"Insomnia sufferers can benefit from therapy," reports The Guardian.
US and European guidelines already advise GPs to refer patients with insomnia for a type of talking therapy called cognitive behavioural therapy (CBT) before trying sleeping pills.
But until now there's been insufficient evidence about how well CBT works in primary care settings, such as GP surgeries, rather than in specialist sleep clinics.
Researchers reviewed studies looking at the effect of cognitive behavioural therapy for insomnia (CBT-I), a type of cognitive therapy specifically tailored to tackle sleep problems, when delivered in primary care.
CBT-I involves 4 to 6 sessions with a therapist teaching 3 tactics to improve sleep:
Researchers found 13 studies, involving more than 1,500 patients, which overall reported "medium to large" positive effects on how well people said they slept.
Time taken to get to sleep and time spent lying awake after waking in the night showed the biggest improvements.
Difficulty with sleep over the long term is linked to problems such as depression.
Find out more about sleep and tiredness
The researchers who carried out the study were from Queen's University in Canada. They had no specific funding for the study.
It was published in the peer-reviewed British Journal of General Practice.
The Guardian and the Mail Online both carried accurate and balanced reports of the study.
This was a systematic review of randomised controlled trials (RCTs) and case series.
RCTs are the best type of study to show whether a treatment works, and systematic reviews are a good way of summarising the research into a particular topic.
Case series are less reliable than RCTs.
Researchers looked for studies published between January 1987 and August 2018 that reported the results of CBT-I in a general population of adults.
They reviewed and summarised the results, commenting on the strength of the studies, differences between them, and whether or not they also included help for people wanting to stop taking sleeping pills.
They chose to include not just RCTs, which are the standard way of testing treatments, but also "before and after" case studies, which look at what happened to people before and after being offered treatment, but do not randomly assign people to treatment.
The RCTs included control groups, in which people were put on waiting lists or given advice about how best to improve sleep, rather than CBT-I.
The researchers found 10 RCTs including 1,418 people, 2 case series of 96 people and 1 RCT of 80 people where the doctors, rather than the patients, were randomly assigned to refer patients for the treatment.
CBT-I was mostly carried out in these studies by a nurse, psychologist, counsellor or social worker.
In the studies that used CBT-I, 4 looked at adults of mixed age and 4 at older adults.
The 3 mixed aged studies that used variations on CBT-I, such as only 2 sessions, use of a self-help book or sessions that included stretching exercises and sleep education, showed smaller sleep benefits.
The case series showed large sleep benefits from CBT-I.
The researchers said: "The results of this review provide evidence that CBT-I (group or individual) is effective at improving sleep onset and maintenance in patients in primary care with chronic insomnia."
They added that "the best methods of integrating CBT-I in primary care services need to be identified", but said it was likely to involve interdisciplinary team working.
This study suggests that CBT delivered in primary care does indeed help people with insomnia get to sleep more quickly and spend less time lying awake after waking in the night.
These effects last for several months to a year.
But the study has some limitations. Case studies are less reliable than RCTs. Also, the study only looked at night-time symptoms of insomnia, so we do not know whether people having CBT-I felt less tired during the day.
There were also no details about the overall length of time people were asleep, which did not change with CBT-I. This and quality of life would have been important outcomes.
And the study assumed that people had been checked for other causes of sleep problems, so we do not know whether the results would apply for people who had not had other sleep problems ruled out.