Document [original]

ZuberA, etal. BMJ Open 2022;12:e045327. doi:10.1136/bmjopen-2020-045327

Open access

Does adherence to a quality indicator

regarding early weaning from invasive

ventilation improve economic outcome?

A single- centre retrospective study

Alexander Zuber,1,2 Oliver Kumpf,2 Claudia Spies ,2 Moritz Höft,2 Marc Deffland,2

Robert Ahlborn,3 Jochen Kruppa,1 Roland Jochem,4 Felix Balzer 1,2

To cite: ZuberA, KumpfO,

SpiesC, etal. Does adherence

to a quality indicator regarding

early weaning from invasive

ventilation improve economic

outcome? A single- centre

retrospective study. BMJ Open

2022;12:e045327. doi:10.1136/

bmjopen-2020-045327

►Prepublication history and

additional supplemental material

for this paper are available

online. To view these files,

please visit the journal online

(http://dx.doi.org/10.1136/

bmjopen-2020-045327).

Received 30 September 2020

Accepted 26 August 2021

1Institute of Medical Informatics,

Charité Universitätsmedizin

Berlin, Berlin, Germany

2Department of Anesthesiology

and Operative Intensive

Care Medicine, Charité

Universitätsmedizin Berlin,

Berlin, Germany

3IT Department, Charité

Universitätsmedizin Berlin,

Berlin, Germany

4Departments of Machine Tools

and Factory Management, TU

Berlin, Berlin, Germany

Correspondence to

Professor Felix Balzer;

felix. balzer@ charite. de

Original research

employer(s)) 2022. Re- use

permitted under CC BY- NC. No

commercial re- use. See rights

and permissions. Published by

BMJ.

ABSTRACT

Objectives To measure and assess the economic impact

of adherence to a single quality indicator (QI) regarding

weaning from invasive ventilation.

Design Retrospective observational single- centre study,

based on electronic medical and administrative records.

Setting Intensive care unit (ICU) of a German university

hospital, reference centre for acute respiratory distress

syndrome.

Participants Records of 3063 consecutive mechanically

ventilated patients admitted to the ICU between 2012 and

2017 were extracted, of whom 583 were eligible adults

for further analysis. Patients’ weaning protocols were

evaluated for daily adherence to quality standards until ICU

discharge. Patients with<65% compliance were assigned

to the low adherence group (LAG), patients with≥65% to

the high adherence group (HAG).

Primary and secondary outcome measures Economic

healthcare costs, clinical outcomes and patients’

characteristics.

Results The LAG consisted of 378 patients with a median

negative economic results of −€3969, HAG of 205

(−€1030), respectively (p<0.001). Median duration of

ventilation was 476 (248; 769) hours in the LAG and 389

(247; 608) hours in the HAG (p<0.001). Length of stay

(LOS) in the LAG on ICU was 21 (12; 35) days and 16 (11;

25) days in the HAG (p<0.001). LOS in the hospital was 36

(22; 61) days in the LAG, and within the HAG, respectively,

26 (18; 48) days (p=0.001).

Conclusions High adherence to this single QI is

associated with better clinical outcome and improved

economic returns. Therefore, the results support the

adherence to QI. However, the examined QI does not

influence economic outcome as the decisive factor.

INTRODUCTION

In the last decades, the need for quality

management (QM) in the hospital has been

growing. On one hand costs have been rising

and on the other patients, health insurance

and public pressure urge hospitals to improve

outcome and services by cutting or tying

reimbursement to valid quality indicators

(QIs).1 This is why in the medium and long

run quality- oriented reimbursement (pay

for quality) might change the hospital land-

scape.2 Economics of health have been estab-

lished widely in order to curb costs for the

national healthcare system. Many countries

introduced diagnosis- related groups (DRGs)

in order to pay on averaged costs and on a

generalised financial reimbursement per case

(fixed prices). Reimbursement for inpatients

is linked to DRG accounting and updated

annually based on reported data from

hospitals. The fee- for- service system induces

hospitals to improve internal processes as

reimbursement is predefined and to work

goal- oriented towards therapeutic aims.3

In modern medicine, a considerable part of

hospital costs arises from intensive care. The

cost structure of a tertiary German hospital

shows that ca. 20% of costs are generated

in intensive care units (ICUs).4 Especially,

mechanical ventilation is the main cost driver

in ICUs.5 Approximately 6% of the patients

in intensive care are affected by prolonged

Strengths and limitations of this study

►This is the first study evaluating whether a quality

indicator on weaning has effects on the economic

outcome parameters on a per case basis.

►Results of the cost unit accounting practice is well

established and is thus representative for a detailed

examination of unit costs.

►The test and validation sample was taken from a

reference centre specialised on acute respirato-

ry distress syndrome in adult patients with severe

medical conditions.

►Control for interactions with other quality indicators

is necessary as the examined quality indicator is po-

tentially connected with other ones.

►The study results are based on German reimburse-

ment system and might be typical for a tertiary

university hospital rather than German hospitals in

general

2ZuberA, etal. BMJ Open 2022;12:e045327. doi:10.1136/bmjopen-2020-045327

Open access

mechanical ventilation and weaning from mechanical

ventilation represents an essential element in the treat-

ment of critically ill patients as it can take up to 50% of the

ventilation time.6 As a consequence, up to 37% of all ICU

resources are allocated to these patients.7 This means that

weaning patients from mechanical ventilation is not only

essential for clinical outcomes like duration of ventilation

or length of stay (LOS),8 9 but also a critical step from an

economic perspective as costs can be reduced. Therefore,

this process is a critical phase in intensive care. However,

the ideal weaning process is still subject to debate.10

About 40% of patients receiving mechanical ventila-

tion will experience a complicated weaning process.11

Patients categorised in prolonged weaning, failing at least

three spontaneous breathing trials (SBTs) or receiving

more than 7 days of weaning after the first SBTs, have an

increased risk in developing hospital mortality, mainly

through ventilator- associated pneumonia (VAP),6 but also

through postintensive care syndrome (PICS) or chronic

critical illness (CCI).12 Due to demographic changes and

technological advances in intensive care, the number

of older patients with complex diseases or comorbidi-

ties needing ventilation is increasing.13 14 This generates

growing costs, as the cohort of patients requiring respira-

tory support accounts for a disproportionate percentage

of the resources available in intensive care.15

With the purpose of managing quality throughout the

difficult framework conditions of hospital care, a proactive

and structured QM is essential.16 In general, QM focuses

on securing and improving clinical services economi-

cally, performed by physicians or nurses according to the

patient’s needs.17 In Germany, in the context of European

and national QM initiatives, consensus- based standardised

QIs were developed for intensive care medicine since

2010—third version in 2017—by the German interdisci-

plinary society for intensive and emergency care (DIVI)

in order to simplify the measurement of relevant quality

data, to record timely and to allow transparent compari-

sons of patient data. The according quantification of QM

helps measuring effectiveness and efficiency of ICUs.18 19

QIs enable a descriptive picture of the actual condition

and are an indispensable instrument for comparisons

between different states of quality.18 Potentially, widely

accepted QIs can progress hospital economics and

support the reduction of the national budget for health-

care, even though a recent study has shown that cost–

quality relationships are difficult to generate.20

QIs empower advances in intensive care medicine to be

measured and evaluated on a regular basis.19 QIs can be

defined as representative figures for quality of structure,

processes or outcome within the medical care process.

Thus, indicators are useful for measuring improvement

in the context of QM and should be developed in line

with evidence- based literature.21 Ideally, measures for QIs

can be extracted from routine patient data to avoid excess

documentation work. Therefore, patient data manage-

ment systems (PDMS) are pivotal for measuring complex

quality figures.18 The economic aspects for the whole

hospital of the introduction of QIs are not well investi-

gated. However, the is evidence that the application of

QIs is a value- creating instrument.12

The objective of this study was, to determine the

economic impact of adherence to a single QI evalu-

ating the weaning process from invasive ventilation. We

analysed this by comparing economic results per case

and clinical outcome parameters like LOS between two

groups of either high- quality or low- quality adherence.

Additionally, we sought to determine factors that would

influence a potential interaction between economic and

outcome parameters.

METHODS

This original research is in accordance with the Consoli-

dated Health Economic Evaluation Research Standards.

Patient and public involvement

Patients and the public were not directly involved in this

observational study.

Study centre

We conducted this single study- centre in a university

hospital (Charité-Universitätsmedizin Berlin). This obser-

vational analysis was performed at a 14- bed ICU (refer-

ence centre), specialised in treatment of acute respiratory

distress syndrome in adult patients. All patients at our

ICU were treated according to guidelines and internal

standard operating procedures for clinical practice.22

Study design

This was a retrospective descriptive study, using data from

multiple electronic databases used in routine patient

care and for routine administrative purposes. All patients

admitted to and discharged from the ICU between

1 January 2012 and 31 December 2017 who received

invasive ventilation during their stay were eligible to be

included in this study. Furthermore, duration of ventila-

tion <95 hours, receiving no invasive ventilation, terminal

status, incomplete patient record or missing readiness

to be weaned were defined as exclusion criteria (see

figure 1).

Confidentiality was guaranteed, no interventions were

performed and only clinical routine data were collected.

Data were retrieved from a PDMS called (Computer

Organised Patient Report Assistant; COPRA System,

Berlin, Germany). Data are recorded both automatically

by patient monitors and manually by caregivers. The ICU

staff validates all information manually. However, the

design of the PDMS prevents manual alterations to the

data, for example adding missing values after discharge

from the ICU. PDMS data are also transferred to the clin-

ical information and accounting system (SAP, Walldorf,

Germany). Based on this administrative system, cost unit

accounting is performed annually. In addition to basic

demographic data, we assessed clinical and administra-

tive parameters of in- patient cases (eg, LOS). Data were

ZuberA, etal. BMJ Open 2022;12:e045327. doi:10.1136/bmjopen-2020-045327

Open access

retrieved using a structured query. No patient identifiers

were extracted in order to secure anonymity of patients’

data. Data related to diagnoses were not retrieved from

the administrative systems.

PDMS data of patients included in the study were trans-

ferred to the study database, where we also collected the

administrative and cost accounting data, respectively. We

contrasted patient, intensive care and economic param-

eters of the two adherence groups (see table 1). Then,

we calculated the profits per case by subtracting costs

of reimbursement per case. In order to generate an

economic outcome per case for the dependent variable

in multivariate linear regression. Besides administrative

data, we used different scores for assessing the QI for eligi-

bility. Selection criteria were: (1) no additional workload

required for documentation, (2) the availability within

the PDMS system, (3) standardised values for all patients

and the existence of standard operating procedures for

each indicator and (4) the relevance of the indicator for

clinical decision making.19 23

Procedures

In this study, we used present key performance indica-

tors (KPIs) in order to examine the adherence to the

QI ‘Early Weaning from Invasive Ventilation’ until ICU

discharge.8 A small set of evidence- based KPIs was estab-

lished in 2009, providing indicators that were already

available within the PDMS. The KPIs in intensive care

medicine proved helpful for practical use and compliance

with standard operating procedures. A description of the

KPI is provided in online supplemental material. Within

the weaning therapy, fast visual feedback for ‘readiness

to wean’ and ‘weaning protocol compliance’ were imple-

mented. If both KPIs were positive, the according result of

the SBT was recorded.23 Once the patient was assessed to

be ready to wean since the primary disease showed clinical

improvement, the standard weaning protocol activities

were conducted on a daily basis according to standard

operating procedures. Congruent with clinical guidelines

in place, weaning protocols were adapted to evaluate the

progress of respirator therapy.22 The subsequent result

was recorded in the weaning protocol. For each patient,

we monitored the daily weaning protocol compliance

between readiness to wean and ICU- discharge in order

to evaluate the percentage of adherence. Within the

weaning process, the SBT represents the major diagnostic

test to evaluate if the patient can be extubated success-

fully.10 The SBT is successful if the patient succeeded

the trial and does not have to be re- intubated within 48

hours.24 This process is directly linked to a specific QI for

weaning derived from the DIVI- QI.19 A definition of the

indicator is presented in online supplemental material.

Outcome parameters

In this study, we investigated for economic results, clin-

ical outcome parameters per case and the respective

adherence to quality. Economic results were defined as

the profit or loss per case, by subtracting all assigned

costs from the reimbursement on a case level. Clinical

outcomes as a representative for clinical effectiveness

were measured in order to set economic outcomes in

relation to the purpose of medicine. Adherence to quality

was calculated on a per case level in order to categorise

the patients into groups.

We used the adherence level of the examined QI in

order to create two quality groups. We calculated the final

quality level by averaging the daily indicator results for the

duration with equal weights per day. In order to set the

optimal cut- off point for dichotomously distinguishing

between high- adherence and low- adherence of weaning

quality, we combined recommendations from literature

with our institutional standards. A cut- off value of 70%

deemed as a suitable fulfillment- threshold for QIs.25

However, due to partially high workload under certain

circumstances in intensive care, we decided to lower the

cut- off for 5% tolerance in order to account for missing

values in documentation. Therefore, we inserted a cut- off

for weaning protocol compliance at 65% adherence. The

low adherence group (LAG) was defined as adherence to

QI of less than 65%. The high adherence group (HAG)

was defined as adherence to QI of equal or more than

65%. Once this threshold was reached, the QI was charac-

terised as high adherence.

Statistical analyses

Descriptive analyses and statistical testing were

performed using SPSS, V.14.0 (SPSS) for Windows.

Results are expressed as median (IQR) or frequency

(%). We controlled data for risk and severity by exclusion

as patients and therapies in intensive care are hetero-

genic, as studies have shown.18 Differences between the

adherence groups in terms of outcome parameters were

tested using the univariate unpaired t- test and χ2 statistics

for independent variables as appropriate with a p<0.05

regarded as significant.

Figure 1 Patient inclusion and exclusion criteria. Flow chart

of the process used in this study for patient record inclusion.

Numbers listed are number of patients in each group. ICUs,

intensive care units; KPIs, key performance indicators.

4ZuberA, etal. BMJ Open 2022;12:e045327. doi:10.1136/bmjopen-2020-045327

Open access

In order to investigate the influencing factors in more

detail, parameters that were found to be statistically

significant on univariate analysis or out of discussion

among the experts underwent stepwise multivariate anal-

yses. We used multiple linear regression analyses to model

the relationship between the independent variables and

the outcome of profitability. Regression coefficients (95%

CI) and the corresponding p values were calculated for

each factor. Testing the dataset for outliers was performed

using the cook distance test, based on the model. The

test did not indicate the need to dismiss cases from the

sample. Due to an exploratory character of the research,

no adjustments for multiple testing were made.

RESULTS

All patients with complete electronic patient records

(n=3063 patients) were screened for eligibility. After

selection regarding inclusion and exclusion criteria, 583

patients were included in the final analysis (figure 1). Of

these patients, 378 showed low- adherence if the indicator

was below 65% and 205 showed high adherence. The

median age of admitted patients was 57 (40; 70) years;

64.7% of patients were male. There were significantly

(p=0.038) more male patients within the HAG (70.2%)

than in the LAG (61.6%). As reflected by a median Acute

Physiology and Chronic Health Evaluation II (APACHE

II) admission score of 21 (14; 27), a Simplified Acute

Physiology Score II (SAPS II) admission score of 47 (34;

61) and a Sequential Organ Failure Assessment (SOFA)

admission score of 9 (7; 12), the study population was

characterised by severe medical conditions. Patient

Table 1 Patient demographics and outcome parameters in comparison between QI adherence groups

All patients

LAG

QI <65%

HAG

QI ≥65%

P valuen=583 n=378 n=205

Demographics

Age (years) 57 (40; 70) 57 (40; 70) 55 (42; 69) 0.770

Gender (male) 377 (64.7%) 233 (61.6%) 144 (70.2%) 0.038

ICU score on admission

APACHE II 21 (14;27) 21 (15;27) 21 (14;27) 0.986

SAPS II 47 (34;61) 47 (35;60) 47 (34;62) 0.860

SOFA 9 (7;12) 9 (7;12) 9 (7;13) 0.526

Average SOFA 8.2 (6.6;10.3) 8 (6.5;10.1) 8.4 (6.8;10.7) 0.140

Type of admission to Study- ICU 0.651

Medical 290 (49.7%) 190 (50.3%) 100 (48.8%)

Emergency surgery 232 (39.8%) 146 (38.6%) 86 (41.9%)

Elective surgery 61 (10.5%) 42 (11.1%) 19 (9.3%)

Outcome parameter

Duration of Ventilation Study- ICU (hour) 431 (250;709) 476 (248;769) 389 (247;608) <0.001

Total duration of ventilation hospital

(hour)

578 (338;924) 597 (310;992) 535 (361;821) 0.017

No spontaneous breathing trials 1 (0;2) 1 (0;2) 1 (0;2) 0.456

No reintubation 0 (0;1) 0 (0;1) 0 (0;1) 0.531

Type of discharge of study- ICU <0.001

ICU 161 (27.6%) 100 (26.5%) 61 (29.8%)

Intermediate/ward 260 (44.6%) 172 (45.5%) 88 (42.9%)

Rehabilitation 110 (18.9%) 56 (14.8%) 54 (26.3%)

ICU mortality 52 (8.9%) 50 (13.2%) 2 (1.0%)

LOS Study- ICU (days) 19 (11;32) 21 12;35) 16 (11;25) <0.001

LOS hospital (days) 33 (20;54) 36 (22;61) 26 (18;48) 0.001

Profit (€)−2999 (−15 946; 7730) −3696 (−21 170; 6828) −1030 (−11 134; 9449) <0.001

Discrete variables are presented as a total number of encounters and were analysed with χ2 test for non- parametric samples.

APACHE II, Acute Physiology and Chronic Health Evaluation II; HAG, high adherence group; ICU, Intensive care unit; LAG, low adherence

group; LOS, Length of stay; QI, quality indicator; SAPS, Simplified Acute Physiology Score; SOFA, Sequential Organ Failure Assessment.

ZuberA, etal. BMJ Open 2022;12:e045327. doi:10.1136/bmjopen-2020-045327

Open access

demographics are displayed in table 1. Along the line, at

discharge patients generated an average daily SOFA score

of 8.2 (6.6.; 10.3) indicating resource- intensive moni-

toring and treatment of the patient.

In order to account for the remaining clinical patient

outcomes after grouping, we analysed the ventilation

parameters. Overall in the median, patients were venti-

lated for 431 (250; 709) hours on the ICU and 578 (338;

924) throughout their hospital stay. Following the divi-

sion into two adherence groups, there was a significant

reduction in duration of ventilation on ICU from 476 to

389 hours (p<0.001). Overall in- hospital duration of venti-

lation was decreased from 597 to 535 hours (p=0.017).

Concerning the number of SBTs and reintubations, there

was no significant finding (p=0.456 and p=0.531). In addi-

tion to the significant decrease in ventilation parameters

seen between the differences in adherence, the LOS was

decreased by 5 days from 21 to 16 (p<0.001) and overall

in- hospital LOS decreased from 36 to 26 days per patient

(p=0.001) in the median, indicating strong arguments

for QI adherence. With regard to economic outcome,

the overall median economic results (loss) per case was

−€2999. There was an increase in profitability from a

median loss of €3696–€1030 (p<0.001).

Considering the discharge of the patients, there was

a highly significant difference (p<0.001) between both

groups. Most patients were discharged to intermediate

care (44.6%), other ICUs (27.6%) or rehabilitation

(18.9%). Within the LAG, 50 (13.2%) patients died on

the ICU compared with 2 (1.0%) in the HAG. This gives

room to assume a certain impact of weaning quality on

mortality. However, since we did not include diagnosis

data, we cannot exclude an influence from this fact.

Multiple linear regression

The results of the multivariate linear regression analysis of

the complete study population of 583 patients are given

in table 2. The parameters were not adjusted for severity

of illness. The fixed variables age, sex and percentage of

QI adherence examined did not show significant effects

on profitability.

In the linear regression analysis, the LOS on the

study- ICU (p<0.001), the LOS in the hospital (p<0.001),

the averaged daily SOFA score (p<0.001) and the aver-

aged daily costs per patient (p<0.001) were shown to have

significant effects on the profitability (table 2). Strong

effects were found for the averaged daily SOFA score,

which increased profits per case by €1608 (95% CI €892

to €2323) for each SOFA point. Furthermore, the LOS

on the ICU decreased profits per case for €529 for every

day longer on the ICU. To the best of our knowledge,

multivariate regression for economic outcome has not yet

been conducted for these factors. The regression model

was performed without the admission scores for SAPS II,

SOFA and APACHE II. When these scores were included,

the statistical significances remained unchanged for the

remaining variables that were analysed (see table 2).

Comparing the cumulative parameters of weaning

patients along the years (see table 3), a higher number

of patients weaned as well as a higher average SOFA-

score can be associated with a higher number of median

economic result. The observation over time supports the

outcome parameters of table 1. Considering the develop-

ment since 2012, there is an increase in the number of

patients weaned per year and a decrease in the median

hours of ventilation per patient.

DISCUSSION

The most important finding was that clinical and

economic results were better within the HAG than the

LAG. We sought to evaluate whether adherence above

a certain quality threshold leads to a better economic

result per case for the hospital. Our univariate model

confirmed our hypothesis that higher quality leads to

better LOS and hospital costs of intensive care patients.

However, an improvement of the QI ‘early weaning’ was

not directly associated with a significant impact on the

profitability per case. In the regression model, we were

not able to prove that more quality lead to higher earn-

ings. Instead, significant factors were clinical outcome

parameters (LOS ICU, LOS Hospital and averaged daily

SOFA score), which had direct effects on profitability.

Moreover, these parameters were also superior within the

HAG, indicating a certain quality effect. This sequence of

effects shows that quality affects the economic results indi-

rectly via clinical outcome. This means that quality leads

to clinical efficiency. Literature already proposes a more

effective use of the costly resource ICU.26 Thus, from

an economic perspective it is recommended to transfer

patients as early as possible from ICU downstream (eg,

intermediate care) since a prolonged ICU- stay might be

inappropriate, dangerous and costly.23 25

Highly specialised ICUs are resource- intensive and cost-

intensive and not universally available. By implementing

QM as a method to constantly eliminating the factors

Table 2 Multiple linear regression analysis of factors

affecting the profit of 583 intensive care patients who

underwent the weaning process

Variable B (95% CI) SE P value

Age (years) −16 (−119 to 87) 52 0.765

Gender (male) 1139 (−2628 to 4906) 1918 0.553

Quality (%)* 3732 (−2457 to 9920) 3151 0.237

LOS Study- ICU (days) −529 (−671 to −387) 72 <0.001

LOS hospital (days) −143 (−213 to −71) 36 <0.001

Reintubations −928 (−2.457 to 602) 779 0.234

Average SOFA 1608 (892 to 2323) 364 <0.001

Daily costs (€)−7.6 (−11 to −4) 2 <0.001

*Quality, adherence to the quality indicator ‘early weaning from

invasive ventilation’.

ICU, intensive care unit; LOS, length of stay; SOFA, Sequential

Organ Failure Assessment.

6ZuberA, etal. BMJ Open 2022;12:e045327. doi:10.1136/bmjopen-2020-045327

Open access

of chance, hospitals are trying to reduce complexity in

defining, measuring and learning from QIs. Further-

more, QM is associated as a necessity for certification

processes and therefore incremental part of critical care

concepts.1 The importance of weaning protocols and

according adherence is based on studies that have proven

between 70% and 80% of all patients receiving >24 hour

invasive ventilation could already be weaned after the first

SBT.8 27 28 This is why in 2011, a study at our institution

investigated that the support of fast visual feedback for

adherence to standard operating procedures within the

PDMS led to decreased duration of mechanical ventila-

tion and higher documentation compliance, supporting

our findings.29 The approach of measuring and steering

quality with indicators carries several direct and indirect

economic incentives. First, less loss per patient due to

better clinical outcome has positive effects on the general

economic results of the department. Second, decreased

LOS on the ICU gives room to available beds earlier and

therefore other patients to fill in the existing resource.30

Third, because of public reporting and potential pay for

quality structures, indicators are important methods for

measuring quality and safety in healthcare, resulting in

better outcome.31 In particular, transparent QIs allow

department leaders to identify weak spots and initiate

improvement in a structured and measurable way.2 Our

matched with a study performed in 2008, showing positive

clinical outcome effects of ventilator weaning protocol

measures.32 Patients spent less time on mechanical venti-

lation, and thus less time in intensive care and in the

hospital. We found that the more patients that could be

weaned per year, the less time they spent on the ventilator

and better the economic results followed, since more

patients generating contribution margins covered fixed

costs. This effect shows that redundant capacities can be

used for new admissions and thus higher throughput,

similar to a former study at our institution.33

This study is the first to find that high adherence to

the QI ‘early weaning from invasive ventilation’ above

a proven threshold of 65% showed higher economic

returns (or less losses) than low adherence. Furthermore,

the study is unique in using a case defined data set to

examine the economic effect of a single QI. Current

economic prediction models in intensive care usually

describe interventions of entire QM programmes30 or

changes in staffing.33 Overall, we found that the median

financial return for a hospital is negative when focusing

on weaning from ventilation. This is independent of their

QI adherence results. In Germany, insurance companies

reimburse hospitals using the G- DRG System (German

DGRs System) based on a performance- oriented

compensation for inpatients. Within DRG- Systems,30 the

casemix of weaning patients does not provide adequate

economic incentives for quality based critical care since

the reimbursement is mainly focused on procedures, for

example, duration of ventilation. This is consistent with

other studies that found higher process quality led to

decreased ventilator dependence and reduced reimburse-

ment.25 26 34 To avoid wrong incentives, reimbursement

should potentially be tied to patient- centred outcomes.

For example, the prevention of VAP, PICS and CCIs. In

this study, we used comprehensive per- patient cost data,

based exclusively on the DRG- system. At our institution, a

case- related cost calculation is well established and highly

accurate for reimbursement per case and costs since we

have been substantial cost- accounting reference centre

since the beginning of the G- DRG- system. Therefore, we

used this administrative data to calculate the economic

outcome per case.35 In Germany, a representative mix of

hospitals gather case- related treatment costs on a yearly

basis in order to report them to the Institute for the

Hospital Reimbursement System for continuous develop-

ment.36 On an annual basis, cost weights are adjusted for

each DRG, potentially leading to higher reimbursement

per case. Hospitals can also benefit from economies of

scale, considering more cases per year with fixed reim-

bursement values. This may explain why in 2015 and 2016

profits per case were higher.

The results of this study can inform policy- makers on

the following points: In Germany, the application of QIs

in critical care is so far not mandatory.12 Since positive

effects of clinical and economic parameters can be found

Table 3 Financial demographics in median over time of 583 patients who underwent the weaning process

Variable 2012 2013 2014 2015 2016 2017

Weaning

patients

65 82 100 114 125 97

Average

SOFA

7.5 (5.6; 9.3) 8.3 (6.7; 11.0) 8.2 (6.5; 10.1) 8.1 (6.6; 9.6) 8.9 (7.0; 10.7) 8.3 (6.7; 11.0)

Duration of

ventilation

(hour)

660 (480; 977) 451 (230; 667) 400 (206; 673) 439 (261; 720) 374 (239; 602) 364 (210; 619)

Case- Mix

Index*

22.7 (19.1; 30.1) 18.0 (11.0; 23.9) 19.6 (11.6; 28.1) 18.8 (10.9; 23.8) 17.7 (11.6; 29.1) 23.2 (13.9; 32.2)

Profits per

case (€)†

−12 517 (−24 848; −806) −11 011 (−28 547; 999) −945 (−14 141; 8843) 390 (−11 340; 12 201) 3439 (−7494; 8784) −3136 (−22 012; 8284)

*Case- Mix Index, Averaged case- mix per case according to German DRG- system.

†Averaged financial result per case.

DRG, diagnosis- related group; SOFA, Sequential Organ Failure Assessment.

ZuberA, etal. BMJ Open 2022;12:e045327. doi:10.1136/bmjopen-2020-045327

Open access

measuring the adherence to only one indicator of the

DIVI set (n=10), it is recommended to establish QIs widely

and combine patient- centred outcomes with economic

outcomes systematically. Over the years examined, we

found that weaning and the according QI have developed

positively as the number of patients receiving weaning

increased while the duration of ventilation per patient

decreased. The relation between these two parameters

shows that the quality of care increased and the organisa-

tion for the volume effect became more efficient, which

is a dominant economic factor according to Nguyen et

al.37 However, in order to evolve further in this direction,

intensive care needs adequate reimbursement. Higher

assessment scores as SAPS II or SOFA play an important

role in ICU reimbursement and might induce higher

DRG reimbursement. Considering QM, contrary to the

majority of ward care, which benefits from shorter LOS

within the flat- compensation system, a decrease in LOS

in intensive care is not rewarded with higher reimburse-

ment. Literature confirms our analyses.36 This is why we

recommend that efforts for quality should be shifted in

the centre of reimbursement in intensive care for better

clinical outcomes, following the approach of valued-

based payment (pay for quality), where ICUs are checked

on costs and quality of service.38 Furthermore, because

keeping patients on the ICU and on mechanical ventila-

tion economically- incentivised is proven to be dangerous

for the patient8 and inefficient for the organisation.30 This

structural change can ensure the incentives for intensiv-

ists to adhere to quality standards instead of collecting

ventilation hours. Our argument is supported by a recent

publication of a group of experts in intensive care. They

argue in favour for a reform in hospital reimbursement,

away from flat- compensation towards progressive levels

of intensive care. Moreover, they suggest a central plan-

ning of all system relevant intensive care infrastructures

and according criteria for quality standards.39 In the end,

hospitals benefit from investments in quality, as clinical

quality has subsequent effects on economic returns.

Thus, not only hospitals, insurance companies and policy-

makers profit from adherence to QIs, also the patient

who should be in the centre of healthcare does.

Unanswered questions and future research

As noted previously, the study was conducted in a tertiary

university hospital, which is characterised by specific

and well- established medical processes and structures. A

transfer of our observations to other ICUs or reimburse-

ment systems is not feasible. The current study is subject

to its retrospective design and potential selection bias, as

some of the cases with incomplete data or special diag-

noses were not detected during the observation period.

We could have used neurological and neurosurgical diag-

noses to exclude patients with low chances for weaning

outcome, but in our administrative system there is no time

point matched to it accordingly as diagnoses are often

added just before discharge. For example, patients devel-

oping specific neurological conditions after their stay

on the study- ICU. Some aspects of our analysis deserve

comment on limitation. First, the weaning process has

constantly evolved during the years between 2012 and

2017. Since the importance of the weaning protocol

emerged throughout the years, the focus on measures

hereof and according documentation improved over the

years as documentation became mandatory at our insti-

tution.8 Furthermore, it was not possible matching the

qualifications of staffing as a determinant of adherence

to quality and curbing of costs. There is supposed to be

a connection between experience and cost awareness.40

Second, even though indicators and our study- ICU can

be examined independently for research purposes, the

QI and its progression are substantially connected to

other intensive care indicators.19 For further research,

the interactions between the QIs and the progression on

other ICUs need to be considered. Our results provide a

robust assessment of the impact of changes of the quality

adherence and robust evaluation of their effects.

CONCLUSION

While the need for critical care increases constantly for

various reasons (eg, demographic change or pandemic

crisis), the challenge to provide high- quality but cost-

effective services will only become more important. Avail-

able resources differ among the various hospital sizes

and types. Although we examined a single indicator for

quality in a university reference centre and found proof

that high adherence to it lead to significantly better clin-

ical outcome, we think patients and hospitals in general

benefit from high adherence to quality measures. Within

the univariate analysis, major clinical parameters were

significantly better in the HAG. Furthermore, we showed

that adherence for 65% or higher generated significantly

higher median earnings within our univariate analysis.

However, we also showed that the investigated QI does

not significantly affect economic results in our multi-

variate analysis. Instead, by using clinical parameters as

proxies for clinical outcome, they were found to be the

main drivers for according economic success. The reason

for this is the increased number of patients who could be

treated due to more total capacity, when LOS decreased

due to higher quality. This is why the focus of this study

is not only on reimbursement and on costs, but also on

the direct effect of quality on the clinical outcome, which

subsequently influences economic results.

Overall, quality matters for reimbursement, but reim-

bursement is not adjusted to the costs of providing quality.

Since there is no central, structured and timely publication

of comparable quality data in Germany, it is difficult for poli-

tics and assurances to reimburse on a pay for quality model

as the basis for comparisons is missing as not mandatory. Still,

as quality in treatment is decisive for the patient’s hospital

choice and the results of the treatment, QIs will be essential

for public information and health economics as the patient

decides where to be treated.

8ZuberA, etal. BMJ Open 2022;12:e045327. doi:10.1136/bmjopen-2020-045327

Open access

Contributors CS introduced quality indicator based treatment for critically ill

patients at Charité hospital in terms of both research and implementation in patient

care. She perceived the underlying idea for this study. CS and FB set the aims and

design of this study. They accept full responsibilty for the work, had access to the

data and controlled the decision to pubilsh. AZ and RA performed data collection.

AZ conducted statistical analysis supervised by JK. AZ shared responsibility for

the study design, had full access to the data and drafted the manuscript. CS and

OK contributed to the interpretation of data from a medical point of view, and

specifically from the perspective of quality indicators. MH and MD contributed

from the perspective of economics, RJ from the perspective of quality science.

FB supervised the overall coordination of the study and contributed from the data

science perspective. All authors critically reviewed and advised with their expertise

on the manuscript.

Funding This analysis is part of a quality improvement effort from the Department

of Anesthesiology and Operative Intensive Care Medicine of the Charité

Universitätsmedizin – Berlin, Campus Mitte and Virchow- Klinikum.

Competing interests CS reports grants from Aridis Pharmaceutical, grants

from B. Braun Melsungen AG, grants from Drägerwerk AG & Co. KGaA, grants

from Deutsche Forschungsgemeinschaft/German Research Society, grants

from Deutsches Zentrum für Luft- und Raumfahrt e. V. (DLR)/German Aerospace

Centre, grants from Einstein Stiftung Berlin/Einstein Foundation Berlin,

grants from European Society of Anaesthesiology, grants from Gemeinsamer

Bundesausschuss/Federal Joint Committee (G- BA), grants from Inneruniversitäre

Forschungsförderung/Inner University Grants, grants from Projektträger im DLR/

Project Management Agency, grants from Stifterverband/Non- Profit Society

Promoting Science and Education, grants from WHOCC, grants from Baxter

Deutschland, grants from Cytosorbents Europe, grants from Edwards Lifesciences

Germany, grants from Fresenius Medical Care, grants from Grünenthal, grants

from Masimo Europe, grants from Pfizer Pharma PFE, personal fees from Georg

Thieme Verlag, grants from Dr. F. Köhler Chemie, grants from Sintetica, grants from

European Commission, grants from Stifterverband für die deutsche Wissenschaft

e.V./Philips, grants from Stiftung Charité, grants from AGUETTANT Deutschland,

grants from AbbVie Deutschland GmbH & Co. KG, grants from Amomed Pharma

GmbH, grants from InTouch Health, grants from Copra System, grants from Correvio,

grants from Max- Planck- Gesellschaft zur Förderung der Wissenschaften e.V., grants

from Deutsche Gesellschaft für Anästhesiologie & Intensivmeidzin (DGAI), grants

from Stifterverband für die deutsche Wissenschaft e.V./Medtronic, grants from

Philips Electronics Nederland BV, grants from BMH, outside the submitted work; In

addition, CS has a patent 10 2014 215 211.9 licensed, a patent 10 2014 215 212.9

licensed, a patent 10 2018 114 364.8 licensed, and a patent 10 2018 110 275.5

licensed. MH has nothing to disclose. MD has nothing to disclose. RA has nothing to

disclose. JK has nothing to disclose. RJ has nothing to disclose. FB reports grants

from Einstein Foundation, personal fees from Axon Publishing, grants from Vifor

Pharma, personal fees from Elsevier Publishing, grants from Federal Ministry of

Health, Germany, grants from Berlin Institute of Health, outside the submitted work.

Patient consent for publication Not applicable.

Ethics approval The study was approved by the Ethics Commission of Charité –

Universitätsmedizin Berlin (EA2/139/20). The need for patient’s consent was waived

due to the retrospective nature of the study.

Provenance and peer review Not commissioned; externally peer reviewed.

Data availability statement Data are available on reasonable request.

Supplemental material This content has been supplied by the author(s). It has

not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been

peer- reviewed. Any opinions or recommendations discussed are solely those

of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and

responsibility arising from any reliance placed on the content. Where the content

includes any translated material, BMJ does not warrant the accuracy and reliability

of the translations (including but not limited to local regulations, clinical guidelines,

terminology, drug names and drug dosages), and is not responsible for any error

and/or omissions arising from translation and adaptation or otherwise.

Open access This is an open access article distributed in accordance with the

Creative Commons Attribution Non Commercial (CC BY- NC 4.0) license, which

permits others to distribute, remix, adapt, build upon this work non- commercially,

and license their derivative works on different terms, provided the original work is

properly cited, appropriate credit is given, any changes made indicated, and the use

is non- commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.

ORCID iDs

ClaudiaSpies http://orcid.org/0000-0002-1062-0495

FelixBalzer http://orcid.org/0000-0002-6789-8471

REFERENCES

1 Holtel M, Roßmüller T, Frommhold K. [Quality Management in

Medicine: What the Surgeon Needs to Know]. Zentralbl Chir

2016;141:583–90.

2 Magunia P, Keller M, Rhode A. [Effects of quality- oriented

remuneration]. Unfallchirurg 2016;119:454–6.

3 Busse R, Geissler A, Aaviksoo A, etal. Diagnosis related groups

in Europe: moving towards transparency, efficiency, and quality in

hospitals? BMJ 2013;346:f3197.

4 Martin J, Neurohr C, Bauer M, etal. [Cost of intensive care in a

German hospital: cost- unit accounting based on the InEK matrix].

Anaesthesist 2008;57:505–12.

5 Cooper LM, Linde- Zwirble WT. Medicare intensive care unit

use: analysis of incidence, cost, and payment. Crit Care Med

2004;32:2247–53.

6 Funk G- C, Anders S, Breyer M- K, etal. Incidence and outcome of

weaning from mechanical ventilation according to new categories.

Eur Respir J 2010;35:88–94.

7 Cox CE, Carson SS, Govert JA, etal. An economic evaluation of

prolonged mechanical ventilation. Crit Care Med 2007;35:1918–27.

8 Deutsche Gesellschaft für Anästhesiologie & Intensivmedizin. S3-

Leitlinie invasive Beatmung und Einsatz extrakorporaler Verfahren bei

akuter respiratorischer Insuffizienz, Langversion, stand 04.12.2017,

2017. Available: www.intensivmedizin-leipzig.de [Accessed 21 Nov

2018].

9 Ely EW, Baker AM, Dunagan DP, etal. Effect on the duration of

mechanical ventilation of identifying patients capable of breathing

spontaneously. N Engl J Med 1996;335:1864–9.

10 Pham T, Brochard LJ, Slutsky AS. Mechanical ventilation: state of the

art. Mayo Clin Proc 2017;92:1382–400.

11 Cohen IL, Booth FV. Cost containment and mechanical ventilation in

the United States. New Horiz 1994;2:283–90.

12 Kumpf O. [Quality indicators in intensive care medicine : Background

and practical use]. Med Klin Intensivmed Notfmed 2021;116:17–28.

13 Ihra GC, Lehberger J, Hochrieser H, etal. Development of

demographics and outcome of very old critically ill patients admitted

to intensive care units. Intensive Care Med 2012;38:620–6.

14 WeanNet Study Group. [WeanNet: The network of weaning

units of the DGP (Deutsche Gesellschaft für Pneumologie und

Beatmungsmedizin) - results to epidemiology an outcome in

patients with prolonged weaning]. Dtsch Med Wochenschr

2016;141:e166–72.

15 Navalesi P, Frigerio P, Patzlaff A, etal. Prolonged weaning: from the

intensive care unit to home. Rev Port Pneumol 2014;20:264–72.

16 Braun J- P, Bause H, Bloos F, etal. Peer reviewing critical care:

a pragmatic approach to quality management. Ger Med Sci

2010;8:Doc23.

17 Bundesministerium für Gesundheit. Qualitätssicherung im

Krankenhausbereich, 2018. Available: https://www.bundesge

sundheitsministerium.de/qualitaet-krankenhausversorgung.html

[Accessed 3 Sep 2018].

18 Brinkmann A, Braun JP, Riessen R, etal. [Quality assurance

concepts in intensive care medicine]. Med Klin Intensivmed Notfmed

2015;110:575–80. 582- 583.

19 Kumpf O, Braun J- P, Brinkmann A, etal. Quality indicators in

intensive care medicine for Germany - third edition 2017. Ger Med

Sci 2017;15:Doc10.

20 Søgaard R, Enemark U. The cost- quality relationship in European

hospitals: a systematic review. J Health Serv Res Policy

2017;22:126–33.

21 Donabedian A. The role of outcomes in quality assessment and

assurance. QRB Qual Rev Bull 1992;18:356–60.

22 Spies C, Kastrup M, Kerner T. SOPs in Anästhesiologie und

Schmerztherapie. Thieme 2013.

23 Kastrup M, von Dossow V, Seeling M, etal. Key performance

indicators in intensive care medicine. A retrospective matched cohort

study. J Int Med Res 2009;37:1267–84.

24 Fadila M, Regunath H. Ventilator weaning. Treasure Island, FL, 2018.

25 Nachtigall I, Tamarkin A, Tafelski S, etal. Impact of adherence

to standard operating procedures for pneumonia on outcome of

intensive care unit patients. Crit Care Med 2009;37:159–66.

26 Rapoport J, Teres D, Zhao Y, etal. Length of stay data as a guide

to hospital economic performance for ICU patients. Med Care

2003;41:386–97.

27 Brochard L, Rauss A, Benito S, etal. Comparison of three

methods of gradual withdrawal from ventilatory support during

ZuberA, etal. BMJ Open 2022;12:e045327. doi:10.1136/bmjopen-2020-045327

Open access

weaning from mechanical ventilation. Am J Respir Crit Care Med

1994;150:896–903.

28 Vitacca M, Vianello A, Colombo D, etal. Comparison of two methods

for weaning patients with chronic obstructive pulmonary disease

requiring mechanical ventilation for more than 15 days. Am J Respir

Crit Care Med 2001;164:225–30.

29 Kastrup M, Nolting MJ, Ahlborn R, etal. An electronic tool for visual

feedback to monitor the adherence to quality indicators in intensive

care medicine. J Int Med Res 2011;39:2187–200.

30 Smyrnios NA, Connolly A, Wilson MM, etal. Effects of a

multifaceted, multidisciplinary, hospital- wide quality improvement

program on weaning from mechanical ventilation. Crit Care Med

2002;30:1224–30.

31 Lindenauer PK, Remus D, Roman S, etal. Public reporting and pay

for performance in hospital quality improvement. N Engl J Med

2007;356:486–96.

32 Girard TD, Kress JP, Fuchs BD, etal. Efficacy and safety of a paired

sedation and ventilator weaning protocol for mechanically ventilated

patients in intensive care (awakening and breathing controlled trial): a

randomised controlled trial. Lancet 2008;371:126–34.

33 Kastrup M, Seeling M, Barthel S, etal. Effects of intensivist coverage

in a post- anaesthesia care unit on surgical patients' case mix and

characteristics of the intensive care unit. Crit Care 2012;16:R126.

34 Soo Hoo GW, Wen YE, Nguyen TV, etal. Impact of clinical guidelines

in the management of severe hospital- acquired pneumonia. Chest

2005;128:2778–87.

35 InEK GmbH. Kalkulation von Behandlungskosten, 2016. Available:

https://www.g-drg.de/Kalkulation2/DRG-Fallpauschalen_17b_KHG/

Kalkulationshandbuch

36 Riessen R, Hermes C, Bodmann K- F, etal. [Reimbursement of

intensive care services in the German DRG system : Current

problems and possible solutions]. Med Klin Intensivmed Notfmed

2018;113:13–23.

37 Nguyen Y- L, Wallace DJ, Yordanov Y, etal. The volume- outcome

relationship in critical care: a systematic review and meta- analysis.

Chest 2015;148:79–92.

38 Blumenthal D, Abrams MK. Tailoring complex care management for

High- Need, high- cost patients. JAMA 2016;316:1657–8.

39 Riessen R, Markewitz A, Grigoleit M, etal. [Discussion paper

for ahospital financing reform in Germany from the perspective

of intensive care medicine]. Med Klin Intensivmed Notfmed

2020;115:59–66.

40 Wunsch H, Gershengorn H, Scales DC. Economics of ICU

organization and management. Crit Care Clin 2012;28:25–37. v.