Editorials

Building evidence on chronic disease in old age

Standardised assessments and databases offer one way of building the evidence

Evidence based medicine, clinical effectiveness, and performance indicators are the topics of the moment, but their applicability to frail elderly people in community and institutional settings is problematic because of insufficient relevant evidence. About 2-15% of the population aged over 65 of industrialised countries are residents in nursing homes and related facilities. These elderly people have multiple impairments and chronic diseases, features that generally exclude them from randomised controlled trials. There are, however, other methods that can be used to gather the evidence on which we can base our interventions for these older people.

Standardised assessment measures go beyond diagnosis to include physical and psychosocial function and must be included in an evaluation of health care for older people. The creation of databases containing such data in conjunction with data on treatments could begin to fill the gaps of missing evidence. ^{1 2} Under certain conditions and for certain applications, research based on clinical databases has been favourably compared with randomised controlled trials.³

Before they can be used to evaluate care, data on functional status must be routinely available. In 1990 the Royal College of Physicians and the British Geriatrics Society recommended the use of standardised assessment scales to "enhance the communication between those who care for older people in different settings, encouraging the development of a common clinical language and descriptors of disability."⁴ The recommended scales covered activities of daily living, communication, visual and hearing disability, cognitive function, depression, quality of life, and assessment of social status. However, their widespread systematic adoption in hospital practice has not been achieved, and use even of the most widely used measuresthe Barthel index and abbreviated mental testis highly variable.⁵ Outside hospital, there is high variability in the assessment tools used by social services departments⁶ and poor documentation in nursing homes.⁷

A uniform, comprehensive, standardised assessment for routine long term care of older people, the minimum data set-resident assessment instrument (MDS-RAI), has been introduced into all nursing homes in the United States, Iceland, and three provinces in Canada. A US research group has combined data from the minimum data set-resident assessment instrument, including detailed drug use information, with data from Medicare enrolment and hospital discharge claims files, enabling the study of drug treatment effects using valid measures of outcome in this frail population.

These data have been used to show that 73% of US nursing home residents are women and that they have a mean age of almost 83 years, have many diagnoses, and are prescribed six drugs on average.⁸ Treatment for hypertension in older patients and those with marked impairment of physical and cognitive function does not follow recommended guidelines,⁹ and 26% of residents with cancer who have daily pain receive no analgesia.¹⁰ The first report comparing hospitals providing long term care in Ontario, using 24 quality indicators derived from aggregations of patient level data from the minimum data set-resident assessment instrument, has recently been released.¹¹

Early reports of implementation of the minimum data set-resident assessment instrument found improvements in care, and many unanswered questions about the use of selected medical treatments, ranging from antipsychotic medication to the use of restraints in nursing homes, are now beginning to be addressed.¹²

In the United Kingdom the advent of clinical governance and performance indicators reflects the trend to an evidence based ideal. In the US the Health Care Financing Administration and organisations such as the Joint Commission on Accreditation of Healthcare Organisations and the National Committee for Quality Assurance are committed to developing aggregated quality indicators to characterise the performance of health maintenance organisations, hospitals, nursing homes, and home healthcare agencies. In Canada reporting mechanisms at both federal and provincial government levels are being established to improve the quality and accountability of health care. Without the relevant data, we are likely to exclude frail elderly people from these initiatives to raise standards.

The US National Committee for Vital and Health Statistics has proposed a set of "core health data elements" to accompany any electronic record of a healthcare service event, including standardised codes for classifying functional status. The code sets in the minimum data set-resident assessment instrument should be considered to be the prime candidate.

Standardised assessment scales using the same core set of items as the minimum data set-resident assessment instrument have now been developed for community care, mental health, acute and postacute care of older people. Systematic evaluations of these instruments are under way in many countries through the interRAI collaborative research network (www. inter rai.org). Systematic analysis of data from uniform comprehensive standardised assessments in routine practice aggregated into high quality databases should contribute to evaluating the effectiveness and outcome of care provided to frail elderly people with chronic disease.

G I Carpenter, senior lecturer in health care of the elderly. 

Centre for Health Services Studies, University of Kent, Canterbury, Kent CT2 7NF (G.I.Carpenter@ukc.ac.ik)

R Bernabei, director. 

Facoltà di Medicina e Chirurgia "Agostino Gemelli," Università Cattolica del Sacro Cuore, 00168 Rome, Italy

J P Hirdes, director. 

Canadian Collaborating Centre InterRAI, Providence Centre, Scarborough, ON M2L 3G1, Canada

V Mor, director. 

Center for Gerontology and Health Care Research, Brown University, Providence, RI 02912, USA

K Steel, director. 

Homecare Institute, Hackensack University Medical Center, Hackensack, NJ 07601, USA

Acknowledgments

The authors are fellows of the interRAI collaborative research network.