The share of home visits in all NP consultations was 17% in Practice A and 51% in Practice B. Both NPs had a higher share of autonomously conducted consultations during home visits than in the office. In Practice A, the NP was completing her continuous postgraduate education, and the proportion of consultations in which she was autonomous increased from 0% in the first month of her employment to 19% after 13 months of GP supervision. In Practice B, the NP had completed her postgraduate education and had been working for two years as a NP before data was collected. On average, she was autonomous in about three-quarters of her consultations.
In the 1940s, home visits accounted for 40% of GP consultations in the US whereas nowadays, this proportion has decreased to less than 1% . In Switzerland, most GPs still provide home visits, but the total numbers are relatively low compared to other European countries and have declined in recent years [3, 27]. Mueller et al.  found that home visits accounted for 2.5% of all GP consultations in the Canton of Vaud between 2006 and 2015. Our results showed that home visits represented 17% and 51% respectively of all NP consultations. This might indicate the potential of NPs to counter the decreasing numbers of GP home visits in Switzerland. This assumption is supported by data from countries at a more advanced stage of NP role implementation. In the US, for instance, around 3300 NPs performed over one million home visits (accounting for 22% of all visits) to Medicare beneficiaries and thus more than the physicians did in 2013 .
The difference between the two NPs included in our study most likely reflects their unequal level of training and experience. The numbers might also be influenced by the different approaches and stages of the two projects and the demand for home visits. In a study from the Netherlands, Dierick-van Daele et al.  found that the extent to which NPs could focus on home visits was influenced by the needs of the individual family practice as well as financial incentives. Moreover, the attitudes of patients, GPs and nursing home staff towards the NP role may have affected the numbers. Literature shows that patients usually report high satisfaction with NP home visits and care [4, 29, 30]. However, the acceptance of the NP role among health professionals can be ambivalent, especially in the early stages, when role clarification is not yet achieved and the NP might be perceived as a threat that could take over tasks traditionally attributed to other established professionals [31, 32].
Mueller et al.  also found that about 85% of GP home visits were routine follow-up visits to people aged 65 years and older, a group of patients who often suffer from multiple chronic diseases. In a systematic review, Martinez-Gonzalez et al.  reported that nurses in advanced roles in primary care are able to provide at least equivalent care for patients with chronic conditions compared to doctors. A Swiss study from Imhof et al.  showed that home visits to elderly people (age 80+) provided by nurses in advanced roles appear to be effective and could lower the number of falls, acute events, and hospitalization. However, the nurses included in this intervention study received a specific training to provide in-home consultations and stem from different health care settings (e.g. home care, nursing homes).
In Practice A, clinical supervision was provided by the two GPs as part of the structured preceptorship in the NP’s postgraduate education. In this context, the proportion of the different levels of supervision over time could be interpreted as a learning curve. The initial period of three months, in which the NP was mainly observing and being observed, might reflect the lack of her experience in primary care as well as the unfamiliarity of the GPs with her role. After this initial phase, the NP’s autonomy increased more or less steadily but in most cases, she still consulted the GP before making a final decision. Consulting the GP before discharging the patient may have various reasons. It is possible that the NP wanted the GP to confirm her findings (“second look”) or to teach her a certain procedure or examination as she was still in training. This level of supervision may have also appeared due to legal restrictions in the NP’s scope of practice. For instance, Swiss NPs are not (yet) allowed to prescribe drugs independently without consulting a GP. In this regard, Barnes et al.  found that restrictive state regulations in terms of scope of practice as well as reduced reimbursement rates decrease NP participation in primary care.
In Practice B, the NP showed greater autonomy compared to Practice A. On average, she was autonomous in about three-quarters of all her consultations. This might be because the NP was more experienced as she had been working as a NP in this family practice for two years when the data was collected. Furthermore, the GPs were used to working with other non-medical health professionals and to handing over tasks. In this practice, the NP often did not consult the GP directly during the consultation but at an earlier or later point and reported about several patients collectively. This appears as a more efficient and less direct approach of GP supervision. The results from Practice B show the potential autonomy of a well-trained, more experienced and well-integrated NP in a Swiss family practice despite lacking regulations.
The NPs in our study showed higher autonomy during home visits compared to office-based consultations. This might be because direct GP supervision on home visits is more difficult and requires additional effort; time- and site-wise. It is also possible that many of the home visits were routine follow-up visits to chronically ill elderly, as opposed to consultations in the practice, where younger people with acute, purely medical problems may show up. Lastly, it is worth mentioning that a share of 100% autonomous consultations is neither realistic nor desirable as the exchange between health professionals appears to be important in complex cases .
There are several limitations in our study. The NPs collected data on their own activities; hence, we cannot exclude self-report bias. Our results might not be generalizable as we only gathered data from two practices. Furthermore, we could not differentiate between home visits at patient’s home and in nursing homes, and we did not collect data about patient characteristics or the content of these visits. It was also not possible to identify the reasons why which level of GP supervision was applied in which occasion. Lastly, the duration of data collection in Practice B was much shorter compared to Practice A.
In this pioneering phase, our method allowed to get first data on NP home visits and autonomy. Yet, separate identification numbers (or global location numbers) for NPs could help to conduct studies on a bigger scale. In countries in which NPs are well-established and registered providers, studies on NPs are usually based on billing or health claim data [12, 23, 24, 36]. In order to promote the NP role in Swiss primary care, educational programs need to include preceptorships with sufficient numbers of supervised, clinical learning hours. This requires closer collaboration with family practices and might need additional efforts from the GPs. However, NPs could help reducing the GPs’ workload, e.g. by offering time-consuming home visits to multimorbid elderlies. In any case, more research is needed in order to explore patient characteristics and content of NP consultations, and to specify the NP role in Swiss family practices. Lastly, the safety and quality of care provided by NPs has to be addressed by looking at specific health outcomes.