Introduction

Top Abstract Introduction Methods Results Discussion References

More than a quarter of the workload of general practitioners arises from consultations with children, and about 90% of children are taken to see their general practitioner every year. ^{1 2} Factors which predict consulting behaviour in children include previous experience of child care by the parents, the extent of illness, the age of the child, having an unemployed father, material deprivation, living in rented accommodation, and attendance at nurseries.^3-5 Previous studies, however, have usually relied on parental recall of consultations with general practitioners by using cross sectional or retrospective data on selected populations. Moreover, although many of these factors are related to individual socioeconomic circumstances, little is known about how overall socioeconomic differences affect childhood illness and consultations with general practitioners. The few studies that have examined childhood illnesses specific for social class have often had contradictory results. For example, some studies have found that children from social classes I and II have a higher prevalence of asthma, whereas other studies have found that the prevalence of severe asthma was highest among children from lower socioeconomic groups. ^{6 7}

Social class differences in morbidity and in the use of health services remain important. There are large differences in mortality and morbidity between social classes for the major causes of illness.^8-10 Lifestyle patterns which affect health, such as smoking and material circumstances, also vary according to socioeconomic status.¹¹ Deprived areas with high morbidity often receive poorer healthcare services, and users of preventive services are often those least in need of such care. ^{12 13} Furthermore, relative social class differences in health have widened in recent years.¹⁴

The fourth national survey of morbidity in general practice reported some preliminary findings on socioeconomic differences in consultation rates.¹ We used data from the survey to examine further the association between childhood consultations in general practice and socioeconomic status in children aged 0 to 15 years registered with 60 general practices in England and Wales.

	Methods

Top Abstract Introduction Methods Results Discussion References

The fourth national survey of morbidity in general practice was conducted between September 1991 and August 1992.¹ The main objective of the study was to examine the patterns of disease seen by general practitioners by the age, sex, and socioeconomic status of the patients.

Practices in study
Sixty volunteer general practices took part in the survey. The practices had a special interest in the collection of morbidity data and were not typical of all practices in England and Wales. For example, they were more likely to record clinical information on practice computers (100% v 34%) and had a larger mean list size (7700 v 5200).

Patients in study
The study population comprised a 1% sample of the population of England and Wales. The sample was representative of the population for age, sex, social class, and housing tenure, but there was under-representation of ethnic minority groups because relatively few inner city practices participated in the survey.

Recording and validation of morbidity data
Before the survey started doctors and staff from each practice attended three 2 day training sessions on the recording of morbidity data. Practices then collected data for 2 to 4 weeks before the start of the survey. These data were analysed and any errors or inconsistencies reported to the practices. Once the morbidity survey started general practitioners and nurses recorded information on all face to face contacts with patients. Each reason for consulting and the place of contact was directly entered into patient records on the practice computer and defined as one consultation. Every consultation was given a diagnostic Read code and the data then transferred on disk to the Office of Population Censuses and Surveys where an international classification of diseases, ninth revision (ICD-9) code was assigned. The underlying disease for each episode of illness was in turn mapped to a category of severity: serious (for example, diabetes), intermediate (for example, iron deficiency anaemia), or minor (for example, upper respiratory tract infections) (see box). The categories were predefined and independent of doctors' opinions of the clinical condition of the patient at the time of presentation.

Socioeconomic data
Socioeconomic data for all patients registered were collected by interviewers during the year of the study. Occupation of the parent or guardian of the child was recorded and converted to social class with the registrar general's classification (class I the highest and class V the lowest). Other socioeconomic data collected included housing tenure, ethnicity, whether the child was living with a sole adult, and economic position last week of the head of the household. Data for children under 16 years of age were usually provided by a parent or close relative. The response given at interview was assumed to apply for the entire year. Age was grouped as 0 to 4, 5 to 9, and 10 to 15 years. Because of the relatively small number of children in classes I and V, social class was grouped as I-II, III non-manual, III manual, and IV-V.

Statistical methods
The consultation and socioeconomic data supplied by the practices for each child in the survey were linked to produce a record for each child. Consultations in subgroups of social class were examined by tabulating the mean rates of consultation. To overcome clustering dependencies at the individual level a mean consultation rate for each child was calculated and each of these summed over the number of children in each social class category to produce a person based mean consultation rate with 95% confidence interval.¹⁵ All rates were corrected to take into account the fact that not all the children were followed up for the entire year of the study. Because adjustment for age and sex did not make any significant difference to social class differences, unadjusted rates are presented throughout. Rate ratios were calculated as the ratio of mean rates; confidence intervals were estimated with the  method.¹⁶ The outcomes were consultation rates for any reason, illness (ICD-9 chapters 1 to 17), infections (ICD-9 chapter 1), asthma (ICD-9 493), accidents and poisonings (ICD-9 chapter 17), home visits, and prevention (immunisation, screening, surveillance, or antenatal care). Linear trends across categories of social class were compared by constructing a linear regression model. For this last comparison those children whose social class was not known were excluded from the analysis. The trends are presented as  coefficients of the line of best fit, which shows the rate of change in unadjusted mean consultation rate across each social class category. For ease of interpretation these changes were expressed as percentage change from the mean rate across each category.

	Results

Top Abstract Introduction Methods Results Discussion References

Baseline characteristics
A third of the 106 102 children were from social classes I-II, 10% from III non-manual, 27% from III manual, and 17% from IV-V (table 1). When we compared our figures with the results of the 1991 census a greater proportion of households had parents or guardians in full time employment (67% v 45%), and there was some under-representation of non-white ethnic groups. Other characteristics were similar to those found in the 1991 census.

Consultation rates
There were 324 064 consultations, and the corrected mean annual consultation rate was 3.7 consultations per child per year (95% confidence interval 3.64 to 3.75; table 1). Of these consultations, 87% were for illness episodes, 11% for preventive episodes, and 2% for other reasons. Of the episodes of illness, the underlying disease was classed as of minor severity in 50%, moderate in 43%, and serious in 7%. Infectious diseases and respiratory episodes made up over 40% of the diagnoses.

Differences in consulting rates across social class
Overall consulting rates were 18% higher in children from social classes IV-V than in children from social classes I-II (table 1). A larger increase of 23% was seen for consultations for episodes of illness (table 2). Home visiting rates doubled from social classes I-II to IV-V. There was a small decrease in consultation rates for preventive episodes from social class I-II through to IV-V, but this was not significant (table 2).

	Discussion

Top Abstract Introduction Methods Results Discussion References

Overall consultation rates in children increased linearly from social classes I-II to classes IV-V. Children from social classes IV-V were more likely to consult their general practitioner for an episode of illness, including for disorders such as infections, asthma, and injuries and poisonings. While these children have higher consulting rates for minor illness they also seem to have poorer health, with higher consultation rates also for intermediate and serious categories of illness. The one exception to this pattern was for preventive care, where children from social class IV-V consulted less frequently when compared with children from social classes I-II.

Comparison with other studies
The findings reported in our paper contradict those from another recently published study. In an analysis of data from the general household survey, Cooper et al found no evidence of socioeconomic differences in childhood consultation rates.¹⁷ Their sample, however, was substantially smaller than our own (20 473 v 106 102) and covered a wider age range (0-19 years v 0-15 years). Furthermore, the consultation rates were based on parental recall of consultations during a 2 week period, whereas our own study used validated data that were prospectively collected over 1 year. Hence, the findings of our study are likely to be a more accurate reflection of the association between childhood consultation rates and social class. Cooper et al did, however, find a higher consultation rates in children classed as south Asian, and this is consistent with the findings of our own study.

Generalisibility of findings
The general application of our findings is potentially limited because the practices taking part in the survey were volunteers. This resulted in fewer practices from inner city areas and hence lower rates of unemployment and ethnicity among the patients in the study. The ecological fallacy, however, was avoided as social class was recorded at the individual level. The data were also collected prospectively, and validation studies suggested that there was good recording. The study sample was also reasonably similar to the population of England and Wales for most socioeconomic characteristics. Hence, it seems unlikely that biases could account for the large differences seen in consultation rates between children from different social classes. Because of the well known association between social class and ill health the differences in consultation rates are unlikely to be entirely due to inappropriate use of general practitioner services by children from social classes IV and V. This conclusion is supported by the findings on consultation rates by severity of underlying illness. The higher consultation rates for illnesses of serious and moderate severity suggest that morbidity levels in children in this study were higher in children from social classes IV-V than from social classes I-II.

Conclusions
This study has highlighted the importance of the socioeconomic background of children when the use of primary care services is examined. Our findings have implications for targeting children for preventive practices such as immunisation and the prevention of injury and poisoning.¹⁴ Members of the primary healthcare team and planners of health services need to be aware of the impact of socioeconomic circumstances on morbidity when planning health services.

	Acknowledgments

We thank Rosie McNiece for her help and Karen Dunnell and Sean Hilton for their comments.

Contributors: SS, AM, and MJ planned the analysis of data from the morbidity survey. SS analysed the data with help from MJ. SS and AM wrote the paper and received comments from MJ. SS and AM are the guarantors for the paper.

Funding: SS is funded by Research and Development Directorate of the South Thames Regional Office of the NHS Executive. The fourth national study of morbidity in general practice was funded by the Department of Health and carried out under the supervision of a project board with representatives from the Department of Health, the Royal College of General Practitioners, and the Office of Population Censuses and Surveys (now the Office for National Statistics).

Competing interests: None declared.

	References

Top Abstract Introduction Methods Results Discussion References

1. Office of Population Censuses and Surveys, Department of Health, Royal College of General Practitioners. Morbidity statistics from general practice. Fourth national study 1991-1992. London: HMSO , 1995.
2. Drever F, Whitehead M, eds. Health inequalities. London: Stationery Office, 1997(Decennial supplement No 15.)
3. Gabriel J, Campion P. Illness behaviour in mothers with young children. Soc Sci Med 1985; 20: 325-330.
4. Hewison J, Wyke S. Respiratory illness in children: what makes parents decide to consult? Br J Gen Pract 1990; 40: 226-229[Medline].
5. Gilmour H, Morrison JM. Children seen frequently out of hours in one general practice. BMJ 1991; 303: 1111-1114.
6. Strachan DP, Anderson HR, Limb ES, O'Neill A, Wells NA. A national survey of asthma prevalence, severity and treatment in Great Britain. Arch Dis Child 1994; 70: 174-178[Abstract].
7. Ernst P, Demissie K, Joseph L, Locher U, Beckland M. Socio-economic status and indicators of asthma in children. Am J Crit Care Med 1995; 152: 570-575[Abstract].
8. Smith GD, Bartley M, Blane D. The Black report on socioeconomic inequalities in health 10 years on. BMJ 1990; 301: 373-377.
9. Blaxter M. Health and lifestyles. London: Tavistock-Routledge , 1990.
10. Office of Population Censuses and Surveys. Socio-demographic mortality differentials 1971-1975. London: HMSO , 1982(Longitudinal Study Series LS No 1.)
11. Office of Population Censuses and Surveys. Health survey for England. London: HMSO , 1992.
12. Davidson N, Townsend P. Inequalities in health: the Black report. Harmondsworth: Penguin , 1980.
13. Majeed FA, Chaturvedi N, Reading R, Ben-Schlomo Y. Monitoring and promoting equity in primary and secondary care. BMJ 1994; 308: 1426-1429[Free Full Text].
14. Acheson D. Independent inquiry into inequalities in health. London: The Stationery Office , 1998.
15. Donner A, Klar N. Methods for comparing event rates in intervention trials when the unit of allocation is a cluster. Am J Epidemiol 1994; 140: 279-289[Abstract/Free Full Text].
16. Armitage P, Berry G. Sampling. In: Statistical methods in medical research. 3rd ed. Oxford: Blackwell Scientific Publications , 1987.
17. Cooper H, Smaje C, Arber S. Use of health services by children and young people according to ethnicity and social class: secondary analysis of a national survey. BMJ 1998; 317: 1047-1051[Abstract/Free Full Text].
18. Carr-Hill R, Rice N, Roland M. Socioeconomic determinants of rates of consultation in general practice based on the fourth national morbidity survey of general practices. BMJ 1996; 312: 1008-1012[Abstract/Free Full Text].
19. Charlton JRH, Fleming DM. Morbidity and healthcare utilisation of children in households with one adult: comparative observational study. BMJ 1998; 316: 1572-1576[Abstract/Free Full Text].
20. Aylin P, Majeed FA, Cook D. Home visiting by general practitioners in England and Wales. BMJ 1996; 313: 207-210[Abstract/Free Full Text].

(Accepted 18 December 1998)