Evaluation of telephone first approach to demand management in English general practice: observational studyBMJ 2017; 358 doi: https://doi.org/10.1136/bmj.j4197 (Published 27 September 2017) Cite this as: BMJ 2017;358:j4197
All rapid responses
This study has been widely reported as showing that telephone-first GP triage doesn't work. This interpretation is only valid if the questions the study failed to ask are ignored.
Some news sources reporting this study used headlines that were, in context, reasonable "Phone-first GP consults 'no panacea for reducing workload' (eg http://ct.moreover.com/?a=31981498170&p=1pl&v=1&x=0VZh2oPqbuCUQmJO_ZG4jQ) is a common phrasing and technically accurate even when the tone of the report itself is mostly negative. "Offering phone consultations as standard could increase GP workload" (see http://ct.moreover.com/?a=31977777830&p=1pl&v=1&x=DCJM9g3LznUWixkRF2YF4g) is slanted towards the negative but could be considered technically accurate. Most of the news reports tended towards skepticism that telephone-triage works in ways that were more negative than the headline. But the tone of most reports miss an important fact reported in the research: the variability of the results.
The research understates the significance of the large degree of variation in the results and fails to ask the obvious question: "why do some practices see large reduction in GP workload when others do not". In fact it would be equally valid to rewrite the negative Pulse headline as "Offering phone consultations as standard could DECREASE GP workload" as that is what the results say.
The BMJ editorial headline is relatively balanced "Policy makers should reconsider their unequivocal support for these systems" but also failed to ask the obvious question about why some practices saw large benefits and others did not. There is no point in either recommending an intervention unequivocally or condemning it as ineffective when you don't know why it sometimes works and sometimes does not.
I have a theory that might help with future research but would also help temper the negative reporting of the study. Telephone-first triage only works well if you manage it well.
Two factors matter a lot if the GP is to gain any benefit from lower workload. One is the length of the typical phone call (it needs to be much shorter than a face-to-face consultation. The other is the proportion of calls that need a face-to-face follow up. These are both strongly dependent on the skill of the GP (skilled practitioners have short calls and few follow-ups). While it is true that some patients need more time than others it is likely that, all other things being equal, a skilled GP will have much shorter calls and fewer follow-ups. A well-managed practice will monitor these metrics (as well as some quality metrics to ensure time isn't traded for poor quality decisions) and will optimise them over time.
This implies that phone-first triage by itself is not an effective tool for anything but that phone-triage plus competent management processes can yield big improvements. The study ignored this factor (or lumped it into the "more research needed" category). It also shows that lumping together the results by reporting the average across all participants is misleading (this is how the negative headlines are derived). A more accurate headline would be "some practices see big benefits while others do not".
It is also worth noting that some of the conclusions are very dependent on dubious data about GP workload. The largest supplier of GP clinical systems does not routinely report how long telephone-calls take making it essentially impossible to judge whether a system using many calls saves GP time or not. The second largest vendor can report this, but local setups often prevent the data being reliable. If GPs are to operate a phone-first process effectively they need to be able to record the length of calls reliably.
In short we should not dismiss telephone-first triage as a failure. Nor should we accept it as a magic-bullet. We should pursue an better understanding of why it sometimes yields large benefits and sometimes doesn't.
Competing interests: I provide analysis services to GP Access, one of the firms participating in the work described in the paper (though I was not involved in the study and was not providing services at the time it was completed).
The report by Newbould et al  is important as the telephone first system has been introduced into many general practices without a proper evidence base.
Writing as representatives of a major patient charity, we think the authors have not paid enough attention to the adverse effects patients experienced and the extra costs incurred by hospitals.
A 4% immediate increase in ambulatory care admissions (“conditions for which admission could, in principle, be avoided by good primary care”) and increasing thereafter is bad news for patients (Table 5). This fits well with the finding the ESTEEM study  that after doctor triage there was an increase of 22% in patients having to seek medical help out of hours . Telephone first essentially converts many GP consultations from face to face to telephone consultations and it is now clear that GPs are not as effective in these much shorter consultations where they cannot observe or examine patients .
Whilst good and bad comments from patients are reported, the overall ratings from all the patients in the intervention group are what matter. For these, despite a 20% gain in immediate access, “the GP communication composite” is significantly negative (P<0.001). “Would you recommend your GP surgery?” Significantly negative (P<0.01), and “Seeing preferred GP” is also significantly negative (P<0.035). And all this incurs extra costs for secondary care at the rate of £1.12 per patient registered with the GP. The last thing NHS hospitals want now is thousands of extra admissions and additional costs of about £50 million pa if applied nationally.
In general practice, as patients, we do regret the significant shortening of GP consultations as longer duration is associated with greater quality  and being more patient-centred 
Telephone first is a doctor dominant system that reduces patient autonomy, which is why patients see the GP of their choice significantly less often. It can force patients to discuss embarrassing symptoms on the telephone with a doctor they have never met.
Newbould et al. conclude that there is “clear evidence that a considerable part of the GP workload can be dealt with through phone consultations.” This statement could only be justified if increasing the proportion of telephone consultations led to outcomes as good as usual care, delivered at the same or lower costs and their own findings show that they are not.
N.A.P.P therefore opposes this system.
Competing interests: No competing interests