The third step weighs the NRGs according to time and skills required for care administration determined by Delphi studies.

Transcription

1 Development and use of Nursing Related Groups in the Belgian Budget of Financial Means for hospitals. Delphine Beauport, Arabella D Havé, Federal Public Service of Health, Food Chain Safety and Environment Introduction The Belgian budget of financial means for hospitals (BFM) takes into account nursing care independently from the DRG system. It does so by adding a nationwide Belgian Nursing Minimum Dataset (BNMDS) to its hospital discharge dataset. The BFM is a closed-end budget determined at the national level and allocated to individual hospitals consisting of three major components (A, B, C). The BNMDS is used, amongst other parameters, to calculate and allocate B2 - the budget for clinical costs of the BFM. B2 is based on a points system consisting of basic (to cover basic activities) and supplementary points (depending on activity and care profile). The current BFM currently uses one version of the BNMDS whilst gradually introducing a revised version. A new calculation method, called the Nursing Related Group (NRG) has been developed and applied to the revised BNMDS. An NRG consists of care episodes that are homogeneous for nursing care. The objective of the NRG is to assign a predefined care profile to each nursing care episode in order to enable a financing system that takes into account differences in intensity of nursing care between hospitals and thus determines the supplementary points. Methods NRGs are developed based on 78 nursing interventions of the revised B-NMDS, the type of nursing unit, age, occurrence of surgery, time in hospital stay relative to surgery/delivery. The development consists of four steps: creation, assignment, weighing and application. The first step creates groups of care episodes that meet predefined criteria using the BNMDS of a full registration year. For each group of care episodes, an average care profile is calculated, based on the scores for 78 nursing interventions within the group, using RIDIT-analysis. The predefined criteria are determined by location, moment, duration of the care episode, age and medical and nursing aspects of the patient. The second steps assigns a NRG to a care episode using two indicators, a main indicator an a secondary indicator. Not all NRGs can be assigned to all care episodes. Restrictions such as the duration of the care episode, the number of interventions registered during a care episode (e.g. the number of interventions registered during a care episode needs to be higher than 3) and a list of criteria for specific NRGs need to be taken into account (e.g. the NRG care to newborn can only be applied to patient younger than 1 year). The third step weighs the NRGs according to time and skills required for care administration determined by Delphi studies.

2 The fourth step is the application of the NRGs within the BFM. All scores (i.e. NRG-weights for each care episode) are added and the share of each hospital in the whole of care profiles (type and volume) is calculated. Supplementary B2-points are assigned based on the share of weighted patient days with a higher weight than the median national weight per patient day. The median functions as a proxy for recorded and performed basic nursing activities. Results 31 NRGs are created, of which 21 NRGs apply to care episodes that last 24 hours (i.e. an entire inpatient day) and 20 apply to care episodes lasting less than 24 hours. 11 NRGs apply only to episodes lasting less than 24 hours. Each NRG is weighted based on required staffing levels. All NRG-weights for each care episode are added and the share of each hospital in the whole of care profiles (type and volume) is calculated. Because the impact of this new methodology on the budget of the individual hospitals is uncertain and in order to avoid too large fluctuations in the BFM, calculations of supplementary points are based for 70% on the original BNMDS calculation method and for 30% on NRGs. In order to shift to a higher share of the NRGs in the BFM, the quality of the data was further analysed and audited by the Federal Public Service of Public Health. As a result of this, 5 interventions were removed from the revised BNMDS after outlier-analysis and supplementary PCA-analysis. The scores for these interventions were impacting the homogeneity of the NRG (due to coding practice) and Principal Components Analysis showed clinically illogic correlations between interventions. Furthermore, no correcting measures for coding practice could be identified. The revised B-NMDS was tested on inter-rater reliability and validity when both instruments were developed and it met the international accepted psychometric properties. Conclusions In order to shift to a 100% share of NRGs, further refinements of the system are planned, such as further improvement of the quality of the data, decoupling the NRG from structural parameters such as type of nursing unit, the creation of intermediate NRGs between low nursing intensive and high nursing intensive NRGs and calculation of adjusted weights for care episodes lasting less than 24 hours and lasting 24 hours.

3 1 Development and use of Nursing Related Groups (NRG) in the Belgian Budget of Financial Means for hospitals. Arabella D Havé Chief of Terminologie, Classification, Grouping & Audit Federal Public Service of Health, Food Chain Safety and Environment

4 2 Background Costs of nursing staff account for approximately 50% of the total personnel budget and 20 to 30% of the hospital running costs (Welton et al., 2006) High impact on the hospital budget Knowledge gap about the actual relationship between cost of nursing and reimbursement for nursing care DRGs are not very homogeneous to nursing care DRGs only explain 20% to 40% in the variability of nursing care Coefficients of variation for nursing care per DRG are reported varying from 0.22 tot 2.56 But: completing DRG-data with nursing data can improve the prediction for total hospital length-of-stay, total ICU-days and total charges with about 30% (Welton et al., 2005)

5 3 Financing of hospital care in Belgium Introduction Source:

6 4 Introduction 01 Budget of Financial Means (BFM) 02 Payment per medical activity (nomenclature) Budget for drugs Agreements hospital - insurance Patient 2% 10-15% 5% 05 40% 40%

7 5 Introduction 01 Budget of Financial Means (BFM) Fixed budget Part A: capital costs Part B: operational costs (88%) B1: general operational costs Variable budget Supplementary Budget B2: clinical costs (48%) B3: medical-technical departments B4: fixed clinical costs B5: pharmacy costs Part C: corrective measures 40%

8 6 Problem statement Variable budget Fixed Budget C/D-wards E-Wards Intensive Care 20% Nomenclature Deciles 70% 20% 40% Nomenclature Nomenclature (REA) APR-DRG Deciles Deciles Deciles 80% 30% 40% B-NMDS-I (70%) B-NMDS-II (30%) B-NMDS-I (70%) B-NMDS-II (30%) B-NMDS-I (70%) B-NMDS-II (30%) Deciles Market Share Deciles Market Share Deciles

9 Problem statement Basic nursing care B-NMDS-I map Continuum of intensity of care Zones with intensive care profile (ZIP) Technical nursing care

10 8 Problem statement Belgian hospital financing system takes nursing care data into account, however: Complex and non-transparent system: different criteria, costweights, deciles according to the type of beds Some nursing wards left out of the system: e.g. geriatric nursing wards Calibration of the NMDS-zones is based on actual staffing ratios, which favours nursing wards with high nurse staffing levels Use of the deciles: global increase in nursing intensity in the Belgian hospitals is not taken into account Small differences between hospitals belonging to different deciles can result in big differences in budget Big existing differences between hospitals within the same decile aren t taken into account.

11 9 Problem statement There is no link to APR-DRG The use of B-NMDS version I (B-NMDS-I) within this system is also questioned: This version was developed in 1985 and implemented in 1988 To which degree this version is still a correct measure to differentiate hospital nursing care between settings in its staffing and resource needs?

12 10 Method Development of B-MNDS-II 78 nursing interventions Implementation of new dataset 2008 Nursing Related Groups Budget of Financial Means 2014 Type of nursing unit Age Occurence of surgery LOS

13 11 Four step development The first step creates groups of care episodes that meet predefined criteria using the BNMDS of a full registration year NRGs The second steps assigns a NRG to a care episode using two indicators, a main indicator an a secondary indicator. 04 The third step weighs the NRGs according to time and skills required for administration of care determined by Delphi studies. The fourth step is the application of the NRGs within the Budget of Financial Means for hospitals.

14 12 Method Creation of groups of care episodes meeting predefined criteria and using B-NMDS-II of a full registration year Age Medical and nursing characteristics of the patient Location Moment Duration of the care episode 24 hours or an entire inpatient day < 24 hours For each group of care episodes, an average care profile is calculated, based on the scores for 78 nursing interventions within the group, using RIDIT-analysis.

15 13 Method Moment Type of unit Periods of 24h NO YES (21) Periods of <24h YES (11) YES (20) 31 NRGs are created, of which 21 NRGs apply to care episodes that last 24 hours 20 apply to care episodes lasting less than 24 hours 11 apply only to episodes lasting less than 24 hours

16 14 Examples NRGs Basis groep Type VPE Andere criteria NRG 1 CI, HI Type DRG, passage op OPR 1 - Chirurgische intensieve zorgen 2 DI, HI 2 Niet-chirurgische intensieve zorgen A ward is characterized by its main bedindex Label of NRG 3 EI Leeftijd 3 - Pediatrische intensieve zorgen 4 NI 4 - Neonatale intensieve zorgen 5 BR 5 - Zorgen aan gecompliceerde wonden Intensive NRG s 6 AR, OB, M, MI Dag van de bevalling, items W (Moeders) 6 - Obstetrische zorg/kraamzorg op de dag van de bevalling (Moeders) 7 M 7 - Kraamzorg (Moeders) 8 MI 8 - Zorg bij hoogrisicozwangerschappen (Moeders) 21 NRGs Type unit Other criteria can be necessary to identify the NRG 9 M Leeftijd (Pasgeborenen) e.g.: DRG-type - surgical or medical; performance of procedure or intervention 10 N 10 - Neonatale zorg 9 - Zorg op de materniteit (Pasgeborenen) 11 E Leeftijd 11 - Pediatrische zorg

17 15 Graphical representation of an NRG Items (N=78) V200 Pressure ulcer prevention: Repositioning MEAN RIDIT = 0,565 RIDIT scale (From 0 to 1) RIDIT = 0,5 Reference Population Red: MEAN RIDIT NRG geriatric care (G 24u) Green: MEAN RIDIT NRG care for contagious patients (L 24u)

18 16 Method An NRG is assigned to each care episode Two indicators: a primary indicator and a secondary indicator. Primary indicator Calculation: Adjusted MEAN RIDIT of all 78 items IF MEAN RIDIT of item within NRG > = 0. 5, THEN RIDIT used to calculate primary indicator ELSE 1 RIDIT used to calculate primary indicator Secondary indicator Calculation: Adjusted MEAN RIDIT of specific items Items with exceptionally high or low RIDIT (25% treshold) If NRG cannot be assigned using the primary indicator, the NRG is assigned using the secondary indicator

19 17 Example primary indicator NRG Geriatric Care K300: Mechanical ventilation Occurence less probable -> adjusted or 1-RIDIT K300 V200 Pressure ulcer prevention: Repositioning Occurence more probable -> no adjustment or RIDIT V200 Primary indicator = MEAN (Ridit A100,, (1 Ridit K300),, Ridit V200,, Ridit Z400)

20 18 Example secondary indicator NRG Geriatric Care Orange: RIDIT >= upper treshold; GREY: RIDIT <= lower treshold K300: Mechanical ventilation Occurence less probable -> adjusted or 1-RIDIT K300 V200 Pressure ulcer prevention: Repositioning Not specific for NRG -> excluded from calculation of secondary indicator Secondary indicator = MEAN (Ridit A100, (1 Ridit B100), Ridit B200, Ridit B400, ( 1 Ridit Z200), Ridit Z300)

21 19 Weighing All items scores are weighted by time and competence needed to provide this type of care Time and competence are determined through Delphi-approach Conversion of score in points for time: 1 point corresponds with approx. 5 minutes of care Points of time are weighted by competence : Weight from 1 to 5

22 20 Application C/D-wards 20% Nomenclature Deciles 80% B-NMDS-I (70%) B-NMDS-II (30%) Deciles Market Share Variable budget Fixed Budget E-Wards Intensive Care 70% 20% 40% Nomenclature Nomenclature (REA) APR-DRG Deciles Deciles Deciles 30% 40% B-NMDS-I (70%) B-NMDS-II (30%) B-NMDS-I (70%) B-NMDS-II (30%) Deciles Market Share Deciles

23 21 Application NRG-points for 3 groups of care episodes are used: For each episode of care the number of NRG-points is calculated Points are added up for every finance group (CD, E or I) for every patient day -> point per patientday Median of points per patientday of finance group CD sets the threshold for what is financed through the variable budget Part above the threshold is taken into account for the variable budget

24 22 Application Only points per patientday > median CD are taken into account for variable budget: Points per patientday median CD This patientday receives no variable budget

25 23 Conclusion Points of improvement: There is no link with DRGs A major problem in linking both datasets is the different design of data-collection. The B-HDDS is a summary of the hospital stay, collected for all hospitalized patients at time of discharge. The B-NMDS uses a cross-sectional data collection method for a balanced sample of inpatient days. As a consequence, not all DRG s have sufficient nursing data and not all nursing data are representative for the stay of a DRG Nursing intensive departments such as geriatrics are not included in the complementary financing scheme Many financial incentives are focused on reducing the length of stay without considering the compression of nursing care during that stay

26 24 Conclusion Further improvements of the system Further improvement of the quality of the data 5 items have been removed which impacted the homogeneity of the NRG Audit The creation of intermediate NRGs between low nursing intensive and high nursing intensive NRGs Calculation of adjusted weights for care episodes lasting less than 24 hours and lasting 24 hours

27 25 THANK YOU FOR YOUR KIND ATTENTION