• By Laura Whybra
  • Posted on Tuesday 09th December, 2014

Why are social work RCTs so unpopular?

strong>Almost everyone agrees that programs for children in need should be based on the best possible evidence. In many cases, the ideal evidence would come from a randomized controlled trial (RCT). But randomization can clash with the day-to-day concerns of local authorities and social workers. A recent article offers a rare behind-the-scenes look at a social work RCT. RCTs offer the best chance of being able to say with confidence that group differences in outcomes are a reflection of the treatment, rather than other factors. They have become widely used in health and psychology. Despite this, RCTs are rarely used in social work. A small trial in England of a foster care program for children with extremely challenging behavior aimed to take advantage of the power of randomization. Versions of Multi-dimensional Treatment Foster Care for Adolescents (MTFC-A) had produced good results in the US and Sweden. A solid RCT could provide evidence about how it worked in the UK context. The researchers intended to recruit 130 children. But after three years of concentrated recruitment, only 34 children participated in the RCT. Moreover, in eight of these cases, the local authority decided to override the randomization and assign the child to a different program. Such small numbers and high cross-over rates make it hard to draw firm conclusions from the trial.So what went wrong? While the researchers identified a number of challenges, one of the most fundamental – and one of the most important for future trials – was the tension between the requirements of an RCT and the needs and views of social workers and managers. Different perspectives from research and practiceThe RCT design requires that all children in the study should have an equal chance of participating in the program. But the researchers soon found that their RCT referral process was “often far removed from the reality of social workers’ and managers’ day-to-day decision making about placements.”First, the original government funding required evaluation, but not randomization. Of the 18 local authorities that participated, only six agreed to assign children to the program randomly. Most local authorities were not convinced of the potential value of an RCT approach, especially when the evaluation had to begin immediately, as soon as the first placements were made. The local authorities, understandably, had other initial priorities – such as sorting out the implementation of MTFC-A so it could be viable in their area. The implementation challenge was particularly important given the difficulties of putting in place a new program for children with severe emotional difficulties, challenging behavior, and in unstable care situations. Second, MTFC-A is an intensive and expensive program, so it was easier to justify for children who seemed more likely to end up in an expensive placement regardless. Some stakeholders were concerned that randomization risked assigning “low-cost” children to MTFC-A and “high-cost” ones to treatment as usual.As a result, some children were effectively earmarked for MTFC-A before random assignment.Third, and perhaps most important, many managers and program staff preferred to rely on professional judgment, rather than random assignment, in making placements. Especially those in small authorities who knew the children very well wanted to make decisions based on whether particular children seemed well-suited to the program.The researchers said, “Although they freely admitted that resource shortages meant that many placements were, in effect, already allocated by chance, they were unconvinced by our argument that randomization could make a virtue of this necessity.” These three issues combined with other concerns, including worries that the study would add to already-heavy workloads for social workers, and fears that it was a mistake to ask vulnerable children to participate in decisions about research. Although some social workers and managers were enthusiastic about the RCT, many were not. The result, overall, was a stunning lack of demand for MTCF-A. Lessons for the futureSo what can be done to meet the immediate needs and concerns of social workers, managers, and local authorities, while conducting research that may help children in the long run?First, the authors suggest that incentives should be provided for local authorities and social workers to take part in all research, including RCTs. At the moment, many do not see it as in their interest to take part. Financial incentives, reports and individual feedback are some ways that their involvement can be encouraged, and this should be costed into proposals. Second, RCTs are expensive and should only be undertaken when a program is really ready to be tested. This may mean waiting for the program to “bed in,” so that those who are responsible for implementing it can work out the inevitable teething problems and recruit an adequate number of participants. Third, RCTs are just one of many approaches to evaluation, and effective studies may combine more than one research method. One of the advantages of the MTCF-A trial was that it included an observational arm as well as an RCT. An additional 185 children participated in the parallel observational study. Fourth, there are optimal conditions in which a RCT can be implemented; the intervention and the population it is targeted at must be clearly defined, and there must be equal access to receive the intervention. There must also be an agreement at the start among those involved about what type of research is required. Fifth, there are always ethical considerations to consider with conducting a RCT. Careful thought must be taken about what participants in the control group receive, and how the difference between treatment and control programs will be seen by participants and those implementing the intervention.Finally, a bigger issue perhaps is the reluctance among social work staff to embrace experimental research methods, especially when these require them to relinquish control over the assignment process. To date, the main body of research in this field is qualitative. However, RCTs can add another valuable research tool to the toolkit, since they can demonstrate whether an intervention works and for whom it works. The verdict is still out on how researchers can best address the legitimate and understandable concerns of social workers and managers to make RCTs more attractive to those who will implement them. *********ReferenceDixon, J., Biehal, N., Green, J., Sinclair, I., Kay, C., & Parry, E. (2014). Trials and Tribulations: Challenges and prospects for randomized controlled trials of social work with children. British Journal of Social Work, 44, 1563-1581.

Back to Archives