LNSLNS

The aim of any quality-based coordination of care has to be high-quality, supraregionally homogeneous healthcare. If no quality data are available then agreeing a minimum caseload is at the very least a protective measure against the risks entailed in ad-hoc provision of care, which in critical situations provides neither the necessary experience and routines nor the necessary structures and processes. This is confirmed by the significantly better quality of care offered by larger departments. However, wherever quality data are available thanks to mandatory documentation by neonatal units, it is not clear why one should rely on caseload alone. The results show how differentiated the outcome quality above a minimum caseload—using the example of mortality—has to be assessed. Concentration alone does not seem a convincing concept. Including three years of results permits a classification of the quality of care delivered by individual centers that is robust enough to accommodate accidental fluctuations.

We agree with Pfeifer as well as Vogtmann et al. (1) that the necessary risk adjustment should not be limited to the parameters gestational age and severe malformations. Our analyses showed nine risk factors that are relevant for the adjustment model. These include, after thorough consideration, administration of FiO2 and base excess at admission—parameters that at this early stage reflect the condition of the preterm infants rather than the care given by the center and can therefore be counted as risk factors.

Using a logistic regression model in the non-hierarchical form aimed to adjust for all factors that reflect exclusively the risk profile of the preterm infants and are not affected by the care given in the centers. By including the centers in a multilevel model we would have corrected for the non-specific but present effect of hospital care. Such an approach would have run counter to our study’s objective—namely, to conduct a risk-adjusted comparison between hospitals. Undoubtedly it would be interesting to investigate the way in which factors such as staffing structures, equipment, the organization, process and internal quality management, and the quality culture affect mortality in preterm infants and how these correlate to the caseload. But this information was not available to us. However, one objective of our study was to evaluate, for the given data, the potential use of neonatal data. Looking for a threshold value would make sense only if a minimum quality standard was defined and if, in the context of theoretical modeling to date, a monotonically decreasing or increasing effect on the result could be assumed. The direct evaluation of individual centers according to the outcome quality that was actually achieved on the basis of the available neonatal data seemed the more pragmatic approach. The extent to which the data for 2010—when the new survey of neonatal data started—are sufficiently complete is not known to us as we do not have access to these data. This will need to be clarified in the coming years by taking birth records into account.

The exclusion of preterm infants admitted to neonatal units more than 24 hours after birth and of infants referred to other units was unavoidable. Children admitted 24 hours after birth were likely to be pretreated. Therefore, a clear assignment of treatment quality to a single neonatal unit was not possible. Moreover, including children admitted more than 24 hours after birth would have been likely to give rise to the accusation of artificially increasing mortality rates in large centers with these high-risk infants. However, mortality cannot be calculated for infants referred to other units. Therefore, the risk-adjusted comparison between preterm infants exclusively treated in one center seemed to us the most clearly defined and therefore fairest approach.

DOI: 10.3238/arztebl.2013.0118

Dr. rer. medic. Marcus Kutschmann, Dipl.-Stat., Dr. med. Christof Veit
BQS Institut für Qualität und Patientensicherheit,
m.kutschmann@bqs-institut.de

Conflict of interest statement

The BQS Institute for Quality and Patient Safety (BQS Institut für Qualität & Patientensicherheit) received financial support from the German Hospital Association (DKG, Deutsche Krankenhausgesellschaft e. V.) for the analysis on which the authors’ reply is based.

Dr Kutschmann and Dr Veit are employees of the BQS Institute for Quality and Patient Safety.

1.
Vogtmann C, Koch R, Gmyrek D, Kaiser A, Friedrich A: Risk-adjusted intraventricular hemorrhage rates in very premature infants – towards quality assurance between neonatal units. Dtsch Arztebl Int 2012; 109(31–32): 527–33 VOLLTEXT
2.
Kutschmann M, Bungard S, Kötting J, Trümner A, Fusch C, Veit C:
The care of preterm infants with birth weight below 1 250 g: risk-adjusted quality benchmarking as part of validating a caseload-based management system. Dtsch Arztebl Int 2012; 109(31–32): 519–26 VOLLTEXT
1.Vogtmann C, Koch R, Gmyrek D, Kaiser A, Friedrich A: Risk-adjusted intraventricular hemorrhage rates in very premature infants – towards quality assurance between neonatal units. Dtsch Arztebl Int 2012; 109(31–32): 527–33 VOLLTEXT
2.Kutschmann M, Bungard S, Kötting J, Trümner A, Fusch C, Veit C:
The care of preterm infants with birth weight below 1 250 g: risk-adjusted quality benchmarking as part of validating a caseload-based management system. Dtsch Arztebl Int 2012; 109(31–32): 519–26 VOLLTEXT

Info

Specialities