Accepted Manuscript Title: Policy encouraging earlier hip fracture surgery can decrease the long-term mortality of elderly patients Author: Kobi Peleg Michael Rozenfeld Irina Radomislensky Ilya Novikov Laurence S. Freedman Avi Israeli PII: S0020-1383(14)00133-8 DOI: http://dx.doi.org/doi:10.1016/j.injury.2014.03.009 Reference: JINJ 5677 To appear in: Injury, Int. J. Care Injured Received date: 9-12-2013 Revised date: 6-2-2014 Accepted date: 14-3-2014 Please cite this article as: Peleg K, Rozenfeld M, Radomislensky I, Novikov I, Freedman LS, Israeli A, Policy encouraging earlier hip fracture surgery can decrease the long-term mortality of elderly patients, Injury (2014), http://dx.doi.org/10.1016/j.injury.2014.03.009 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Transcript
Accepted Manuscript
Title: Policy encouraging earlier hip fracture surgery candecrease the long-term mortality of elderly patients
Author: Kobi Peleg Michael Rozenfeld Irina RadomislenskyIlya Novikov Laurence S. Freedman Avi Israeli
Received date: 9-12-2013Revised date: 6-2-2014Accepted date: 14-3-2014
Please cite this article as: Peleg K, Rozenfeld M, Radomislensky I, NovikovI, Freedman LS, Israeli A, Policy encouraging earlier hip fracture surgerycan decrease the long-term mortality of elderly patients, Injury (2014),http://dx.doi.org/10.1016/j.injury.2014.03.009
This is a PDF file of an unedited manuscript that has been accepted for publication.As a service to our customers we are providing this early version of the manuscript.The manuscript will undergo copyediting, typesetting, and review of the resulting proofbefore it is published in its final form. Please note that during the production processerrors may be discovered which could affect the content, and all legal disclaimers thatapply to the journal pertain.
Address: National Center for Trauma and Emergency Medicine Research, Gertner
Institute, Tel-Hashomer, Israel, 52621.
Keywords: hip fracture; earlier surgery; DRG; long-term mortality.
Page 2 of 22
Accep
ted
Man
uscr
ipt
2
Policy encouraging earlier hip fracture surgery can decrease the long-term mortality of elderly patients
Abstract Background: In April 2004 the Israeli Ministry of Health decided to condition DRG
payment for hip surgery by time between hospitalization and operation, giving a fine
for every day’s delay beyond 48 hours. An evaluation study performed two years after
the reform has shown the positive influence of the reform on patient's survival in the
hospital. This study evaluates the impact of the reform on the longer-term mortality of
patients.
Methods: A retrospective study based on data from 9 hospitals of the national trauma
registry available for the years 2001-2007, with surveillance on two-year survival
through data of Ministry of the Interior. The study population includes patients aged
65 and above with an isolated hip fracture following trauma. Mortality curves and
Cox Regression were utilized to compare the influence of different parameters on
long-term mortality.
Results: Earlier surgery had a significant positive impact on survival through the
whole length of the study period. In the period after the introduction of the new
reimbursement system for hip fracture surgeries, a significant decrease in the longer-
term mortality was observed up to 6 months of follow-up, even when adjusted by
patients' age, gender and the receiving hospital. After 6 months there was no further
decrease in relative risk, though the survival advantage remained with patients
hospitalized after the reform.
Conclusions: The reform appears successful in decreasing the longer-term patient
mortality after hip fracture through influencing surgical practice.
Page 3 of 22
Accep
ted
Man
uscr
ipt
3
Introduction
Hip fractures are frequent in the elderly population and are associated with much
suffering, decrease in quality of life and increased mortality, while the cost to the
healthcare system is high1,2. According to many studies, patients who sustained a hip
fracture are at much higher risk of dying as compared to other representatives of their
age group1-3.
The recommended method of treating a hip fracture is either replacement or fixation
surgery. Most sources agree that surgery should be performed in the first 24 hours and
not later than 48 hours after hospitalization4-7. Many studies have shown that delay of
surgery can lead to increased morbidity, mortality and length of stay in the hospital4-7.
The growing awareness of the effects of delayed surgery of hip fractures on medical
outcomes brought a change in the system of reimbursement of Israeli hospitals for
these surgeries. In April 2004 the Director of the Israel Ministry of Health (IMH)
issued a directive to impose a differential pricing on hip fixations, conditioned by time
of surgery. The directive defined full DRG payment for hip fixations of all patients
with diagnosis of isolated hip fracture only if the surgery was performed in the first 48
hours of hospitalization. In cases of later surgery not justified by medical
considerations such as clear contraindications, each day of delay would further
subtract from the payment. This new policy was introduced for all Israeli hospitals
performing hip fracture surgeries, while demanding a consistent registration of
surgery times.
In 2007 the Israeli National Center for Trauma and Emergency Medicine Research
performed a study, sponsored by the National Institute for Health Services Research,
whose purpose was to measure the impact of this policy change on volume of delayed
surgeries and in-hospital mortality8. The study found an increase of 35% in volume of
operations in the first 48 hours of hospitalization and a decrease of one day in median
waiting time for hip fixations in the period after the change in the reimbursement
system compared to the period before it. In the later period, in-hospital mortality
decreased by more than 30%8.
The drop in in-hospital mortality shown by that study demonstrated the short-term
advantages that occurred after introduction of the new method of reimbursing
hospitals for hip fracture surgeries, but many other studies stress the importance of
considering the effects of earlier surgery for hip fracture on longer-term mortality6,9-11.
Page 4 of 22
Accep
ted
Man
uscr
ipt
4
Recent studies have shown that in different spans of follow-up after hip fracture
surgery, there are variations in the importance of different factors influencing
mortality3,12. These factors such as patients' age, gender and the hospital environment
should be considered when trying to establish a relationship between time of surgery
and longer-term mortality.
In this study we investigate the longer-term effects of the change in reimbursement
system on mortality following hip fracture, and to what extent the possible changes in
mortality after the reform could be ascribed to decreased time of surgery.
2. Objectives
• To examine whether introduction of differential pricing of hip fracture surgery
conditioned by time of surgery lowered the long-term mortality of patients
aged 65+.
• To find out whether shorter waiting times for hip fracture surgery of patients
aged 65+ are associated with lower long-term mortality.
Materials and Methods
This is a retrospective study of patients included in the Israeli National Trauma
Registry (ITR) during the period 2001-2007 (with year 2004 omitted as the year of the
policy change). Patients aged 65 and above with an isolated diagnosis of hip fracture
(ICD-9-CM 820)13 who were hospitalized at all 6 level I trauma centers in Israel and
at 3 regional trauma centers were included in the study. These 9 hospitals were chosen
because they were included in the Registry during all 7 years of the study. The data on
patients in the registry were linked to mortality data from the population database of
the Ministry of the Interior in order to obtain two-year follow-up on their survival.
Altogether 10,900 patients met the inclusion criteria.
The main outcome measure was longer-term mortality, with emphasis on 6-month,
one-year and two-year mortality. Comparisons of interest were: between patients
operated in less than 48 hours since arrival, those operated after 48 hours and
unoperated patients; and between the period before the change of reimbursement
policy (2001-2003) and the period after (2005-2007). In addition to survival analysis
of patients over the full two-year follow-up period, we also performed survival
analysis of patients over the 6 months to 1 year period and from the one-year to two-
Page 5 of 22
Accep
ted
Man
uscr
ipt
5
year period. This was done in order to analyze whether the factors influencing
mortality change over time.
Chi-square tests were used to compare categorical variables between the period before
and after the reimbursement policy change. Survival curves were estimated by the
Kaplan-Meier method and converted to mortality curves by subtracting survival
percentages from 100%; Cox regression was used to estimate Hazard Ratios (HR)
comparing the periods before and after reimbursement policy and adjusted for
patient's age, gender and the receiving hospital. The time of surgery was then inserted
into the model in order to estimate to what extent the difference between the two
policy periods was explained by the decreased waiting time for hip-fracture surgery.
A value of p<0.05 was considered to be statistically significant.
Results
To compare the long-term mortality before and after the change of reimbursement
policy, we began our analysis with a comparison of the patients in the two periods
(Table 1).
Females represented the majority (71%) in both periods. The age distribution was
slightly more heterogeneous in the first period, but in both periods almost half of the
patients belonged to the 75-84 age-group, with mean age of 81 (2001-2004:
Me=81.11/Std=7.39; 2005-2007: Me=81.27/Std=7.29). More patients required ICU in
the second period and 55% of patients in the second period entered a rehabilitation
unit, compared to only 45% in the first period. The change in reimbursement policy
led to more patients being operated (88% v 86%) and substantially more patients
operated in the first 48 hours since hospitalization (53% v 39%). However, even after
the reform, about 12% of the patients received conservative treatment. The patients in
the second period had greatly reduced LOS and lower in-hospital mortality (2.4% v
3.5%).
The difference in mortality between the two periods reduced after one-month follow-
up, but increased again after three months, the absolute difference in percentage
remaining at a stable level of ~2% over 6 to 24 months (which is more than 100 actual
patient lives saved just in the hospitals participating in the study). The difference after
6-months follow-up was the most statistically significant of these comparisons. Figure
1 shows the sharp decline in previously stable levels of 6-month mortality one year
after the reform stabilizing again at a lower level.
Page 6 of 22
Accep
ted
Man
uscr
ipt
6
The association of early surgery with patient's survival over two-years of follow-up in
both study periods is summarized by mortality curves (Figure 2). Expectedly, patients
who did not undergo hip fracture surgery had consistently higher mortality than
operated patients. The figure also shows a clear separation between the mortality
curves according to waiting time to surgery, with patients operated in the first 48
hours since hospitalization having the lower mortality. The influence of these factors
was identical in both study periods. The significance of differences between the
curves was additionally proven by Cox Regression, predicting mortality by time to
surgery, adjusted for receiving hospital and patients' age and gender in a 2-year
follow-up.
We have also found that both among patients operated in less than 48 hours and
among those operated later there was no difference in mortality between the two study
periods (Figure 3).
In order to evaluate the impact of the change in reimbursement policy on patients'
longer-term mortality, we calculated the Hazard Ratios (HR) over a 6-months follow-
up interval in the post-reform period in reference to the pre-reform period, while
adjusting for patient's age, gender, and by the receiving hospital (inserted sequentially
into the regression model). The results are presented in Table 2, where each
consecutive model signifies expansion of the independent variables list by one more
factor.
We have found that 6-month mortality significantly decreased in the period after the
change in reimbursement policy. The magnitude of this difference was largest when
all adjusting factors were present in the equation, with an estimate that patients
hospitalized in the second period were 13% less likely to die after hip fracture than in
the first period.
As expected, mortality increased with age and was significantly associated with the
receiving hospital; males had significantly higher mortality than females. When the
surgery variable was inserted into the model, the effect of period on mortality became
insignificant, suggesting that the difference between the before and after
reimbursement policy change periods was, in fact, explained by the change in surgery
practice (decrease in waiting time and increase in percentage of patients operated
upon).
When the model was applied to 1-year and 2-year mortality, it showed mixed results.
When the period 0-1 year or 2-year was analyzed, the results were similar to the
Page 7 of 22
Accep
ted
Man
uscr
ipt
7
previously described 0-6 months analysis. Nevertheless, the influence of policy
period on mortality decreased in the longer spans of follow-up (HR 0.91 (CI: 0.84-
0.98) over 0-1 year; HR 0.94 (CI: 0.88-1.01) over 0-2 years). The length of follow-up
had a similar effect on the influence of gender on mortality (HR 1.75 (CI: 1.61-1.91)
over 0-1 year; HR 1.69 (CI: 1.58-1.82) over 0-2 years).
However, when the model was applied over the 6-month to 1-year period and over
the 1 year to 2 years period, there were no differences found in the hazard ratios of
patients in the two policy periods (HR 1.03 (CI: 0.87-1.21) over 6 months to 1 year;
HR 1.04 (CI: 0.91-1.17) at 1 year to 2 years). The receiving hospital was also found to
have no association with mortality over these periods.
In order to account for possible gender-related bias in the impact of the
reimbursement reform we analyzed the interaction between gender and waiting time
to hip fracture surgery in both periods in addition to access to rehabilitation care
(Table 3).
The results show that women received significant preference for earlier surgery (as
well as for the surgery itself) in both periods. They were also better influenced by the
reform as their percentage of early surgeries increased by 14.8% as opposed to a
13.0% increase among males. In the second period, women also had greater access to
rehabilitation care (56.8% vs. 53.1%). Males were found to be somewhat younger
than females in the second period, but the percentage of patients older than 85 was
similar in both genders.
Discussion
The goal of our study was to assess the impact of the new reimbursement system, that
was aimed at stimulating reduced waiting time for hip fracture surgery, on long-term
mortality of patents older than 65. For that goal we sought to compare the clinical
outcomes of patients hospitalized due to isolated hip fracture before the new system
was implemented to the patients hospitalized in the later period in a two-year follow-
up.
Our results have clearly shown that following the reform of reimbursement for hip
fracture surgeries in Israel not only did the percentage of earlier surgeries increase, the
LOS decrease and in-patient mortality decrease as shown by the previous study8, but
also longer-term mortality of patients decreased. The period after the reimbursement
reform may have seen some clinical improvements, such as enhanced medication
Page 8 of 22
Accep
ted
Man
uscr
ipt
8
regiments, advances in surgical implants and, as we have shown in this study – greater
access to postoperative rehabilitation. However, the fact that when patient survival
was analyzed separately among "operated <48 hours" and "operated 48< hours"
populations, no differences were found between the two periods, can suggest that the
reimbursement reform and the subsequent decrease in waiting times for hip fracture
surgery could provide an explanation for the general decrease in long-term mortality
in the second period.
Many previous studies have shown the benefits of earlier hip fracture surgery for
elderly patients' survival and functional recovery4-7. However, there is still no clear
consensus on the subject, as other studies demonstrated that waiting time for surgery
loses its importance when adjusted for other factors, such as patent's age and gender,
the nature and the severity of previous comorbidities and the variability in levels of
receiving hospitals14-17. From our results it can be clearly determined that hip fracture
surgery in the first 48 hours since hospitalization was associated with a decrease in
longer-term mortality of elderly patients even when adjusted for demographic and
organizational factors. Though our database did not provide us with information on
patients’ comorbidities, the apparent influence of earlier surgery on survival was very
strong even after adjustment for the available factors.
We also found that the association of the period before and after the new
reimbursement system with patients' survival varied between different periods of
follow-up. After 3-months follow-up, the reform regained its positive impact on
patients' survival, but after adjusting by patient's age, gender and the receiving
hospital, the impact after 6 months of follow-up was no longer seen. The most likely
explanation for this finding is that as time passes since surgery, the advantages of
early surgery dwindles and other external factors that could influence mortality, such
as co-morbidity, come into play1,9,10. Previous research has already pointed to the fact
that when considering the long-term mortality of hip fracture patients, different
factors become dominant at different periods of follow-up3,12.
Special attention was drawn in the literature to the impact of age on mortality risk at
different follow-up periods among women, who usually comprise the majority of hip
fracture patients3. It was found that the risk for long-term mortality after hip fracture
is the greatest among females aged 65-69 years, while in the first year of follow-up
the risk is similar for the whole age span of women older than 653. Our study has
found that age had a monotonic influence on mortality among both genders regardless
Page 9 of 22
Accep
ted
Man
uscr
ipt
9
of the length of follow-up period. On the other hand, the influence of gender on
mortality decreased through the length of follow-up following the first 6 months after
hospitalization, during which males had a hazard ratio of 1.84 as compared to
females. The almost two-fold disadvantage of males in survival after hip-fracture is
well documented in the literature and is usually explained by their inferior previous
health condition and higher risk of infection11,18,19. The higher volume of co-
morbidities among male hip fracture patients can also explain the preference that
females in our study received for earlier hip fracture surgery, as patients who would
have greater benefit from the surgery could have been prioritized by the treating
hospital11,15,17,19. Female patients in our study were also found to have greater access
to rehabilitation care in the period after the change in the reimbursement policy. This
could also be the result of conscious hospital policy change after the reform of the
reimbursement system and one of the factors contributing to the observed decrease in
mortality12.
Future research should target the specific policy of the treating hospitals regarding the
influence of demographic factors on their priorities for treating hip fracture patients
and the possibilities of influencing this prioritization by policy.
The understanding of driving forces behind hospital policies is especially important
when trying to influence them by reform20-22. There is always a concern that
conditioning the financial reimbursement of hospitals for earlier hip fracture surgeries
would induce the surgeons to operate patients with contraindication for such surgery.
If this was to happen, the reform could lead to an increase rather than decrease in
mortality. In population aged 65 years and older it would be common to encounter co-
morbidities and use of medications that would provide a contraindication for surgery 9,15-17,23. The fact that even after the conditioning of DRG payments by time of surgery
only about 53% of patients in our study were operated early and about 12% received
conservative treatment, may indicate that hospitals still give more weight to clinical
considerations even when provided with economic stimuli.
Limitations
This study had several limitations that should be taken into consideration when
interpreting the results. Our main limitation was the lack of data on co-morbidities in
the Israeli Trauma Registry. If that information was available for the study, the found
Page 10 of 22
Accep
ted
Man
uscr
ipt
10
differences in mortality could have much more definitively be attributed to the change
in reimbursement policy and consequent decrease in waiting times for surgery.
An additional limitation is the fact that during the years of the study not all Israeli
trauma centers were included in the registry, thus limiting the scope of our
conclusions to hospitals actually analyzed in the study.
Conclusions
Our study indicates that a reform of reimbursement for hip fracture surgery
significantly increased the proportion of patients receiving hip-fracture surgery within
48 hours of admission and thereby decreased the longer-term mortality of elderly
patients up to 6 months after surgery. The reform did not influence the mortality rate
beyond 6 months following surgery, though the advantage in the total percentage of
survivors was preserved up to 2 years after surgery. The evidence suggests also that as
a result of the reform, hospitals may change their policies of prioritizing patients for
treatment, while trying to find a balance between the preference for earlier surgeries
and the clinical contra-indications for such surgeries.
Page 11 of 22
Accep
ted
Man
uscr
ipt
11
References
1. Brauer CA, Coca-Perraillon M, Cutler DM. Incidence and mortality of hip
fractures in the United States. JAMA. 2009;302(14): 1573-1579.
2. Tsuboi M, Hasegawa Y, Suzuki S, Wingstrand H, Thorngren K. Mortality and
mobility after hip fracture in Japan: a ten-year follow-up. J Bone Joint Surg Br.
2007; 89(4) 461-6.
3. LeBlanc ES, Hillier TA, Pedula KL, Rizzo JH, Cawthon PM, Fink HA, et al. Hip
Fracture and Increased Short-term but Not Long-term Mortality in Healthy Older
Women. Arch Intern Med. 2011;171(20):1831-1837.
4. Smith EB, Parvizi J, Purtill JJ. Delayed surgery for patients with femur and hip
fractures-risk of deep venous thrombosis. J Trauma. 2011 Jun;70(6):E113-6.
5. Simunovic N, Devereaux PJ, Sprague S, Guyatt GH, Schemitsch E, Debeer J, et
al. Effect of early surgery after hip fracture on mortality and complications:
systematic review and meta-analysis. CMAJ. 2010 Oct 19;182(15):1609-16.
6. Shiga T, Wajima Z, Ohe Y. Is operative delay associated with increased mortality
of hip fracture patients? Systematic review, meta-analysis, and meta-regression.
Can J Anaesth. 2008 Mar;55(3):146-54.
7. Carretta E, Bochicchio V, Rucci P, Fabbri G, Laus M, Fantini MP. Hip fracture:
effectiveness of early surgery to prevent 30-day mortality. Int Orthop. 2011
Mar;35(3):419-24.
8. Peleg K, Savitsky B, Berlovitz Y, ITG, Israeli A. Different reimbursement
influences surviving of hip fracture in elderly patients. Injury. 2011
Feb;42(2):128-32.
9. Söderqvist A, Ekström W, Ponzer S, Pettersson H, Cederholm T, Dalén N, et al.
Prediction of mortality in elderly patients with hip fractures: a two-year
prospective study of 1,944 patients. Gerontology. 2009;55(5):496-504.
10. Kim SM, Moon YW, Lim SJ, Yoon BK, Min YK, Lee DY, et al. Prediction of
survival, second fracture, and functional recovery following the first hip fracture
surgery in elderly patients. Bone. 2012 Mar 6. [Epub ahead of print]
11. Wehren LE, Hawkes WG, Orwig DL, Hebel JR, Zimmerman SI, Magaziner J.
Gender differences in mortality after hip fracture: the role of infection. J Bone
Miner Res. 2003 Dec;18(12):2231-7.
Page 12 of 22
Accep
ted
Man
uscr
ipt
12
12. Castronuovo E, Pezzotti P, Franzo A, Di Lallo D, Guasticchi G. Early and late
mortality in elderly patients after hip fracture: a cohort study using administrative
health databases in the Lazio region, Italy. BMC Geriatr. 2011 Aug 5;11:37.
13. Supplementary classification of external causes of injury and poisoning (E880-
E999). In: Puckett CD, ed. The educational annotation of ICD-9-CM. 4th ed.
* Curves estimated by the Kaplan-Meier method; no significant difference was found.
Page 18 of 22
Accep
ted
Man
uscr
ipt
18
Table1: Comparison of patients in two study periods (% (n)).
Period
Parameter 2001-2003 (N=5,538)
2005-2007 (N=5,362)
Male 28.93 (1602) 28.95 (1589) Gender
Female 71.07 (3936) 71.05 (3773)
65-74 19.75 (1094) 18.35 (984)
75-84 45.92 (2543) 49.01 (2628)
Age**
85+ 34.33 (1901) 32.64 (1750)
Rehabilitation** Yes 45.02 (2493) 55.71 (2987)
0-3 5.27 (391) 6.58 (352)
4-6 17.95 (991) 31.18 (1668)
7-13 53.60 (2959) 49.36 (2641)
LOS** (29 values are missing)
14+ 23.18 (1280) 12.88 (689)
<48 hours 38.59 (2133) 52.80 (2826)
>48 hours 47.24 (2611) 35.24 (1886)
Time to surgery** (21 values are missing)
No surgery 14.17 (783) 11.96 (640)
In-hospital mortality** 3.54 (196) 2.35 (126)
1-month mortality 5.40 (299) 5.02 (269)
3-month mortality* 11.11 (615) 9.88 (530)
6-month mortality** 15.78 (874) 13.84 (742)
12-month mortality* 20.98 (1162) 19.12 (1025)
24-month mortality* 29.79 (1650) 27.68 (1484) * Significant difference between periods (p value<0.05). ** Significant difference between periods (p value<0.01).
Page 19 of 22
Accep
ted
Man
uscr
ipt
19
Table 2: Cox Regression models: 6-month mortality in two policy periods adjusted
by receiving hospital, patient's age and gender, and by time of surgery.
HR, 95% CI Variable Model 1 Model 2 Model 3 Model 4 Model 5 Period 2005-2007 0.91,
Age (adjustment only)* p-value - - 0.0001 0.0001 0.0001
Gender Male - - - 1.84,
1.66–2.03 1.78,
1.61–1.96 Female (Reference group)
- - - 1 1
Time to surgery No surgery - - - - 3.88,
3.41–4.41 >48 hours - - - - 1.51,
1.34–1.70 <48 hours (Reference group)
- - - - 1 + All 9 hospitals that contributed data to the study. * Age was inserted as a continuous variable; the mortality grew with age monotonically.
Page 20 of 22
Accep
ted
Man
uscr
ipt
20
Table 3: Gender comparison: age structure, access to Rehabilitation Care and waiting
time for hip fracture surgery in both policy periods (% (n)).
2001-2003 2005-2007 Parameter Male Female Male Female
