Length of hospital stay after hip fracture and short term risk of death after discharge: a total cohort study in Sweden
BMJ 2015; 350 doi: https://doi.org/10.1136/bmj.h696 (Published 20 February 2015) Cite this as: BMJ 2015;350:h696All rapid responses
Rapid responses are electronic comments to the editor. They enable our users to debate issues raised in articles published on bmj.com. A rapid response is first posted online. If you need the URL (web address) of an individual response, simply click on the response headline and copy the URL from the browser window. A proportion of responses will, after editing, be published online and in the print journal as letters, which are indexed in PubMed. Rapid responses are not indexed in PubMed and they are not journal articles. The BMJ reserves the right to remove responses which are being wilfully misrepresented as published articles or when it is brought to our attention that a response spreads misinformation.
From March 2022, the word limit for rapid responses will be 600 words not including references and author details. We will no longer post responses that exceed this limit.
The word limit for letters selected from posted responses remains 300 words.
Dear Editor,
We read with interest the paper by Nordström P et al., and we believe that some considerations may help the interpretation of the results from abroad, and maybe a better dissemination of the results [1].
First, length of in-hospital stay (LOS) is largely dependent on the features of the Healthcare System, and often related to local organizational factors. Considering hip fracture, post-acute pathways differ significantly in the different Countries mainly due to a dissimilar rationalization of the resources.
In general, there is an inverse relationship between LOS and rate of transfer to rehabilitation services in the community. The models characterized by very short in-hospital stay, where the patients undergo early discharge and community rehabilitation, are supported by the presence of adequate post-discharge rehabilitation services, able to take care of the patients. This situation is depicted by the services implemented in the USA, where, the LOS has decreased dramatically to a national average of 5-6 days, but more than 80% of hip fractured subjects is discharged to inpatient rehabilitation or community skilled nursing facilities to continue rehabilitation [2]. The opposed scenario is typically represented by the UK services, where patients undertake a longer rehabilitation during in-hospital stay. These services were set up to commence and continue the early phase of rehabilitation into the hospital. Therefore, LOS usually exceed the 20 days, most of the patients complete functional recovery during in-hospital stay, and only a small proportion of them (less than 30%) needs to be transferred to other community services for further rehabilitation [3].
In the middle of these two scenarios, there are the majority of European Countries with LOS between 10 and 15 days. In these European models, usually, the rehabilitation is broken down in two phases: early rehabilitation that occurs during in-hospital stay, and late rehabilitation that takes place after the discharge. The proportion of hip fracture patients discharged to community rehabilitation facilities is quite variable, depending on the availability of these services in the community and by the local organization of the Healthcare System.
Second, almost all developed Countries implemented, in the last two decades, strategies to reduce in-hospital stay in order to reduce costs associated to chronic diseases and to improve resources re-distribution and utilization. However, for acute conditions, such as hip fracture, the pathway of care does not terminate after the acute phase, as most of the patients require a post-acute phase for clinical stabilization and rehabilitation. Therefore, the strategies to reduce length of in-hospital stay in hip fracture older adults cannot simply be founded on the optimization of the clinical and surgical management of the patient in the acute phase, but should be based on a comprehensive re-organization of the Healthcare System, planning the definition/implementation of post-acute clinical pathways able to guarantee a continuity of care from discharge to final return to pre-fracture living situation. In this context, it is well established that different post-acute care settings may differ in long-term outcomes [4,5].
Finally, the LOS should not be regarded as the sole variable influencing clinical outcomes, since it is the overall pathway of care that determines the final result. It is actually absolutely important to consider all the steps of the care pathway (including the post-acute discharge allocation) in defining the relation between LOS and mortality.
Recent data, supporting this last point, have demonstrated that while implementing an Orthogeriatric Care pathway is absolutely necessarily that all those interventions and strategies that have been shown to improve short- and long-term outcomes became part of it [6]. Several steps of the care pathway, such as early surgery, immediate weight bearing, prevention of common geriatric syndromes, discharge planning, may affect the final result, and a service with several efficient phases, but even a single ineffective phase, could really penalize the final outcome.
In conclusion, we feel that the results by Nordström P et al. are extremely interesting, since they support the concept that while implementing strategies to reduce LOS in hip fracture older adults, the overall care pathway should be re-organized in order to face the complex needs of frail older adults that does not terminate with the acute phase. Particularly, the Healthcare System should guarantee subsequent and integrated phases characterized by different levels of intensity, ensuring the continuity of care, to produce acceptable clinical and rehabilitative results [7].
We are wondering whether a shorter length of stay could also affect negatively functional recovery, having the same impact that Authors have clearly demonstrated on survival.
References
1. Nordström P, Gustafson Y, Michaëlsson K, Nordström A. Length of hospital stay after hip fracture and short term risk of death after discharge: a total cohort study in Sweden. BMJ 2015;350:h696.
2. Gehlbach SH, Avrunin JS, Puleo E. Trends in hospital care for hip fractures. Osteoporos Int 2007;18:585-91.
3. The National Hip Fracture Database. National Report 2012. Available at: www.nhfd.co.uk. Accessed October 1, 2014.
4. Bachmann S, Finger C, Huss A, Egger M, Stuck AE, Clough-Gorr KM. Inpatient rehabilitation specifically designed for geriatric patients: systematic review and meta-analysis of randomized controlled trials. BMJ 2010; 340:c1718.
5. Mallinson T, Deutsch A, Bateman J, Tseng HY, Manheim L, Almagor O, Heinemann AW. Comparison of discharge functional status after rehabilitation in skilled nursing, home health, and medical rehabilitation settings for patients after hip fracture repair. Arch Phys Med Rehabil 2014;95:209-17.
6. Pioli G, Frondini C, Lauretani F, Davoli ML, Pellicciotti F, Martini E, Zagatti A, Giordano A, Pedriali I, Nardelli A, Zurlo A, Ferrari A, Lunardelli ML. Time to surgery and rehabilitation resources affect outcomes in orthogeriatric units. Arch Gerontol Geriatr 2012;55:316-22.
7. Beaupre LA, Binder EF, Cameron ID, Jones CA, Orwig D, Sherrington C, Magaziner J. Maximising functional recovery following hip fracture in frail seniors. Best Pract Res Clin Rheumatol 2013;27:771-88.
Competing interests: No competing interests
The recent paper by Nordström et al. [1] appeared to demonstrate increased risk of death in people discharged from hospital within 10 days of a hip fracture. However, we would echo Martyn Parker's words of caution [2]; it is potentially dangerous to draw conclusions from even the most elegant statistics, unless these interpreted in a clinical context.
The National Hip Fracture Database (www.nhfd.co.uk) is a clinically led quality improvement programme. Originally developed by the British Orthopaedic Association and British Geriatrics Society, it is now commissioned as a national clinical audit by the Healthcare Quality Improvement Partnership and managed by the Royal College of Physicians. It collates clinical information on all over 60 year olds presenting with hip fracture in England, Wales and Northern Ireland - using these to monitor outcomes and performance, including those aspects of care central to Best Practice Tariff in England.
In 2013 we recorded data on 65,535 people. As in the Swedish study we recorded a higher risk of death within 30 days of fracture for those who were discharged before 10 days. There appeared to be 104 'excess deaths' in this group, when compared with patients who remained in hospital for more than 10 days.
However, this pattern reflects a number of confounding factors which we addressed by examining mortality among people who were admitted from their own home and who were successfully returned there. People discharged home before 10 days actually showed lower mortality than those who remained in hospital for longer (figure).
The appearance of 'excess deaths' among people discharged before 10 days was entirely accounted for by other patient subgroups.
A third occurred in people who were admitted from their own home and transferred to care in an acute hospital, rehabilitation unit or hospice. Half occurred in the frail group of people who were admitted from care homes and who had returned there within 10 days. The remainder occurred among people admitted from their own home who were discharged to care homes (table).
Patients discharged to care homes before 10 days were at increased risk of dying, but this is a frail and complex group of individuals. The small absolute numbers of such deaths within a population of over 65,000 would not justify our cautioning against allowing people to return to their care home when the patient, their family and the Hip Fracture Programme's multidisciplinary team agree this is appropriate.
NHFD data for the health and social care system in England, Wales and Northern Ireland suggest that it is safe for patients to return to their own home within 10 days of hip fracture.
References
1. Nordström P, Gustafson Y, Michaëlsson K, Nordström A. Length of hospital stay after hip fracture and short term risk of death after discharge: a total cohort study in Sweden. BMJ 2015;350:h696
2. Parker M. BMJ. Rapid response, 28th February 2015
Competing interests: No competing interests
FAO: JK Anand.
1. The surgeon.
2. We try and instil confidence in the patient such that they will manage well at home. A lot of our patients are very keen to be at home as soon as possible. Discharge planning is often a joint decision. If a patient was very distressed about going home they would stay a few more days.
3. Yes.
4. Rarely. Our readmission data is better than the national average.
5.No.
Thank you for your interest. If you have any more queries please don't hesitate to ask.
Competing interests: No competing interests
We are grateful for the responses with respect to our article "Length of hospital stay after hip fracture and short-term risk of death after discharge".1 With this letter we would like to respond to some of these comments. In the paper we admitted several limitations, and some have additionally been commented. In the linked editorial,2 the authors suggests that the association found between LOS and risk of death could have been influenced by survivor bias. Accordingly, since the highest rates of death occur early after a hip fracture, the risk of death would automatically increase after discharge if LOS decreases. However, this possibility was actually tested and the results presented in our article. Thus, in a sensitivity analysis we evaluated the association between LOS and the risk of death within 11-30 days of admission for patients with a LOS of 10 days or fewer, i.e., all patients had survived to at least 10 days after admission. The significant association between death and shorter LOS also in this model confirmed the main results of our study.
A limitation in our study was the lack of proper evaluation of the underlying cause of death, as acknowledged by us and commented by Anne Mwirigi, and perhaps especially any deaths attributable to pulmonary embolism. Thus, previous studies have shown that pulmonary embolism (PE) is a common cause of death following a hip fracture.3 Prevention includes early mobilization and use of low-molecular weight heparin or drugs with similar effects.4 5 Notably, complications such as PE may be of special interest with respect to early discharge, as the responsibility for mobilisation, rehabilitation and providing adequate thromboprophylaxic drugs are shifted to non-hospital care providers. Irrespectively, a proper evaluation of specific causes of death would clearly provide important information useful for implementing measures with the aim to reduce the risk of complications and risk of death after hip fracture.
Louise Southern writes in her response that we "appear to make the assumption that patients who are likely to die should necessarily be cared for in hospital." This is not correct. In our article we conclude that shorter length of stay in hospital (LOS) after a hip fracture is associated with a progressively increased risk of death for those with a LOS of 10 days or less. We also speculate whether this association could reflect a situation where patients are discharged early to rehabilitation in community settings, which results in patients' exposure to fewer care providers with adequate education. Unfortunately, we could not test this hypothesis in the present study as we did not have access to discharge location. This would be a very important topic for further research as pointed out in another comment by Martin J Parker.
Louise Southern also suggests that institutions providing opportunities for palliative care should not be penalized because of high post-discharge mortality rates. We agree, and to the best of our knowledge there is no such systems in Europe or in the rest of the world. In contrast, we have economic systems in several European countries and the United States that favour a short LOS in hospital and subsequent early discharge.6 Preferentially, LOS should primarily be adapted to the patient's medical conditions after careful and experienced clinical decision making.
Peter Nordström, Karl Michaelsson, Yngve Gustafson, and Anna Nordström
1. Nordstrom P, Gustafson Y, Michaelsson K, Nordstrom A. Length of hospital stay after hip fracture and short term risk of death after discharge: a total cohort study in Sweden. Bmj 2015;350:h696.
2. Cram P, Rush RP. Length of hospital stay after hip fracture. Bmj 2015;350:h823.
3. Perez JV, Warwick DJ, Case CP, Bannister GC. Death after proximal femoral fracture--an autopsy study. Injury 1995;26(4):237-40.
4. Handoll HH, Farrar MJ, McBirnie J, Tytherleigh-Strong G, Milne AA, Gillespie WJ. Heparin, low molecular weight heparin and physical methods for preventing deep vein thrombosis and pulmonary embolism following surgery for hip fractures. Cochrane Database Syst Rev 2002(4):CD000305.
5. Eriksson BI, Bauer KA, Lassen MR, Turpie AG, Steering Committee of the Pentasaccharide in Hip-Fracture Surgery S. Fondaparinux compared with enoxaparin for the prevention of venous thromboembolism after hip-fracture surgery. N Engl J Med 2001;345(18):1298-304.
6. Burwell SM. Setting value-based payment goals--HHS efforts to improve U.S. health care. N Engl J Med 2015;372(10):897-9.
Competing interests: No competing interests
Re: the Manchester analysis
Could Mr Sharrock and colleagues please also tell us:
1. Who decided that the patient should leave the unit? The surgeon in charge? The GP named as the patient's GP?
2. Did the patient have the final say about his departure date?
3. Did the patient's next of kin have any say?
4. How often were the patients discharged from the ward, only to be readmitted or to be admitted to say, the geriatric wards?
5. Did the funding bodies exert any pressure, overtly or covertly, to get the patient out?
Thank you
Competing interests: Old chap. Might need such care one day. Was involved in a Hospital at Home scheme which commenced in 1978, with a very explicit, supreme veto on admission and discharge in the hands of the patient/next of kin. I do not know what happens there these days.
We have analysed our fractured neck of femur data collected between September 2008 and December 2014. Data was initially collected for a local audit but later for National Hip Fracture Database (NHFD) requirements. We sub-analysed our data because of concern that patients admitted from some Clinical Commissioning Groups took much longer to be discharged than patients from elsewhere. The aims of the audit were to discover whether the concern was valid, and if so, what the reasons may be. Details of 1813 patients were available. The audit found that the mean, length of stay of patients from one CCG was 3 days more than the others. Few reasons for this were found in patient demographics including age, mental and physical state and promptness of surgery. Total extra bed days for the 660 CCG A patients were 1980 days over the 6.25 years. Our mean length of stay is poor at 27 days but we get 67% of patient discharged to their home.
A simple analysis of NHFD data that there is a disparity between length of stay and successful home discharge, it can be seen from the graph in 2013's and 2014's report that the top 10 performing hospitals in terms of length of stay only achieved a 29% home discharge rate. We note that the report considered it an improvement that the mean had shifted between 46.4 to 48% between 2013 and 2014 and a good rate of home discharge is commended. We noted that our worst performing CCG in terms of days in hospital did have more patients discharged to their own homes and for this reason should find some congratulations. We do however note that their excess stays of around 3 days has lead to 1980 extra bed days over the 6.25 years.
National Reports 2013/14 (2014). London: National Hip Fracture Database. Available at: http://www.nhfd.co.uk
Competing interests: No competing interests
To our knowledge, length of stay and risk of death are outcome measurements of quality of hospital care, in which better quality is linked to shorter length of stay and lower risk of death. In this study, shorter length of hospital stay after hip fracture is related to higher risk of death among patients with a length of stay of 10 days and less [1]. Another study about hip fracture patients found that lower quality of hospital care is associated with longer length of hospital stay, and higher mortality rates [2]. The issue raised is how to measure quality of care outcome when these two measurements are here and there not consistent.
For future research, besides considering diagnoses and type of operation, investigators should take into account the hospital volume or hip fracture unit volume as one of the important factors in analysis. In addition, it is necessary to determine the minimum length of stay to guarantee the quality of care and reduce the risk of death.
References
[1] Nordström P, Gustafson Y, Michaëlsson K, Nordström A. Length of hospital stay after hip fracture and short term risk of death after discharge: a total cohort study in Sweden. BMJ 2015;350:h696.
[2] Kristensen PK, Thillemann TM, Johnsen SP. Is bigger always better? A nationwide study of hip fracture unit volume, 30-day mortality, quality of in-hospital care, and length of hospital stay. Medical care 2014;52:1023-9.
Competing interests: No competing interests
We read with interest the cohort study by Nordstrom and Gustafson1, reporting that in a cohort of Swedish patients between 2006-12, a short hospital stay following hip fracture surgery was associated with increased mortality. Patients with a length of stay of 1-5 days had twice the mortality at 30 days compared with >15 days admission (OR 1.97). Lack of evaluation of the underlying cause of death was cited as a major limitation of the study. The authors stated that the cause of death was ‘exposure to non-specific factor-home’ (ICD-10 code X590) in 21% of cases, suggesting an unknown cause of death in these patients. The other causes of death reported were cardiovascular disease (13%), cancer (10%), dementia (7%), falls (6%) and stroke (4%).
The lack of information on the incidence and duration of venous thromboprophylaxis, or any deaths attributable to VTE is striking. Thus an important preventable cause of mortality was missed, for which a simple intervention could potentially change the reporting and outcome of this narrative. For the rate of deep venous thromboembolism (symptomatic and asymptomatic) is approximately 40% following proximal femoral fracture, and that of symptomatic pulmonary embolism (PE) is 6%. PE has been reported as a fourth most common cause of death following hip fracture, with one autopsy study of 581 cases reporting it as the cause of death in 14% of cases 2.Hospital-associated VTE can occur up to 90 days post discharge, with the highest risk in the first 20 days post-operatively3. And yet appropriate use of thromboprophylaxis reduces the risk of hospital acquired VTE after hip surgery by 60-70% 4.
In 2009 NHS England started a thromboprophylaxis improvement programme which had financial penalties to ensure 90% (later increased to 95%) of all adult hospital admissions had VTE risk assessment and appropriate thromboprophylaxis in line with NICE Clinical Guideline 144. Since this was launched there has been an 8% reduction PE deaths5,6. In a subgroup analysis of surgical patients, primary VTE deaths within 90 days of discharge were significantly reduced after introduction of thromboprophylaxis 6.
The lack of appreciation of the global scale of VTE is an international phenomenon. Amazingly the WHO monitoring of the global disease burden has no section at all on VTE. The ISTH World Thrombosis Day steering committee team have shown that VTE causes a major burden of disease across low-income, middle income and high income countries, with about 10 million hospital-acquired VTE occurring every year7. and we are campaigning to ensure in future the global disease burden of VTE is monitored.
1. Nordstrom, P, Gustafson, Y, et al. Length of hospital stay after hip fracture and short-term risk of death after discharge: a total cohort in Sweden. BMJ 2015: 2015; 350:h696
2. Perez, JV, Warwick, DJ, Case, CP, et al Death after proximal femoral fracture: an autopsy study. Injury. 1995; 26, 237-240
3. White, RH, Romano, PS, Zhou, H, et al Incidence and time course of thromboembolic outcomes following total hip or knee arthroplasty.Arch Intern Med1998;158,1525-1531.
4. Geerts W H, Bergqvist D et al; American College of Chest Physicians. Prevention of venous thromboembolism: American College of Chest Physicians Evidence-Based Clinical Practice Guidelines (8th Edition). Chest 2008;133(6 Suppl):381S-453S
5. Catterick E, Hunt BJ. Impact of national versus thromboembolism risk assessment tool in secondary care in England: retrospective population-based database study. Blood Coagulation and Fibrinolysis. 2014, 25
6. Lester W, Freemantle N, et al. Fatal venous thromboembolism associated with hospital admission: a cohort study to assess the impact of a national risk assessment target. Heart 2013; 0:1-6
7. ISTH commitee for World Thrombosis Day. Thrombosis a major contributor to the global disease burden. JTH 2014; 12: 1580-90
Competing interests: No competing interests
In an era of pressure to reduce length of stay both to avoid harm and to reduce healthcare costs, we should remain wary of the potential adverse effects of early discharge, and carefully consider the very useful data reported in this article. However, whilst the authors have acknowledged some of the limitations of drawing conclusions from their results, they have not acknowledged the many alternative explanations for the association between early discharge and death, and appear to make the assumption that patients who are likely to die should necessarily be cared for in hospital. Where death is a likely inevitable outcome of the acute illness, patients and their families should be given the option of palliative care in their home or nursing home. Indeed as this practice is itself a marker of high quality care (http://www.goldstandardsframework.org.uk/cd-content/uploads/files/Primar...), then it is important that institutions providing such opportunities are not penalised for their subsequently higher post-discharge mortality rates. The fact that the incidence of dementia was higher in the 'early discharge' population in this study indicates that this may have been a factor, as the mortality rate for hip fracture in patients with dementia is higher [55% at six months, compared to 12% for cognitively intact patients, JAMA. 2000 Jul 5;284(1):47-52.]. Furthermore, the clinical teams may have judged that some patients with advanced dementia were more likely to accept rehabilitation if nursed in a familiar environment, or in a less acute setting.
Let's not propagate the myth that all patients at the end of their life need to be in hospital; early expert assessment is required, after which it may be apparent that discharging the patient to their own environment is more appropriate, especially if intervention is felt likely not to change, or even worsen, the outcome. Therefore whilst it is vital that patients with preventable and treatable illnesses receive the acute care that they require, high post-discharge mortality rates do not necessarily equate to inappropriate discharge. In particular it would be interesting to investigate the effects of the place of discharge, whether a comprehensive geriatric assessment was performed pre-discharge, whether the deaths were 'expected', and whether palliative care interventions were delivered, before drawing conclusions. Patients require services that are well-designed and comprehensive; it is important that the evaluation of such services meets a similarly high standard.
Competing interests: No competing interests
Re: Length of hospital stay after hip fracture and short term risk of death after discharge: a total cohort study in Sweden
The authors note the fact that the highest rates of death occur early after a hip fracture. They also acknowledge that this fact could bias the association between length of stay and death after hospi-tal discharge. We examined the extent to which this potential source of bias could have affected the results.
A negative correlation between length of stay and early post-discharge mortality is induced by the fact that the time-frame of measurement is not fixed. This form of time-frame or “discharge” bias is well established in the literature (eg, Pouw et al, BMJ 2013 http://www.bmj.com/content/347/bmj.f5913; Wolkewitz et al J Clin Epidemiol 2012).
We used survival data on 5,000 hip fracture patients by randomly selecting records from the English Hospital Episode Statistics (HES) linked to Office for National Statistics (ONS) Mortality Data. We simulated values for length of hospital stay for each patient, independent of their survival. Using the simulated data, we calculated post-discharge mortality within 30 days.
The attached figure illustrates the problem of time-frame or discharge bias. If patients were dis-charged from hospital at 5 days, the post-discharge 30-day mortality would be 7.1% (344 / 4875). In contrast, the rate for those discharged at 10 days would be 6.5% (312 / 4805). The difference is purely down to the difference in time-frame.
We identified a significant spurious association between length of stay and post-discharge mortality. Our estimate was of a similar magnitude to the estimate presented in the study, i.e. for each 1 day decrease in length of stay, we observed a 2% increase in post-discharge mortality (odds ratio 1.023, 95% confidence interval 1.023 to 1.033, P value <0.0001).
The authors note that they identified a significant association between length of risk and in-hospital mortality. A positive association could be a sign that such a bias is present.
There are options for analysing the data without introducing bias, such as treating in-hospital deaths as a competing risk (Putter et al, Statist Med 2007).
Competing interests: No competing interests