have plenty of those! Rather the main aim is to improve the manner in which COVID-19 data is reported to the public.

该分析吨他的目的不是为了提供另一个COVID 19度，我们有足够的那些！ 而是主要目的是改进向公众报告COVID-19数据的方式。

The current obsession with case totals, incident rates, and prevalence rates, while scientifically relevant, may not help the general public to calibrate its behaviour, because the average person simply does not have an intuitive understanding of whether a particular number of cases per 100k population, for example, suggests more or less caution is warranted.

目前对案件总数，事件发生率和患病率的痴迷，尽管在科学上是相关的，但可能无法帮助公众校准其行为，因为普通人根本不了解每10万人口中是否有特定案件数量，例如，建议应或多或少谨慎。

In this article we have proposed an alternative approach which anchors such precise metrics with respect to certain landmark reference points, in this case the high prevalence during lockdown and the low prevalence during the Summer; obviously, in different countries and regions other landmark periods may be appropriate. The point is, that since we all lived through the prevalence highs of lockdown and the corresponding lows of Summer, we retain a memory of how we felt and behaved, and by anchoring current levels to these landmarks we may be able to better calibrate our current behaviour.

在本文中，我们提出了一种替代方法，该方法针对某些地标性参考点锚定此类精确指标，在这种情况下，锁定期间的流行率较高，而夏季期间的流行率较低； 显然，在不同的国家和地区，其他里程碑时期可能是合适的。 关键是，由于我们都经历了锁定的普遍高点和夏季的相应低点，所以我们保留了对自己的感觉和行为的记忆，并且通过将电流水平锚定到这些标志性位置，我们可能能够更好地校准电流行为。

翻译自: https://medium.com/data-in-the-time-of-corona/calibrating-behaviour-infection-risk-5c294d0f510e

地磁校准 椭球校准