Achieving Minimum Caseload Requirements
an Analysis of Hospital Discharge Data From 2005–2011
Background: The German Federal Joint Committee (the highest decision-making body of physicians and health insurance funds in Germany) has established minimum caseload requirements with the goal of improving patient care. Such requirements have been in place for five types of surgical procedure since 2004 and were introduced for total knee endoprosthesis surgery in 2006 and for the care of low-birth-weight neonates (weighing less than 1250 g) in 2010.
Method: We analyzed data from German nationwide DRG statistics (DRG = diagnosis-related groups) for the years 2005–2011. The procedures that were performed were identified on the basis of their operation and procedure codes, and the low-birth-weight neonates on the basis of their birth weight and age. The treating facilities were distinguished from one another by their institutional identifying numbers, which were contained in the DRG database.
Results: In 2011, there were 172 838 hospitalizations to which minimum caseload requirements were applicable. 4.5% of these took place in institutions that did not meet the minimum requirement for the procedure in question. The percentage of institutions that did not meet the minimum caseload requirement for complex pancreatic surgery fell significantly from 64.6% in 2006 to 48.7% in 2011, and the percentage of pancreatic surgery cases treated in such institutions fell over the same period from 19.0% to 11.4%. A significant reduction in the number of institutions treating low-birth-weight neonates was already evident before minimum caseload requirements were introduced. For all other types of procedure subject to minimum caseload requirements, there has been no significant change either in the percentage of institutions meeting the requirements or in the percentage of cases treated in such institutions.
Conclusion: After taking account of the potential bias due to the identification of institutions by their institutional identifying numbers, we found no discernible effect of minimum caseload requirements on care structures over the seven-year period of observation, with the possible exception of a mild effect on pancreatic procedures.
Many observational studies have shown that the outcomes of some surgical procedures are correlated with caseload (1–10). For this reason, the German Federal Joint Committee (G-BA) has established minimum caseload requirements for certain types of elective procedure for which this has been demonstrated. Minimum caseload requirements for hepatic and renal transplantation, complex esophageal and pancreatic surgery, and stem cell transplantation went into effect in 2004. Total knee replacement surgery has been subject to minimum caseload requirements since 2006, and the care of neonates weighing less than 1250 g at birth (whether born prematurely or at term) since 2010 (11, 12). Thus, such requirements have been established for seven different types of procedure.
The minimum caseload requirements for total knee replacement surgery were suspended in 2011 after legal battles (13). The minimum caseload requirement for the care of low-birth-weight neonates was to be increased in 2011 from 14 to 30 cases per year, but this planned increase was also suspended, and the old requirement of 14 cases per year is still in effect (14). The minimum caseload requirements are formally binding on all hospitals that are permitted to charge the German statutory health-insurance carriers for their services. The relevant regulation states: “If an institution is not expected to meet the minimum caseload requirement for an elective procedure, then the procedure may not be performed in that institution, starting in the year that the requirement goes into effect” (§5 Mm-R). The regulation also provides for certain general exceptions to the rule. When a particular type of procedure is established in an institution and the staff performing the procedure in that institution changes, the new team is allowed up to two years to meet the caseload requirement; three years are allowed when the procedure is new to the institution. Moreover, comprehensive coverage of the population must be provided (15).
Research on this topic to date has centered on changes in the care situation from 2004 to 2006. Over this period, there was no major change in the number of hospitals providing care, and 1% to 31% of patients (depending on the procedure) were operated on in hospitals that did not meet the minimum caseload requirement for the procedure in question (16).
Further changes in the care situation from 2006 onward have not been assessed until now. We therefore studied developments with respect to the affected procedures over the years 2005–2011. This study is based on hospital billing data; the billing units were identified on the basis of the institutional identifying numbers that the hospitals used for billing.
Materials and methods
In this study, we performed a monitored remote data analysis on the complete federal DRG statistics (DRG = diagnosis-related groups) for individual cases that was obtained from the research data centers of the German federal and state statistical offices (17). The DRG statistics covered all hospitalizations billed according to the DRG system. Each of the ca. 17 million inpatient treatments carried out in Germany per year can be evaluated in this way. The main exceptions are psychiatric and psychosomatic cases, which are not covered by DRG statistics but are, in any case, irrelevant to the issue of minimum caseload requirements (for more on this, see eMethods). The documentation of every case included a main diagnosis, additional diagnoses, the procedures performed (coded according to the Operations and Procedures Classification—Operationen- und Prozedurenschlüssel, OPS), the reason for discharge, and the anonymized institutional identifying number of the hospital (18–21) (further details in eMethods).
Cases of treatment that were subject to minimum caseload requirements were identified in these data on a case-by-case basis by means of the OPS codes for each type of procedure that were established by the G-BA for the procedure and year in question. For hepatic transplantation and pancreatic surgery, the G-BA definition also includes the OPS codes for post-mortem organ removal (5–503.0 hepatectomy, post mortem; 5–525.4 pancreatectomy, post mortem [for transplantation]). These codes were included in the analysis to determine whether the minimum caseload requirements, as set by the G-BA, were actually reached if the institutions also performed transplantations. Treating institutions that performed only post mortem hepatectomy and pancreatectomy were not considered in the analysis, however, because a reduction in the number of organ donations cannot be a goal of regulations that establish minimum caseload requirements. In the OPS, the category of liver transplantation contains both post-mortem and living-donor transplantation, and these were both tallied to determine the caseload, as is stipulated in the minimum caseload regulation. Institutions that performed only living-organ donations without any coded transplantations were counted as independent institutions.
Neonates weighing less than 1250 g at birth were identified in the database by their birth weight and an age of 28 days or less (22).
Treating (or, more precisely, billing) institutions were distinguished from one another by their institutional identifying numbers (Institutionskennzeichen, IK numbers). This was the basis for the tallying of case numbers for each institution. The German Federal Statistical Office considers a billing institution for the purpose of DRG statistics to be a “hospital providing inpatient services that are reimbursable in the DRG system (hospitals that are spread over multiple sites but have a single institutional identifying number are counted only once)” (23). The number of such billing institutions differs slightly from the number of hospitals contained in the register of basic data for German hospitals (Grunddaten der Krankenhäuser, ), which is usually cited as the number of hospitals in Germany.
In 2011, the register of basic data for German hospitals counted a total of 2045 hospitals, of which 1736 were general hospitals and the remaining 309 were either hospitals with exclusively psychiatric and psychotherapeutic (and also neurological) beds or else day clinics and night clinics, which are not subject to minimum caseload requirements and generally do not bill in the DRG system.
The DRG statistics for 2011 are derived from a total of 1601 billing institutions with different IK numbers. This number is a bit lower than the number of hospitals in the basic data register, although it should be remembered that not all general hospitals bill according to the DRG system. For reasons of data protection, we cannot determine what institutions account for the difference, as it is not permitted for individual hospitals to be identifiable.
Both collections of statistics revealed a drop in the number of hospitals over the course of the study. The basic hospital data showed that, in 2005 and 2011, there were 1846 and 1736 general hospitals, respectively; the DRG statistics for the same years contained 1,725 and 1,601 IK numbers, respectively). The current study is based on ‘billing institutions’ in the sense used in the DRG statistics. It should be borne in mind that one hospital can have multiple IK numbers, while, conversely, a single IK number can refer to treatment at multiple, geographically separated sites.
We used generalized linear regression models weighted by the annual total number of billing institutions and cases to search for linear trends in the number of treating institutions in each year, the percentage of such institutions that did not meet the minimum caseload requirements, and the number of cases treated in them. For the determination of trends, only the statistics from 2006 onward were considered, because different caseload requirements and definitions were in effect in 2005.
In eMethods, we describe additional subgroups of the types of procedures that were subject to minimum caseload requirements (eTable).
In 2011, 172 838 cases, or 1.0% of all inpatient cases in the DRG statistics, were subject to minimum caseload requirements. Depending on the type of procedure group, 16.1% to 68.4% of the performing institutions did not meet the minimum caseload requirements for the procedure in question; 4.5% of the 172 838 cases were performed in such institutions.
The number of cases per year rose for all types of procedure studied except the care of low-birth-weight neonates (Table 1). There was a continuous, significant decline in the number of institutions treating low-birth-weight neonates (from 465 to 347); this was the only such decline noted for any of the types of procedure subject to minimum caseload requirements. Accordingly, the percentage of institutions that met the minimum requirement of 14 low-birth-weight infants treated per year rose significantly from 31.6% in 2005 to 43.8% in 2011 (p<0.0001).
For all other types of procedure, the number of treating institutions varied only slightly over time, without significant trends (Table 1). No concentration of care in a smaller number of institutions was seen. The observed increases in the mean case numbers per institution and per year were accounted for in large part by increases in the total number of cases treated, with a relatively stable number of treating institutions. Only in the case of pancreatic surgery the increase in total delivered care resulted in a significant decrease in the percentage of institutions that failed to meet the minimum caseload requirement (p = 0.012) (Figure 1). The number of pancreatic operations performed in such institutions decreased significantly, as did the number of low-birth-weight treatments in institutions that did not meet the minimum caseload requirement for such treatment (Figure 2).
The percentage of institutions that met the minimum caseload requirements rose significantly from 2006 to 2011 both for pancreatic surgery (35.4% to 51.3%) and for the treatment of low-birth-weight neonates (33% to 44%). There was no significant rise for any of the other five types of procedure. The figures for 2006 and 2011, respectively, were: for hepatic transplantation, 45.7% and 54.8%; for renal transplantation, 74.5% and 73.5%; for esophageal surgery, 29.6% and 31.6%; for stem cell transplantation, 64.6% and 64.8%; for knee replacement surgery, 79.5% and 83.9% (Table 1).
There were no systematic differences in patient sex ratios between institutions that did and did not meet the minimum caseload requirements. There were, however, systematic differences with respect to age: in institutions that did not meet the minimum caseload requirements for hepatic transplantation, esophageal surgery, and stem cell transplantation, the average age was slightly higher (Table 2).
The major strength of this study is the completeness of the database. The analyses presented here are based on complete billing data for each of the procedures in question, and, therefore, the parent population of all relevant inpatient cases was evaluated, rather than only a sample. The evaluation was thus hospital-independent and unaffected by the hospitals’ own determinations or self reporting as to whether they met the minimum caseload requirements (as found, for example, in the hospitals’ legally mandated quality-assurance reports).
On the other hand, the use of this extensive set of data also results in a major limitation of the study. In the DRG statistics of the German Federal Statistical Office, the distinction of one institution from another is made on the basis of IK numbers, i.e., the institutions considered in this study are single units with respect to billing. This definition does not correspond exactly to that of the approved German hospitals that can be found in other statistical databases.
A billing institution, as indicated in the DRG statistics, may correspond to only a part of a hospital. This can lead to underestimation of the number of institutions that met the minimum caseload requirements. The opposite, however, may also be the case: multiple hospital sites can share a billing (IK) number, and this can lead to overestimation of the number of institutions that met the minimum caseload requirements.
The overall number of billing institutions in the DRG statistics turns out to be slightly lower than the number of hospitals in the basic hospital data. Moreover, the number of institutions has declined over time in both statistics, perhaps because of mergers or a preferential loss of smaller institutions. Either mechanism would tend to result in an overestimation of the number of institutions that met the minimum caseload requirements.
As “hospitals” are defined differently in each database, studies based on different databases can yield divergent results not only regarding the overall number of hospitals, but also regarding the percentage of hospitals that met minimum caseload requirements for any particular type of procedure (see, for example, the article by von de Cruppé et al. in this issue ).
When interpreting the findings of this study and similar studies, one must bear in mind that the definitions of individual procedures covered by minimum caseload requirements may have changed over the period of observation in such a way as to affect the number of procedures subject to the requirement (e.g., hepatic transplantation, from 2005 to 2006); that some of the minimum caseload thresholds were increased by the G-BA between 2005 and 2006; and that hospitals performing pancreatic and esophageal surgery exclusively on children are exempt from such requirements. All of these things were considered in the present study.
With the use of these methods, this study of the seven types of procedure that were subject to minimum caseload requirements from 2005 to 2011 revealed that the percentage of patients treated in institutions that did not meet such requirements dropped significantly in response to the requirements for only one of the seven types of procedure, namely, complex pancreatic surgery. A comparable, steady drop over the years 2005 to 2011 was observed for the care of low-birth-weight neonates but cannot be considered an effect of the minimum caseload requirement, which, in this case, was not introduced until 2010.
As for the number of institutions performing each of the procedures in question, slightly more institutions performed total knee replacement surgery in 2011 than in 2006 (a rise from 999 to 1024 institutions), but, in 2011, 16.1% of them failed to meet the minimum caseload requirement. On the other hand, the number of institutions caring for low-birth-weight neonates fell over the same period from 433 to 347, even though a minimum caseload requirement was only introduced in 2010. No statistically significant increases or decreases in the number of performing institutions were seen for any of the other five types of procedure.
Previously published findings on the implementation of minimum caseload requirements in the years 2004–2006 are consistent with those reported here. Geraedts et al. concluded at that time that their observation period had been too short to permit any robust conclusion about the efficacy of minimum caseload requirements (16). The present findings clearly show that minimum caseload requirements have had no discernible effect on care structures over a seven-year period, except for a possible, mild effect on pancreatic surgery, as revealed by an analysis of the treating and billing institutions. The observed changes in preterm neonatal care cannot be due to the minimum caseload requirements, as they mostly took place before the requirements went into effect in 2010. For none of the other types of procedure was there any decline in the number of institutions performing it, although this was intended when the minimum caseload requirements were introduced and might have been expected to occur. One can only speculate about why the minimum caseload requirements appear to have had so little effect.
A possible weakness of this study is that the evaluation was based on billing institutions (IK numbers). As recent years have seen more hospital mergers than hospital split-ups, one might have expected to see an artefactual decrease in the number of institutions that failed to meet the minimum caseload requirements. It is also conceivable that the institutions that failed to meet the minimum caseload requirements were actually exempt from them under the stated exceptions to the minimum caseloads regulation. This would then have to have been true for the entire period of observation (because of the lack of observed change after the minimum caseload requirements were introduced). This, in turn, would indirectly imply that there was no need for the minimum caseload requirements in this case.
It is, however, also conceivable that the monitoring and implementation mechanisms for minimum caseload requirements are currently inadequate because too little attention is paid to this matter in the annual budget negotiations of hospitals and health insurance carriers, and on the state governmental level.
If minimum caseload requirements are wanted, they should also be implemented. The current state of regulation seems not to have achieved this aim. The causes of failure to meet the minimum caseload requirements should be examined in each of the affected institutions, or in a sampling of them, so that appropriate conclusions can be drawn about monitoring mechanisms, the legal formulation of minimum caseload requirements, and, possibly, a more accurate definition of the procedures that are to be covered by them.
Conflict of interest statement
The academic department of structural advancement and quality management in health care in which the authors work is funded by Helios Kliniken GmbH.
Manuscript submitted on 12 December 2013, revised version accepted
on 2 July 2014.
Translated from the original German by Ethan Taub, M.D.
Dr. rer. medic. Dirk Peschke
Technische Universität Berlin
Fachgebiet Strukturentwicklung und
Qualitätsmanagement im Gesundheitswesen
Steinplatz 2, 10623 Berlin, Germany
lieferung (last accessed on 12 April 2014).
infant_newborn (last accessed on 12 April 2014).
|1.||van Heek NT, Kuhlmann KF, Scholten RJ, et al.: Hospital volume and mortality after pancreatic resection: a systematic review and an evaluation of intervention in the Netherlands. Ann Surg 2005; 242: 781–8, discussion 8–90. CrossRef MEDLINE PubMed Central|
|2.||Gooiker GA, van Gijn W, Wouters MW, et al.: Systematic review and meta-analysis of the volume-outcome relationship in pancreatic surgery. Br J Surg 2011; 98: 485–94. CrossRef MEDLINE|
|3.||Wouters MW, Gooiker GA, van Sandick JW, Tollenaar RA: The volume-outcome relation in the surgical treatment of esophageal cancer: a systematic review and meta-analysis. Cancer 2012; 118: 1754–63. CrossRef MEDLINE|
|4.||Markar SR, Karthikesalingam A, Thrumurthy S, Low DE: Volume-outcome relationship in surgery for esophageal malignancy: systematic review and meta-analysis 2000–2011. J Gastrointest Surg 2012; 16: 1055–63. CrossRef MEDLINE|
|5.||Yasunaga H, Horiguchi H, Matsuda S, et al.: Relationship between hospital volume and operative mortality for liver resection: Data from the Japanese Diagnosis Procedure Combination database. Hepatol Res 2012; 42: 1073–80. CrossRef MEDLINE|
|6.||Birkmeyer JD, Siewers AE, Finlayson EV, et al.: Hospital volume and surgical mortality in the United States. N Engl J Med 2002; 346: 1128–37. CrossRef MEDLINE|
|7.||Soohoo NF, Zingmond DS, Lieberman JR, Ko CY: Primary total knee arthroplasty in California 1991 to 2001: does hospital volume affect outcomes? J Arthroplasty 2006; 21: 199–205. CrossRef MEDLINE|
|8.||Katz JN, Barrett J, Mahomed NN, Baron JA, Wright RJ, Losina E: Association between hospital and surgeon procedure volume and the outcomes of total knee replacement. J Bone Joint Surg Am 2004; 86-A: 1909–16. MEDLINE|
|9.||Wehby GL, Ullrich F, Xie Y: Very low birth weight hospital volume and mortality: an instrumental variables approach. Med Care 2012; 50: 714–21. CrossRef MEDLINE PubMed Central|
|10.||Heller G, Gunster C, Misselwitz B, Feller A, Schmidt S: Jährliche Fallzahl pro Klinik und Überlebensrate sehr untergewichtiger Frühgeborener (VLBW) in Deutschland – Eine bundesweite Analyse mit Routinedaten [Annual patient volume and survival of very low birth weight infants (VLBWs) in Germany-a nationwide analysis based on administrative data]. Z Geburtshilfe Neonatol 2007; 211: 123–31. CrossRef MEDLINE|
|11.||Gemeinsamer Budesausschuss: Anpassung der Anlage 1 der Mindestmengenvereinbarung nach § 137 Abs. 1 Satz 3 Nr. 3 SGB V. Köln: Bundesanzeiger Verlag 2006.|
|12.||Gemeinsamer Bundesausschuss: Bekanntmachung eines Beschlusses des Gemeinsamen Bundesausschusses zur Versorgung von Früh- und Neugeborenen. Köln: Bundesanzeiger Verlag 2009.|
|13.||Gemeinsamer Bundesausschuss: Bekanntmachung eines Beschlusses des Gemeinsamen Bundesausschusses über eine befristete Außervollzugsetzung einer Regelung der Mindestmengenvereinbarung: Mindestmenge für Kniegelenk-Totalendoprothesen. Köln: Bundesanzeiger Verlag 2011.|
|14.||Gemeinsamer Bundesausschuss: Bekanntmachung eines Beschlusses des Gemeinsamen Bundesausschusses über eine befristete Außervollzugsetzung einer Änderung der Mindestmengenregelungen: Mindestmenge für Früh- und Neugeborene Perinatalzentren Level 1. Köln: Bundesanzeiger Verlag 2012.|
|15.||Gemeinsamer Bundesausschuss: Mindestmengenregelungen. Köln: Bundesanzeiger Verlag 2013.|
|16.||Geraedts M, de Cruppe W, Blum K, Ohmann C: Implementation and effects of Germany’s minimum volume regulations: results of the accompanying research. Dtsch Arztebl Int 2008; 105: 890–6.|
|17.||Forschungsdatenzentren der Statistischen Ämter des Bundes und der Länder: Datenangebot | Fallpauschalenbezogene Krankenhausstatistik (DRG-Statistik). www.forschungsdatenzentrum.de/bestand/drg/index.asp (last accessd 19. March.2014).|
|18.||Statistisches Bundesamt: Fallpauschalenbezogene Krankenhausstatistik (DRG-Statistik). Qualitätsbereicht. Wiesbaden: Statistisches Bundesamt 2011.|
|19.||Mansky T, Nimptsch U: Notwendigkeit eines ungehinderten Zugangs zu sozial- und krankheitsbezogenen Versichertendaten für die Bundesärztekammer und andere ärztliche Körperschaften sowie wissenschaftliche Fachgesellschaften zur Optimierung der ärztlichen Versorgung. Expertise im Rahmen der Förderinitiative zur Versorgungsforschung der Bundesärztekammer. Technische Universität Berlin; 2010. www.bundesaerztekammer.de/downloads/Datenzugang-2.pdf (last accessed on 12 April 2014).|
|20.|| Institut für das Entgeltsystem im Krankenhaus: Datenlieferung gemäß § 21 KHEntgG. Dokumente zur Datenlieferung. www.g-drg.de/cms/Datenlieferung_gem._21_KHEntgG/Dokumente_zur_Daten|
lieferung (last accessed on 12 April 2014).
|21.||Institut für das Entgeltsystem im Krankenhaus: Vereinbarung über die Übermittlung von DRG-Daten nach § 21 Abs. 4 und Abs. 5 KHEntgG. www.g-drg.de/cms/Datenlieferung_gem._21_KHEntgG/Dokumente_zur_Datenlieferung/21-Vereinbarung (last accessed 12. April 2014).|
|22.||World Health Organization: Infant, Newborn. www.who.int/topics/|
infant_newborn (last accessed on 12 April 2014).
|23.||Statistisches Bundesamt: Fallpauschalenbezogene Krankenhausstatistik (DRG-Statistik) 2011. Fachserie 12, Reihe 6.4. Wiesbaden: Statistisches Bundesamt 2012.|
|24.||Statistisches Bundesamt: Grunddaten der Krankenhäuser 2011. Fachserie 12, Reihe 6.1.1. Statistisches Bundesamt: Wiesbaden 2012.|
|25.||de Cruppé W, Malik M, Geraedts M: Achieving minimum caseload requirements: an analysis of hospital quality control reports from 2004–2010. Dtsch Arztebl Int 2014; 111: 549–55.|
|26.||Nimptsch U: Leistungserbringerbezogene Merkmale. In: Swart E, Ihle P, Gothe H, Matusiewicz D: Routinedaten im Gesundheitswesen. Handbuch Sekundärdatenanalyse: Grundlagen, Methoden und Perspektiven. Bern: Verlag Hans Huber 2014; in press. PubMed Central|
|27.||Deutsche Stiftung Organtransplantation (DSO): Organ Donation and Transplantation in Germany. Annual Report 2011. Frankfurt am Main: Deutsche Stiftung Organtransplantation 2012.|
|28.||AQUA – Institut für angewandte Qualitätsförderung und Forschung im Gesundheitswesen GmbH: Qualitätsreport 2011. Göttingen: AQUA-Institut GmbH 2012.|