Whenever a child is born in England or Wales, whether the birth occurs in hospital or at home, the local health visitor will contact the mother and arrange to visit the mother and child at home a week or two after the birth. She will check the baby’s health and offer the mother advice on feeding, sleeping, and other aspects of care. She will usually follow this visit up with another, but then will concentrate on those who request or appear to need more attention. She will remain available for consultation until the child reaches school age. This system was set up in the early years of the twentieth century as a result of high and apparently stagnating infant and child mortality, and has continued to operate with only relatively minor variations for most of the ensuing century.

Maternal education through health visiting was established as a national strategy to combat high infant and child mortality in the early years of the century because of worries about race deterioration and the persistently high infant mortality rate (Dwork, 1987; McCleary, 1933; Szreter, 1996, chapter 5). The practice of health visiting was already well established in some areas of England and Wales at the turn of the century but was to become systematised and rationalised by the infant and child welfare movement (McCleary, 1933, chapter VI). The earliest organised health visiting scheme was set up by the Ladies’ Sanitary Reform Association of Manchester and Salford. It was originally staffed by volunteers but soon employed salaried visitors, and in 1890 it was taken charge of by the Manchester medical officer of health (MOH). Towards the end of the nineteenth century a number of other local authorities began to appoint women sanitary inspectors (or lady health missioners as they were called in Worcestershire), whose duties included health visiting. The first working model for systematised health visiting was established in Huddersfield in the early years of the twentieth century [1]. Early notification was the key issue, enabling the prompt visitation of babies in the first ten days when they were at their most vulnerable and when instruction would produce most benefits. A local Act of parliament was introduced to assure compulsory notification of a birth within 48 hours to the MOH. This proved to be highly successful and a similar Act relating to the entire country was passed in 1907 with the period of notification reduced to 36 hours (Dwork, 1987, 139). The legislation was initially permissive but was adopted by the majority of sanitary authorities, metropolitan boroughs and large towns, and in 1915 the Notification of Births (Extension) Act made early notification compulsory for the entire country.

Given that the mother and child health movement was one of the major initiatives of the early twentieth century in the battle against high infant and child mortality, there have been remarkably few attempts to quantify its success or failure. This can fairly be ascribed to the peculiar difficulties inherent in trying to measure its impact. At first sight, comparison between the numbers of health visitors in different districts and infant mortality rates would appear to provide a plausible way of measuring the impact of health visitors. A coherent source of information about health visitors is provided by a comprehensive report by the Carnegie Trust, published in 1917 (Hope, 1917). This report, based on information furnished by local Medical Officers of Health (MOHs), records the numbers of health visitors employed by each county and the number of visits made. However, since many of the health visitors were only partially engaged in maternal and infant welfare, and the proportion of their time was not stated, a cross-sectional analysis is unrewarding (Reid, 1997). Details of the percentages of births visited and re-visited in urban areas with populations over 20,000 are provided in a report to the Local Government Board in 1913 (Newsholme, 1913). This shows wide cross-sectional variation in visiting patterns, but gives little information on the relative development of services in the process of rapid expansion. Smith (1979, 114) concluded from the chronology that health visiting must have been important, but some of the claims made by the MOHs of temporal coincidences between improvements in infant mortality and the development of health visiting have since been discredited. The pioneers of the Huddersfield scheme claimed startling successes but Marland (1993, 45-8) has shown that this was a product of statistical acrobatics and that infant mortality did not seem to respond to the introduction of the scheme. Infant mortality showed no greater progress in Huddersfield after the introduction of health visiting than in England and Wales as a whole, and other towns even caught up with Huddersfield’s already low infant mortality rates. Attempts to link temporal trends in the level of health visitor provision with changes in mortality rates is also made difficult by the influence of other factors. Mein Smith (1993), for example, argued cogently that a simple temporal coincidence between the decline in mortality from infantile diarrhoea and the rise of the infant and child welfare movement is not enough to prove a causal link.

The data set is based on the records made by the health visitors employed by the Derbyshire County Council in conjunction with District Councils, who visited children born in rural and small town Derbyshire between the years 1917 to 1922 [5]. The women who acted as school nurses for the education committee and as health visitors for the TB committee were appointed as health visitors, “therefore obviating dual visits to the homes, preventing overlapping and saving travelling expenses” (Derbyshire MOH Report, 1915, 9).

The 1907 Act of Parliament introduced provision for health visitors to attend to the health and welfare of children under the age of five, but the application of the Act varied widely throughout the country. Peretz (1992) has shown that the provision of maternal and child welfare services was moulded by local political and social circumstances, giving rise to a wide variety of service provision, not only in extent, but also in philosophy and outlook. The percentage of births visited varied greatly (Newsholme, 1913, 106-117): 90 per cent of births in the Durham districts studied by Buchanan (1983, 285-6) were visited, while the figure was as low as 28 per cent in the Scottish districts. In Tottenham infants were visited repeatedly until the age of five, but in Merthyr Tydfil often only one visit was paid before the child started school (Peretz, 1992). It was understood, although seldom spelt out, that health visitors were for mothers in poorer neighbourhoods, not for those with a doctor, nurse or nanny (Marks, 1996, 174-5; Peretz, 1992, 107). In some areas targeting was overt and finely tuned, concentrating on the poorer classes. In the nearby city of Derby only a decade or two earlier, the six per cent of infants from whom no information was obtained by health visitors were reported to be the most affluent (Howarth, 1905). In Kensington, although between 75 and 80 per cent of births were deemed suitable for attention by health visitors, a means test was used to select suitable families and those revisited tended to be “those who were regarded as more ignorant and inexperienced, and where homes were seen to be more dirty and neglected” (Marks, 1996, 172-7). This is likely to have been the direction of bias everywhere, but the degree of targeting is likely to have varied from place to place, depending on resources and the concerns of the local authority.

Local MOH reports and committee minutes do not elicit much information on visiting policy in Derbyshire, and unfortunately there is not even a systematic record of how many health visitors were assigned to each district. Council minutes and reports do suggest that some district councils were anxious for more work to be done and agreed to pay a higher rate for the services of additional health visitors (Derbyshire MOH Report, 1916, 6). Furthermore, both the 1915 and 1918 MOH reports state that infant welfare work was to be concentrated on mining districts where the infant mortality was highest, despite the low proportion of women going out to work, the higher wages and better housing of coal miners. “Their wives”, wrote Dr Barwise in the 1918 report, “on whom the health of the children largely depends, have in the past received no instructions in home management” (Derbyshire MOH Report, 1915, 9; 1918, 38). In addition, health visitors were engaged in campaigns against specific diseases and conditions, such as opthalmia neonatorum. This is a condition of the eyes which develops when mucus is inadequately cleaned from the eyes at birth, and was responsible for 25-50% of infantile blindness in the early twentieth century (Buchanan, 1983, 259; Gale, 1945, 15). Sometimes babies died from it, and others were left with impaired sight. It was blamed on the competence of the midwife, and has also been associated with venereal disease in the mother (Cugali, 1984). Infants afflicted were in need of careful attention at their most vulnerable age. There was a specific campaign to reduce opthalmia in Derbyshire at this time; midwives were required to notify all cases, and were even paid 1 shilling per case for notifications received immediately (Derbyshire County Council Public Health Committee, 14 September 1915).

Measurement of visiting patterns is fraught with difficulty. A simple measure of the number of visits is inappropriate since higher risk children are likely to have had fewer visits because they died or moved out of observation for some other reason. The number of visits per month, a measure of the frequency of visiting which controls for the length of time in observation, is further complicated by the fact that visits were more frequent when children were younger. Children who died will therefore have had more visits per month on average than those who survived and were visited for longer, simply because their period of observation was restricted. For this reason, the analysis has been confined to the numbers of visits in three separate age groups, 0-6 months, 6-12 months and 12-24 months, considering only the children who continued to be in observation beyond the period in consideration [11].

Table 2 shows the average numbers of visits in the three age periods, according to the factors identified above (some other variables, such as sex, were investigated but found not to be significant). Children received an average of 1.96 visits during the first 6 months, 1.24 visits during the second six months and 1.87 visits in the following year. As expected, children in high risk or socially disadvantaged groups did receive slightly more visits, but by and large the differences were not large. The biggest difference was to children who suffered from opthalmia neonatorum, who received 6.53 visits in the first six months in comparison to the average of 1.96. Other differences were much smaller, although twins, illegitimate children, those with unsatisfactory homes and those artificially fed at a young age also received more visits than average. The working class bias was evident: the higher social classes (I and II) receiving significantly fewer visits than the lower classes in every age period. Interestingly, social class III received the most visits in every age period, and this can be attributed to the fact that miners were mainly placed in class III.

It has been concluded that there was some modest targeting by Derbyshire health visitors. As explained initially, it is not possible without a great deal more data to assess the actual impact of health visitors. It is however possible to assess whether the targeting was focused on the correct groups and, by assessing the nature of the advice, at least offer opinions about whether the health visitors themselves are likely to have made a difference. This can be approached using four questions. The first asks whether the health visitors targeted the right women, the second asks what advice was given, the third asks whether that advice was likely to have improved health and survival, and the fourth asks whether more targeting would have had a greater effect.

Table 5 shows a hazards model of post neonatal mortality [15]. This confirms that some of the factors which the health visitors appear to have been using to target their visits were those associated with higher risk of death. Twins, illegitimate children, higher parity children, those whose mothers had had a previous child death (indicated by the maternity history variable, for parities two and over only), the artificially fed, those in denser (urban) districts and mining districts were all at higher risk of death. It appears that the factors the health visitors appear to have selected as indicating higher risk of death were by and large correct. One major exception is class; although there the higher classes were more likely to survive, this advantage disappears once other factors were controlled. Although health visitors did concentrate more on the working classes, it may not have been correct to dismiss higher class families as “not necessary to visit”.

There is relatively little information on the mechanics and details of visiting, exactly what advice was offered and how, and the extent to which that advice was acted upon. Any discussion is bound to enter the realm of speculation to some degree.

It is difficult to speculate as to whether the actual advice will have had any specific effect on children’s health and mortality. Campaigns against specific diseases such as opthalmia neonatorum and long tube feeding bottles were likely to have had an effect, but it appears that health visitors were often stymied in their efforts to do good by the poor conditions mothers were living in which inhibited them from carrying out the advice (Marks, 1996, 267; Newsholme, 1935, 374). However it is likely that as Newsholme argued, the removal of ignorance may have made a significant contribution towards the reduction of infant mortality, but “only […] when the disabilities of the mother in the poor family […] are removed or reduced”(Newsholme, 1935, 374). Health visitors are therefore unlikely to have made major short-term changes, but the knowledge and awareness of health and hygiene they spread is likely to have interacted with other forces towards better health over the longer term.

Could health visitors have prevented more deaths had they targeted more effectively? To do so would presume that their knowledge could be used to combat specific causes of death or conditions. Newsholme’s sensible suggestion above, that it was not knowledge alone, but a combination of knowledge and improving conditions which decreased mortality, militates against this. It also assumes that a good reception from the mothers would have accompanied more efficient targeting, an assumption which is highly speculative. Health visitors occupied a difficult position as intermediaries between family and the official medical service, and there is evidence that health visitors were resented by some mothers (Lewis, 1980, 107). However other sources suggest that after initial resistance they were welcomed and listened to (King and O’Brien, 1995; Marks, 1996, 265-8; Peretz, 1992,107-10). Dr Sidney Barwise, the MOH for the County of Derbyshire, wrote (maybe optimistically) in his 1918 Annual Report: “It is extremely gratifying and fills one with hope to know that the Health Visitors have been extremely well received. The Health Visitors feel that they are the servants of the mothers of the County. It is a common thing for them to be stopped in the street by women who ask for leaflets which they think will be useful to friends and neighbours. The only difficulty that we have met with is that as the children grow bigger the mothers feel hurt that the more frequent visits which were necessary during the first year of life are not maintained.”

This examination of the targeting practices of health visitors in Derbyshire has demonstrated that there was some modest targeting on factors which were generally linked to higher mortality. Specific medical conditions such as opthalmia neonatorum overshadowed other indications of high risk and social selection. The latter produced no greater differentials in the number of visits than resource driven constraints governed by local circumstances. The fairly universal nature of the health visiting service in Derbyshire does not necessarily detract from the value of health visitors in improving and maintaining children’s health, however. This is because their impact was less connected to averting individual deaths than improving knowledge about child care and hygiene which could interact with independent improvements in sanitation, housing and the environment to improve health in the long term.