SEARCH

SEARCH BY CITATION

Keywords:

  • FEMORAL FRACTURE;
  • PATIENT DISCHARGE;
  • NERVOUS SYSTEM DISEASES;
  • RESPIRATORY TRACT DISEASES;
  • CARDIOVASCULAR DISEASES

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References
  10. Supporting Information

Previous studies found that the risk of a femoral fracture in residents newly admitted to nursing homes was highest during the first months after admission and declined thereafter. Many nursing home admissions are preceded by a hospitalization. Therefore, the present study aimed to analyze if a similar risk pattern of fall-related fractures could also be observed in community-dwelling people at home after discharge from the hospital. Routine data of more than 690,000 German people aged 65 years and older with more than 2 million hospital discharges were used to calculate fracture rates in the first 6 months after hospitalization, for people discharged to live in the community. Incidence rates of femoral fractures as a function of time since discharge from hospital were analyzed. Analyses were stratified by sex, age, the need for care, and diagnostic groups. For femoral fractures the incidence was highest during the first months after discharge and declined thereafter. This pattern was observed in women and men, in different age-groups, in different diagnostic groups, and in people with and without the need for care. For example, rates for femoral fractures in women declined from 17.4 to 11.0 per 1000 person years over the first 6 months after admission, and in men over the same time period from 8.2 to 4.5 per 1000 person-years, respectively. We conclude that the first weeks at home after discharge from the hospital are associated with an increased risk for femoral fractures. © 2013 American Society for Bone and Mineral Research.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References
  10. Supporting Information

Two previous studies demonstrated that the first time after admission to a nursing home is a high-risk situation for femoral fractures.1, 2 In these studies the fracture risk after admission was highest during the first months and declined thereafter. For example, femoral fracture risk during the first month was nearly twice as high as the risk 9 months later. A potential cause of the observed pattern may have been the new environment, which is a challenge to many of the new, and often cognitively impaired, residents. They are not used to the bedroom, the way to the toilet, or may have difficulties finding the light switch. These aspects may have been responsible for an increased risk of falling. In addition, nursing home admissions are frequently preceded by a hospitalization. A morbidity-related weakness with a deterioration of gait and balance, and a (subacute) delirium may have also contributed to an increased risk of fall-related fractures immediately after transition to the nursing care facility. The latter reasons, however, could be also present when community-dwelling people are discharged home after a hospitalization. This is of relevance because a discharge to the patient's home is much more frequent than a discharge to a nursing home.

Our a priori hypothesis was that a time-dependent risk pattern for femoral fractures after hospitalization can be observed in patients admitted from the community and discharged to the community and that this risk pattern is particularly pronounced in frail people and in people with nervous system diseases like dementia or Parkinson's disease. Therefore, we aimed: (1) to analyze the distribution of femoral fracture rates occurring at patients' homes as a function of time since discharge from hospital; (2) to perform the analyses separately in people with the need for care and without the need for care; and (3) to analyze if the pattern of risk was associated with specific discharge diagnostic groups.

For the analyses we used data for more than 690,000 community dwelling people with more than 2 million hospitalizations in Bavaria, a federal state of Germany.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References
  10. Supporting Information

Study population

The basic dataset consisted of all people aged 65 years and over, and insured with the Allgemeine Ortskrankenkasse Bayern (AOK Bavaria) between January 1, 2004 and June 30, 2009. Bavaria is a federal state with 12.5 million inhabitants in the south of Germany. Health insurance, including coverage for care at home or in a nursing home, is statutory in Germany. The AOK is Germany's largest statutory health insurance company. This nonprofit health insurance company covers nearly 50% of the Bavarian population aged 65 years and over.

Data source

We used the routine data collection systems of the health insurance company to gain data on gender, age, dates and main diagnoses of admission to and discharge from a hospital, status of the need for care (see Long-term care insurance and level of care, below), residency in a nursing home or date of admission to a nursing home, and, if appropriate, date of death for each individual. All data are held by the same health insurance company. Therefore, we did not have to link data from different sources.

Fractures

The objective of the study was to analyze femoral fractures occurring at patients' homes after discharge from hospital as a result of a nonfracture diagnosis. To identify fractures of the femur (S72 in the 10th revision of the International Classification of Diseases [ICD-10]), the information about (re)admission diagnoses to hospital after a prior hospitalization was used. Femoral fractures that were treated in outpatient clinics were not captured in the data base and are not included in the analyses.

Disease groups

Hospital discharge diagnoses were used to identify specific disease groups that were the cause of hospitalization, such as neoplasms (ICD-10: C00–D48), nervous system diseases (G00–G99), cardiovascular diseases (I00–I99), respiratory diseases (J00–J99), digestive system diseases (K00–K93), and genitourinary system diseases (N00–N99).

Long-term care insurance and level of care

In 1995, long-term care insurance was introduced in the German social insurance system and is compulsory for all citizens.3 In order to claim for long-term care benefit, people must need a minimum of 90 minutes of assistance with basic activities of daily living (ADL) such as washing, eating, or dressing, and of instrumental activities of daily living (IADL) such as cleaning or shopping, per day. Depending on the extent of care required, recipients are categorized into three levels after an assessment by a nurse or a physician of the medical service of the German statutory health insurance system. Long-term care benefits for community-dwelling persons with the need for care are either financial support for informal care performed by relatives or volunteers, or reimbursement for professional care. In this study, the need for care was defined as a categorization in one of the three levels of care, and fracture rates were calculated for people with and without the need for care, separately and combined.

Statistics

All persons insured by the AOK Bavaria, aged 65 years and over on January 1, 2004, and living in the community were eligible for the study cohort. Persons who attained the age of 65 years during the observation period were also eligible for the study cohort at the date of their 65th birthday. For older people in Germany, there are three main living arrangements: living at home (with or without professional care); assisted living in a residential facility (with or without professional care); and nursing home (with professional care). In our study community-living people were defined as all people not living in a nursing home. People residing in “assisted living” were included in the group of community-living people. The data structure of the long-term care insurance permitted differentiation between community-living people and residents of nursing homes. Analyses were restricted to patients with at least one hospitalization during the observation period who were admitted from the community and discharged to the community. A stay in a rehabilitation clinic was treated in the same way as an acute hospital stay. If a patient was transferred from an acute hospital directly to a rehabilitation clinic, follow-up started after discharge from the rehabilitation clinic. Person-years at risk were accumulated between the date of discharge from hospital or rehabilitation clinic and the end of the study (June 30, 2009), date of death, nursing home admission, or date of hospital admission due to a fracture or another reason.

If a person had several hospital stays during the observation period, each hospitalization was handled in the same way and as a separate case; ie, all hospitalizations of the participants were usually considered. However, participants with a femoral fracture as discharge diagnosis of their first hospitalization were excluded because it is known that readmissions with a related diagnosis to the index stay are increased during the first months.4, 5 A second reason was to avoid double coding of the same fracture (eg, readmission due to complications or transfers between hospitals). For the same reason the observation time of the participants was restricted up to a first femoral fracture and following hospitalizations were no longer considered. Hospitalizations with discharge diagnoses of fall-related fractures (ICD-10) other than femoral fractures (pelvis [S32 except fractures of the lumbar spine], lower leg [S82], humerus and shoulder [S42], forearm [S52], hand [S62], and skull [S02]) were also excluded in order to analyze femoral fracture rates after discharge from hospital restricted to diagnoses other than fall-related fractures. If the admission date due to a femoral fracture was identical with the discharge date due to another diagnosis, participants were excluded in order to avoid cases in which patients with a fracture during their hospitalization were transferred from one unit to another.

The crude incidence rate of femoral fractures was calculated by dividing the number of fractures by the total number of person-years. The rates are presented as fractures per 1000 person-years with 95% confidence intervals for each month after discharge. The follow-up period was restricted to the first 6 months after discharge because changes in fracture rates after this time were probably no longer caused by the preceding hospital stay. For the analyses stratified by disease groups at discharge, women and men and 2 months of follow-up at a time were combined in order to receive robust fracture rates. The analyses were further stratified by age groups (65 to <75 years; 75 to <85 years; and ≥85 years), and by persons with and without the need for care.

Because the hypothesized pattern of fracture rates may have been more pronounced in patients with longer hospital stays, a sensitivity analysis was performed that included only those hospitalizations with a length of at least 7 days.

All calculations were carried out with SAS version 9.2 (SAS Institute Inc., Cary, NC, USA).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References
  10. Supporting Information

A total of 6309 incident femoral fractures occurred at home within the first 6 months after 2,013,674 discharges from the hospital. The most frequent discharge diagnosis group was “cardiovascular diseases” (24.3%), followed by “neoplasms” (11.0%) and “digestive system diseases” (10.5%). At discharge about one-fifth of the study population was categorized as community-dwelling persons with the need for care (22.8% in women and 18.2% in men) (Table 1).

Table 1. Discharge Characteristics and Diagnoses
 WomenMen
  1. Q1 = quartile 1; Q3 = quartile 3.

Persons, n404,049288,298
Absolute number of discharges, n1,109,859903,815
Age at discharge, years, median (Q1–Q3)76.3 (70.8–82.1)73.5 (69.2–78.7)
Age at discharge, n (%)
 65 to <75 years488,122 (44.0)524,539 (58.0)
 75 to <85 years464,814 (41.9)314,850 (34.8)
 ≥85 years156,923 (14.1)64,426 (7.1)
Need for care at discharge, n (%)  
 Yes253,338 (22.8)164,381 (18.2)
 No856,521 (77.2)739,434 (81.8)
Discharge diagnosis, n (%)
 Neoplasms121,657 (11.0)139,365 (15.4)
 Nervous system diseases40,706 (3.7)38,473 (4.3)
 Cardiovascular diseases270,306 (24.3)235,037 (26.0)
 Respiratory diseases54,205 (4.9)66,045 (7.3)
 Digestive system diseases116,927 (10.5)94,176 (10.4)
 Genitourinary system diseases51,597 (4.6)54,317 (6.0)
 Others454,461 (41.0)276,402 (30.6)

For femoral fractures the incidence was highest during the first month after discharge and declined thereafter. This pattern was observed in women and men and in people with and without the need for care. For example, rates for femoral fractures in women declined from 17.4 to 11.0 per 1000 person years over the first 6 months after admission, and in men over the same time period from 8.2 to 4.5 per 1000 person-years, respectively. The relative risk reduction between the first month and the mean of the fifth and sixth month was about 30% and was similar in all subgroups (between 22% and 36%). Because the absolute femoral fracture rates were higher in women than in men, and higher in people with the need for care than in people without the need for care, the corresponding absolute decline of fracture rates was also considerably more pronounced in women and in people with the need for care. In people without the need for care the excess risk of a femoral fracture after hospitalization was limited to the first month after discharge, whereas in people with disabilities excess risk seemed to continue for a longer period (Table 2). In people with a hospital stay of at least 7 days femoral fracture rates were somewhat higher than in the total study population. In this sensitivity analysis an even more pronounced risk reduction was observed for women (from 20.7 to 12.2 per 1000 person-years) but not for men (Supplementary Table A).

Table 2. Incidence Rate of Femoral Fractures as a Function of Time Since Discharge From the Hospital in Community-Living People in Bavaria Between 2004 and 2009
Month after dischargeAllWithout the need for careWith the need for care
Number of fracturesPerson-yearsFractures per 1000 person-years (95% CI)Number of fracturesPerson-yearsFractures per 1000 person-years (95% CI)Number of fracturesPerson-yearsFractures per 1000 person-years (95% CI)
  1. CI = confidence interval.

Women and men
 11833137421.913.3 (12.7–14.0)948111560.58.5 (8.0–9.0)88525861.434.2 (32.0–36.5)
 21266111825.811.3 (10.7–11.9)63293165.46.8 (6.3–7.3)63418660.434.0 (31.3–36.6)
 398899048.410.0 (9.4–10.6)55183754.96.6 (6.0–7.1)43715293.428.6 (25.9–31.3)
 484389670.69.4 (8.8–10.0)48176587.36.3 (5.7–6.8)36213083.727.7 (24.8–30.5)
 574882200.39.1 (8.5–9.8)43270766.46.1 (5.5–6.7)31611433.927.6 (24.6–30.7)
 663175902.88.3 (7.7–9.0)39965776.66.1 (5.5–6.7)23210126.222.9 (20.0–25.9)
Women
 1133476850.817.4 (16.4–18.3)68560865.811.3 (10.4–12.1)64915985.040.6 (37.5–43.7)
 293863526.414.8 (13.8–15.7)47151807.59.1 (8.3–9.9)46711718.939.9 (36.2–43.5)
 373556743.613.0 (12.0–13.9)40547030.08.6 (7.8–9.5)3309713.634.0 (30.3–37.6)
 463151668.612.2 (11.3–13.2)36643278.58.5 (7.6–9.3)2658390.131.6 (27.8–35.4)
 553047578.111.1 (10.2–12.1)30440191.87.6 (6.7–8.4)2267386.430.6 (26.6–34.6)
 648744117.211.04 (10.1–12.0)30837529.18.2 (7.3–9.1)1796588.127.2 (23.2–31.2)
Men
 149960571.08.2 (7.5–9.0)26350694.75.2 (4.6–5.8)2369876.423.9 (20.9–26.9)
 232848299.56.8 (6.1–7.5)16141357.93.9 (3.3–4.5)1676941.524.1 (20.4–27.7)
 325342304.86.0 (5.2–6.7)14636724.94.0 (3.3–4.6)1075579.919.2 (15.5–22.8)
 421238002.05.6 (4.8–6.3)11533308.83.5 (2.8–4.1)974693.220.7 (16.6–24.8)
 521834622.16.3 (5.5–7.1)12830574.64.2 (3.5–4.9)904047.522.2 (17.6–26.8)
 614431785.64.5 (3.8–5.3)9128247.53.22 (2.6–3.9)533538.115.0 (11.0–19.0)

A similar pattern with highest fracture rates just after discharge from hospital was found when different age groups were analyzed separately (Fig. 1, Supplementary Table B).

thumbnail image

Figure 1. Incidence rate of femoral fractures as a function of time since discharge from the hospital in community-living people at the age of 65 to <75 years, 75 to <85 years, and ≥85 years.

Download figure to PowerPoint

Femoral fracture rates were also analyzed stratified by specific discharge diagnostic groups. For nervous system diseases and respiratory diseases the pattern in the previous paragraphs was particularly pronounced and consistently found both in people with and without the need for care (Fig. 2; Supplementary Table C).

thumbnail image

Figure 2. Incidence rate of femoral fractures as a function of time since discharge from the hospital, stratified by diagnostic groups at discharge in community-living people with and without the need for care.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References
  10. Supporting Information

We found an increased risk for femoral fractures during the first few months after discharge from hospital. This pattern was present in women and men, in different age groups, in people with and without the need for care, and for most discharge diagnoses. In community-living people without the need for care, excess risk was limited to the first month after discharge whereas the excess risk in people with the need for care declined gradually and resembled the risk pattern of residents newly admitted to a nursing home.1, 2 To our knowledge, there exists no data so far about fracture incidence as a function of time after discharge from hospital.

Potential causes of the observed pattern

The interpretation of our findings remains speculative. However, it seems to be plausible that a morbidity-related weakness after hospitalization may contribute to an increased risk of falls and fall-related femoral fractures immediately after transition. In addition, some of the patients may have had delirium during hospitalization, which is a common and serious problem in older adults with an acute illness or after an operation. In some cases, symptoms of delirium may last for months or even for years and can therefore also operate during the time after hospitalization.6 In the clinical setting, delirium is a strong risk factor for falls7, 8 and this may be also the case during the first weeks at home.

People with diseases of the nervous system such as dementia or Parkinson's disease are particularly vulnerable to delirium6 and falls. In our analyses the risk pattern was particularly pronounced in patients discharged with nervous system diseases and respiratory diseases. In these disease groups the most frequent diagnoses in our data were transient ischemic attacks, sleep disorders, epilepsy, extrapyramidal and movement disorders such as Parkinson's disease, and other degenerative diseases of the nervous system such as dementia, and influenza and pneumonia or other acute lower respiratory infections such as acute bronchitis for respiratory diseases.

Clinical relevance

Compared with other reasons for (re)admission to hospital, such as infections or heart failure, femoral fractures are a relatively rare event.5 The magnitude merely of the excess risk during the first month in all women or in all men aged 65 years and older, however, was comparable to the average risk of a 75-year-old woman or man in the German population.9 In absolute numbers, about 83 and 35 additional femoral fractures per 100,000 discharges can be expected during the first 2 months in women and men, respectively. In high-risk groups such as women and men aged 85 years and older, the additional number of fractures increases to about 200 and 100 femoral fractures per 100,000 discharges, respectively.

The major strengths of the study are its large number of hospital discharges and incident femoral fractures, the exact documentation of the time under risk, and the information about discharge diagnoses and the need for care at an individual level.

Several limitations have to be considered. Misclassification in the coding of fractures cannot be ruled out. For femoral fractures, however, the case ascertainment is systematic and the outcome misclassification should be low because nearly all femoral fractures are brought to the hospital. Despite the large number of discharges we had to combine diseases to diagnosis groups, women and men and 2 months of follow-up at a time in order to obtain sufficient fractures in the categories of the subgroups. Furthermore, the data did not allow us to detect operations during the index hospitalization. Some of the patients may have had an episode of “short-term care” in a nursing home for up to 4 weeks after discharge from hospital, which may have been an additional risk factor for falls and fractures due to the unknown environment. The data structure did not allow to identify these people. Therefore, a misclassification of femoral fractures in people having “short-term care” in a nursing home may have contributed to the observed risk pattern. However, in Germany, it is very unusual to discharge patients to a nursing home after hospitalization while their mobility and self care ability is improving prior to a return to the community. We defined the need for care as being categorized in one of the three levels of care. This definition may be arbitrary but the criteria are clear, the assessments are done by experts, and the method has been shown to have good levels of interrater reliability.10 Our data were derived from only one health insurance company and may not be representative of the whole German population. About 90% of the German population is a member of one of the statutory health care insurances and about 10% are privately insured. The AOK is by far the largest statutory health care insurance company and covers nearly 50% of the Bavarian population aged 65 years and older. The AOK is open to all people. Compared with people insured by other statutory or private insurance companies, persons insured at the AOK represent lower rather than higher socioeconomic levels. This may have influenced the absolute number of fractures because femoral fractures have been shown to be associated with person's income or socioeconomic status.11, 12 It cannot be excluded that the patients' socioeconomic status had also an influence on the observed risk pattern, for example by a difference in the availability of social support at home.

The biological mechanisms hypothesized to be the reason for the observed pattern of femoral fractures probably apply in all countries. However, other factors that are different in different countries and different health care systems may influence the pattern to some extent. Examples are the average length of hospital stay, the frequency of an episode of “short-term care” after hospitalization, the availability of a comprehensive home care service system, and the extent of social support by the family. The direction in which these factors may change the observed pattern remains speculative. A shorter hospital stay, for example, reduces the time of immobilization but may increase the risk of a delirious state being present after discharge.

Implications for practice

The best evidence to reduce falls and possibly fractures in community-dwelling people has been shown for strength and balance training.13 Medical treatment of osteoporosis has been shown to reduce femoral fractures.14 But it seems plausible that both approaches are of minor benefit during the first weeks after discharge from the hospital. During this period, an optimization of contextual factors which are part of a good discharge planning may play a more important role. There is not sufficient evidence to recommend hip protectors in community-living people.15 In persons with a high basis risk such as people with the need for care, however, hip protectors may be a useful measure at least during the first months after discharge from the hospital.

In summary, we found that the first weeks after discharge from hospital were associated with an increased risk for femoral fractures.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References
  10. Supporting Information

The analysis was supported by the Bundesministerium für Bildung und Forschung (Förderkennzeichen: 01EC1007A). We thank Regina Merk-Bäuml, Ralf Brum, and Stefanie Büttner from the Allgemeine Ortskrankenkasse (AOK) for their support of our analyses.

Authors' roles: KR contributed to the evaluation, analysis, and interpretation of data, and drafting the article. IDC, ME, and HHK contributed to the interpretation of data and revising the article. CB contributed to the evaluation and interpretation of data, and revising the article. AK and JK contributed to the analysis and interpretation of data.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References
  10. Supporting Information
  • 1
    Rapp K, Becker C, Lamb SE, Icks A, Klenk J. Hip fractures in institutionalized elderly people: incidence rates and excess mortality. J Bone Miner Res. 2008;23:182531.
  • 2
    Rapp K, Lamb SE, Klenk J, Kleiner A, Heinrich S, Konig HH, Nikolaus T, Becker C. Fractures after nursing home admission: incidence and potential consequences. Osteoporos Int. 2009;20:177583.
  • 3
    Becker C, Leistner K, Nikolaus T. Introducing a statutory insurance system for long-term care (Pflegeversicherung) in Germany In: Michel JP, Rubenstein LZ, Vellas BJ, Albarede JL, editors Geriatric programs and departments around the world. Paris-New York: Serd-Springer; p 5564. 1998.
  • 4
    Teixeira A, Trinquart L, Raphael M, Bastianic T, Chatellier G, Holstein J. Outcomes in older patients after surgical treatment for hip fracture: a new approach to characterise the link between readmissions and the surgical stay. Age Ageing. 2009;38:5849.
  • 5
    Epstein AM, Jha AK, Orav EJ. The relationship between hospital admission rates and rehospitalizations. N Engl J Med. 2011;365:228795.
  • 6
    Inouye SK. Delirium in older persons. N Engl J Med. 2006;354:115765.
  • 7
    Corsinovi L, Bo M, Ricauda AN, Marinello R, Gariglio F, Marchetto C, Gastaldi L, Fissore L, Zanocchi M, Molaschi M. Predictors of falls and hospitalization outcomes in elderly patients admitted to an acute geriatric unit. Arch Gerontol Geriatr. 2009;49:1425.
  • 8
    Stenvall M, Olofsson B, Lundstrom M, Svensson O, Nyberg L, Gustafson Y. Inpatient falls and injuries in older patients treated for femoral neck fracture. Arch Gerontol Geriatr. 2006;43:38999.
  • 9
    Icks A, Haastert B, Wildner M, Becker C, Meyer G. Trend of hip fracture incidence in Germany 1995–2004 : a population-based study. Osteoporos Int. 2008;19:113945.
  • 10
    Kliebsch U, Brenner H. [Interrater reliability of expert assessment by the health insurance medical service in determining eligibility for disability benefits]. Gesundheitswesen. 1995;57:63844. German.
  • 11
    Zingmond DS, Soohoo NF, Silverman SL. The role of socioeconomic status on hip fracture. Osteoporos Int. 2006;17:15628.
  • 12
    Quah C, Boulton C, Moran C. The influence of socioeconomic status on the incidence, outcome and mortality of fractures of the hip. J Bone Joint Surg Br. 2011;93:8015.
  • 13
    Gillespie LD, Robertson MC, Gillespie WJ, Lamb SE, Gates S, Cumming RG, Rowe BH. Interventions for preventing falls in older people living in the community. Cochrane Database Syst Rev. 2009; C D007146.
  • 14
    Reginster JY. Antifracture efficacy of currently available therapies for postmenopausal osteoporosis. Drugs. 2011;71:6578.
  • 15
    Gillespie WJ, Gillespie LD, Parker MJ. Hip protectors for preventing hip fractures in older people. Cochrane Database Syst Rev. 2010; C D001255.

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References
  10. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
jbmr_1809_sm_SupplTabs.doc155KSupplementary Tables

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.