Home Visiting Trial and Follow-up
||2001-2004 - Main Trial
2004-2007 - 3 Year Follow-up
||Department of Health, Nuffield Foundation
|Collaborators||Health Services Research Unit, University of Oxford;University of Warwick|
The Oxfordshire Home Visiting Study was one of the first UK-based randomised controlled trials to evaluate the effectiveness of a professionally delivered, intensive home visiting programme, in improving parenting and child outcomes including the prevention of abuse and neglect. The home visiting trial was completed in 2004. The objective of the economic study was to evaluate the cost-effectiveness of an intensive home visiting programme directed at vulnerable families during the antenatal and postnatal periods in reducing the risk of abuse and neglect in the first year of life. The design was an economic evaluation alongside a multicentre randomised controlled trial in which 131 eligible women were randomly allocated to receive 18 months of intensive home visiting (n=67) or standard services (n=64). A cost effectiveness analysis of this public health intervention was undertaken from a societal perspective.
Trial Results: The mean costs in the control and intervention arms were £3,874 and £7,120 respectively, a difference of £3,246 (p<0.000). The mean ‘health service only’ costs in the control and intervention arms were £3,324 and £5,685 respectively, a difference of £2,361 (p<0.000). One of three independent objective assessments, predictive of infant abuse and neglect, showed improvements in maternal sensitivity (p<0.04) and infant cooperativeness (p<0.02) in the intervention arm. There was also a non significant increase in the likelihood of intervention group infants being removed from the home due to abuse and neglect. The results of the study provide tentative evidence to suggest that, within the context of regular home visits, specially trained health visitors can increase maternal sensitivity and infant cooperativeness and are better able to identify infants in need of removal of the home for child protection. These potential benefits were delivered at an incremental societal cost of £3,246 per woman.
The main trial paper was published online in October 2006 and appeared in the Archives of Diseases in Childhood: in March 2007. The economic evaluation of the trial is currently under review with The Journal of Public Health. The costing study was published as a separate article: See- Netten, A and Curtis L (Eds), Unit Costs of Health and Social Care 2006.
Three-year follow up study
A three year follow-up of the trial has recently been completed. The aims of the follow up were as follows: to establish the effectiveness of the intervention at 3-year follow-up in terms of improving a range of outcomes associated with poor or abusive parenting, and the early identification of infants in need of removal from the home; to establish the views of service recipients and providers concerning the value of the intervention and its impact and to evaluate the cost-effectiveness and cost-benefit of the intervention.
Currently the results of this study are in report form and have not yet been submitted to a peer –reviewed journal. A summary of the results are provided below.
Follow-up study - methods
The resource utilization questionnaire administered at 2, 6 and 12 months to women participating in the main Home Visiting trial was administered at 36 months. This questionnaire asked the women about their resource utilization over the last 2 years since the 12 month follow up point. Unit costs (2007/8) adjusted by appropriate quantities were attached to the items of resource-use to obtain a study cost. Costs were summed for each participant and the mean difference in costs between the two arms of the trial estimated. Unit costs were attached to resources to allow the reporting of variance in cost arising through economic significance as well as statistical significance.
As this was a follow up study, it was anticipated that the data would suffer from a proportion of loss to follow up i.e. attrition. As a consequence, missing data were analyzed to identify type of ‘missingness’ in the first instance to assess the most appropriate method of data handling7. Where the data were missing completely at random (MCAR) due to random drop out (i.e. where the drop-out was not a function of allocation group) then appropriate statistical methods were employed to account for the missingness. In addition to this, tests of normality in the cost data distributions were carried out using non-parametric one sample KS tests and where such tests were rejected skewness reported and median values reported alongside mean values. Further the 95% confidence intervals for the mean cost difference were computed using non-parametric bootstrapping methods.
Follow-up study - economic results
The three year data set contained 102 women, a loss to follow up of 22%. Exploring these drop out data to test for type of missingness revealed no statistically significant difference in drop out as a function of either trial allocation (p=0.73) or as a Pearson’s correlation incorporating the sum of risk factors (p=0.884). Based on these results and given the presence of unit non-response missing data, IDs for these data were reinstated and regression algorithms were employed to impute missing data for key cost variables as well as the total cost variables.
A mean cost estimate per woman per arm of the trial was computed. Tests of normality on the cost data distributions were rejected (one same KS tests, p<0.000). The 95% confidence interval for the cost difference between arms was therefore obtained using non-parametric bootstrapping methods.
The mean health service only costs in the control and intervention arms at 36 months were: £1,826 v’s £3,329, a difference of £1,503, p = 0.083 (bootstrapped 95% confidence interval for the cost difference: -£609 - £2,565). The mean societal costs in the control and intervention arms at 36 months were: £3,963 v’s £4,196, a difference of £233, p=0.812 (bootstrapped 95% confidence interval for the cost difference: -£2,740 - £1,454). The total health and societal costs of the intervention arm are statistically significantly greater for home start visits for both study mother and infant (although we can assume these were the same visits). Additional greater although non-significant health and social service costs incurred in the intervention arm were for: GP appointments for the study infant; social work home visits for the study mother; alcohol and drug support for the study mother; Paediatric visits for the study infant; Psychiatrist costs for the study mother and housing department appointments. A number of resource costs however were non-statistically greater in the control arm, these were for: social worker home visits for the study infant; obstetric appointments for the study mother; family centre visits for the study mother and infant; local advice centres; citizen’s advice bureau visits and private child care.
Combining the societal 12 month cost data with the 36 month cost data by inflating the 12 month cost data to 2007/8 levels reveals that at 36 months the total societal cost of the home visiting arm is £12,427 compared to the control arm of £8,441, a difference of £3,985 (95% bootstrapped CI for the cost difference: £192 - £5,297). Looking at the ‘health service only’ costs, at 36 months the cost of the home visiting arm is £9,901 while the control arm is £5,669, a difference of £4,232 (95% bootstrapped CI for the cost difference: £1,949 - £5,709).
The main trial results combined with the follow up results suggest that intensive home visiting improved maternal sensitivity and better enabled health visitors to identify infants in need of further protection at an incremental cost of £3,985 (95% bootstrapped CI for the cost difference: £192 - £5,297) per woman at 36 months. Looking at the ‘health service only’ costs, at 36 months the incremental cost is £4,232 (95% bootstrapped CI for the cost difference: £1,949 - £5,709). The extent to which these potential benefits are worth the costs, however, is a matter of judgment.
McIntosh E, Barlow J, Stewart-Brown S, Davis H (2008). Economic evaluation of an intensive home visiting programme: A cost-effectiveness analysis from a societal perspective. Under revision: Journal of Public Health.
Barlow J, Davis H, McIntosh E, Jarrett P, Mockford C, and Stewart-Brown, S (2007). The role of home visiting in improving parenting and health in families at risk of abuse and neglect: Results of a multicentre randomised controlled trial and economic evaluation. Arch Dis Child, 92(3):229-33.
McIntosh, E, and Barlow, J (2006). The costs of an Intensive Home Visiting Intervention for Vulnerable Families. In: Netten, A and Curtis L (Eds), Unit Costs of Health and Social Care 2006, Personal Social Services Research Unit, University of Kent, Canterbury.