Abstract. Manchester, UK The Study was conducted from the Centre for Health Services Studies

Transcription

1 Age and Ageing 2002; 32: Identification of registered nursing care of residents in English nursing homes using the Minimum Data Set Resident Assessment Instrument (MDS/RAI) and Resource Utilisation Groups version III (RUG-III) IAIN CARPENTER 1,MICHELLE PERRY 2,DAVID CHALLIS 3,KEVIN HOPE 4 1 Centre for Health Service Studies, University of Kent, Canterbury, GKT School of Medicine and Dentistry, London, and East Kent Hospitals NHS Trust, UK 2 CHSS, George Allen Wing, University of Kent, Canterbury CT2 7NF, UK 3 PSSRU, Faculty of Medicine, Dentistry, Nursing and Pharmacy, University of Manchester, Dover Street Building, University of Manchester, Oxford Road, Manchester M13 9PL, UK 4 School of Nursing, Midwifery and Health Visiting, Coupland III Building, University of Manchester, Oxford Road, Manchester, UK The Study was conducted from the Centre for Health Services Studies Address correspondence to: Dr G. I. Carpenter, Centre for Health Services Studies, George Allen Wing, University of Kent, Canterbury CT2 7NF, UK. Fax: (q44) g.i.carpenter@ukc.ac.uk Abstract # Age and Ageing Vol. 32 No. 3 # 2003, British Geriatrics Society. All rights reserved. Aim: to determine if a combination of Minimum Data Set/Resident Assessment Instrument (MDS/RAI) assessment variables and the Resource Utilisation Groups version III (RUG-III) case-mix system could be used as a method of identifying and reimbursing registered nursing care needs in long-term care. Method: the sample included 193 nursing home residents from four nursing homes from three different locations and care providers in England. The study included assessments of residents care needs using either the MDS/RAI assessments or RUG stand-alone questionnaires and a time study that recorded the amount of nursing time received by residents over a 24-h period. Validity of RUG-III for explaining the distribution of care time between residents in different RUG-III groups was tested. The difference in direct and indirect care provided by registered general nurses (RGN) and care assistants (CA) to residents in RUG-III clinical groups was compared. Results: the RUG-III system explained 56% of the variance in care time (Eta2, P=0.0001). Residents in RUG-III groups associated with particular medical and nursing needs (enhanced RGN care) received more than twice as much indirect RGN care time (t-test, P-0.001) and 1.4 times as much direct RGN and direct CA time (t-test, P-0.01) than residents with primarily cognitive impairment or physical problems only (standard RGN care). Residents with enhanced RGN care received an average of 48.1 min of RGN care in 24 h (95% CI ) compared with an average of 31.1 min (95% CI ) for residents in the standard RGN care group. A third low RGN care group was created following publication of the Department of Health guidance on NHS Funded Nursing Care. With three levels, the enhanced care group receives about 38% more than the standard group, and the low group receives about 50% of the standard group. Conclusions: the RUG-III system effectively differentiated between nursing home residents who are receiving low, standard and enhanced RGN care time. The findings could provide the basis of a reimbursement system for registered nursing time in long-term care facilities in the UK. Keywords: elderly, assessment, long-term care, reimbursement, case-mix 279

2 I. Carpenter et al. Introduction In the NHS Plan: The Government s Response to the Royal Commission on Long Term Care (2000), the Department of Health proposed to change the long-term care payment system in the UK and stated that in the future, the NHS will meet the costs of registered nurse time spent on providing, delegating or supervising care in any setting [1]. However, this statement left unanswered the question of how needed registered nursing time could be defined, quantified and appropriately reimbursed. Although various methods exist by which researchers have attempted to measure the need for nursing care [2], there is no nationally accepted system for determining the total amount of registered nursing resources required by residents in long-term care. The Joseph Rowntree Foundation commissioned this study to determine if assessment variables from the Minimum Data Set/Resident Assessment Instrument (MDS/RAI) [3, 4] in combination with the Resource Utilisation Groups version III (RUG-III) [5] case-mix system could be used as a method of identifying and reimbursing registered nursing care needs in long-term care. Methods The sample included all of the 193 nursing home residents of four nursing homes from three different locations and care providers in England, one in the South West, one in the North and two in Liverpool. A total of 24 registered general nurses (RGNs) and 56 care assistants (CAs) participated in the study. The study, undertaken between February and March 2001, included two components: assessments of residents care needs using either the MDS/RAI assessments or RUG standalone questionnaires; and a time study that recorded the amount of nursing time received by residents over a 24-h period. The MDS/RAI was designed to be used by nursing staff to assess residents needs and to assist in the development of individual care plans [3, 4]. This information can be used to present a comprehensive evaluation of a resident s clinical characteristics, such as functional ability, cognitive status, health conditions and psychosocial well-being. The MDS/RAI is listed under the Department of Health s guidance of tools and scales that localities may consider implementing in their approach to the single assessment process [6]. RUG-III is a case-mix classification system that uses a subset of MDS/RAI assessment items to determine nursing home resource use and staff time costs [5]. Residents are grouped into one of seven main categories based on the relationship between their clinical characteristics and the staff time (and therefore cost) involved in providing their care. The seven categories in descending order are: rehabilitation, extensive services, special care, clinically complex, impaired cognition, behaviour problems, and physical function. The first group uses the most resources while the last group uses the least. ADL (activities of daily living) scores, presence of depression and nursing rehabilitation needs are then assessed to allow subdividing of each of these clinical groups, to give a final RUG group. Staff time was measured over a 24-h period using a time recording sheet. All care staff were trained in the completion of the time sheet and a member of the research team was present in each home during the time data collection period. Carrying the sheet for the duration of their shift, care staff recorded the time spent with individual residents on direct and indirect care. Direct care time was defined as time spent in hands-on care. Indirect care time included all time spent on care planning, staff supervision, discussion with carers, relatives or other professionals etc. when that time could be allocated to a specific resident [7]. At the end of each shift, the total recorded time was reconciled with total shift time. Staff grades and local pay scales were recorded. This permitted staff time to be wage weighted and allocated between RGN and CA time. Full MDS/RAI assessments completed within 4 weeks of the time study were used in two homes in which the MDS/RAI was being used. Where MDS/RAI assessments within this time period were not available, and also for the residents of the remaining two nursing homes, stand-alone RUG-III assessments were completed. RUG-III groups were computed from the nursing home s MDS/RAI records or by direct entry of completed RUG questionnaires into a dedicated RUG-III group software program. RUG-III group and recorded time were entered into an SPSS database together with resident, staff grade and nursing home identification data. The extent to which RUG-III explained the variation in wage-weighted time per resident was computed using comparison of means and the eta statistic. Mean RGN and CA direct and indirect time was compared by RUG-III clinical group to differentiate groups associated with greater RGN care time. Following publication of the Department of Health (DoH) guidance on NHS Funded Nursing Care [9], the data were further examined to identify high, medium and low intensity users of total direct and indirect RGN time.* Results The number of residents in each of the seven main RUG-III clinical groups is shown in Figure 1. There were no residents in the three most resource-intensive rehabilitation groups and only two residents in each of the special rehabilitation low-adl group (1.04%) and extensive care group (1.04%). In fact, as was expected * Supplementary material is available to subscribers with the online version of the journal at the journal website. 280

3 Registered nurse contribution to care Figure 1. Distribution of residents by RUG-III clinical groups. for a nursing home population, the largest group of residents consisted of those in the reduced physical function group (43.01%). The distribution of wage-weighted time between residents in this study followed a similar pattern to that of the original UK hospital-based validation study [8] (Figure 2). The per cent of the variance in care time between residents explained by the RUG-III was 56% (Eta 2, P=0.0001; Table 1). Any figure above 30% is considered useful. There was a clear difference between the amount of care provided to residents in the RUG-III clinical groups that included those with particular medical and nursing needs (clinically complex and above or enhanced nursing care) compared with residents with primarily cognitive impairment or physical problems only (impaired cognition and below or standard nursing care). Residents in the enhanced nursing care group received more than twice Figure 2. Comparison of case-mix indices (CMI) of RUG- III groups from Joseph Rowntree Study and Inpatient RUG Validation Study [8]. Table 1. Variance explanation of distribution of wageweighted care time between RUG-III groups in each home Home Home 1 (n=55) Home 2 (n=29) Home 3 (n=53) Home 4 (n=56) Total Eta 2, P as much indirect RGN care time (t-test, P-0.001) and 1.4 times as much direct RGN and direct CA time (t-test, P-0.01) than residents in the standard nursing care group (Table 2). Residents with enhanced nursing care received an average of 48.1 min of RGN care in 24 h (95% CI ) compared with an average of 31.1 min (95% CI ) for residents in the standard nursing care group (Figure 3). Following publication of the DoH guidance [9], a third low care group was created by further subdividing the standard nursing care group according to the least physically dependent residents in the reduced physical function RUG-III group. Comparison of mean RGN time per group with 95% confidence intervals is shown in Figure 4. With three levels, the enhanced care group receives about 38% more than the standard group, and the low group receives about 50% of the standard group. The cut points in the RUG-III system are shown in Figure 5. When this time is adjusted for skill mix, an estimation of cost can be achieved. The number of residents in each group and the mean wage-weighted RGN time compared with the proposed DoH funding levels of each group is shown in Table 3. The ratios of minutes per 24 h between these groups matches remarkably closely to the ratios of proposed funding for the groups. Discussion Variance explanation The variance in time spent with residents explained by RUG-III in this study is in line with other published studies. The RUG-III case-mix system showed a clear difference in care time, with more care provided by Table 2. Mean (SD) minutes of care per 24 h for each resident by RUG-III clinical group RUG-III clinical group care RGN direct care RGN indirect care CA direct care CA indirect care Total clinically complex and above (n= 79) 35.3 (23.3) 12.9 (13.4) (74.3) 3.0 (5.9) Total below clinically complex (n =114) 25.4 (19.3) 5.7 (10.1) 73.1 (57.6) 2.6 (4.7) Ratio of 0clin complex: -clin complex t-test P-0.001, t-test P-0.01, t-test n.s. 281

4 I. Carpenter et al. Figure 3. Mean RGN time per 24 h for residents in the enhanced RGN group compared with those in the standard RGN group (showing 95% confidence intervals). Figure 4. Mean RGN time per 24 h for residents in the enhanced, standard and low RGN group (showing 95% confidence intervals). RGNs to residents with complex medical and nursing conditions than to those with less complex conditions. There are three main approaches to collecting the data for workload measurement, they are: retrospective task analysis by staff, work sampling of staff by outside observer, and self-reporting by staff. Work sampling by an outside observer and self-reporting by staff are more frequently referred to as time studies. In retrospective task analysis, a list of tasks is completed by nurses at the end of the shift, recording each task performed. In work sampling, an outside observer makes random observations at regular intervals during all shifts to determine how much time is spent on various tasks. In self-reporting, nurses document the amount of time they spend on each type of activity as they work, accounting for each minute of their shift. A number of studies have reported the utility of workload measurement systems, or time studies, to better understand the roles and activities of nurses [2, 7, 10 15]. Of these, only six detailed the activity differences between nurse grades, one in a long-term care facility. Three studies used self-report, and three used work sampling observation. Different terms were used to define the nursing grades and different interventions were recorded, making comparisons difficult. Comparisons are further complicated as the studies were conducted in different environments, e.g. psychiatry, rehabilitation, and long-term care. Nonetheless, with the exception of one, a consistent pattern emerged: more senior nurses spent a greater proportion of their time providing indirect care or carrying out unit-related activities than direct care, and nursing assistants spent a greater proportion of their time providing direct care. The RUG-III system has been extensively tested for validity and reliability in various care settings around the world using self-recorded time sheets [5, 8, 16 19] and in one using task analysis [20]. Use of the time sheet system compares favourably with alternative systems. Several workload studies have reported the effectiveness of task analysis based on a needs assessment. Others have cautioned against concentrating on the tasks performed for a number of reasons: decision-making is a key part of the nurse s role but is difficult to quantify as it is not a directly observable task; workload analysis systems may not account for variations due to environmental factors such as geographical situation, case-mix, facility size and services available; they ignore a nurse s ability to meet several needs simultaneously when caring for individual residents. Furthermore, although they describe how different grades of nurse provide various kinds of care, they do little to explain the variations involved in caring for specific individuals. Advantages of a self-report design, such as that used in RUG-III studies, is that environmental factors can be controlled by using differential care time between residents as the basis of analysis and time can be attributed to specific grades of nurse and between direct and indirect care time. Using patients records to determine need for nursing care according to stability and predictability has also been tested. The more unstable and unpredictable the resident s needs, the more complex the care required [21]. The authors found that while this method helped determine the level of complexity, decision making, and clinical judgement required to meet individual needs, it was not designed for determining required staff numbers. Ford and McCormack [22] have also suggested the use of stability and predictability using the RCN Assessment Tool for Older People. The tool was tested for acceptability, internal consistency, and inter-rater reliability in a pilot study of seven nursing homes where nurses received detailed training. The tool s three scales, essential care, stability/predictability and nursing assistance, were found to be internally consistent. The essential care scale had good inter-rater reliability. However the nursing assistance and stability/predictability scales showed disagreement between assessors, perhaps indicating a need for further training and further development of the tool. 282

5 Registered nurse contribution to care Figure 5. The RUG-III system showing proposed RNCC cut points. The tool proposed by the DoH for determining the registered nurse contribution to care (RNCC) [9] is based on the assessment of risk, stability and predictability by a nurse assessor reviewing a care plan. We believe that this approach will be prone to subjective variation between assessors. It is also possible that nurses will feel 283

6 I. Carpenter et al. Table 3. Mean (SD) wage-weighted registered nurse care time (min) by proposed RNCC funding level RNCC level (RUG-III) Wage-weighted time in 24 h Enhanced (n=79) (66.64) 110 Standard (n= 77) 72.6 (44.17) 70 Low (n=37) 36.5 (38.10) 35 Proposed DoH funding per week t-test between enhanced and standard P-0.01, t-test between standard and low P pressurised by conscience, home proprietors, residents or their family in a way that will influence their assessment as their decision will determine the contribution to their care that will be paid for by the NHS. It also places a high reliance on the quality of care plans that, in practice, are likely to be of variable standard and design, rendering them difficult to use in an equitable fashion. Our findings suggest that the RUG-III case-mix system effectively differentiated between nursing home residents who are receiving low, standard and enhanced RGN care time. These findings could provide the basis of a reimbursement system for registered nursing time in long-term care facilities in the UK, while it is important to note that the system does not state how much registered nurse care time a resident should receive. We believe that this approach, based on a validated objective assessment system, would be a more equitable approach to the task and would not place undue pressure on the assessors. In particular it meets the requirements of government policy for greater consistency in the provision of long-term care [23]. As the ratios of RGN care time between groups in this study match so closely with the DoH proposed funding levels, we believe there is a compelling argument to adopt this approach for widespread use. Key points. The RUG-III system has been well validated and explanation of distribution of care time by RUG in this study is in line with other published studies.. The system can differentiate groups of Nursing Homes residents who receive different amounts of Registered Nurse Contribution to their Care.. The RGN time provided to groups in this study relate very closely to the re-imbursment levels set by the DoH.. The proposed RUG-III system is an objective assessment that can inform the Assessor s decision and lessen the difficulties that arise from having to make a subjective judgement of existing care plans which may vary in their quality and content.. Final allocation to an RNCC level should be by the professional judgement of the Nurse Assessors appointed to conduct the determination. Acknowledgements The authors would like to thank the many people who have supported this work. In particular, the Joseph Rowntree Foundation for funding this piece of work and Cedric Dennis, Director of Care Services and Chrysa Apps, Practice Development Manager of the Foundation for their support given to this study. Thanks are also extended to the members of the Project Advisory Group for their support, advice and guidance throughout the work: Ian Ireland, Head of Quality Assurance BUPA Care Services; Stephen Sharp, Registration and Inspection Manager, North Yorkshire Health Authority; George Wood, Deputy Chief Executive, York Health NHS Trust; Professor Jan Reed, University of Northumbria at Newcastle; and Chris Vellenoweth, Independent consultant. Finally the authors would like to thank the staff and residents at the homes that took part in this study. References 1. Department of Health. The NHS Plan: the Government s response to the Royal Commission on Long Term Care. London: Department of Health, Fagerstrom L, Rainio AK. Professional assessment of optimal nursing care intensity level: a new method of assessing personnel resources for nursing care. J Clin Nurs 1999; 8: Morris JN, Hawes C, Fries BE et al. Designing the National Resident Assessment Instrument. Gerontologist 1990; 30: Hawes C, Morris J, Phillips C, Fries B, Murphy K, Mor V. Development of the nursing home Residents Assessment Instrument in the USA. Age Ageing 1997; 27 (Suppl 2): Fries BE, Schneider DP, Foley WJ, Gavazzi M, Burke R, Cornelius E. Refining a case-mix measure for nursing homes: Resource Utilisation Groups (RUG-III). Med Care 1994; 32: Department of Health. Single Assessment Process: Assessment tools and scales. 2002: scg/sap/toolsandscales/index.htm. 7. Flynn E, Heinzer MM, Radwanski M. A collaborative assessment of workload and patient care needs in four rehabilitation facilities. Rehabil Nurs 1999; 24: Carpenter GI, Main A, Turner GF. Case-mix for the elderly in-patient. Resource Utilisation Groups (RUGs) validation project. Age Ageing 1995; 24: Department of Health. NHS Funded Nursing Care. Practice guide and workbook. 2001: jointunit/freenursingcare/practiceguide.pdf. 10. Vanputte AW, Sovie MD, Tarcinale MA, Stunden AE. Accounting for patient acuity: the nursing time dimension. Nurs Manage 1985; 16:

7 Registered nurse contribution to care 11. Hagerty BK, Chang RS, Spengler CD. Work sampling. Analyzing nursing staff productivity. J Nurs Adm 1985; 15: Robichaud AM, Hamric AB. Time documentation of clinical nurse specialist activities. J Nurs Adm 1986; 16: Hendrickson G, Doddato TM, Kovner CT. How do nurses use their time? J Nurs Adm 1990; 20: Bridel JE. Why measure workload? Prof Nurse 1993; 8: Freeman S, Shelley G, Gay M, Ingram B. Measuring services: a district nursing dependency tool. Nurs Stand 1999; 13: Bjorkgren MA, Fries BE, Shugarman LR. A RUG-III case-mix system for home care. Can J Aging 2000; 19 (Suppl 2): Fritjers D, Van der Kooij C. Resource Utilisation Groups for nursing home patients in the Netherlands. SIG, Dutch Centre for Health Care Information, Utrecht, Ikegami N, Fries BE, Takagi Y, Ikeda S, Ibe T. Applying RUG-III in Japanese long-term care facilities. Gerontologist 1994; 34: Ljunggren G, Fries B, Winblad U. International validation and reliability testing of a patient classification system for longterm care. Eur J Gerontology 1992; 1: Carillo E, Garcia-Altes A, Peiro S et al. Sistema de clasificacion de pacientes en centros de media y larga estancia: los Resource Utilisation Groups, version III. Validacion en Espana. Rev Gerontol 1996; 6: Fulton TR, Wilden BM. Patient requirements for nursing care: the development of an instrument. Can J Nurs Adm 1998; 11: Ford P, McCormack B. Determining older people s need for registered nursing in continuing healthcare: the contribution of the Royal College of Nursing s Older People Assessment Tool. J Clin Nurs 1999; 8: Cm4169. Modernising social services: Promoting independence, improving protection, raising standards. London: Department of Health, Received 3 May 2002; accepted in revised form 30 October