Assessing the causes inducing lengthening of hospital stays by means of the Appropriateness Evaluation Protocol

Health Policy 99 (2011) 66–71

Contents lists available at ScienceDirect

Health Policy journal homepage: www.elsevier.com/locate/healthpol

Assessing the causes inducing lengthening of hospital stays by means of the Appropriateness Evaluation Protocol Pierre Fontaine a,∗ , Jessica Jacques b , Daniel Gillain a , Walter Sermeus c , Philippe Kolh a , Pierre Gillet a a b c

University Hospital of Liège, Liège, Belgium University of Liège, Liège, Belgium Centre for Health Services & Nursing Research, Catholic University Leuven, KULeuven, Belgium

a r t i c l e

i n f o

Keywords: Appropriateness Evaluation Protocol Belgium Length of stay Inappropriate patient stay Inappropriate hospitalisation day Management audit Medical records Retrospective studies

a b s t r a c t Objectives: The objective is to evaluate the use of the Appropriateness Evaluation Protocol (AEP) as a screening tool for determining the causes of the non-justified days to help hospitals to decrease the length of stay while preserving the quality of care. Methods: Three successive cross-sectional surveys were conducted from 2003 till 2005, in 23 Belgian hospitals. During this period, 10 921 days were audited by means of the AEP. This study is focused on adult acute non-intensive care units. The appropriateness of each day of the sample was assessed, and for those considered as inappropriate, the reasons explaining the prolongation of the stay were investigated. Results: The proportion of inappropriate days was 24.61%. There is a high variability across specialties and hospitals. Regarding inappropriate days, the analysis of causes of prolongation, globally, by bed index or by hospital, indicated clearly internal and external factors that lengthen stays. The most frequent reasons are waits for an examination (22%) and the lack of extra-hospital structures (31%). Conclusions: The use of AEP as a tool of internal audit to measure the proportion of nonjustified days and their causes turns out to be possible and the obtained results has provided some accurate and useful information for the participating, and allowed them to take concrete decisions which lead to shrinking of the length of hospital stay. © 2010 Elsevier Ireland Ltd. All rights reserved.

1. Introduction The evolution of population health in industrialized countries increased the life expectancy of newborns of 8.6 years between 1960 and 2000 [1]. Similar evolution has been seen in Belgium. It has a health system that has an easy access to care and is of a quality level that is recognized as widely satisfactory, without generating long waiting lists for certain treatments [2].

∗ Corresponding author at: Département d’Information Médicale, CHR de Huy, rue des Trois Ponts 2, 4500 Huy, Belgium. Tel.: +32 85 848230; fax: +32 85 848236. E-mail address: [email protected] (P. Fontaine). 0168-8510/$ – see front matter © 2010 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.healthpol.2010.06.011

As in other industrial countries, health spending in Belgium grows faster than its Gross Domestic Product (GDP). According to the Organisation for Economic Co-operation and Development (OECD) [1], Belgium was ranked 11th among the countries of the EC in 1980, and 9th in 1990 with regard to expenditure for health care as part of the GDP. In 2003, Belgium reached the 3rd rank after Germany and France that alongwith Belgium, it is true, are the only countries of the EC where no waiting list has appeared [2]. To protect a fair access to the hospital care and avoid an uncontrolled explosion of the healthcare spending, the federal authorities had developed a mixed system of hospital financing. It is based on two main axes: a fee-for-service system for medical activities and a lump-sum financing system for the operations of the hospital (known as Budget

P. Fontaine et al. / Health Policy 99 (2011) 66–71

Financial Means – BFM) including the budget that would allow the hospital to budget the functioning of hospital departments (nursing units, operating theatre, emergency rooms and the central sterilisation). At present, adaptations have essentially influenced this second axis. Since 2002, the BFM is determined based on a measure of “justified” activity, which is calculated by applying to the hospital stay a standard length of stay according to the Allpatient Refined Diagnosis Related Groups (APR-DRG), the severity-of-illness category and the age group (below and above age 75). Based on these algorithms, a theoretical number of “justified” days are calculated that will be reimbursed. If the actual length of stay is longer, the hospital has to finance the related costs itself. If the actual length of stay is shorter, some profit is made. It creates an incentive for the hospital administrator to influence the medical teams to reduce the length of stay. This decrease in the length of stay however may result in lowering of the quality of care. For that reason, it is essential for the administrator to be able to determine the reasons for avoidable inpatient days. In this perspective, the Appropriateness Evaluation Protocol (AEP) is one of the most valid tools [3–9]. The AEP assesses the relevance of an admission or of a hospitalisation day through explicit, predefined criteria, which are related to care but independent of the pathology [3]. It is an instrument applicable to all adult patients admitted in a medical, surgical or gynaecological service [4–9]. 2. Methods Since the aim of this study is to determine the causes inducing an additional day in hospital, the relevance of admissions is not considered here. The AEP algorithm usually consists of about 20 explicit and objective criteria. If at least one criterion is met, the inpatient day is justified. The protocol consists of three kinds of criteria: the criteria relating to medical activity (part A), the criteria relating to nursing activity (part B), and the criteria relating to the state of the patient (part C). Several adaptations of AEP have been made in French, Dutch and other languages using a fourth section with a list of explanatory criteria of inappropriate stays. It identifies the reasons associated to the days estimated by the first part as inappropriate. It allows to understanding why the patient is hospitalised although based on the objective AEPcriteria in part A–C there was no justification [10–16]. This section D is more variable than the three earlier sections due to the differences of health policies between the considered countries. The inappropriate presence of the patient at the hospital can be due to endogenous causes (dysfunction of the hospital organization) or exogenous (external dysfunction, related to the patients, to the medical-social context or to the unavailability of extramuros beds). Within the framework of the study, the questionnaire AEP was adapted to fit the characteristics of Belgian healthcare. The AEP-version used in this study is based on the original AEP form from the United States (US-AEP) [4]. Other versions available in the literature, and especially

67

the French AEP [10], the Dutch AEP [11], and the European AEP [13–14,16], were also analysed to develop a protocol of evaluation of the inpatient days adapted to the requirements of the Belgian health care. We wrote it both in French and Dutch, which are the main languages in Belgium. Thus an expert group took out of each AEP what they considered as the best criterion of the parts A–C in order to take into account the reality of the everyday work in the nursing units, and the availability or unavailability of some services out of the hospital structure. Nevertheless, it seemed important to us to preserve the initial structure to be able to compare the results obtained in this study with those appearing in the literature. The modifications consisted essentially of some precisions of quantitative order (thresholds, standards, values associated to the criteria). The care at present assumed by hospitals, nursing homes and chronic care facilities (RHHRC), and community care as well as medical practice in Belgium served as base for the identifying these values. To make AEP operational, a user manual was written with a precise description of every criterion. In particular, it resumed the definitions clarifying the contents of the criterion, the standards, the characteristics, the exceptions which are to be taken into account for the data collection. This manual proved to be most helpful to get a uniform interpretation of items so that similar care situations are registered everywhere in the same way. A specific training was given to assessors by the team of the study. Parts A–C of the AEP form have the objective to assess if inpatient day is appropriate. A fourth section was added to highlight the reasons associated with the inappropriate days: the D part. In the case of an inappropriate day, either the patient still needs acute care and is waiting for them, or the patient is at the end of his stay waiting for the discharge. In the first case, it is necessary to identify the service awaited by the patient and the reason for which this service is not given. In the second, it is necessary to define the reason why the patient stays one more day. This section must be filled in if none of the clinical criteria from part A to C would justify the inpatient day. Its content was defined through a pilot study. The identified reasons can be associated with various organizational levels of the health system: waiting for the results of an examination, delays for appointments, waiting for diagnostic procedures or treatment, gaps in the continuity from hospital care to home care. Next to healthcare organizational issues, some reasons for unjustified hospital day are related to the willingness of the family to accept the patient to return home or home not being ready yet for his return. Several open questions were added to control the relevance and the exactness of the answers without having to check the patient record later again. We increased the accuracy by adding sub-divisions to the existing items. Each reason was thus reviewed to identify all the situations which can be met by the patient during his hospital stay. Every main group of reasons was detailed in a precise way to facilitate its use and gain time. Their inventory took into account the characteristics and the functioning of the

68

P. Fontaine et al. / Health Policy 99 (2011) 66–71

hospital. Several open questions also allowed the assessor to add specific reasons if the list did not include the one adapted to a specific situation.

2.1. Sample A first phase of the study was conducted in one hospital. 611 inpatient days were reviewed by both medical doctors and nursing staff, and analysed to test and to adapt the questionnaire. The data of this test are not taken into account in the results described below but the study showed an acceptable inter-rater comparison rate (Cohen’s kappa = 0.7). Thanks to this conclusion, the assessors of the survey described here were only nursing staff members, since the validation study showed that there was no interest in requesting a contribution from Medicine doctors. The following phases took place from December 2002 till March 2005. Sixty-five hospitals were contacted for participating in this study. There was no additional funding for these hospitals but every participating hospital was promised an individualized feedback being benchmarked anonymously to all other participating hospitals. Globally, 23 hospitals participated. The days of AEP-assessment were selected randomly. The weekends were excluded because hospital activity is weak during these days. An interval of 48 h between two periods of assessments was taken. All the present patients in the care units during the days of inquiry were selected. To be able to analyse in the factors justifying a stay, and although AEP only requires a single criterion, the assessors were asked to check all the criteria corresponding to clinical condition of the patient.

2.2. Analysis For the statistical treatment, all the collected data were grouped in a single Microsoft Excel® database. The file was imported to SPSS (SPSS® 13.0, SPSS Inc. Headquarters, 233 S. Wacker Drive, 11th floor Chicago, Illinois 60606). The comparisons of percentages were realized by the 2 test. In case of a number of cases lower than the acceptable values, a Fisher exact-test was preferred.

3. Results The sample is composed of 12 978 inpatient days collected in 23 hospitals. According to the protocol, each inpatient day corresponding to the day of discharge (n = 945) was eliminated from the sample. In the final sample, only admissions on surgical, internal medicine, geriatrics and mixed surgical and internal medicine wards were retained what corresponds to a sample of 10 921 days. The number of inpatient days (ID) in surgery (n = 3460) is widely lower than that of medicine (n = 5165) although the number of admissions was equal. It is explained by longer length of stay in medicine than in surgery. The rate of justified days is the ratio between the number of justified days and the total number of days. The proportion of justified days is globally 75.39%. The surgical and internal medicine beds have rather similar rates of justified days, around 75%. The inpatient days in mixed wards are proportionally slightly more justified (80.8%). This is due to the fact that the cases which are assessed come from oncological (66%) and cardiological wards (24%). We also note a particularly low score for geriatric beds (63.3%). We notice that the proportion of justified days decreases for the patients older than 70 years. Analysis of the causes of inappropriateness is particularly sensitive on elderly population. Likewise, the comparison between the rates of justified days according to whether the measurement day is a Friday or another day of the week showed that Friday is characterized by a particularly high rate of justified days with 81.27% against 74.56% on the other days (p2 < 0.001). The analysis of the sample reveals that Fridays are usually characterized by a particularly high proportion of discharge of patients. The sample for this day is thus rather weak with regard to the other days of the week. It does not seem unreasonable to think that both effects are associated. Actually, the hospital will be probably pushing to discharge patients for whom a supplementary day is not absolutely necessary if we are on the eve of the weekend. The analysis of the part D of the non-justified days is the most relevant in term of hospital management. Part D is divided into 7 types of explanation, the relative frequencies are given in Table 1.

Table 1 Rate of each criterion explaining inappropriate days. Explanation criteria

Surgery

Internal medicine

Pediatry

D1 waiting for an operating room medical or surgical procedure D2 waiting for a non-operating room therapeutic or diagnostic procedure D3 waiting for a medical opinion D4 procedure or examination could have taken place today but has been delayed D5 waiting for examination results D6 patient could get out but the exit is delayed D7 other explanations

31%

4%

29%

9%

32%

5% 2%

Geriatry

Mixed (surgery and internal medicine)

Total

2%

11%

12%

7%

22%

21%

22%

11% 4%

14% 0%

16% 1%

8% 3%

10% 3%

4% 22%

8% 31%

21% 7%

10% 49%

3% 13%

7% 31%

4%

4%

21%

2%

3%

4%

P. Fontaine et al. / Health Policy 99 (2011) 66–71

69

Fig. 1. Detail of D6 explanation criterion: patient could exit but exit delayed.

Globally, we observe a predominance of the D2 criteria (patient in wait of a surgical or medical operation not requiring the surgical unit) and D6 (the patient could be discharged, but his discharge is delayed). These are mostly true for internal medicine beds, whereas, in geriatric beds, it is clearly D6 which is dominating. With regard to surgery beds, it is D1 (patient in wait of a surgical or medical operation requiring the surgical unit) that are met most frequently. After further analysis, it turns out that D1 corresponds to the admission day which is still often the day before the surgical operation, though it is not medically needed. We particularly notice a difference between hospitals depending on whether they developed a consultation of ambulatory preoperative anaesthesia, or not. On the other hand, we shall note, in internal medicine, a problem concerning the level of the waits for examinations, and more particularly the ultrasound (D2-5), doppler ultrasound (D2-8) and scintigraphies (D2-14). The waits for a scanner (D2-13), are seen as well in medicine as in geriatrics. Concerning the unjustified days due to the wait of a medical opinion (D3), we note in geriatric care units that the waits of opinion of the cardiologists (D3-1) and the neurologists (D3-8) generate proportionally the most unjustified days. The criteria corresponding to the postponement of an examination or a procedure (D4) and to the wait of medical results (D5) remain globally rare. Fig. 1 shows very clearly the impact of a lack of extrahospital capacity on the length of stay in the hospital. We note more particularly a percentage of the unjustified days of 21% in geriatrics because of the difficulties to get places in a rehabilitation centres or nursing homes.

Furthermore, we note that this problem of placement seems more important in the Dutch-speaking region (44% of the unjustified days) than in the French-speaking region (17%). 4. Discussion These explanation criteria bring to the hospitable administrators relevant elements relating to the factors which generate inappropriate days. These are both connected to the internal organizational gaps and to the broader healthcare environment. From then on, the interpretation of the D part of a given hospital can only be made by those who know the respective internal and external environments of the hospital. Some hospitals, on the basis of this diagnosis, were able to bring an adequate answer to factors generating nonjustified days. On the other hand, the sample is so important that the global results provide a good indicator of the main causes, and thus the control levers on which the public authority can act. Recruitment of enquired hospitals was on a voluntary basis, which may induce a bias of selection. On reading these results, we realize that the reasons for which a patient who could leave the hospital but stays at least one more night are bound not only to the organization of the hospital itself, but also to other elements on which it has no or little influence. It is of course the case of extramural facilities, but that of the patient himself or his family as well. It is thus important to distinguish the endogenous causes, for which corrective actions can be taken by the hos-

70

P. Fontaine et al. / Health Policy 99 (2011) 66–71

Table 2 Percentages of unjustified days explained by endogenous or exogenous criterions per bed index. Surgery Endo-endogenous Endo-exogenous Exogenous Undetermined Without

4% 51% 18% 18% 14%

Internal medicine

Pediatry

Geriatry

5% 55% 24% 18% 6%

0% 73% 9% 27% 9%

3% 46% 33% 25% 6%

pital, from the exogenous causes on which it cannot act. So, we notice that with the exception of the criteria D6-0-1 to D6-0-3, the explanatory factor D6 is essentially bound to the place of residence and to the extramural facilities. This low proportion of justified days for older patients is connected with the higher proportion of criteria D6. Now, the hospital has not enough influence on these elements. On the contrary, the D2 criterion, which explains about a quarter of the unjustified days, is directly bound to the hospital organization and, more exactly, to the coordination between the care units and the ancillary care units. The detail of this criterion provides the hospital administrator an objective and relevant vision of the sources of problems. Among endogenous factors, we make a difference: those pertaining directly to the organization of the care unit (the “endo-endogenous” causes), and those dependent on the organization of the other services, in particular the ancillary care units (the “endo-exogenous” causes). Besides these three classes, there are two other categories of criteria: the “indefinite” ones, for the criteria of explanation for which it is impossible to determine if they recover or not from the organization of the hospital, and the “without” ones for which no criterion D was validated. These classifications result from a consensus, and do not have, for some at least, an absolute character. For example, the criterion “the patient refuses to return” is considered as exogenous, because it results from the will of the patient to continue the stay. However, preparing the patient for the discharge is partially the duty of the hospital staff too. Table 2 shows that a small majority of the unjustified days (51%) are due to a lack of coordination of the care unit with the various ancillary services, or also with the medical profession. This observation infers that there is, for the hospital administrator, a wide action scope to improve the length of stays, while protecting the quality of cares provided to the patient. 5. Conclusion The tool chosen to judge the opportunity of the stays (admission and hospitalisation day) is the AEP. This instrument allows to assess the appropriateness of an admission or of an inpatient day through explicit, predefined criteria, which are relative to the care but independent from the treated pathology [3]. This choice was motivated by the wide availability of literature, the acknowledged reliability and validity of this tool, its relatively easy adaptation to the Belgian context and, naturally, the perfect adequacy between the purpose of this instrument and the objectives

Surgery/internal medicine 4% 45% 9% 6% 37%

Total 4% 51% 23% 19% 11%

of the survey. One of the main qualities of this tool is that it is independent of the pathology, and is thus going to justify the stay only by objective criteria related to the care supplied to the patient, or to his state. However, this characteristic is both a strength and a weakness of the tool. If the AEP allows to justify a day through the provided care or services absolutely requiring an acute hospital environment, at no time it gives information about the opportunity of these care and services themselves. For example, we cannot imagine to perform a surgical operation out of the hospital, but was this operation absolutely necessary? The survey was realized on a large scale to constitute a representative database. To do so, the recording of the protocol was realized for all the present patients during the days of inquiry (cross-sectional). The data were processed to supply at every participating hospital a feedback analysing its own situation and placing it with regard to all the other hospitals. This allowed to achieve one of the objectives of the inquiry, namely the constitution of a tool of internal audit to be used by the hospital managers. The relevance of the results, the reactions of the hospital managers having participated, as well as the concrete measures taken thanks to the AEP audits plead all in favour of the utility of these inquiries. Indeed, AEP provides not only an analysis of their level of performance for the justified days and admissions, but also determines the causes leading to unjustified stays. According to their own statements, this last part proved to be particularly interesting for the administrators of the participating hospitals. This is showing that length of stay is not only related to endogenous organizational factors but also the exogenous factors, such as lack of sufficient extramural capacity. The latter is underestimated in the hospital financing regulations, which put an incentive on reducing length of stay as a hospital responsibility and not as a regional responsibility. This last point, taken on the global inquiry level, estimates the needs of non-hospital structures and takes on a particular interest for authorities, or for hospital associations. The analysis of the D part allowed us to assess dysfunctions, lacks of coordination between care units and ancillary care services, and to quantify them in supplementary days. It would allow to assess the impact on the length of stay, which was not done before. The corrective actions often concerned the medical and nursing rituals and were not always easy to implement. For example, this is the case for hospitals systematically admitting surgical patients a day before surgery, or planning some examinations only at a fixed day. Other corrections were more of technical order. We remember a hospital where they received the results of clinical biology with one day delay.

P. Fontaine et al. / Health Policy 99 (2011) 66–71

In every case, these actions undertaken on the basis of the AEP diagnoses allow a gain of efficiency leading to a better organization of the hospital stay, and a decrease in the length of stay while the health interests of the patient are preserved, or even improved because the risk of iatrogenic infection will be decreased. Besides, the observations could be linked with the analysis of “clinical pathways”. Indeed, within the framework of a planning and of a standardization of some stays, it can only be beneficial to determine a profile of the sources of inappropriate days as well as their causes. Reducing the time spent in the hospital is a heavy, inevitable trend. This fact, if not managed can turn out seriously harmful in the quality of care. Conversely, a detailed analysis of the non-justified days will allow to identify the internal and external causes, and to bring corrective actions, to the advantage of the patient. Acknowledgement Sources of support: Federal Public Service (FPS) Health, Food Chain Safety and Environment, Department of Care Institutions.

[6]

[7]

[8]

[9]

[10]

[11]

[12]

[13]

References [1] Organisation for Economic Co-operation and Development, Écosanté OCDE. Paris: OCDE; 2003. [2] Marchand M. L’assurance maladie à la croisée des chemins. Reflets et Perspectives de la Vie Economique 1992;31:69–92. [3] Winterhalter G, Blanc T, Kulczyki E. Importance et causes de l’utilisation inappropriée identifiée à l’hôpital de St-Loup. Rapport N◦ 1/version du 11 mars 1991. Hôpital de zone de St-Loup/Orbe. Lausanne [unpublished]. [4] Gertman PM, Restuccia JD. The Appropriateness Evaluation Protocol: a technique for assessing unnecessary days of hospital care. Medical Care 1981;19(8):855–71. [5] Kalant N, Berlinguet M, Diodati J, Dragatakis L, Marcotte F. How valid are utilization review tools in assessing appropriate use of acute care

[14]

[15]

[16]

71

beds? Canadian Medical Association Journal 2000;162(13):1809– 13. Kaya S, Vural G, Eroglu K, Sain G, Mersin H, Karabeyoglu M, et al. Liability and validity of the Appropriateness Evaluation Protocol in Turkey. International Journal for Quality in Health Care 2000;12(4):325–9. Peiro S, Meneu R, Rosello ML, Portella E, Carbonell-Sanchis R, Fernandez C, et al. Validity of the protocol for evaluating the inappropriate use of hospitalization. Medical Care (Barcelona) 1996;107:124–9. Smeets P, Verheggen F, Pop P, Panis L, Carpay J. Assessing the necessity of hospital stay by means of the Appropriateness Evaluation Protocol: how strong is the evidence to proceed? International Journal for Quality in Health Care 2000;12(6):483–93. Strumwasser I, Papanjpe NV, Ronis DL, Share D, Sell L. Reliability and validity of utilization review criteria. Appropriateness evaluation protocol, standardized medreview instrument, and intensity-severity-discharge criteria. Medical Care 1990;28:95–111. Lombard I, Lahmek P, Diène E, Monnet E, Logerot H, Levy Soussan M, et al. Etude de la concordance inter-observateurs des raisons de la non-pertinence des journées d’hospitalisation identifiée par la version franc¸aise de l’Appropriateness Evaluation Protocol (2ème partie). Revue d’Epidemiologie et de Sante Publique 2001;49:367–75. Panis L, Verheggen F, Pop P. To stay or not to stay. The assessment of appropriate hospital stay, a Dutch report. International Journal for Quality in Health Care 2001;13(4):55–67. Esveld S. The Appropriateness Evaluation Protocol. A study of the reliability and validity of the AEP. Maastricht University; 1995 [in Dutch]. Lang T, Liberati A, Tampieri A, Fellin G, Gonsalves Mda L, Lorenzo S, et al. A European version of the Appropriateness Evaluation Protocol. Goals and presentation. The BIOMED I group on appropriateness of hospital use. International Journal of Technology Assessment in Health Care 1999;15(1):185–97. Libarti A, Apalone G, Lang T, Lorenzo S. A European project assessing the appropriateness of hospital utilization: background, objectives and preliminary results. International Journal for Quality in Health Care 1995;7:187–200. Lorenzo S, Sunol R. An overview of Spanish studies on appropriateness of hospital use. International Journal for Quality in Health Care 1995;7:213–8. Lorenzo S, Lang T, Pastor R, Tampieri A, Santos-Eggimann B, Smith H, et al. Reliability study of the European appropriateness evaluation protocol. International Journal for Quality in Health Care 1999;11(5):419–24.