General Practice

Morbidity and healthcare utilisation of children in households with one adult: comparative observational study

BMJ 1998; 316 doi: (Published 23 May 1998) Cite this as: BMJ 1998;316:1572
  1. D M Fleming, directora (Bill{at},
  2. J R H Charlton, statisticianb
  1. a Birmingham Research Unit, Royal College of General Practitioners, Birmingham B17 9DB
  2. b Office for National Statistics, London SW1V 2QQ
  1. Correspondence to: Dr Fleming
  • Accepted 11 November 1997


Objective: To identify and consider differences in morbidity in children in households with one adult presenting to general practitioners compared with children in households with more than one adult.

Design: Observational study; data analysed with logistic regression controlling for age, sex, and practice.

Subjects: 93 356 children aged 0–15 years included in the fourth national study of morbidity in general practice and for whom data about household structure were available. Among them 10 983 (11.8%) were living in households with a sole adult.

Methods: Morbidity data were recorded from each consultation as the assessment diagnosis made by the general practitioner.

Main outcome measures: Number of consultations and consultations per person for any illness, infections, acute respiratory infections, asthma, and accidents; number presenting and mean consultations per person for immunisation; number receiving home visits and home visits per person visited; average annual frequency of consultation among those consulting.

Results: Compared with children in other households, a higher proportion of children in households with one adult consulted for infections and accidents. The proportion consulting for immunisation was lower and the proportion receiving home visits greater. Mean numbers of consultations per person consulting were also generally higher for all conditions. For infections, accidents, and home visits, the differences were evident in all age groups.

Conclusions: The study confirms the importance of single parent families as an indicator of deprivation. Children in such families should be targeted for immunisation and accident prevention.

Key messages

  • Children in households with one adult consult general practitioners more frequently than those in households with two or more adults and receive increased numbers of home visits

  • They are more likely to consult for accidents, and they attend less frequently for immunisation

  • Single parent households are an appropriate indicator of deprivation

  • Children from households with one adult require specific targeting by general practitioners, health visitors, and primary healthcare workers for accident prevention and immunisation uptake


Between 1961 and 1994, the proportion of households made up of a lone parent with dependent children increased from 2% to 7%, and the number of households increased from 16.2 million to 23.1 million—thus the actual number of households with lone parents and dependent children in Britain increased fivefold.1 In 1991, 19.4% of children were living in a one parent family situation, mostly (18%) with their mother. There is an excess of lone parents in black African and Caribbean ethnic groups, in social class III non-manual (assessed in women), and in metropolitan areas such as Greater London, south Wales, and the western part of Scotland.1

The Committee on One Parent Families identified finance and housing as major problems2: Bradshaw noted that two thirds of lone parents received supplementary benefit.3 Studies of the health of children in single parent households have generally found that it differs little from the health of children in two parent situations.47 However, a common definition of single parent household was not always used. Behavioural problems,8 accidents,911 and non-accidental injury12 have been found more commonly in children of single parent households, and an Australian study reported reduced rates of polio immunisation.13 Roberts and Pless drew attention to the twofold difference in rates of injury between the children of lone mothers and those in two parent households and related this difference to elements of social deprivation.10

Kai studied 95 parents of preschool children, including 29 sole parents in a disadvantaged inner city community, and drew attention to parental anxiety about the gravity of feverish illnesses.14 The impression of general practitioners that young single mothers bring their babies more readily to the doctor with comparatively minor problems was part of the consensus assessment which led to the establishment of the Jarman index as a determinant of a deprived area.15 This paper examines some of the problems for health and health care delivery associated with childhood in a single parent household and assesses both the needs of such children and the implications for general practitioners.


Data collected in the fourth national study of morbidity in general practice were used.16 The participating general practitioners and practice nurses in 60 practices recorded their assessment of the problems at each face to face encounter between September 1991 and August 1992. The problems were entered onto the practice computer, using conventional medical terms, and stored as Read codes.17 A consultation or episode type was assigned to each entry, distinguishing “first ever” diagnoses, “new” episodes of illness, and “ongoing” consultations. The total study population of approximately half a million was representative of the national census population by age, sex, marital status, tenure of housing, economic position, occupation, and whether they lived in an urban or rural area. There were small differences in distribution by social class and by ethnic composition.16

Socioeconomic data were collected by trained field workers.18 For children under 16 years of age, data obtained included housing tenure, ethnic group, country of birth, whether they were living with one or more adults, economic position of parent one year ago, and current or most recent occupation and employment status of parent. These data were obtained in a single interview from each person registered during the course of the study year. In most cases the data for children were provided by the mother; for a minority they were provided by the father and occasionally by a grandparent. Unlike the census, the definition of social class was not restricted to people who had been employed at some stage during the past 10 years; thus, more people were assigned social classes in the survey than in the census. The occupation of the head of the household (usually the mother when the household had only one adult and the father in other households) was as described by the respondent. Answers applicable at the time of interview were applied to the data for the entire year.

For this study, comparisons were made between children (aged <1 year, 1–4 years, and 5–15 years) living in a household with two or more adults (other household) and children in households with only one adult. Children were counted on the first occasion they consulted for the specified reason, and rates were expressed per 10 000 person years at risk, calculated from the number of days each child was registered in the practice during the survey. Children were grouped by age at the midpoint of the study; hence those aged 17 months or less at the end of the study year were aged less than 12 months at the midpoint of the study and were included in the group <1 year.

Comparisons were made for children consulting with: any illness (international classification of diseases ninth edition, chapters 1–17 inclusive); infectious diseases (ICD chapter 1); acute respiratory infections (ICD numbers 460-466); asthma (ICD 493); accidents excluding medical misadventure (ICD E800-E869 and E880-E949); immunisation (ICD V03-V06); and home visits.

Separate analyses of differences in terms of odds ratios were made using logistic regression. 19 20 The reference population was children living in households with two (or more) adults and the analyses took into account age (month of birth in those aged <1 year), sex, and practice. Odds ratios and 95% confidence intervals were derived relative to the reference populations. In further regression analyses, we included urbanisation (derived from patients' post code), ethnic origin, social class, housing tenure, and distance to practice (distance between patient's and practice postcodes treated as a continuous variable). For reference purposes, the index population was white, living in an urban area, social class I or II, in owner occupied housing. Odds ratios (and confidence intervals) were derived for the effect of each factor independent of all others.

We also calculated the average annual number of consultations per child for each condition, and the 95% confidence interval, and compared these for children in households with one adult and those in other households. After normality of distribution was tested for, differences were evaluated by t test.


The study population included 93 356 children aged 0–15 years for whom we had relevant socioeconomic information: 10 983 (11.8%) were living in households with one adult. Table 1 summarises the distribution by ethnic group, social class, and housing tenure. The proportions of children in households with one adult were greatest among people of black (African or Caribbean) origin, in social class IIIN, and living in council housing.

Table 1

Population by urbanisation, ethnic group, social class and housing tenure

View this table:

Table 2 shows rates for children consulting with illnesses, reporting accidents, receiving immunisation, and visited at home. Rates for any illness in children under 1 year exceed 10 000 per 10 000. This apparent anomaly relates to the use of a denominator based on person years at risk. New babies frequently attend the general practitioner soon after birth and on average would only be at risk for 6 months of the study year. Rates for any illness were slightly greater in children of all ages living in households with one adult, but rates differed for infections (ICD chapter 1), acute respiratory infections, and asthma. Rates for asthma in boys exceeded those for girls.

Table 2

Rates of children consulting (per 10 000 population) for selected illness groups, for accidents, immunisation, and receiving home visits: by sex and age groups for children in households with one adult and other households

View this table:

Rates for children consulting with accidents were considerably higher in households with one adult: for boys under 1 year they were 50% higher, and for girls they were 35% higher. In all age groups, accident rates for boys were higher than those for girls. Boys under 1 year and aged 1–4 years and girls aged 1–4 years were less likely to present for immunisation if living in households with one adult. Overall, one third more children in households with one adult were visited at home.

The primary regression analyses (with adjustments for age, sex, and practice only) showed that children in the three age groups in households with one adult were more likely to present with infections, more likely to present with accidents, less likely to present for immunisation (except age group 5–15 years), and more likely to have received a home visit (table 3). Odds ratios were generally similar to those derived using the fuller analysis model, with the exception of the result for immunisation in children <1 year, where the odds ratio in the fuller model was 0.96 (95% confidence interval 0.69 to 1.34). Residence in council housing was the single most important adverse factor for achieving immunisation (0.54; 0.41 to 0.71). Social class and ethnic origin by themselves were not associated with poor immunisation uptake.

Table 3

Odds ratios (95% confidence intervals) of children from households with one adult (relative to those in other households) presenting to general practitioner. Values are adjusted for age, sex, and practic

View this table:

To assess the impact on general practitioners' workload, we examined the mean numbers of consultations per child for each condition studied (table 4). We first checked to ensure that there were no important differences in the registration period of children in households with one adult and of those in other households. Differences in mean numbers of consultations were mainly found among children aged under 1 year and included increased mean numbers of consultations for any illness, acute respiratory infections, and home visits. Decreased mean numbers were found for immunisation; mean numbers for infections and for accidents were similar.

Table 4

Consultations per person consulting for selected illness groups, accidents, immunisation, and receiving home visits by sex and age group in households with one adult and other household group

View this table:



This study has shown higher rates for children consulting and increased mean numbers of consultations per child consulting by children in households with one adult compared with children in other households, particularly in the first year of life. Increased rates of home visits were seen throughout childhood. A household with one adult (“sole adult household”) is not quite the same as “single parent family,” though the implications of the findings of this study are the same. The interviewers were members of practice staff seconded to the study and familiar with the household composition of many of the families involved. It is unlikely that bias could be introduced from variations in the responses of interviewees from the two household groups. Some children effectively were reared by a sole parent but because they were living in households with other adults (in a grandparental home, for example) they were included in the “other household” category.

Practices recruited to the study were well distributed geographically and by practice characteristics; objective recording lasted 12 months; the population was reasonably representative of the national population by age and sociodemographic characteristics16; large numbers of children (95 000) were included. For these reasons we believe the findings can be generalised to England and Wales.

In spite of very detailed information available for each child, we cannot standardise for variables such as passive smoking, sibling order, or number of siblings, all of which might influence results for respiratory disease. One conclusion from this study therefore is a plea for the use of a family or household identifier in any future major morbidity survey.

The study included children for whom household data were available, in total exceeding 95% of the entire childhood population surveyed. As Judge and Benzeval have pointed out, the group of children with an unoccupied social class status is dominated by children of lone mothers and carries twice the risk of death by accident.11 Social class was not used as a variable for the primary analyses of the data because of the limitations of using the occupation of the head of household as a determinant of social class, dependent on which parent is heading the household.

Smith reported rates of pregnancy in young women (aged under 20) six times as high in the most deprived areas as in the most affluent.21 One in four teenage pregnancies in deprived areas ended in abortion, compared with two in three in the most affluent areas. These studies suggest that if there is any selection bias in our study, it tends to underestimate the problems of children in sole parent situations.

Children in households with one adult consulted more frequently, which may reflect the insecurity of a sole adult with no opportunity to share responsibility for a sick child, substantiating the observations of Kai's focus group study.14 There is also the transference effect whereby stress in one person (the sole parent) is manifest in problems presented by those immediately around him or her (the child).


The number of children presenting with accidents was greater in households with one adult regardless of age, as in other studies. 911 22 In contrast, the results for immunisation indicate that fewer boys were brought for immunisation; among those who came there was a reduced mean number of consultations, suggesting that fewer completed the immunisation course. However, these results should be seen in the context of other relevant factors such as residence in council housing. The interrelationships between poverty and health contain several components, among which housing is perhaps the most significant. 2 3 The links between low income and poor housing on the one hand and childhood accidents on the other have already been identified.10 It is not simply a matter of the type of housing; access to nursery places and child care may be equally important. Supportive care for socially disadvantaged people has been shown to improve pregnancy outcome as measured by birth weight.23

The study results support the inclusion of single parent status as a determinant of deprivation for providing income supplements for general practitioners.15 Children from households with just one adult create extra work for doctors, especially extra home visits, and they are more difficult to immunise, thus making it more difficult for general practitioners with large numbers of these children to achieve immunisation targets. The higher rates for home visiting bear on arrangements for out of hours primary care surgeries, especially because a “sole adult parent” is less likely to have a car or carer available for other children.

These findings provide a challenge to society and the health service. The prevention of accidents is a key area in the Health of the Nation strategy.24 Carter and colleagues examined general practitioners' attitudes to preventing injury in children and felt that members of the primary care team might do more towards preventing injuries during suitable consultation opportunities.25 Roberts and colleagues reviewed 11 randomised control trials of home visiting programmes and concluded that such programmes have the potential to reduce significantly the rates of childhood injury.26 On the basis that accidents are preventable and immunisation prevents disease, perhaps more could be done to minimise the risks faced by children in households with one adult. The responsibilities lie only partly with healthcare providers; the causes of and minimisation of poverty are for political initiatives. Single parents and their children live at the bottom end of the income scale.27


Contributors: DMF and JRHC jointly initiated this study. The database was assembled from data provided by practices contributing to the fourth morbidity study in general practice. JRHC was responsible for the statistical input to the study and was assisted by Judith Charlton in undertaking the computer searches. DMF was chiefly responsible for the preparation of the manuscript and the related secretarial work was undertaken by Joan Dainty. DMF and JRHC are guarantors of the content of this report.

Funding: The morbidity survey was funded by the Department of Health, who have given permission to publish.

Conflict of interest: None.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
View Abstract