A closer look at relationship continuity and the GP Patient Survey
According to reports published today, patients in England are finding it harder to get to see their own GP. Various media sources have reported a ’27.5% fall’ in “the number of patients… able to get an appointment with their [GP]“ between 2012 and 2017. From a cursory scan on Twitter, this seems to support the anecdotal findings of many people using primary care: that it feels hard to get appointments, and that long waits are becoming more common. But is it true that access to preferred GPs has fallen so dramatically, and, if so, what does this mean for the quality of care in general practice?
Today’s media coverage is based on a new paper from Dr Louis Levene & colleagues published in the British Journal of General Practice[i]. Their study is based on secondary analysis of data collected via the GP Patient Survey (GPPS)[ii] from 2012 to 2017 inclusive. GPPS is a very large dataset and should provide good insight into trends over time, and the study uses good quality statistical methods to address potential confounding factors like differences in practice populations.
Despite the strengths of the study, there are some problems in the reporting of the results. The biggest issue is in the measure of ‘relationship continuity’ used in the study. This is based on the product of two separate GPPS questions:
- Q8 – Is there a particular GP you usually prefer to see or speak to?
- Q9 – How often do you see or speak to the GP you prefer?
The second question is only answered by patients who said that they did have a preferred GP. In 2017, for example, 46.2% of people said that they had a preferred GP; of these, 55.6% said that they saw or spoke to that person ‘always’, ‘almost always’, or ‘a lot of the time’. Multiplying those two figures together gives a “patient-perceived relationship continuity” of approximately 25.7%[iii].
The trouble here is that combining these two items conflates what patients want – ie whether or not they have a preference (Q8) – with what they get – ie how often they see a preferred GP (Q9). The resulting measure tells us the proportion of patients who both had a preferred GP and were able to see or speak to that person most of the time. The proportion of people in this category has fallen by 27.5% from 2012-2017 – but this is only partly down to people not getting what they want.
From 2012 to 2017, Levene et al report (p3; table 1) that the proportion of patients with a preferred GP fell from 56.2% to 46.9%: a change of 9.3% points or a proportionate decline of 16.5%. The proportion of these people who were able to see that GP at least ‘a lot of the time’ also fell, from 65.2% to 55.5%: a 9.7% points or 14.9% proportionate fall. Combining these figures creates a much larger reported fall in “patient-perceived relationship continuity”, particularly when the proportional change is reported instead of the absolute change[iv].
Reviewing the results for these two questions separately presents a more nuanced picture. Demand for access to a particular GP appears to have fallen over time, and this is not necessarily a bad thing; seeing the same professional time and time again is only one way for practices to deliver continuity, and decreasing demand may indicate that patients feel more comfortable that their needs can be consistently managed by the entire practice team.
Continuity of care is sometimes described as comprising three types of continuity – not just relational but also management and informational continuity[v] – and improvements in those other elements or in the way practice teams work together with a common goal and purpose may decrease reliance on the relationship between individual patients and GPs.
So, has the proportion of patients that are able to get an appointment with their GP fallen by 27.5% from 2012 to 2017? No: this is the wrong way to describe the statistic reported, and as a consequence it overstates the problem. A more readily interpretable measure is derived by multiplying the proportion of people with a preferred GP by the inverse of Q9 – that is, the proportion who cannot see their preferred GP at least ‘a lot of the time’. This gives the proportion of people who have a preferred GP and can’t see them regularly. This figure has risen from 19.6% in 2012 to 20.9% in 2017: a 1.3% point or 6.7% proportional increase in the percentage of people who cannot see their preferred GP at least a lot of the time. The fact that we can produce such different answers depending on exactly how we describe the statistic show how much of a need for caution there is in reporting this kind of data. For reporting to be accurate, it is vital that the statistic is described clearly and precisely.
Although the scale of change is not as dramatic as has been reported, any increase in this alternative metric is cause for concern – and Levene et al’s paper is an important study highlighting a worrying trend. The findings show that nearly half of us still prefer to see a particular GP, and the likelihood of being able to see that preferred GP where it matters is decreasing. The net effect is that around one in five of us have a preferred GP who we can’t regularly see.
For whatever reason, then, some patients are finding that their preferences are less often met. Not only is this contrary to the spirit of person-centred care, but good relationship continuity has been found to be associated with better satisfaction, reduced costs, and better health outcomes[vi]. Whilst pressures on primary care providers are likely to continue, ensuring high quality, person-centred care in general practice will require a focus on allowing patients to access care in the manner they prefer.
[i] Levene, L. S., Baker, R., Walker, N., Williams, C., Wilson, A., & Bankart, J. (2018). Predicting declines in perceived relationship continuity using practice deprivation scores: a longitudinal study in primary care. British Journal of General Practice, bjgp18X696209. https://doi.org/10.3399/bjgp18X696209
[iii] Please note that the figures in this paragraph are based on the 2017 GPPS National Report; they differ from figures quoted in Levene et al and elsewhere in this blog because they cover different populations (Levene et al exclude single-GP practices, for example) and because of differences in modelling.
[iv] This figure has declined from 37.5% in 2012 to 27.2% in 2017 – a 10.3% point drop, representing a 27.5% proportionate change over time.
[v] eg Haggerty, J. L., Reid, R. J., Freeman, G. K., Starfield, B. H., Adair, C. E., & McKendry, R. (2003). Continuity of care: a multidisciplinary review. BMJ : British Medical Journal, 327(7425), 1219–1221.
[vi] Freeman, G., & Hughes, J. (2010). Continuity of care and the patient experience. London, UK: The King’s Fund.