Disease Registers Dr John Newton Consultant Epidemiologist, Unit of Health-care Epidemiology University of Oxford

Quality in the NHS depends on competence Knowledge Competence Structures

High Quality Health Care Education, Training, CPD, Information, Libraries, R&D High Quality Health Care Clinical Governance, Controls Assurance, Learning Organisation, IT, Clinical Audit, User involvement, Learning Hospital, etc…...

Source: BMJ 2003;326:274-276

Ophthalmology Waiting List: Analysis using survival analysis waiting time 800 600 400 200 -200 Cum Survival 1.2 1.0 .8 .6 .4 .2 0.0 -.2 procedure group other procedures DCR trabeculectomy squint surgery vitrectomy lens surgery

“We will strengthen the information base on chronic diseases in the population by establishing a series of disease registers in different parts of the country.” Saving Lives: our healthier nation July 1999

maintain registers of those at greatest risk “primary care trusts will identify and maintain registers of those at greatest risk from serious illness - concentrating particularly on areas where ill health is most prevalent – so that people can be offered preventive treatment” (NHS Plan, paragraph 1.6).

Registers identified 234 registers Contacted 117 of which 70 provided complete data

Types of register identified Specific diseases 31 Cancer 29 Diabetes 27 CHD 25 Congenital anomalies 22 Cerebral palsy 8 HIV 6 Cystic fibrosis 5 Interventions 5 Mental health 5

Some specific disease registers Ankylosing Spondylitis Register Brittle Bone Disease Register Central Register of Systemic Sclerosis Patients Childhood Fatigue Register Conn's Syndrome Register (primary hyperaldosteronism) Cystinosis Register Duchenne Muscular Dystrophy European Neuronal Ceroid Lipofusinosis (Battens disease)

Generic registers Biobank Target population

Uses of registers 1 Patient care regular review and recall structured care programmes monitoring high risk groups managing demand / regulating access communication risk stratification

Uses of registers 2 Public Health surveillance planning the provision of health care monitoring the burden of ill health monitoring the impact of prevention

Uses of registers 3 Technology assessment Research descriptive studies improving the performance of clinical trials studies of process hypothesis testing when trials are not available

National audit of thalassaemia screening using UK registers Ethnic Group No. of cases (%) offered screening Cypriot 119 (92) Indian 46 (63) Pakistani 76 (51) Bangladeshi 7 (35) Modell B, BMJ 2000

Service and research registers compared Service use Accuracy and timeliness are important Value determined by number of cases Internal consistency is important Patients can opt out if they wish Subjects need access to their own information on demand Research use Accuracy and completeness are important Value from a representative group and adequate power Internal and external consistency important Non-consent is a problem for the whole register costs of analysis and reporting should be considered

Issues raised in the study Definitions Need for evaluation Standards of good practice Costs and funding sources Privacy and confidentiality Service or research or both? Appropriate population levels

The “IEA definition” Register: In epidemiology the term register is applied to the file of data concerning all cases of a particular disease or other health-relevant condition in a defined population such that the cases can be related to a population base. Last, 1995

Features of registers People not events People registered having something in common Information held is systematically updated Register is based on a defined population Donaldson, 1992

Characteristics of a good register appropriate multidisciplinary team stable funding focused aims data collection systems and design that relate well to function relevant leadership Pryor, 1985

Evaluation of a register should include public health importance systems and objectives uses and outputs evaluation against eight attributes resources used system meeting its objectives

Assessing validity Completeness Validity death certificate method independent case ascertainment historic comparisons Validity diagnostic criteria re-abstracted data internal consistency

I want to set up a register Expert Group & Steering Group Notify the Information Commissioner Ethics approval Arrange access to data Data security Accountability Publicity

“The temptation to collect information is hard to resist.” Weddell, 1973

Fundamental considerations Case definition Case ascertainment Identification of duplicates Follow up Consent

Number of Tayside residents with known diabetes who were identified independently by each data source used in electronic record linkage Type of treatment Source of data Insulin Oral Diet Encashed prescriptions 1868 2690 583 Hospital diabetes clinics 1715 2050 858 Young adult/paed. clinics 193 0 0 Mobile eye unit 1593 2578 1313 Biochemistry database 1077 1675 896 Hospital discharge diag. 1278 820 465 Total 2228 3419 1949

Satisfying the DPA 98 General awareness of uses of personal health data Cogent reasons for not obtaining consent Necessary for a legitimate “medical” purpose Data controller should be a medical practitioner or owe an equivalent duty of confidence to the data subject

Satisfying the common law duty of confidentiality Personal data should only be disclosed with consent where the law requires it if it is in the public interest to do so

Consent “any freely given specific and informed indication of his wishes by which the data subject signifies his agreement to personal data relating to him being processed.” EC Directive 95/46/EC

Human beings have a right to privacy in a humane society but this right is not absolute in either an ethical or a legal sense. Adapted from Higgs, 1996

Why not get consent? Disproportionate effort bias against ill disabled disadvantaged busy

Registers need identifiable data To avoid double-counting To link cases over time - follow up For validation To link to other databases

PERIC results 3921 adults in GB Asked to assess a selection of vignettes (n=10) How happy to allow access to their records? 2.1% of respondents were unhappy with all vignettes 11.6% of all responses were “unhappy”

“The only correct method of handling persons lost to follow-up is not to have any.” Dorn, 1950

How much do registers cost? from CJD - £1.2 million pa Cardiac intervention database - £500,000 pa to Osteogenesis Imperfecta - research fellow half a day per week

Which registers to fund? adequate quality important problem avoid duplication exploit economies of scale generic registers - “just in case” funding according to function long term funding

Recommendations National strategy recognition of different purposes establish agreed standards system for appraisal rational resource allocation Public Health Observatories to collate and evaluate registers

Recommendations National co-ordinating centres to: develop methodology provide advice to local registers develop data standards for common diseases practical support (questionnaire design etc.) quality assurance collate information at national level

Recommendations The NHS should establish a research infrastructure funding programme, together with other sponsors, that could fund appropriate research registers. How much is needed? £200k pa x 10 £50k pa x 60 i.e. about £5 m pa in total

Conclusions Useful but obscure literature Registers have a lot to contribute Over 250 “registers” in England Terminolgy confused, quality variable Fragmentation of policy & strategy Need for expert resource No rational funding system Uncertainty over data protection law