Sharing client information in public services:
the management of blame
Charles D. Raab (University of Edinburgh)
Perri 6 (Nottingham Trent University)
Christine Bellamy (Nottingham Trent University)
Paper presented at the ECPR Joint Sessions of Workshops, Granada, 14-19 April 2005
Workshop: Blame-avoidance and Blame-management: Institutional and Policy Implications (Directors: Christopher Hood and Moshe Maor)
13
Abstract
This paper addresses issues of blame-deflection, avoidance, dissolution, distribution and management for errors in situations where professionals’ decisions whether or not to share personal information about vulnerable clients in social and health services may be fateful for the lives and interests of the persons concerned.
The abilities and strategies of different actors to handle the blame associated with United Kingdom cases of child abuse, mental health care, post-discharge care for older people and child murder vary between services. So too does the nature and evaluation of the risks involved in information-disclosure decisions, as well as the configuration of policy instruments available for managing potential blame. The risks are judgment errors resulting in false negatives or false positives; each type may have serious consequences which lead the media and the public to allocate blame for tragic outcomes.
The paper offers an analysis that might, if empirically validated, help to explain the variation noted above. It considers normative, organisational, cultural and other factors involved in the success or failure of attempts by professionals, managers, policy-makers and regulators to manage blame. The nature of available policy instruments mandating or guiding procedures for processing and disclosing clients’ information also bear on explanations of the structure of opportunities and constraints upon blame-avoidance strategies. The paper works towards a synthesis with explanatory power in terms of theories of risk and blame where ostensibly opposing values or imperatives are at stake (e.g., protecting vulnerable persons through information exchange whilst also respecting confidentiality and protecting privacy).
13
Introduction: blame for judgment errors in different structural and information conditions
The political science literature on blame has tended hitherto to focus mainly on two kinds of empirical phenomena. The first comprises the strategies of politicians facing imperatives for retrenchment of welfare provision but fearful of electoral defeat for imposing losses upon large and politically powerful constituencies (Pierson, 1994; Weaver, 1986). The second concerns the choices facing policy-makers when certain types of decisions need to be made in situations where their own motives might be suspect and where, therefore, the credibility of the decisions might be undermined. In the first case, ‘blame avoidance’ strategies have been identified. These strategies may involve solutions that are sub-optimal to the point of deferring retrenchment and paying compensation to groups facing losses above levels that would be economically efficient or even justifiable on grounds of fairness. In the second type of case, the literature emphasises ‘blame deflection’ strategies, of which delegation to technocratic, scientific or independent organisations is the most important (Hood 2002); Ellis, 1994).
The present paper, however, is concerned with neither of these phenomena, but discusses instead some common and crucially important situations in public administration and social policy, where:
- there is a risk of blame arising not simply because those upon whom losses might be imposed as a result of particular decisions will protest, but because, in the prevailing conditions of limited and unreliable information, whatever decision rule is applied, the risk of committing one or other type of judgment error (see below) is high; unlike the welfare retrenchment case, then, there is no available strategy that can reliably be expected to reduce exposure to blame;
- there is limited scope for compensating those upon whom losses are imposed as a result of judgment errors; and
- decision-making responsibility cannot be delegated any further with the expectation of a different decision; the risk of blame arises from the nature of the decision rule and the information conditions under which it is taken: the same problem will arise whoever is charged with making the decisions.
We discuss the issues raised by such situations with reference to examples drawn from four fields within the policy sectors involved in health and social care and the criminal justice system in the United Kingdom (UK). These are all fields in which questions of blame, including the attribution or denial of responsibility, have been aired in the glare of publicity surrounding dramatic and tragic events that have befallen especially vulnerable persons, including young children. We seek to understand and explain variations in the patterns of blame-management behaviour in these fields, by adducing a theoretical framework that should lend itself to empirical testing. Although we cannot yet provide systematic empirical evidence, preliminary findings from our current programme of research into information sharing and privacy in social policy[1] may allow better insights into the processes and understandings involved when professionals, managers and others in these four fields are exposed to blame by being required to make what may prove literally to be life-or-death decisions.[2]
Decisions in conditions specified above are prone to two kinds of error of judgment. False negative judgment errors (‘under-reacting’) occur when a person is deemed not to be in danger or dangerous, but when it subsequently transpires that they were. Conversely, false positive judgment errors (‘over-reacting’) occur when a person is deemed to be in danger or dangerous, but when it subsequently transpires that they were not. In this paper we are particularly concerned with decisions about whether or not to share – or even to retain – client information or to intervene or not to intervene on the basis of that information. Where information indicates the greater or lesser possibility that someone presents a risk or is at risk, any decision rule specifying the conditions for sharing, retention or intervention unavoidably risks one or other type of error to a greater or less degree. In many areas of social policy, the risks attaching to both false positive and false negative errors are both palpable and serious. As recent child protection scandals have demonstrated, they may, on the one hand, involve potential physical harm to, or even the murder of, a child. Conversely, a child may be wrongly taken into care with devastating effects upon the family. In either case, the perceived miscarriage of justice may generate a heated public reaction. Blame attaches to the professionals and other service personnel for the outcomes of their decisions, and not just for procedural aspects: having acted reasonably or with due attention to professional codes may offer no adequate defence when a child dies needlessly. Reputations, careers, and self-respect hang in the balance when the professional judgment of frontline workers and their managers is called into question in this way. Whole services, locally or nationally, may come under critical fire as formal inquiries are set in train and then report their findings.
In such situations it is far from clear what defensive practices decisions makers should best use to protect themselves against the threat of blame. What counts as the most rational defensive strategy may well be modulated by post hoc judgments about the relative risks associated with each type of error, as a result of a particular scandal. For example, the Chief Constable of Humberside may reasonably have considered that, faced with well-acknowledged confusions about the requirements of data protection law, the safest course of action was to avoid retaining or sharing ‘soft’ data about suspicious people who had never been convicted of a criminal offence. However, in the wake of the murders of the Soham schoolgirls by the school caretaker, Ian Huntley, the Inquiry conducted by Sir Michael Bichard (Bichard, 2004) severely castigated the force for failing to exploit data in ways that would have exposed Huntley’s potential for doing serious harm. In the present climate of intense worry about child protection, front line workers may well seek to avoid the risk of subsequent blame by passing on large amounts of information, with less concern for its quality or for the possible overloading of their colleagues. At other times, they may be reluctant to commit themselves to intervention and may communicate too little information, thus failing to appreciate what other workers may need to know in order to offer appropriate and timely care (Reder et al., 1993: 90). Defensive practice can extend to clients too. If individuals know, or suspect, that their confidences will not be safeguarded, they may refuse to disclose or may misrepresent essential facts about themselves, their relationships, their behaviour and their circumstances when professionals compile their case records. High levels of trust placed in caring professionals may be eroded, damaging the effective functioning of relationships between the services and the public. Discourse within the caring professions therefore recognises threats to privacy as as sources of real risks, not just as the overheated imaginings of misguided civil libertarians and privacy advocates. Privacy risks therefore underlie the considerable attention now given to the formulation of guidelines, rules and protocols and other such instruments for handling personal data within and across the agencies involved in decision about the use of case information in all these fields.
How heavily such risks should weigh in the balance against competing concerns for the immediate well-being of vulnerable persons is a dilemma bounded by the Scylla and Charybdis of negative and positive judgment errors. In this paper, we use examples from four fields of public policy, to explore the structural and information factors involved in the management of blame. We argue that these factors are not sufficient to explain the variance observed among these fields. In the second half of the paper, we therefore present a theory of the production and management of blame which we believe may prove more fruitful. At the present stage of our research, we do not have sufficient empirical data from any of these fields adequately to test the explanatory power of our theory. The argument of the paper is therefore necessarily speculative, but aims to explain and justify the theoretical framing of our empirical research.
The changing policy contexts for blame management
The four policy fields we draw on for this paper are: child protection, mental health care, care after discharge for older people, and the identification and management of those convicted of committing, or thought to be likely to commit, sexual offences against children. Together, they provide a useful range of fields for provisional comparison, exemplifying as they do many of the factors involved in the complex decisions that must be taken daily in all these fields to manage risk, danger and blame. As a set they therefore provide good scope for exploring available explanations for observed patterns of blame management.
Decision-making in all these fields is currently being framed by new national policies and, in some cases by, new legislation. The field of child protection currently has the highest salience. Following the publication of Lord Laming’s report on the Victoria Climbié case (Laming, 2003), there have a been a series of official policy statements (Chief Secretary to the Treasury, 2003; Minister for Children, Young People and Families, 2004) culminating in the Children Act 2004. This Act lays the foundations for a new, multi-agency, child protection regime built on the explicit principle that the child’s interests are paramount over all other concerns. This field differs from all others, then, in that this principle at least provides an indication of the respective weights to be accorded by front line workers to stakeholders’ competing interests. In information management terms, the manifestations of this principle are first, the proposed establishment of a universal register of children that would enable each child’s contact with the universal services, that is, health and education, to be tracked, and second, the construction of a register of particularly vulnerable children to which all relevant agencies would have access on specified terms (Department for Education and Skills, 2004). Unsurprisingly, the proposal for a universal register has been heavily criticised on grounds of disproportionate data processing by the state, while the specialist register has attracted considerable opposition for its potential to stigmatise vulnerable children (Information Commissioner, 2004). In other words, controversy around this Act recognises that children’s interests are not always best served by keeping and sharing more information about them, and the paramountcy principle does not, therefore, obviate the essential dilemmas posed by decision-making in this field.
Arrangements for protecting the community against violent and sex offenders have also been put on a statutory basis in the Criminal Justice and Court Services Act, 2000, which compelled the establishment of Multi-Agency Public Protection Arrangements (MAPPA), led by the Probation Service and the Police, in every police force area. The MAPPA are required to assess and manage risks posed by those sex offenders who are required by law to register with the police under the Sex Offenders Act 1997, and to involve a range of other public services, such as mental heath, housing and social services, as necessary. Contrary to common parlance, there has never been a national register of sex offenders, but, as we write, the Police Information Technology Office (PITO) is rolling out a national, computerised database of sex offenders, known as ViSOR. This will be run on the Police National Computer but will also provide access for authorised prison and probation officers. As the case of Ian Huntley demonstrates only too well, however, the most intractable problem in this field is the identification of, and therefore the management of data, danger and risk, associated with people who have not been convicted, who are not, therefore, obliged to register and whose names do not appear on ViSOR. As this case also shows, this problem is the more important, because data collections maintained by the police are increasingly being used for purposes that go well beyond the functional needs of the police themselves, most notably for the purposes of screening by the Criminal Records Bureau of people who wish to work with children or other vulnerable people.
In contrast to the fields described above, arrangements for the care of the mentally ill and older people are not prescribed by statute, but as discussion below shows, these are fields in which there is also increasingly strong assertion of national prescriptions for multi-agency working. It is well recognised, too, that, to be effective, multi-agency working must be supported by common approaches to data management, based on the sharing of client information. An interesting feature of this field is the attempt to encourage changes in working patterns by means of technological changes proposed under the National Programme for IT, which should establish shared electronic care records (ECRs) throughout the UK (Department of Health, 2002). This programme is currently the subject of much worry, criticism and opposition, and may well be delayed. In the meantime, practices continue to vary considerably from place to place, moderated only by national guidelines and local protocols that offer frontline workers uncertain protection from blame.