COVID-19 Testing in Japan: The situation and ramifications


Source: VOX

Basic COVID-19 testing numbers for Japan as of March 16, 2020:

  • Tested: 13,026*
  • Detected: 1,496 (both symptomatic and asymptomatic cases)
  • Deaths: 24

Sources: Toyokeizai, MHLW, Japan Times, Vox

*Note added March 18, 2020: MHLW data gives two sets of testing figures; daily report for Mar 18 now gives 33,212, which is noted as “total number of responses received from local governments and other organizations.”  But the table on the same page gives 15,354, categorized as “tested by PCR.” It is not clear what the difference is.

Compared to countries which have rolled out intensive COVID-19 screening and testing programs, like South Korea, Japan has not done much surveillance testing to date. According to Japan’s health ministry, as of March 6, 13,026 people in total have been tested, compared to over 200,000 in Korea. In fact, in terms of percentage of the population tested, while above the US, at the moment Japan ranks lower than the UK, Netherlands, Israel, Taiwan, Italy, Guangdong, China, and South Korea. Therefore people’s concern that the lack of testing might obscure a much wider incidence of coronavirus which is not being detected is understandable. Recent information suggests that asymptomatic carriers of the virus — those contract the virus but experience no physical symptoms of illness, like coughing or fever — account for 86% of the spread to other people. The testing situation in Japan in terms of percentages and timing seems fairly similar to that in the United States, which appears headed for severe healthcare system overload soon.

Source: Financial Times

It’s been difficult to get a solid understanding of what accounts for the very different outcomes we’ve seen, for instance, between South Korea, which appeared touch and go for a while but seems to have gotten past the worst, and Italy, where we see an ongoing and traumatic health system collapse. Many factors can account for the widely varied outcomes, and the scale, timing, and effectiveness of testing seems to have definitely been playing a large role. Countries that promptly established thorough testing regimes, like Singapore, Korea, and Taiwan, have fared much better so far in “flattening the curve” than those which haven’t. A lot of data is available which visualizes this, including the comparative graph above that has been frequently updated by John-Burn Murdoch of the Financial Times (paywalled). Unlike the steep exponential growth seen in most countries, the plot line curve for Japan is very similar in its reassuring shallow gradient to Singapore, Korea, and Taiwan, which suggests that COVID-19 is largely under control here as it is in those countries. But since, unlike those countries, Japan has not been doing thorough testing, why does the outcome appear similar so far?

The graph notes “isolation of elders” in Japan, positing it as a positive factor for limiting the disease spread. While some elderly care facilities have been closed in a handful of cities, this has not yet been widespread. It may well be that many elderly here have gotten the message and are self-isolating. While some in our network here in Japan have observed this, others have not. There has been other notable social distancing, through the closure of schools and the encouragement of working from home. There is significant evidence that despite the school closures and widespread cancellation of large events, social isolation is not being adhered to strongly in many places. Even now, bars, clubs, and restaurants are fairly full in Tokyo in the evening.

Japan, Singapore, Korea, and Taiwan all have good personal hygiene norms, such as frequent hand washing. Except for Singapore, people in these countries also commonly bow instead of shaking hands. It is easy to imagine that behavioral norms like these might contribute to the apparent similarity of outcomes. Some have suggested that maybe face masks, which are commonly worn in all of these countries, especially now, might be similarly beneficial in limiting the disease spread. These questions require good evidence to either prove or disprove, and unfortunately, we have to admit that for the most part nobody knows for sure.

Until now, testing for the virus in Japan has been limited by policy. There is no official body like the US CDC tasked with handling emergency public health issues. Instead, the National Institute of Infectious Diseases (NIID), under the Ministry of Health, Labor, and Welfare (MHLW) has taken the lead. All requests for COVID-19 testing, we are told, must be approved by this body, which is currently overloaded. As many people have noted, including several we’ve heard from who have enough of the symptoms to strongly suspect they have coronavirus but have been unable to obtain a test in Japan, the NIID has set a high hurdle of conditions which must be met before tests will be approved. In order to qualify for the first medical consultation, the person must have been in close contact with a known infected person, defined as contact over a long period of time or living together, examining or caring for a patient without taking protective measures, or contact with bodily fluids, and currently have fever of 37.5’ Celsius or respiratory symptoms; and/or had visited an endemic area in Hubei or Zhejiang Provinces, and currently have fever of 37.5’ Celsius or respiratory symptoms. Those who call Japan’s Novel Coronavirus Patient Consultation Call Center (03-5320-4592/03-5285-8181) and are recommended for in-person outpatient consultation must then appear at the designated center and be examined. If the doctor suspects COVID-19, he can recommend a PCR (polymerase chain reaction) test, and seek approval from the NIID for it. In practice, we are told, the NIID is so inundated with requests it is often difficult for doctors to get through by phone. This system presents a strong bottleneck to testing.


As is unfortunately often the case, it is difficult for average citizens to understand or navigate this system and its requirements, and many are becoming quite frustrated and alarmed. If they’re feeling ill with similar symptoms, they want to know how serious their prognosis is. A recent article in the Nikkei Asian Weekly attempts to explain the policy. The capacity is there, the writer notes, and the number of tests has been limited because the Ministry of Health, Labor and Welfare wanted the data to support an epidemiological investigation, not as medical care. In particular, the policy focuses on identifying clusters of the disease, to help in the allocation of medical resources. This is also emphasized in official policy documents from February 25th of this year, available at the MHLW website:

“Establish the surveillance system to grasp the epidemic situation in Japan, while switching to use of PCR test for the confirmation of diagnosis necessary to treat pneumonia patients who require hospitalization, in communities where the number of patients continues to increase.”

“The local government shall use active epidemiology surveys to identify a cluster based on individual patient outbreaks in cooperation with the Ministry of Health, Labour and Welfare and experts, and request the necessary measures including closure of the related facilities and voluntary restraint of events if there is a possibility that such a cluster exists.”


“Shift the focus of PCR test to the confirmation of diagnosis for pneumonia patients who require hospitalization, while strengthening the surveillance system to grasp the epidemic situation in Japan.”

Ministry staff felt that surveillance testing should be adequate for understanding the scale and geographical distribution of the outbreak, but that too much testing, particularly of “worried well,” would overwhelm the system. Notably, the government decided against allowing rapid test kits developed by the pharmaceutical company Roche, which have been used with success in Wuhan and elsewhere, to be used in Japan, fearing that the variability of test results would hamper nationwide epidemiological analysis. After criticism, the government agreed to allow the Roche tests to be used and eligible to be covered by Japanese public health insurance from March 6. Beginning March 16, laboratories and medical institutions tasked with testing for COVID-19 have also been able to purchase kits from Kurabo Industries Ltd. that can detect the virus in 15 minutes, as opposed to up to six hours required for PCR tests. Other companies hope to have similarly rapid testing materials available in Japan within a few weeks. We hope that these developments mean that testing in Japan will increase quickly. A few days ago Prime Minister Abe promised that by the end of March Japan’s nationwide COVID-19 sample testing capacity will increase to 8000 per day from the current 6000. Even so, some prefectures only have the capacity to test a little more than 100 samples per day.

Dr. Kiyoshi Kurokawa, a celebrated public health specialist who has served as a science adviser to the Cabinet and chaired the Fukushima Nuclear Accident Independent Investigation Commission (NAIIC), is also a Safecast advisor. In a recent Japan Times article he is quoted as saying, “Any patient who comes to the doctor, if they have a reasonable suspicion (of having coronavirus), and they want this testing, just do it…And what is the price? That comes later anyway.” The epidemiological study is necessary, but it is equally necessary to transparently alleviate people’s legitimate concerns. He notes the reluctance of officials to approve the outsourcing of COVID-19 testing because it’s not clearly “written into law,” a situation which often leads to government paralysis in Japan. Officials may agree in private that it should be done, but no-one wants to take responsibility for breaking the rules even during an emergency. Kurokawa strongly recommends streamlining the outsourcing of testing to university labs and the private sector.

It’s possible, then, that there has been (barely) enough testing in Japan to allow public health specialists to anticipate the eventual spread of the virus and to prepare the necessary hospital beds. But Kurokawa wryly notes, “Nobody knows.” Is it possible that there is a much larger rate of contagion in Japan which is not being detected? Yes, but the corresponding hospitalizations and fatalities have not been evident. If the numbers were growing exponentially like in most countries, it would seem hard to hide that fact. Is it possible that many Japanese, particularly elderly, are dying of pneumonia or other respiratory diseases which are not being diagnosed as COVID-19? It seems possible. Hospitals are not required to share such data, though, and autopsies are normally performed on fewer than 2% of all deaths. Recent funeral rite guidance has indicated, however, that because the deceased are generally not being tested for coronavirus, all who have died of pneumonia should be handled with the same precautions as if they had the virus.

That said, if there was a very large number of highly contagious cases serious enough for hospitalization, we might expect a corresponding increase in respiratory illness among medical workers, which so far has not been evident. Again, however, hospitals are not required to report this. At present, the lack of widespread surveillance testing cannot be reassuring regardless of the comparatively small number of COVID-19 cases reported here so far. We cannot exclude the possibility that Japan might yet experience an exponential growth in cases, and feel that under the circumstances it’s prudent to be prepared for the case numbers to explode in coming weeks. What’s the likelihood of that happening? Without reliable data, nobody knows.