his week, the government announced plans to give those living in England the opportunity to be tested twice a week in a bid to ease the country out of lockdown. Increased testing is a vital tool, but it is only effective if it is part of a wider system aimed at identifying who is infected, their contacts, and getting people to self-isolate. The last step is the crucial one, since that is the way you stop the transmission of infection. So far, this has proved to be the trickiest part. Whereas all the evidence points to the fact most people are by and large adhering to the things that are relatively straightforward – cleaning hands, wearing masks, spatially distancing – it is a very different story when it comes to asking people to stay at home with no contact with others. According to the Covid-19 Rapid Survey of Adherence to Interventions and Responses (Corsair) study, only 52% have self-isolated when required. However, data produced by the Office for National Statistics (ONS) found that between 1 and 13 February this year self-isolation rates were 86%, and in the period 1 to 9 March they had risen still further to 90%. The health secretary, Matt Hancock, enthused: “I am delighted that today’s statistics show the vast majority of people in this country are doing the right thing and following the rules to help protect the NHS and save lives.” Why such a huge discrepancy between Corsair and ONS? Who was right and who was wrong? On closer inspection the answer turns out that both were right, but they were looking at very different groups of people. The ONS approached people who have had a test and provided contacts. Those already in the system, and who adhere to that extent, are very likely to take the final step and self-isolate. Corsair, on the other hand, approached a more general sample of people. Of these, much fewer self-isolated. But the most striking finding was that only one in five people who had symptoms actually had a test in the first place (to be precise, 18.8% over the year rising to 22.2% in the latest wave). Up until now, the barriers to self-isolation have been self-evident: taking time off work means reduced pay, some are unable to self-isolate in crowded accommodation, and then there are logistical issues around caring for children or elderly relatives. The obvious solution is to provide support for people to follow the regulations. In those places that provide comprehensive care packages, such as New York, 90% reported not leaving their homes during the self-isolation period. England’s scheme gives £500 to those on benefits plus some discretionary money to local authorities to meet other needs. However, only one in eight workers is eligible, and of those, only 30% of those who apply are successful. If they do receive the cash, the sum amounts to less than the minimum wage over 10 days. Only one-fifth of those who apply for the discretionary funds are successful and in some areas (like the north-east of England), the figure falls to 10%. We are now realising that the impact of these barriers is more serious than we previously thought. The evidence is clear in the evaluation of mass testing that happened in Liverpool last November – the uptake rate in deprived communities was half that in more affluent areas. That was because if people tested positive they couldn’t afford to self-isolate since they would lose money and possibly their job. But equally, they couldn’t afford not to self-isolate, since that could land them with a £10,000 fine. The only way out was to not get tested. Similarly, if you provide the names of your friends or colleagues you put them in a similar bind, and why would you do that? Again, the solution is to avoid providing contacts. The failure to provide support for self-isolation doesn’t just affect self-isolation figures, it undermines every phase of the testing system. As a consequence, the argument for increasing provision becomes greater than ever. Without extra support, it is largely pointless to offer everyone tests twice-weekly. The policy might result in huge advantages to those who manufacture lateral flow tests, but is unlikely to deliver much advantage in terms of public health. Stephen Reicher is a member of the Sage subcommittee advising on behavioural science