The rollout of Blended Supervision to community-based probation services in April 2022 has already significantly impacted day-to-day operations for the Probation Service. Much like under national standards, Blended Supervision established a baseline delivery model, based on the risks and needs of the Person on Probation (POP), it codifies the minimum engagement expected in the first 12 weeks.
Early feedback is positive and there is hope that it can continue to provide a structure that makes the first three months of every probation journey more predictable, safer and less stressful for both Probation Practitioners and People on Probation.
Probation Practitioners (PPs) and those supporting them can now more clearly identify where the Probation Service’s scarcest resources (time, in-person meetings, support service referrals) are required, based upon current caseload and staffing levels.
If we go a step further and view the realised benefits of the Blended Supervision model through a data analysis lens, there is significant potential for more responsive resource planning.
Data gathering, insight generation and the development of machine learning to inform the Probation journey, have often historically been viewed as more aspirational than deliverable.
Successful rehabilitation is, and always has been, rooted in the relationship between Probation Practitioner and the POP. Utilising expertise and professional judgement to balance both risk management and rehabilitative interventions is not something that can be reduced to a simple algorithm. Using Blended Supervision as an example, the minimum set of expectations doesn’t, and shouldn’t, replace professional judgement. There are, however, huge opportunities in creating an evidence base that identifies best practice and supports development of appropriate and effective applications of data, for both operational and workforce planning.
Given ongoing resourcing pressures, ensuring Probation Practitioners have the most effective evidence to support their decisions will become increasingly valuable. While a computer cannot read a room, it can analyse trends that can flag a hidden risk factor for examination. It could offer historic success rates of sequencing interventions in one order vs. another, or simply offer a snapshot of available services based on keywords in a POP’s file, matrixed by suitability or distance from home address.
There are many possibilities, all of which are impacted by another limiting factor, data quality.
Unwieldly legacy systems, gaps in data post-reunification and data protection concerns are just the beginning of the challenge when it comes to collecting and working with data in the criminal justice space. It is therefore understandable that using data intelligence and modelling to support probation service delivery has sometimes required disproportionately more effort than the results merit.
A year after reunification, changing attitudes and increased availability of technology within probation delivery makes this the right time to revisit the data intelligence conversation.
The codified, time-bound nature of the Blended Supervision model offers a new beginning for data harvesting, effectively passing every individual POP through the same set of conditions over the first three months of their community sentence. If captured, this could provide, in a relatively short time period, a high-quality evidence base upon which to build compelling insights on the early probation journey. The allocation of risk and tier levels is also a simple methodology by which the data could be anonymised whilst maintaining enough specificity to be useful.
Acting now would see the Probation Service empowering itself with data it already owns and organically generates. The time taken to identify what must be extracted, to build a usable data model describing adherence and variation to Blended Supervision, POP compliance and breach patterns, will provide useful insights and management information in a relatively short time. And, in the medium term, it can be linked to ongoing compliance and reoffending data.
Above and beyond the immediate internal uses of these simple insights, making an anonymised version of this data available to trusted partners, such as academic researchers, policy makers and data innovators, could deliver valuable longer-term insights on the local, regional and national level.
Blended Supervision, as it stands, is showing real potential as an effective delivery process, and could be the key to the Probation Service dramatically accelerating its data success story.
Originally posted here