Open Access

Variations in hospital standardised mortality ratios (HSMR) as a result of frequent readmissions

  • Wim F van den Bosch1Email author,
  • Peter Spreeuwenberg2 and
  • Cordula Wagner2, 3
BMC Health Services ResearchBMC series ¿ open, inclusive and trusted201212:91

DOI: 10.1186/1472-6963-12-91

Received: 1 September 2011

Accepted: 4 April 2012

Published: 4 April 2012

Abstract

Background

We investigated the impact that variations in the frequency of readmissions had upon a hospital's standardised mortality ratio (HSMR). An adapted HSMR model was used in the study. Our calculations were based on the admissions of 70 hospitals in the Netherlands during the years 2005 to 2009.

Methods

Through a retrospective analysis of routinely collected hospital data, we calculated standardised in-hospital mortality ratios both by hospital and by diagnostic group (H/SMRs) using two different models. The first was the Dutch 2010 model while the second was the same model but with an additional adjustment for the readmission frequency. We compared H/SMR outcomes and the corresponding quality metrics in order to test discrimination (c-statistics), calibration (Hosmer-Lemeshow) and explanatory power (pseudo-R2 statistic) for both models.

Results

The SMR outcomes for model 2 compared to model 1, varied between -39% and +110%. On the HSMR level these variations ranged from -12% to +11%. There was a substantial disagreement between the models with respect to significant death on the SMR level as well as the HSMR level (~ 20%). All quality metrics comparing both models were in favour of model 2. The susceptibility to adjustment for readmission increased for longer review periods.

Conclusions

The 2010 HSMR model for the Netherlands was sensitive to adjustment for the frequency of readmissions. A model without this adjustment, as opposed to a model with the adjustment, produced substantially different HSMR outcomes. The uncertainty introduced by these differences exceeded the uncertainty indicated by the 95% confidence intervals. Therefore an adjustment for the frequency of readmissions should be considered in the Netherlands, since such a model showed more favourable quality metric characteristics compared to a model without such an adjustment. Other countries could well benefit from a similar adjustment to their models. A review period of the data collected over the last three years, at least, is advisable.

Background

Hospital standardised mortality ratios (HSMRs) are widely used as an indicator to assess and improve the quality of care. HSMRs have been calculated in the UK, USA, Canada, Australia and the Netherlands [1]. HSMR outcomes fluctuate around 100. For example in the UK the outcomes range from 72 to 118 and in the Netherlands from 62 to 142 [1, 2]. The Dutch HSMR model was developed by the Dr. Foster Intelligence Unit in London in co-operation with Kiwa Prismant in the Netherlands. The model adjusts for patient casemix factors including age, sex and diagnostic group. It is based upon the national medical registration data (LMR). In October 2010 the 'Dutch Hospital Data' group (DHD) distributed the hospital specific, 'H/SMR report 2007-2009 with detailed information on diagnostic groups and patient categories', among Dutch hospitals. In this way, each hospital obtained an insight into their own standardised mortality ratios (SMRs) for 50 diagnostic groups and had the opportunity to work on improving patient safety.

HSMRs are also made publicly available in countries, for example in the UK's '2010 Dr Foster Hospital Guide'[2]. However questions have been raised about how reliable, valid and applicable the model really is [310]. It is still not clear to what degree the current HSMR and SMR outcomes are attributable to quality of care and can be used to make a meaningful comparison between hospitals. Bottle et al claim however that variations in adjustment methods have a limited impact on the hospitals position relative to funnel plot control limits [11].

One type of potential distortion in establishing comparable HSMRs is caused by the failure to adjust for variations in readmission frequencies. Van den Bosch, et al [10] have demonstrated that the Dutch 2010 HSMR model favoured hospitals with relatively many frequently readmitted patients, regardless of disease, compared to hospitals with many patients who are not frequently readmitted. They were, however, not able to quantify the effect of this upon the HSMR, because the study was restricted to only six hospitals. In a new study more data was made available enabling us to quantify the possible effect in the Netherlands. We were able to compare the HSMR outcomes of the current HSMR model to an alternative model with an additional adjustment for readmission frequency. In essence we asked: Does the new model represent an improvement in the current model, based on c-statistics, goodness of fit and explanatory power? And if so, to what extent do HSMR and SMR outcomes differ, comparing model 1 to model 2?

Methods

Setting

For the purpose of this model analysis the DHD group gave its formal permission to use an anonymised LMR dataset from 89 Dutch hospitals covering the years 2005 to 2009. Nineteen hospitals were excluded from the calculation because the quality of their patient registration data was insufficient.

This study excluded any experimental research, either on human or animal subjects.

The HSMR model used in this study

We performed HSMR calculations using two models, shown as model 1 and model 2.

For model 1, we used the Dr Foster model [1], applied in the Netherlands in 2010. This Dutch model included clinical admissions only, so no day cases. Furthermore only in-hospital deaths were counted. There are differences between the Dutch and the UK model. The Dutch model used 50 Clinical Classifications Software (CCS) groups based on an ICD-9 coding, whereas the UK used 56 CCS groups based on ICD-10 coding, of which 42 groups were the same as in the Dutch model. Furthermore, the UK model adjusted for palliative care and for the number of previous emergency admissions within one year. The Dutch model did not adjust for either of these.

HSMR in brief

The HSMR ('hospital standardised mortality ratio')is the ratio between the observed number of in-hospital deaths and the predicted number of deaths, determined by comparing the patient casemix with the national average. The outcome is standardised around 100. A value above 100 indicates higher mortality than average, below 100, lower mortality. If the calculation is applied to one of the 50 diagnostic groups, then we speak of the SMR ('standardised mortality ratio').

The Dutch HSMR model of 2010 adjusts for the following casemix properties: year of discharge, sex, age at admission, admission type, comorbidity (charlsonindex), social deprivation, month of admission, source of referral, diagnostic group, and, in part, for the casemix on the primary diagnostic level.

Each diagnostic group is composed of a number of underlying diseases: ICD-9 codes (International Classification of Diseases Ninth edition) as determined by the Clinical Classification Software (CCS). This tool clusters various ICD-9 diagnoses into a manageable number of meaningful categories [12], indicated as 'CCS diagnostic groups'. The selected CCS diagnostic groups have a relatively high mortality and together cover over 80% of the total number of hospital deaths.

For model 2, we took model 1 and added an adjustment for the frequency of readmission, as described by Van den Bosch WF, et al [10]. The authors demonstrated that frequently admitting patients was associated with lower mortality ratios per admission, for which an additional correction would be needed. In their publication the patients were grouped into 'patient view' classes P(m). We have also applied this to our study. Here the admission frequency m was equal to the number of times a patient was admitted to the same hospital during the five-year study period.

For example, a patient who was admitted ten times during the five-year period to hospital X, to be treated for one or more diseases, possibly with different diagnoses, was part of the patient view class P(10) and not part of any other patient view class. This patient contributed ten admissions to P(10) which were all accounted for in the regression calculation.

We were able to retrieve the number of times that a patient was admitted through the unique patient identification number that all of the 70 hospitals included were using. We have grouped the patient view classes into eight categories: m = 1, m = 2, m = 3, m = 4, m = 5-6, m = 7-9, m = 10-20 and m > 20. This is in order to limit the number of categories for the regression calculation and to avoid categories becoming too small.

Agreement, or otherwise of the two models

We have calculated to what extent both models did agree, or did not agree, as follows:
  1. 1.

    Relative change. To what extent did SMRmodel 1 differ from SMRmodel 2 per CCS diagnostic group, per hospital? Using the formula SMRdelta = (SMRmodel2/SMRmodel1 -1)*100%, we calculated the shifts for all of the 3500 SMRs (50 diagnostic groups times, 70 hospitals), occurring when changing from model 1 to model 2. We represented these in a frequency distribution. In a similar way we have calculated and represented the 70 shifts that occurred for the HSMRs of the 70 hospitals. We also compared all pairs of death predictions of model 1 and model 2 per admission by calculating regressions coefficients (R2) on the SMR level and the HSMR level. These metrics were used as a measure of statistical 'distance' between both models.

     
  2. 2.

    Significance scores of SMRs and HSMRs. We determined per SMR how many hospitals with a significantly high SMR score according to model 1 turned out not to be significantly high according to model 2. And vice versa, in other words: how often a significantly high SMR score in model 2 turned out not to be a significantly high score in model 1. In a similar way we calculated these differences on the hospital level for HSMRs.

     

Quality metrics of the two models

We have calculated and compared the quality metrics of the models with respect to:
  1. 1.

    Discrimination, expressed in 'c-statistics' on the SMR level and on the HSMR level. This statistical measure indicates how well a regression model is able to predict mortality. Each predicted outcome per admission is compared to the observed outcome: died or survived. A c-statistic of 0.5 has no predictive value: 50% right, 50% wrong. Values above 0.75 suggest good discrimination. A value of 1 is perfect. The overall c-statistic for the Dutch HSMR (model 1) scored above 0.85.

     
  2. 2.

    Calibration, according to Hosmer and Lemeshow. This represents a statistical test for goodness-of-fit that is frequently used in risk prediction models. The test assesses whether or not the observed event rates match expected event rates in subgroups of the model population. The test specifically identifies subgroups as the deciles of fitted risk values. Models in which expected and observed event rates in subgroups are similar, are called well-calibrated [13].

     
  3. 3.

    Explanatory power, using the pseudo R2 statistic according to the 'Nagelkerke R square', in order to assess the degree to which the additional adjustment for readmission changes the unexplained variance in the data.

     

Sensitivity of the HSMR model to adjustment for readmission

We have investigated to what extent the model was sensitive to variations in the length of the period under review. We took three scenarios: a period of review of one year (2009), two years (2008-2009) and five years (2005-2009). We then examined the statistical distance between model 1 and model 2, and the three model quality metrics on the HSMR level.

Results

We analysed the data from 89 hospitals of which 19 were excluded because they had an incomplete registration of the patient identification numbers. Therefore, we included 70 hospitals for the regression calculation, a total of 2494613 admissions.

Agreement between the two models

Figure 1 shows the frequency distribution of the relative changes of the 3500 SMRs, in going from model 1 to model 2. Similarly the changes of the 70 HSMRs of the hospitals are shown in Figure 2. The SMR changes ranged from -39% to +110% and the HSMR changes ranged from -12% to +11%. The standard deviation of the frequency distribution of the changes amounted to 12.2% (roughly 12 SMR points) for the SMRs and to 4.1% (roughly 4 HSMR points) of the HSMRs.
https://static-content.springer.com/image/art%3A10.1186%2F1472-6963-12-91/MediaObjects/12913_2011_Article_2008_Fig1_HTML.jpg
Figure 1

Frequency distribution of relative changes of all 3500 SMRs when changing from model 1 (Dr Foster model) to model 2 (model with additional adjustment for readmission frequency). * Standard deviation amounts to 12.2%.

https://static-content.springer.com/image/art%3A10.1186%2F1472-6963-12-91/MediaObjects/12913_2011_Article_2008_Fig2_HTML.jpg
Figure 2

Frequency distribution of relative changes of the 70 HSMRs when changing from model 1 (Dr Foster model) to model 2 (model with additional adjustment for readmission frequency). * Standard deviation amounts to 4.1%.

Table 1 shows the agreement that we found between model 1 and model 2, expressed as a statistical distance per SMR. The R2 coefficients varied from 0.60 (COPD) to 0.95 (intracranial injury). For the entire model the coefficient amounted to 0.86.
Table 1

A comparison of SMR outcomes of model 1 and model 2

CCS

code

CCS-diagnostic group title

Number of admissions

Distance between

model 1

and 2 (R2)

Number of hospitals with

SMR significantly high?

c-statistic for

    

Model 1: no

model 2: yes

Model 1: yes

model 2: no

Model 1

Model 2

2

Septicemia (except in labour)

14647

0.83

2

1

0.760

0.783

12

Cancer of oesophagus

9703

0.74

1

3

0.730

0.763

13

Cancer of stomach

13945

0.75

0

2

0.733

0.779

14

Cancer of colon

45116

0.77

2

5

0.773

0.814

15

Cancer of rectum and anus

22246

0.72

1

3

0.771

0.817

17

Cancer of pancreas

9950

0.76

1

0

0.685

0.704

19

Cancer of bronchus, lung

76714

0.75

4

6

0.792

0.822

24

Cancer of breast

64126

0.74

2

1

0.916

0.850

29

Prostate cancer

24092

0.76

3

1

0.886

0.865

32

Cancer of bladder

44288

0.71

1

0

0.872

0.874

38

Non-Hodgkin's lymphoma

19413

0.69

0

5

0.802

0.841

39

Leukaemias

12466

0.77

0

1

0.781

0.820

42

Secondary malignancies

69303

0.77

5

4

0.748

0.770

44

Neoplasms unspec nature or uncertain behaviour

18982

0.63

2

1

0.805

0.840

50

Diabetes mellitus with complications

36837

0.74

1

0

0.824

0.846

55

Fluid and electrolyte disorders

29705

0.81

0

1

0.795

0.824

59

Deficiency and other anaemia

48896

0.61

0

1

0.756

0.802

85

Coma, stupor and brain damage

4286

0.92

0

0

0.801

0.815

96

Heart valve disorders

30708

0.77

1

2

0.775

0.788

100

Acute myocardial infarction

91802

0.72

1

1

0.740

0.781

101

Coronary atherosclerosis and other heart disease

250183

0.66

1

2

0.784

0.805

103

Pulmonary heart disease

26937

0.78

0

1

0.754

0.776

106

Cardiac dysrhythmias

174923

0.72

1

2

0.828

0.860

107

Cardiac arrest and ventricular fibrillation

8118

0.89

2

2

0.739

0.764

108

Congestive heart failure, non-hypertensive

108541

0.63

4

1

0.637

0.702

109

Acute cerebrovascular disease

102295

0.86

3

0

0.744

0.761

114

Peripheral and visceral atherosclerosis

44090

0.94

1

3

0.908

0.914

115

Aortic, peripheral and visceral artery aneurysms

26538

0.93

1

1

0.871

0.887

116

Aortic & peripheral arterial embolism or thrombosis

32881

0.75

1

1

0.882

0.892

117

Other circulatory disease

20452

0.80

0

1

0.849

0.862

122

Pneumonia

128867

0.74

2

0

0.753

0.785

127

COPD and bronchiectasis

82898

0.60

0

2

0.701

0.761

129

Aspiration pneumonitis

4485

0.79

1

1

0.644

0.688

130

Pleurisy, pneumothorax, pulmonary collapse

24798

0.78

0

0

0.812

0.827

133

Other lower respiratory disease

99460

0.77

1

0

0.854

0.866

145

Intestinal obstruction without hernia

34802

0.76

1

1

0.813

0.836

146

Diverticulosis and diverticulitis

35917

0.70

0

0

0.814

0.828

149

Biliary tract disease

135587

0.80

1

0

0.894

0.899

150

Liver disease, alcohol-related

5133

0.69

0

2

0.678

0.758

151

Other liver diseases

14584

0.80

1

0

0.771

0.807

153

Gastrointestinal haemorrhage

37229

0.71

0

1

0.738

0.764

155

Other gastrointestinal disorders

45487

0.91

0

0

0.907

0.908

157

Acute and unspecified renal failure

8804

0.73

0

0

0.723

0.760

158

Chronic renal failure

14570

0.75

0

2

0.846

0.872

159

Urinary tract infections

62382

0.77

0

0

0.817

0.836

226

Fracture of neck of femur (hip)

74557

0.73

0

2

0.719

0.732

233

Intracranial injury

56535

0.95

0

1

0.939

0.941

237

Complication of device, implant or graft

76079

0.69

0

0

0.821

0.833

238

Complications surgical procedures or medical care

67106

0.76

0

0

0.827

0.833

249

Shock

3150

0.80

1

0

0.716

0.733

 

Total all CCS diagnostic groups

2494613

0,86

49

64

0.852

0.867

* Per CCS diagnostic group: number of admissions; statistical distance between model 1 and model 2; number of hospitals with significance differences of SMRs comparing model 1 to model 2, c-statistics for both models. The table includes results of 70 Dutch hospitals over the period 2005 - 2009

Finally, we calculated per SMR for how many hospitals a significantly high SMR score changed to not significantly high from model 1 to model 2 and vice versa; see Table 1. Model 1 indicated 328 SMRs in total as "higher than expected" of which 64 were not indicated by model 2 (20%). On the other hand, model 2 indicated 313 SMRs in total as "higher than expected" of which 49 were not indicated by model 1 (16%). Across hospitals, model 1 indicated that 23 hospitals recorded a higher than expected HSMR of which four were not indicated by model 2 (17%). And, alternatively, model 2 indicated that 23 hospitals were having a higher than expected HSMR of which four were not indicated by model 1 (17%).

H/SMR model quality metrics

For both models the 'c-statistic' per CCS diagnostic group was calculated, see Table 1. Model 2 scored a higher c-statistic compared to model 1 in 47 groups out of 50. For eighteen of these groups the scores of model 2 exceeded the scores of model 1 by between 0.03 and 0.08. Model 1 had one diagnostic group for which the c-statistic was at least 0.03 better compared to model 2. The c-statistic of the HMSR across all hospitals amounted to 0.852 for model 1 and to 0.867 for model 2 (favourable).

The results of the goodness-of-fit test are shown in Table 2, expressed in the quality metrics of the Hosmer and Lemeshow test. It shows that model 2 was better calibrated compared to model 1: 7434 mismatches (5.75% of the total deaths) for model 1 compared to 5474 (4.23%) mismatches for model 2.
Table 2

Contingency table with Hosmer and Lemeshow test for goodness-of-fit

 

Died model 1

Died model 2

Decile

Observed

Expected

Observed

Expected

1

138

268

134

231

2

498

786

469

677

3

1057

1464

997

1295

4

1888

2485

1721

2197

5

3609

4038

3002

3544

6

6181

6399

5299

5599

7

10486

10026

8999

8843

8

16797

15578

14890

14337

9

28064

26026

27112

25084

10

60625

62273

66720

67535

Percentage mismatch

5,75%

 

4,23%

Finally the overall explanatory power expressed in pseudo R2 statistics were calculated. These amounted to 0.259 for model 1 versus 0.282 for model 2 and so favourable to model 2.

Sensitivity of the model to the review period of adjustment

The overall agreement between models 1 and 2 varied, depending on the review period, from 0.92 (one year) to 0.89 (two years) to 0.86 (five years). So a longer review period corresponded with an increasing divergence between the two models.

Table 3 shows how the overall model quality metrics varied for the three review periods. A longer review period corresponded to an increase of model quality metric performance.
Table 3

Comparison of HSMR model quality metrics of model 1 to model 2 for various lengths of review period

 

Model 1

Model 2

Difference

In favour of

Overall c-statistic

    

1 year period

0.857

0.866

0.009

model 2

2 year period

0.856

0.865

0.009

model 2

5 year period

0.852

0.867

0.015

model 2

% decile match hosmer lemeshow test

   

1 y period

93.85%

94.13%

0.28%

model 2

2 year period

93.82%

94.41%

0.59%

model 2

5 year period

94.25%

95.77%

1.52%

model 2

Nagelkerke R square

    

1 year period

0.269

0.285

0.016

model 2

2 year period

0.266

0.284

0.018

model 2

5 year period

0.259

0.282

0.023

model 2

Discussion

How should we interpret the differences in model outcomes?

In this study we have calculated and compared the H/SMR outcomes of two models. The disagreement between the two models was substantial, according to the relative changes in HSMRs and SMRs.

The standard deviation of the frequency distribution of HSMR-change amounted to 4 HSMR points, which was substantial compared to the standard deviation of the HSMR-frequency distribution amounting to 14 HSMR points. Moreover, these shifts introduced an uncertainty into the HSMR outcomes that exceded the uncertainty due to chance as indicated by the 95% confidence intervals (CIs): The 95% CI of a HSMR equalled [-2σ, 2σ]. On average for the 70 hospitals over five years, the standard deviation σ of the CI distribution equalled 2.5 HSMR points.

With respect to the SMRs: the statistical 'distance' between SMRs of model 1 and model 2 expressed in R2 coefficients, varied from 0.60 to 0.95. Low scores indicated susceptibility to adjustment for readmissions. On average chronic diseases scored lower compared to acute diseases as to be expected since chronic diseases in particular often result in readmissions. There was no particular peak among any of the main diagnostic groups. The major diagnostic groups like cancer and heart and lung diseases all showed substantial disagreement between model 1 and model 2 due to adjustment for readmission. The results also showed disagreement between the models in designating hospitals as having higher than expected H/SMRs. The percentages of disagreement varied from 16% to 20%.

Bottom-line: all differences in H/SMR outcomes that we found between the two models, cannot be attributed to differences in the quality of care neither to 'chance'. They have to be attributed to the difference in the type of model, because this is the only distinction applied in producing the results from the same set of admissions. This implies that publicised good and bad practices, linked to hospital rankings based on HSMRs, can differ substantially depending on the choice of model.

How do readmissions affect H/SMRs and how do they occur?

We asked: how did readmissions impact the H/SMR, how did readmissions occur and why did hospitals differ in this respect?

For example: endlessly readmitting the same patient in the same hospital continuously increased the denominator (predicted death) of the calculated HSMR for that hospital. We have observed numerous cases where the contribution to the denominator on behalf of a single patient over the years aggregated to numerical values of 3 to 4. However a patient could only die once and so any patient could maximally contribute a numerical value of 1 to the numerator. Consequently this effect lowered the HSMR ratio and favoured hospitals which had many frequently readmitted patients.

There are many types of mechanisms that explain the variation in the frequency of clinical readmission. These could have been administrative in nature, for example differences in administrating chemotherapy [10]. There were also many systematic differences in treatment practice. For example one hospital applied and recorded a different number of clinical admissions than another hospital did for the same combination of diagnosis and treatment applied to the same patient. Differences also occurred as a consequence of transferring patients back and forth. Poor care during the initial admission may also have triggered readmissions later. As a consequence all these examples of increased readmissions were being 'rewarded' by model 1 while model 2 was correcting for this phenomenon.

The analysis showed that the combined effect of all these mechanisms was not restricted to some of the diagnostic groups, but had an impact upon most of them. Clearly the HSMR model was susceptible to an adjustment for the frequency of readmissions in the way we have defined it.

The analysis also showed that reducing the review period from five to two years and from two to one year, resulted in a substantial reduction of this susceptibility. This phenomenon may be explained as follows: It may take several years before the readmissions of a patient constitute a substantial sequence of admissions. Furthermore a sequence might also have started long before the analysis period started. In these cases apparently one year was not a sufficiently long time period to capture sequences with a substantial amount of admissions. By looking at various years these sequences became better visible.

Choice of model: consequences for usage of H/SMR as indicator

We considered which of the two models would be more favourable to use and why. Since the model characteristics with regard to discrimination, calibration and explanatory power all were in favour of model 2, we recommend using model 2, although this preference does not necessarily invalidate model 1. HSMR models in other countries may need a similar additional adjustment. The impact of readmissions on the current HSMR in the UK is not clear. The UK model adjusts for readmissions for emergency cases. However for non-emergency cases the model does not adjust for readmissions. Moreover, admissions with different primary diagnoses are not counted as readmissions as well. Furthermore the UK model is restricted to a maximum review period of one year, which is too short for the readmission effect to become visible as demonstrated in our study.

A more fundamental issue than the choice of model however, is the fact that two comparable models, such as model 1 and model 2, may deliver such divergent outcomes. Since there is no 'gold standard' one cannot state that the one is false and the other one is true. This finding reveals an uncertainty in HSMR outcomes that exceeds the amount of uncertainty introduced by chance expressed in 95% confidence intervals over the five-year period.

HSMRs and SMRs are increasingly used by physicians to improve the quality of care in their hospitals. In particular diagnostic groups with higher than expected mortality are the subjects of investigation. However, in a lot of cases, the designated significance of these groups may depend on the regression model used, not only on the confidence intervals. The same is applicable for the current HSMR scores in the Netherlands. Physicians and hospital managers should be aware of this phenomenon.

Limitations of the study

A limitation of this study was the fact that we had to exclude 19 hospitals and so the HMSR calculation was based on roughly 80% of the Dutch hospitals. Nevertheless, we think the results of our study were representative of the Netherlands and have clearly demonstrated the impact upon the HSMR of not adjusting for frequency of readmissions.

Another limitation was the fact that the Dutch HSMR only accounts for in-hospital mortality and not for 30-day mortality, which may favour hospitals with shorter lengths of stay [14]. However since the effects, demonstrated in our study, particularly concerned the phenomenon of the frequently returning patient, and so no 30-days deaths, we think this potential bias does not significantly change our conclusions.

Conclusions

The HSMR model for the Netherlands was sensitive to adjustment for frequency of readmissions. Model 1, without adjustment, compared to model 2, with adjustment, produced substantially different HSMR outcomes. The uncertainty introduced by these differences exceeded the uncertainty indicated by the 95% confidence intervals. Since model 2 turned out to be more accurate compared to model 1 with regard to c-statistics, goodness of fit and explanatory power, the authors would prefer to apply adjustment for frequency of readmission. Parties, also in other countries, making use of HSMR-calculations may decide for themselves whether or not it is clinically relevant in their situation to include frequent admissions in the model. In case of adjustment, a review period of at least three years is advisable.

Declarations

Authors’ Affiliations

(1)
St. Antonius Hospital
(2)
NIVEL, the Netherlands Institute for Health Services Research
(3)
VU University Medical Centre

References

  1. Jarman B, Pieter D, van der Veen AA, et al: The hospital standardised mortality ratio: a powerful tool for Dutch hospitals to assess their quality of care?. Qual Saf Health Care. 2010, 19: 9-13.PubMed CentralView ArticlePubMed
  2. The 2010 Dr Foster Hospital Guide. (accessed 12 May 2011), [http://​www.​drfosterhealth.​co.​uk]
  3. Lilford R, Pronovost P: Using hospital mortality rates to judge hospital performance: a bad idea that just won't go away. BMJ. 2010, 340: c2016.View ArticlePubMed
  4. Black N: Assessing the quality of hospitals. BMJ. 2010, 340: c2066.View ArticlePubMed
  5. Penfold RB, et al: Do hospital standardised mortality ratios measure patient safety? HSMRs in the Winnipeg Regional Health Authority. Healthcare Papers. 2008, 8 (4): 8-23.View ArticlePubMed
  6. Van den Bosch WF, Roozendaal KJ, Silberbusch J, Wagner J: Variation in coding patient data impacts hospital standardised mortality ratios (HSMR). Ned Tijdschr Geneeskd. 2010, 154: A1189-in DutchPubMed
  7. Van den Bosch WF, Graafmans WC, Pieter D, Westert GP: Cardiac centres and the HSMR. The impact of special medical procedures upon the hospital standardised mortality ratio. Ned Tijdschr Geneeskd. 2008, in Dutch: 1221-7. in Dutch
  8. Mohammed MA, Deeks JJ, Girling A, Rudge G, Carmalt M, Stevens AJ, Lilford RJ: Evidence of methodological bias in hospital standardised mortality ratios: retrospective database study of English hospitals. BMJ. 2009, 338: b780.PubMed CentralView ArticlePubMed
  9. Lilford R, Mohammed MA, Spiegelhalter D, Thomson R: Use and misuse of process and outcome data in managing performance of acute medical care: avoiding institutional stigma. Lancet. 2004, 363: 1147-1154.View ArticlePubMed
  10. Van den Bosch WF, Kelder HC, Wagner C: Predicting hospital mortality among frequently readmitted patients: HSMR biased by readmission. BMC Health Serv Res. 2011, 11: 57.PubMed CentralView ArticlePubMed
  11. Bottle A, Jarman B, Aylin P: Hospital standardized mortality ratios: sensitivity analysis on the impact of coding. Heal Serv Res. doi:10.1111/j.1475-6773.2011.01295.x
  12. Clinical Classifications Software (CCS) for ICD-9-CM Fact Sheet. [http://​www.​hcup-us.​ahrq.​gov/​toolssoftware/​ccs/​ccsfactsheet.​jsp#what]
  13. Hosmer DW, Lemeshow S: Applied Logistic Regression. 2000, New York: Wiley, ISBN 0471615536View Article
  14. Drye EE, Normand SLT, Wang Y, et al: Comparison of hospital risk-standardized mortality rates calculated by using in-hospital and 30-day models: an observational study with implications for hospital profiling. Ann Intern Med. 2012, 156 (no. 1 Part 1): 19-26.PubMed CentralView ArticlePubMed
  15. Pre-publication history

    1. The pre-publication history for this paper can be accessed here:http://​www.​biomedcentral.​com/​1472-6963/​12/​91/​prepub

Copyright

© van den Bosch et al; licensee BioMed Central Ltd. 2012

This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://​creativecommons.​org/​licenses/​by/​2.​0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Advertisement