We derived a set of local authority-level indicators for each type of health visiting contact (Table 1). For local authorities with multiple quarters included in the study sample, we pooled data across quarters and then described the median and range of these indicators across local authorities. Most of the indicators require individual-level data to be derived and were therefore only available for the 57 local authorities in our study sample. Due to missing data, we could not derive indicators for some local authorities, as highlighted in Table 1.

We mapped the geographical distribution of the coverage of mandated contacts and plotted the number of additional contacts per mandated contact against the estimated number of children under five years in mid-2019 published by the Office for National Statistics (ONS) [22]. We used box plots to visualise variation in the medium (face-to-face or telephone) and locations (at home, health and social settings, or children’s centres) of health visiting delivery for both mandated and additional contacts across local authorities. All counts of contacts from CSDS have been rounded to the nearest 5 to comply with NHS statistical disclosure rules for subnational data [23].

Information on coverage of health visiting contacts was available for all 149 local authorities in England. Other contact characteristics were available for 57 local authorities with complete CSDS data on health visiting (the analysis dataset). These 57 included local authorities were similar to those not included, based on region, urban/rural status, and level of deprivation (Table 2).

There were 4,172,835 mandated contacts delivered to children aged 0-5 years between April 2018 and March 2020 in the 149 local authorities, recorded in the Health Visiting metrics (Supplementary Material A, Appendix Table 4). In the study sample for the same period, we included 1,779,155 contacts from 57 local authorities (Supplementary Material A, Appendix Table 4).

Coverage of the new birth visit was highest (median coverage of 98.5%, range: 89.9%–100.0%), followed by the 6-8-week review (median: 88.6%, range: 17.9%–100.0%) and one-year review (median: 88.8%, range: 19.9%–98.9%; Figure 1). The 2-212 -year review had relatively lower coverage (median: 81.4%, range: 33.1%–97.5%). For 29 (19.5%) local authorities, >20% of eligible children were not recorded as receiving the 6-8-week review. For 26 (17.4%) local authorities, >20% of eligible children were not recorded as receiving the one-year review. For 68 (45.6%) local authorities, >20% of eligible children were not recorded as receiving the 2-2½ -year review and in 5 local authorities, >50% of eligible children were not recorded as having the 2-2½ -year review.

Most local authorities (n = 46/57, 80%) reported conducting more additional contacts than mandated contacts, with a median of 1.6 additional contacts delivered per mandated contact and wide variation across local authorities (range 0.1–8.5, Figure 2). The number of additional contacts per mandated contact varied between 0.6–2.8 in 80% of local authorities, indicating that this variation was not due to a small number of outlying local authorities.

The majority (>96%) of mandated contacts were delivered face-to-face (Figure 3). Most additional contacts were also delivered face-to-face (median 82.7%, range: 47.6%–99.5%). About half of local authorities (n=23, 44.2%) offered >20% of additional contacts virtually. Phone calls were also used to deliver additional contacts: 34 local authorities (65%) delivered >10% of additional contacts through phone calls. Other mediums, including telemedicine, email, and messages, were rarely used for mandated or additional contacts, with <3% delivered through these mediums.

New birth visits and 6-8-week reviews predominantly took place in the home setting, but there was substantial variation in the location of one-year reviews, 2-2½ -year reviews and additional face-to-face contacts (Figure 3).

Mandated contacts were typically longer than additional contacts (Table 3), but high variation existed across local authorities. Face-to-face contacts were longer than phone call contacts. In half of the local authorities with available data on duration (25 out of 50), 65% of new birth visits were at least an hour long, but most local authorities delivered more than half of other mandated contacts in 30-59 minutes. A considerable percentage of mandated contacts were recorded as lasting under 30 minutes (median 4.6%–7.9%) or 30-44 minutes long (median 12.3%–29.6%). Additional contacts were generally short: on average, 36.2% of additional contacts were completed in under 30 minutes. The duration of both mandated and additional contacts showed large variations across local authorities.

Group sessions were rarely used for mandated contacts. Of the 48 local authorities with information on group sessions, only five (10%) local authorities delivered >2% of new birth visits, 6-8-week reviews, and one-year reviews in group sessions. The use of group sessions also varied by local authority and type of contact.

Between April 2018 and March 2020, the new birth visit was delivered with almost universal coverage but variation across England existed in the coverage of the 6-8-week, one-year, and 2-2½ -year reviews. The high coverage of new birth visits and lower coverage of 2-2½ -year reviews are consistent with previous studies using CSDS [13, 14]. However, our results also demonstrate that much of the reported activity was being delivered outside of mandated contacts, with 80% of local authorities delivering more additional than mandated contacts. The duration and way in which these additional contacts are delivered, for example by telephone or face-to-face, also varied across areas, which has not previously been described.

There have been many years of austerity, cuts and a depleted workforce: since 2015 the real term value of the public health grant (from which health visiting is commissioned) has fallen by 26% [24] and the health visiting workforce has decreased by 37% from 11,193 in 2015 to 7,030 in 2022 [25]. Despite this, during our study period (April 2018 to March 2020) health visiting teams reached nearly all babies and most children face-to-face, and conducted over one and a half times the number of additional contacts relative to mandated contacts. This represents a significant public health infrastructure to support the health and development of babies and children and the wellbeing of their families in the critical period before school.

Our study highlights the importance of taking into account additional contacts when measuring health visiting activity. Additional contacts, to address identified needs, were the most frequent type of contact (though shorter and more likely to be on the phone than the mandated reviews) in our analysis. This is consistent with other evidence about significant health visiting activity outside the mandated contacts in England, including a birth cohort study (of children born in 2020) which reported that 27% of 8,628 families had four or more contacts with the health visiting team in the first 9 months of their child’s life [26]. However, we found substantial variation in the frequency, duration and medium of additional contacts, which is unlikely to be completely explained by differences in underlying need of the families across areas. Other factors that may have contributed to this variation include local service specifications and service priority, budgets, personnel capacity, and the availability of other local services [27].

Other evidence suggests that despite substantial additional contacts, not all identified needs can be followed up by the health visiting team: in a survey of 1,186 health visitors, 45% reported that mandated contacts were prioritised over targeted or specialist support and 79% said that the service lacked capacity to offer a package of support to all children with identified needs [28]. Further qualitative work with parents and professionals is needed to understand the experiences and perceived purpose and impact of different types of additional contacts, as well as why models of additional contact are so different across areas.

Our results can inform conversations at the national and local level about the parameters of a ‘good’ health visiting service during a time when early years is a priority area, with targeted funding and innovation (e.g., Start for Life and Family Hubs [29]). For example, our findings raise a question about whether a mandated review can be adequately conducted in less than 45 minutes, given that it should include direct observation of the child and significant health promotion activity. We found that a median of 5%–8% of mandated contacts were recorded as under 30 minutes (Table 3), which should be treated with caution. This is because some of these contacts may have been mislabelled as mandated contacts but were in fact initial conversations to schedule or prepare for a longer contact outside of the 14-day window in which this mandated contact is supposed to take place. However, there was still a median of 12–30% of mandated contacts recorded as lasting between 30–44 minutes (Table 3).

Providers, commissioners and national policy teams should consider how far features of delivery such as <30 minute or <45 minute mandated contacts are consistent with the way health visiting is theorised to work. In other words, is there a minimum time feasibly needed for health visitors or members of their team to build relationships with parents (including those with a mistrust of services), work in partnership with families, undertake health promotion work and conduct a holistic needs assessment of the whole family and family environment? It is possible that some features of service delivery, such as the shorter mandated contacts, are a service response to resource pressure which have the unintended consequence of undermining the mechanisms by which health visiting is thought to work and influence family outcomes. The next steps in building the evidence base for national and local decision-makers include understanding local authority variation in the delivery of health visiting services to different types of families, including those living in deprivation or with other indicators of vulnerability, and the impact of decisions about service delivery on outcomes [30, 31]. As our study involved the use of data that were collected in the pre-pandemic era, we also now need to investigate how far the service delivery has ‘bounced back’ or has been permanently changed following the partial stop of health visiting services and widespread redeployment of staff and a change to virtual contacts [21, 32–34].

A strength of our study is that we used a subset of nationally representative data with high levels of completeness to monitor and measure the medium, location, and duration of health visiting contacts, the delivery of additional contacts, and the use of group sessions. For estimates of coverage of mandated visits, we used aggregate Health Visiting metrics that cover every local authority in England. For the other characteristics of health visiting service delivery, we only included local authorities that correlated with health visiting activity recorded in the Health Visiting metrics (analysis CSDS data). This means that the misclassification of non-health visiting activity as mandated contacts for records with missing values on “Activity Type” is likely to be marginal.

Data quality remains an issue. Firstly, we may not have identified all health visiting contacts in the CSDS, due to missing data in the “Service or Team Type” variable. This was likely a bigger issue for additional contacts because we had no alternative method of identifying these (unlike mandated contacts for which we could use the “Activity Type” variable) [19]. However, our study found a higher number of additional contacts than reported by other analyses of CSDS, which found 0.17 additional contacts per mandated contact for children under three years between April 2019–March 2020 (compared to 1.6 for children under five years in our study sample) [35]. This may be explained by different ways of cleaning and curating the data: our analysis data described health visiting on a subset of local authorities with more complete health visiting data, which may be associated with higher health visiting activity. There might also be misclassification between mandated and additional contacts: some local authorities offer more than five mandated contacts to every child and these extra “mandated contacts” will be identified as additional contacts in our analysis. Results about the use of additional contacts thus need to be interpreted with caution.

Secondly, we could only report on 57 local authorities included in our analysis and we cannot assume that these findings are generalisable to local authorities with incomplete CSDS data [13]. However, these local authorities were similar to the whole of England based on region, urban/rural status, and deprivation quintiles (Table 2). We included data at a local authority-quarter level, meaning our findings are largely based on a few months of data within each area, as more than half of local authorities (n=31/57, 54.4%) included in the study sample only contributed one or two quarters of data. The sample selection largely depended on good reporting of data, which might be associated with better service delivery. However, the coverage of mandated contacts is similar between our study sample and the whole dataset (Supplementary Material A, Appendix Figure 2) and the median duration of 45–60 minutes for mandated contacts and 30 minutes for additional contacts is consistent with the average duration of 42 minutes pooling mandated and additional contacts together reported by health-visiting providers in a survey [36].

Thirdly, we were unable to conduct longitudinal analyses in our study sample, as a result of using the quarterly Health Visiting metrics as a reference to assess the completeness of CSDS. This means we were unable to describe the average number of contacts per child or patterns of contacts over early childhood.

Finally, as this study describes different aspects of service delivery independently, the association between them is not clear. For example, we do not yet know whether local authorities with high coverage of mandated contacts had more additional contacts per mandated contact, or whether those who offered more additional contacts per mandated contact generally had shorter contacts. Further exploration of models of service delivery combining different aspects will help draw a more comprehensive picture of health visiting service delivery in England.