Summary of main findings
General practices in England can be classified into three groups: (1) Small and GP reliant (2) Medium size with a multidisciplinary team (MDT) input, and (3) Large and multidisciplinary. The majority of practices in England are still small and reliant on GPs.
Large and medium size practices perform worse on all patient reported indicators except confidence and trust in healthcare professionals where although medium-size practices with MDT input appear to do better than small GP-led family practices the effect size is small.
Groups performed similarly for incentivised clinical indicators. Medium-size practices performed better on immunisation coverage and worse in asthma reviews compared to small practices, the effect sizes were small, and the association does not extend to large multidisciplinary practices, as one might expect.
Considering non-incentivised clinical indicators, larger practices do better at diagnosing cancer earlier compared to medium and more so smaller practices as measured by the proportion of cancer cases treated that were detected via the two-week wait pathway.
Strengths and limitations
This is the first study in England to have used finite mixture modelling to group practices into different organisational models based on list size and workforce composition and assess the effect of these different organisational models on practice performance. It represents a departure from previous studies where practice size was defined in terms absolute list size, list size per GP or as single-handed versus multiple-handed practices.
We measured practice performance using diverse outcomes ranging from patient reported to clinical and preventive care indicators. There was some uncertainty in class membership, especially for the medium-size practices with MDT input group which had 15% of practices assigned to it with probability < 0.8. This was mitigated by weighting the regression analysis by class membership probabilities.
In the 2022 General Practice Patient Survey, only 29% of targeted participants responded, suggesting significant risk of selection bias, although our results are consistent with other similar studies(11,13,27–30).
We did not control for other confounders in primary care such as prevalence of chronic diseases, patient turnover and fraction of migrants. Nonetheless, previous research demonstrated that these have no effect on clinical outcomes as measured by practice QOF points [26].
Capturing practice level workforce composition is complicated by a number of roles that are employed at PCN level as specified in the Additional Roles Reimbursement Scheme (ARRS) [27]. Staff employed at PCN would not be reported as practice employees in the datasets used in this study, despite working in practices and contributing significantly to the pattern of the workforce. The ARRS roles make-up a significant proportion of the non-GP workforce and future research would be strengthened by inclusion of this data.
Results in relation to other studies
Similar to our findings, previous studies have generally reported that smaller practices outperformed larger practices on patient reported healthcare quality metrics, irrespective of how practice size was defined [11, 13, 28–31].
For clinical outcomes, previous research favours larger practices. Group practices achieved higher QOF points than single-handed practices [26]. Larger practices also had better diabetes control [10, 11, 29], vaccination rates [32], cancer screening [33], depression reviews [34], antibiotic use [35], specialist referrals [36] and use of clinical guidelines [37] than smaller practices. However, no difference was found between smaller and larger practices on blood pressure and cholesterol control [38], use of diagnostic investigations [34, 39, 40] or medication prescription [34, 38–40]. We did not find compelling evidence for better clinical outcomes in larger multidisciplinary practices except that larger multidisciplinary practices appeared to do better at recognising cancer symptoms earlier and referring patients to specialists sooner. This discrepancy may be because most of the quality indicators we used are financially incentivised.
Continuity of care has been associated with better clinical outcomes, especially in chronic diseases such as hypertension[41] and diabets [42]. It is also associated with fewer emegency room attendances [42, 43], fewer hospitalisations [42, 44], high uptake of immunisations [45] and low mortality [42, 46]. It is belived that this is is the case because continuity leads to doctors accumulating more knowledge about their patients and their condition, and develop a sense of responsility towards them which in turn leads to more personalised care [47]. This was not reflected in this study. We believe this has to do with how continuity has been conceptualised. Traditionally, continuity has been defined as repeated contacts with the same doctor over time [48]. Consequently, in GPPS, respondents were asked how often they saw their preferred GP. But chronic disease care is often provided by a multidisciplinary team of practitioners, including nurses and pharmacists, and relationships are built with teams not individuals. Perhaps an alternative definition of continuity that includes nurses and AHPs might capture the relationships built with other practitioners and better reflect the impact on clinical outcomes.
Implications for practice and future research
The lack of significant differences found in clinical outcomes between large, medium and small practices may be a reflection of the fact that large multidisciplinary practice models are newer and are yet to start reaping the benefits of working at scale. Longitudinal studies to assess whether changes in practice’s organisational structure over time produce incremental gains in key indicators would be beneficial. Further, we only included practice-level workforce data, future studies with PCN-level are needed.
More work is needed to accurately code work by different staff groups. Current electronic health record documentation does not distinguish between consultations by GPs and those delivered by other staff [49]. Improving documentation of the activity of these new practitioners is needed to better understand their scope of work and impact in primary care. It is using the wider multidisciplinary team more effectively that has the potential to increase access and provide longer appointments, which are associated with increased satisfaction and positive clinical outcomes [11, 29].
In addition, more understanding of whether different practitioners are being utilised effectively is needed because as new roles evolve there is potential for challenges of integration into the existing primary care team [50]. It is important to clearly define their scope of work and need for supervision so managers can monitor and optimise the working environment for all staff.
Furthermore, to produce optimal results large multidisciplinary practices will require substantial financial and infrastructural investments (estates, medical equipment and information technology) [51].